<?page no="0"?> Gunter NarrVerlag Tübingen Buchreihe zu den Arbeiten aus Anglistik und Amerikanistik Penetrating Language A Critical Discourse Analysis of Pornography von Georg Marko <?page no="1"?> Penetrating Language <?page no="2"?> Buchreihe zu den Arbeiten aus Anglistik und Amerikanistik Herausgegeben von Walter Bernhart, Peter Bierbaumer, Alwin Fill, Gudrun Grabher, Arno Heller, Walter Hölbling, Allan James, Bernhard Kettemann, Christian Mair, Annemarie Peltzer-Karpf, Werner Wolf und Wolfgang Zach Band 23 <?page no="3"?> Georg Marko Penetrating Language A Critical Discourse Analysis of Pornography Gunter Narr Verlag Tübingen <?page no="4"?> Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über <http: / / dnb.d-nb.de> abrufbar. Gedruckt mit Förderung der Karl-Franzens-Universität Graz, des Bundesministeriums für Wissenschaft und Forschung in Wien, des Landes Steiermark sowie der Stadt Graz. © 2008 Narr Francke Attempto Verlag GmbH + Co. KG Dischingerweg 5 · D-72070 Tübingen Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Gedruckt auf säurefreiem und alterungsbeständigem Werkdruckpapier. Internet: http: / / www.narr.de E-Mail: info@narr.de Druck und Bindung: Ilmprint, Langewiesen Printed in Germany ISSN 0939-8481 ISBN 978-3-8233-6380-4 <?page no="5"?> Contents Prologue ............................................................................................... 11 Introduction ........................................................................................ 15 P ART 1: T HEORETICAL B ACKGROUND 17 I Discussing Pornography: Why People are Arguing about Pornography 19 1 Defining pornography ................................................................. 19 1.1 Ontological status ................................................................... 20 1.2 Content ................................................................................... 20 1.3 Intention (on a primary level)................................................. 20 1.4 Intention (on a secondary level) ............................................. 21 1.5 Semiotic modes ...................................................................... 21 1.6 Contextual features................................................................. 21 2 The issue and the camps .............................................................. 22 2.1 Christian moralists.................................................................. 25 2.1.1 Sexual morality ............................................................ 26 2.1.2 Language...................................................................... 26 2.1.3 Regulation .................................................................... 27 2.1.4 Alternatives/ Positive effects/ Sex workers ................... 28 2.2 Liberals................................................................................... 28 2.2.1 Sexual morality ............................................................ 28 2.2.2 Language...................................................................... 29 2.2.3 Regulation .................................................................... 29 2.2.4 Alternatives .................................................................. 30 2.2.5 Positive effects ............................................................. 31 2.2.6 Sex workers.................................................................. 32 2.3 Anti-pornography Feminists .................................................. 33 2.3.1 Sexual morality ............................................................ 33 2.3.2 Language...................................................................... 35 2.3.3 Regulation .................................................................... 37 2.3.4 Alternatives .................................................................. 38 2.3.5 Positive effects ............................................................. 39 2.3.6 Sex workers.................................................................. 40 <?page no="6"?> Contents 6 2.4 Anti-censorship feminists....................................................... 41 2.4.1 Sexual morality ............................................................ 42 2.4.2 Language...................................................................... 42 2.4.3 Regulation .................................................................... 44 2.4.4 Alternatives .................................................................. 44 2.4.5 Positive effects/ Sex workers ........................................ 46 2.5 My own position .................................................................... 46 3 Summing up.................................................................................. 46 II Researching Pornography: Sexology and Hermeneutics 49 1 Sexology ........................................................................................ 49 1.1 Correlational studies............................................................... 50 1.2 Clinical studies ....................................................................... 52 1.3 Experimental studies .............................................................. 53 2 Hermeneutics ................................................................................ 56 2.1 Evaluative surveys ................................................................. 58 2.2 Testimonial studies................................................................. 59 2.3 Content analysis ..................................................................... 61 3 Summing up.................................................................................. 63 III Researching Pornography: Critical Discourse Analysis 67 1 Hermeneutic theory I: What is discourse analytic about Critical Discourse Analysis? ....................................................... 68 1.1 Social constructionism ........................................................... 69 1.1.1 Cognition...................................................................... 70 1.1.2 Society.......................................................................... 74 1.2 Discourse Analysis ................................................................. 78 1.2.1 Discourse as language in use ....................................... 79 1.2.2 Discourse as type of language event ............................ 80 1.2.3 Discourse as concrete language event.......................... 81 1.2.4 CDA versus other linguistically-oriented approaches 82 2 Hermeneutic theory II: What is critical about Critical Discourse Analysis? ..................................................................... 84 2.1 Critical theory......................................................................... 85 2.2 Critical evaluation .................................................................. 89 2.2.1 Political dimension of the research process ................. 90 <?page no="7"?> Contents 7 3 Scientific methodology and metatheory ..................................... 91 3.1 Empirical data ........................................................................ 93 3.1.1 Beyond verbal language............................................... 94 3.1.2 Interpretative practices ................................................. 95 3.2 Quantitative data .................................................................... 96 3.3 Comparative data.................................................................... 97 3.4 Model of the idealized discourse participant.......................... 97 3.5 Hypotheses-based................................................................... 99 4 Summing up................................................................................ 100 IV Analysing Pornography: Concrete Features of the Study 103 1 Corpus analysis: empirical data, comparative data, quantitative data ........................................................................ 103 1.1 Empirical data: the corpus.................................................... 103 1.1.1 The magazines ........................................................... 105 1.1.2 The stories .................................................................. 107 1.2 Comparative data: a corpus for comparison......................... 108 1.2.1 The books................................................................... 109 1.2.2 The stories .................................................................. 110 1.3 Limitations on empirical data: non-verbal signs .................. 111 1.4 Limitations on empirical data: interpretative practices ........ 111 1.5 Computer-assisted analysis of the empirical data ................ 112 1.6 Quantification....................................................................... 115 2 A model of the idealized discourse participant (pornography and erotica consumers) ..................................... 116 2.1 Linguistic model................................................................... 116 2.1.1 Semantic grid ............................................................. 118 2.1.2 Full scenario............................................................... 120 2.1.3 From interpretation to semantic memory................... 124 2.1.4 Background ................................................................ 124 2.2 Social model of the idealized discourse participant ............. 125 3 Combining corpus analysis and the idealized discourse participant model ....................................................................... 128 4 The hypotheses ........................................................................... 130 5 Summing up................................................................................ 133 <?page no="8"?> Contents 8 P ART 2: A NALYSES 135 Introduction to Part 2........................................................................ 137 V Fragmentation 141 1 (Over)Representing body parts: lexeme frequency and variation ...................................................................................... 142 1.1 Technical and practical aspects of the analysis .................... 145 1.2 Presenting and discussing overall results ............................. 149 1.3 Comparing and discussing overall results ............................ 152 1.4 Presenting and discussing specific results............................ 155 1.5 Comparing and discussing specific results........................... 160 2 Immediately attending to body parts: possessives .................. 165 2.1 Technical and practical aspects of the analysis .................... 167 2.2 Presenting and discussing overall results ............................. 167 2.3 Comparing and discussing overall results ............................ 168 2.4 Presenting and discussing specific results............................ 169 2.5 Comparing and discussing specific results........................... 170 3 Casting a side glance: interesting phenomena not examined 172 4 Conclusion .................................................................................. 174 VI Physicalization and Visualization 177 1 Describing people: attributive descriptors .............................. 179 1.1 Technical and practical aspects of the analysis .................... 186 1.2 Presenting and discussing results ......................................... 193 1.3 Comparing and discussing results ........................................ 197 2 Describing body parts: attributive descriptors ....................... 202 2.1 Technical and practical aspects of the analysis .................... 204 2.2 Presenting and discussing results ......................................... 205 2.3 Comparing and discussing results ........................................ 213 3 Casting a side glance: interesting phenomena not examined 219 4 Conclusion .................................................................................. 222 VII Desubjectification 225 1 Representing people’s inner lives: mental process verbs ....... 228 1.1 Technical and practical aspects of the analysis .................... 230 <?page no="9"?> Contents 9 1.2 Presenting and discussing results ......................................... 235 1.3 Comparing and discussing results ........................................ 240 2 Marked perspective: passive voice ........................................... 246 2.1 Technical and practical aspects of the analysis .................... 247 2.2 Presenting and discussing results ......................................... 251 2.3 Comparing and discussing results ........................................ 252 3 Casting a side glance: interesting phenomena not examined 254 4 Conclusion .................................................................................. 256 VIII Passivization 259 1 Positioning sexual partners: semantic role patterns ............... 261 1.1 Technical and practical aspects of the analysis .................... 267 1.2 Presenting and discussing results ......................................... 273 1.3 Comparing and discussing results ........................................ 277 2 Representing women as empty space and men as tools: conceptual metaphors for genitals............................................ 283 2.1 Technical and practical aspects of the analysis .................... 289 2.2 Presenting and discussing results ......................................... 293 2.3 Comparing and discussing results ........................................ 300 3 Casting a side glance: interesting phenomena not examined 308 4 Conclusion .................................................................................. 310 Conclusion: Questions, Answers and Evaluations .................. 313 References ......................................................................................... 323 Index ................................................................................................... 347 <?page no="11"?> Prologue Chorus: What are you doing here? Me: Adding the last strokes to this book, the publication of my PhD dissertation. Chorus: What is it about? Me: The language of pornography. Chorus: Why pornography? Me: Because it is a culturally contested topic. I have always wanted to do research the results of which could have some social and political significance. Secondly, I wanted to write about something considered interesting by a wider audience so as to make people who are normally unaware of the workings of language get interested in a method of linguistic analysis that could prove beneficial for their everyday lives. Chorus: And why language? Me: I think that a linguistic approach to social phenomena such as pornography can shed some new light on old issues. Chorus: Why have you chosen the title Penetrating Language? Me: Because… well, probably because Flying planes can be dangerous. Chorus: Huh? ? ? Me: Who or what is doing the penetrating and who or what is being penetrated? And what is penetration anyway? Chorus: I have seen that your data is taken from magazines bought in 1995 - and this book is published in 2008. Why the large time gap? Me: If you just look at the finished product, it may be hard to see the long and chequered development behind it - even though some of the inconsistencies you will without doubt detect point to this. When I collected data in England and the United States during research stays there in 1994 and 1995, the focus of my <?page no="12"?> Prologue 12 study was much vaguer and broader. Selecting the best research objects (pornographic short stories), the appropriate methodology (corpus analysis), and the ideal analytical framework (Critical Discourse Analysis) then took some time. Preparing the corpus (which meant painstakingly typing hundreds of pornography stories) and developing my own approach to Critical Discourse Analysis - I realized that I had to deviate from its trodden paths - delayed my full dedication to the project further. Chorus: Isn’t it a historical study then? Me: Pornography has certainly changed in the last decade, particularly considering that the time span between the mid-1990s and 2008 includes the rise of the Internet, which especially affected the accessibility of erotic discourses considerably. On the other hand, I suppose that my approach, focusing on fundamental patterns of ideas and attitudes in discourses, can reveal more permanent properties of pornography. Chorus: What is it like to collect data in a project on pornography? Me: Tedious and emotionally exhausting because I do not have the self-confidence to simply walk into an adult bookshop and quickly find what I am looking for. It literally took me hours to enter and then I also spent a lot of time in the shops, always keeping my gaze downwards in order not to acknowledge that I was actually there. I always felt like telling everybody, “Well, I am doing research on pornography, I am not buying it for the sexual value,” but then I was not sure how they would have taken this. I wanted to limit my visits and I therefore bought as much as was available in the field I was interested in and as much as I could afford. Anonymous orders via the Internet were not an option at the time. I subscribed to one magazine, but that was just to get started right at the beginning and seemed not a viable solution for all the material. Chorus: Are you afraid that your research will be misunderstood? Just think of some of the reactions to your talks on the subject and remember the magazine using the headline “First porn doctor.” Me: There are several paths of misunderstanding, some of which I have already experienced. Some people (interestingly mainly men) in the academia seem to equate the analysis of a social phenomenon with the endorsement of the latter, thus deeming <?page no="13"?> Prologue 13 my topic inappropriate. Interestingly, I had similar reactions from non-academics, but here the evaluation was positive. Many males, on learning what I was studying, said, “You’ve come to the right man.” They automatically assumed that I, as a male, would necessarily try to flesh out the juicy details of the language of pornography. It was impossible for them to imagine that what I was doing could be critical and distanced to the object. It gave me a new (and not really appreciated) street credibility. Chorus: Could the book be read as pornography itself so that readers could be disgusted, embarrassed or even aroused? Me: I am positive that the seriousness and rigidity of my analytic approach as well as the style of data presentation will mitigate such effects and that they will also discourage using quotes as porn proper (and, well, there are cheaper forms available). It is important to find a balance between curiosity - I guess most readers will be interested in what pornography may look like - and a scientifically distanced perspective. The curiosity should make it easier to delve into the book. But eventually, it is important to bear in mind that the analyses serve to answer crucial social and political questions and do not simply present the language of pornography, exhibiting it as fascinating and/ or disgusting phenomenon. Chorus: What about the scientific community? I have got the feeling that many will see your approach as half-hearted, doing CDA, but focusing almost exclusively on textual dimensions, and using quantitatively-oriented corpus linguistics without applying proper statistical procedures and without paying too much attention to representativeness. Me: I accept these critical points without denying that I am dissatisfied with how I handled them. To a certain extent, the project was too ambitious right from the start, especially for a single researcher. But I decided to carry on despite these weaknesses because I think that the research still yields very interesting insights into the discourse of pornography and its potential ideological implications and secondly because I wanted to show what a large-scale, strictly data-based study in Critical Discourse Analysis could look like, demonstrating what might go wrong and where there is room for improvement in the process. <?page no="14"?> Prologue 14 The research process was like an incomplete hermeneutic circle: I had a vague idea, I decided on a method and an approach, I collected material, the idea became more concrete, the method and the approach were refined… but at a certain stage it was - for mundane reasons - not possible to go back and collect more material. So particularly in the area of corpus analysis I got stuck at a certain point. I am well aware that compiling a corpus by simply including as much as is easily available within a short time span, accepting that a comparative corpus is only a third of the length of the main corpus et cetera must seem inexcusably naïve, at best. Anyway, I am walking the thin line between originality, ambition and insanity. Chorus: Wasn’t the book longer? Where has all the data gone? And the glossary? Me: I ‘outsourced’ these parts to reduce the length of the book. But you can find all the data and the glossary on my homepage at http: / / www.uni-graz.at/ georg.marko. Chorus: Is this the first book you have written yourself. Me: Yes. I have edited a few volumes, but no work produced entirely on my own. Well, and then this is not really true either because I may have done the research and the writing, but there are a few people who have indirectly contributed by providing valuable critical feedback, most notably my two supervisors Professor Alwin Fill and Professor Bernhard Kettemann - all hat tipping to them - and by supporting me in all possible ways in the publication process, most notably Angelika Pfaller and Susanne Fischer at Gunter Narr Verlag. And then there are of course those who created the social, financial and emotional environment and atmosphere necessary to live and prosper and do such work. So all thanks and more to my parents, my sisters, my grandmother, all my family and friends and the great sunbears of this world. … hey… what are you doing there? Chorus: Don’t worry - we are just turning the page to get it started… <?page no="15"?> Introduction Does pornography objectify women, representing them as passive mindless things at the mercy of male sexual agents, or does it dive into women’s sexual subjectivities, revealing their innermost thoughts, desires and sensations? Does pornography subjugate women farther under the patriarchal yoke, or is it a potential path of female emancipation? Does pornography corrupt the values our society is built on or does it reveal the hypocrisy of traditional morality? Does pornography put sexuality above love, thus destroying relationships, or does it provide a remedy for sexual problems, thus saving relationships? Does pornography cause sex crimes or does it provide a safety valve for potential sexual assaultants? There are a host of questions in the discussion on pornography, questions that are evidence of the fact that it represents an ideal stage for cultural struggles over - among other things - sexuality, gender, morality, and representations. Although answers to these questions are primarily informed by deeply rooted ideological views of sexual morality, the different camps have never been reluctant to cite scientific studies corroborating their positions. Research into pornography therefore plays a seminal role in the debate, even if walking on ideologically swampy ground. This book presents further research into pornography. Like the research that has been done before, it tries to find answers to some of the questions emerging from the socio-political debate, in particular those centring on the issue of objectification. Unlike prior research, however, it starts at the very thing that constitutes pornography, namely at language. It is the first large-scale study of pornography in the tradition of Critical Discourse Analysis (henceforth CDA). CDA assumes that being exposed to language and particular ways of using language has an effect on the receivers’ knowledge and views of the world. By focusing on the ‘nitty gritty’ details of the language used in a large collection of pornographic short stories, I will therefore reveal aspects likely to be contained in the conceptions of the world or, more specifically, of sexuality and female and male sexual roles created in consumers of pornography, partly in comparison to those of consumers of the allegedly softer form of sexual representation, namely of erotica. <?page no="16"?> Introduction 16 The book also has a secondary objective: Critical Discourse Analysis may be solidly founded on a set of social goals but it still sadly lacks a common linguistic metatheory and methodology. Authors of book-length analyses therefore are also confronted with the task of contributing to CDA’s metatheoretical and methodological basis (cf. Fowler 1996: 8f.). Such a contribution involves making fundamental principles explicit and discussing them as well as using the analyses to demonstrate their relevance and feasibility. In the process, the analyses are also intended to evidence how a quantitatively-oriented linguistic discourse analysis of sexual texts may provide an alternative route to the scientific study of sexuality, going beyond the sexological paradigm of laboratory experiments and questionnaire surveys, but without resorting to intuitive speculation. A caveat: the methodological and metatheoretical (side)focus does not imply that I will present the study as a perfect model. On the contrary, I will try to be as critical as possible, making shortcomings, errors, misconceptions, inconsistencies and other types of ‘incubi’ haunting science as explicit as possible. The book is divided into two large parts, the first one dedicated to the theoretical background and the second one to the analyses proper. Part 1 proceeds from a description of the pornography debate and a review of research into pornography and its methodologies to the introduction of Critical Discourse Analysis as a valuable and viable alternative approach. It further outlines the basic features of CDA as used in my research. Part 1 ends in a set of hypotheses. Part 2 takes up the hypotheses and examines whether they can be upheld in an analysis of a corpus of pornographic short stories and a comparative analysis of a corpus of erotica. <?page no="17"?> P ART 1: T HEORETICAL B ACKGROUND <?page no="19"?> I Discussing Pornography: Why People are Arguing about Pornography Their basic query is not what about equality but what about orgasm. Catharine MacKinnon (1992b: 134) There has always been a heated controversy over pornography, raising many issues viewed and approached very differently. A study of pornographic representations cannot afford to ignore these discussions because no matter what my intentions are, simply by studying the phenomenon I am entering the debate. A thorough presentation of why and how pornography has become a topic so fiercely and controversially discussed, which aspects have come to the fore and appear to be particularly relevant and delicate, and which groups have participated in the debate will pave the way for an analysis which is not purely and blindfoldedly academic but which is aware of its social role, its obligations and its responsibilities. This chapter is intended to fulfil these requirements. After presenting my own definition of pornography, I will describe the issues that are at the core of the matter and on which the four main camps strongly disagree. 1 Defining pornography Justice Potter Stewart, unable to come up with his own definition of pornography, resorted to the infamous characterization, “I know it, when I see it” (Johann/ Osanka 1989: 3). This is an indication that, though we may have an intuitive grasp of the concept of pornography, defining it is not as straightforward a task. It no longer simply is the writing (-graphy) of prostitutes (porno-) (cf. Linz/ Malamuth 1993: 2), as its etymology suggests, but today’s pornography is a complex social phenomenon. Although past centuries saw texts and art with aspects that could be classified as pornographic (cf. Kendrick 1987, Hunt 1993, McNair 1996, Tang 1999), these artefacts lacked many of the features of current pornography, which have only emerged in the 20 th century and particularly since the sexual revolution of the 1960s as a result of technological progress and the changing cultural climate. I will restrict myself to pornography as a contemporary phenomenon in my study. <?page no="20"?> Theoretical Background 20 In the following, I will give six defining features of pornography that distinguish it from similar cultural products. 1.1 Ontological status Pornography is not a sexual practice itself, it is just a common way of talking or writing about sexual practices, a common way of representing sexual practices. Pornography is thus, to use the technical term, a sexual discourse 1 (cf. Cameron 1990a, 1992a, Cameron/ Frazer 1992, Hardy 1998), with the headword in pornography being graphy, i.e. the writing, rather than porno, i.e. the whores (cf. Kappeler 1986: 2). Although I consider this an essential feature of pornography, it has to be mentioned that the difference between represented practices and practices proper is becoming more and more blurred and there are many that say a discourse can also become a sexual practice in itself (cf. Baudrillard 1990: 28ff., MacKinnon 1992a: 462, both cit. in Bristow 1997: 145, 153). The borderline becomes particularly fuzzy with telephone sex or interactive sex chats on the Internet. 1.2 Content Pornography’s content is first and foremost sexual explicitness, i.e. the immediate and unmitigated description of the physical, physiological and perceptive (i.e. what it looks like and what it ‘feels like’) details of sexuality (cf. Williams 1979: 103, cit. in Einsiedel 1988: 109). Although other material, e.g. ‘high’ literature on sex or erotica, might also feature unmasked sexuality, pornography usually exceeds it in the sheer quantity of sexually explicit scenes (by three to eighteen times, according to a content analysis by Smith 1976: 19f.). 1.3 Intention (on a primary level) Pornography’s main (and perhaps its sole) intention is to arouse readers (cf. Williams 1979: 103, cit. in Einsiedel 1988: 109, Soble 1985: 8, Christensen 1990: 1, McNair 1996: 57). Unlike other sexual discourses such as educational material or erotica, pornography lacks any culturally revered aspects: it does neither impart scientific knowledge of sexuality nor does it have any aesthetic or literary pretensions (cf. Attorney 1 For the time being, it suffices to define a discourse as a common way of representing things such as sexual practices (but see section III.1.2). <?page no="21"?> Defining pornography 21 General’s Commission on Pornography and Obscenity: Final Report 1986, cit. in McNair 1996: 55). As a consequence, pornography is usually limited to certain forms of use: it is likely to be included in masturbation or in other sexual activity’s initial phase (cf. Report of the Commission on Obscenity and Pornography 1970: 266, cit. in Hunter/ Saunders/ Williamson 1993: 227, Masters/ Johnson/ Kolodny 1992: 354). 1.4 Intention (on a secondary level) On a secondary level, pornography is intended to yield profits for the producer. It thus is a commercial, mass-marketed product, unlike, for example, AIDS information leaflets. (For a detailed analysis of the economic sides of pornography, cf. Hebditch/ Anning 1988, Johann/ Osanka 1989: ch. 2 & 3, Itzin 1992a, Dines 1995, 1998a, Lane 2000.) 1.5 Semiotic modes Having always been fast in colonizing new technologies of communication (cf. McNair 1996: 44), pornography comes in many formats/ media combining verbal, visual and auditory semiotic modes: as books, magazines, comic books, videotapes, audiotapes, interactive CD-ROMs, and Internet sites. This distinguishes it from erotica, which usually are just published as books. 1.6 Contextual features Pornography has typical contextual features concerning appearance and places of dissemination. It usually comes in packaging featuring pictures of nude women, whether book, videotape, or CD-ROM, and it is restricted to particular places, namely to adult shops or to certain areas in general shops, e.g. corner top shelves of newsagents, or specially signed areas in video shops. Particularly conservatives and feminists have mentioned further distinctive features, mainly concerned with potential effects on community morality and the users’ social environment. Pornography is thus defined as perverted and obscene or degrading and humiliating to women (cf. Segal 1990: 30, Dworkin/ MacKinnon 1988: 138f., cit. in Russo 1998a: 14). To include such aspects in a definition, however, weakens any argumentation critical of pornography since the latter’s negative effects are quasi a priori proved. The argumentation is in danger of <?page no="22"?> Theoretical Background 22 becoming circular, and empirical studies would be rendered meaningless and thus redundant. If pornography is defined, for instance, as misogynist, then we cannot find it to be non-sexist through research (a female form would also be inconceivable by definition). Dines and Jensen (1998b: 65) are right in claiming that it is the task of research to show that pornography in the former sense has the features attributed to it by feminists and/ or conservatives. I will not distinguish between different forms of pornography on the basis of explicitness and uncommonness of the practices represented in this book. The distinction between hard-core and soft-core (cf. Easton 1994: xvi, Hardy 1998: 50, Dines 1998a: 63) might be economically and legally justified - hard-core pornography, for instance, is practically excluded from mainstream distribution outlets in the United States (cf. Dines 1998a: 54) and, in its visual form, prohibited in Britain (cf. Hardy 1998: 51). But from my point of view, the differentiation is a matter of degree and therefore not useful in my project. Besides, it is only valid in pictorial pornography, which means the distinctions mentioned do not apply to written material (cf. Ellis 1988, cit. in Hardy 1998: 50). The above definition of pornography should be sufficient for prototypically characterizing the object of discussion and analysis and has informed my selection of the data for my project. Although there are borderline cases, they will not figure prominently in my study. 2 The issue and the camps Opinions are widely divided about many aspects of pornography: about its aesthetic and literary values, its educational usefulness, or its status as a multi-million dollar business. What is at the heart of the debate about pornography, however, are its possible harmful effects: Can pornography trigger patterns of thinking and/ or behaviour harmful to society as a whole or to particular groups within society, undermining public morality and/ or working in favour of the more powerful social groups? The delicate nature of pornography’s harmfulness - and the reason why it has attracted more attention than that of other media supposed to manipulate their consumers - also lies in its unique link to sexual pleasure. Particularly feminists are afraid that sexualizing the negative effects mitigates them in the consumers’ eyes, making them seem less serious - after all, what turns people on cannot be that bad (cf. Kappeler 1986, Itzin 1992a, Russo 1998a). But there is also a second, more general consequence of the connection to sexuality: deprived of its <?page no="23"?> The issue and the camps 23 procreative and purist-religious significance, sexuality has become a major factor in the search for identities in modernity. How I act sexually has become part of how I (choose to) see myself (cf. Baird/ Rosenbaum 1991a, Giddens 1992). Anything sexual will therefore be interpreted as going right to the core of one’s identity, strongly interacting with other features, particularly with gender and religious and moral affiliations, but also with race and age. This entails that there is a lot at stake in talking about sexuality, or, in other words, sexuality assigns a particular urgency and acuteness to a question. Generally speaking, we can distinguish the following four camps in the pornography debate: a. Christian moralists b. Liberals c. Anti-pornography feminists d. Anti-censorship feminists They all agree that pornography is harmful if it can make people think, feel and do what they would otherwise not think, feel and do, what might have severe negative consequences for particular individuals (including sexual violence), what might create a cultural atmosphere hostile to certain social groups, and/ or what might undermine public morality in general (with the last aspect not of equal relevance to all groups). But to have this negative impact, pornography firstly must contain problematic ideas and secondly there must be a way that the latter influence consumers’ minds and acts. The groups’ conceptions of these two aspects are radically different, resting on their views on: i. Sexual morality If we have a clear conception of what constitutes morally good sexuality, e.g. that there should be only two participants, then we will reject depictions of sex that deviate from our norm. If our conception of morality is broad and relativist (‘everybody should do as they please’), then we will consider most representations of sexuality harmless. ii. Language If we do not believe that linguistic and other semiotic representations, i.e. texts in the broadest sense of the word, have any powerful and consistent effects on people’s minds then we consequently will not consider pornography to cause harm. Assuming, on the other hand, that such representations do have an impact on knowledge and action, we <?page no="24"?> Theoretical Background 24 will assign more power to pornography in the dissemination of certain ideas about sexuality and in the formation of certain behavioural patterns in its consumers. The perspectives that the four camps take on these two issues directly lead to conclusions concerning the need for political and legal measures in connection with porn. iii. Regulation The discussion about how to handle pornography legally and politically has mainly centred on the principle of free speech (the First Amendment to the American constitution, cf. Hunter/ Saunders/ Williamson 1993: 199, Easton 1994: ch. 6 and 10; other countries have similar principles, even if not constitutionally manifested, e.g. in Britain, cf. Easton 1994: 122). The question is whether regulating pornography amounts to a violation of the individual’s right to express what s/ he thinks or whether there are aspects that disqualify pornography as a free speech case, comparable to incitement to hate, lying, or libel, where the amount of harm justifies the overruling of the freedom of speech (cf. Baird/ Rosenbaum 1991a: 12, Hunter/ Saunders/ Williamson 1993: 199, Easton 1994: 62-64). The importance of pornography as a political and legal issue is documented by the fact that there have been various commissions to lead investigations into the field (cf. Einsiedel 1988, Johann/ Osanka 1989: ch. 13). The first American commission (Commission on Obscenity and Pornography, cf. Report of the Commission 1970) and the British one (Committee on Obscenity and Film Censorship, cf. Williams 1979) reflected the liberal atmosphere of the time, finding pornography not to be harmful and therefore pleading for a deregulation of laws (cf. Johann/ Osanka 1989: 437ff.). By contrast, the 1986 Attorney General’s Commission on Pornography and Obscenity in the USA (cf. Attorney General’s Commission on Pornography and Obscenity: Final Report 1986), in the conservative climate of the Reagan era, came to a different conclusion (cf. Berger/ Searles/ Cottle 1992: 25, Lisa Duggan, in Carmen et al. 1986: 16). These three aspects capture the central differences between the four groups. But there are three further topics, which partly logically follow from the positions on the first three issues but are still highly relevant. <?page no="25"?> The issue and the camps 25 iv. Positive effects Although the issue in the debate is harmfulness, there has always been mention of possible positive effects that pornography may have, particularly by those groups that are not strongly opposed to it. v. Alternatives Depending on the conception of what is wrong with pornography, there can be suggestions for improving it, mitigating or erasing its negative sides without, however, constraining some of its basic features (e.g. the sexual explicitness). vi. Sex workers Many theorists and activists have voiced concerns about those women and men working in the porn industry or, more generally, in the sex industry and about possible hardships they have to endure in modelling and acting. The four camps have adduced different studies to corroborate their positions. As I will dedicate a whole chapter to what kind of research has been conducted to prove or disprove hypotheses concerning the effects of pornography on users and their environment, I will not mention research issues in the description of the different camps. I will now characterize the positions of the four main camps particularly with regard to the six points described above. 2.1 Christian moralists For those who believe in God, in His absolute supremacy as the Creator and Lawgiver of life, in the dignity and destiny which He has conferred upon the human person, in the moral code that governs sexual activity - for those who believe in these “things,” no argument against pornography should be necessary. Charles H. Keating, Jr. (1991: 28) The Christian moralist or conservative views on pornography are particularly common among those with strong religious affiliations. It was the dominant anti-pornography position up to the 1960s and was regaining ground in the 1980s with the conservative backlash of the Reagan-and-Thatcher era in the US and the UK. <?page no="26"?> Theoretical Background 26 2.1.1 Sexual morality At the core of the conservative argumentation is the belief that moral principles are not relative - i.e. everybody should do as they please - but absolute. These a priori moral laws apply universally, which means they are valid for everybody everywhere at any time, mostly because they are seen as God-given or, secularly speaking, as logically deducible or nature-given (cf. Linz/ Malamuth 1993: 7f.). Christian moralists interpret sexuality as God’s gift and there are absolute moral laws regulating it. Any social practice violating these laws is perverse, obscene and/ or pathological - hence the moralists’ objection to ‘unusual’ sexual practices, to homosexuality, to promiscuity, and to artificial contraception. Generally, positive sexuality consists of penile-vaginal intercourse of husband and wife within marriage with the purpose or possible effect of conception. In less rigid views, sexuality is also positive if it does not (only) serve the satisfaction of immediate bodily desires but is the expression of the love between two human beings. As a consequence of the greater stress on the spiritual dimension of the human being, any overemphasis on bodily matters is rejected and recreational sex is regarded with suspicion. Hence sexual abstinence is highly esteemed (cf. Linz/ Malamuth 1993: 7f., 16ff.). 2.1.2 Language Christian moralists, without much theoretical ado, assume that what we read and watch will affect our thinking and our behaviour in an immediate one-to-one fashion. The conception of the effects is rather mechanistic and passive, as can be seen from the similes adduced in connection with pornographic representations, namely that of infection with a disease or poisoning (cf. LaHaye 1991: 181). Against this background, pornography must be regarded as dangerous and harmful. It features views of sexuality that conservatives do not condone and which they find might, if widely accepted, threaten the social-religious institutions of marriage and the family and thus the very fabric of our society. And widely accepted they will be because representations are considered to be extremely influential on the consumer. He will passively adopt perverse views through his consumption, he will act accordingly, and he will thus become the agent in the moral corruption of the world. Pornography thus practically undermines the concepts of love, affection, commitment and fidelity, even ridicules them and emancipates sex from love and shame (cf. Drakeford/ Hamm 1973, cit. in Johann/ <?page no="27"?> The issue and the camps 27 Osanka 1989: 23, Parker 1991: 184). Or, in John H. Court’s words (1985; cit. in Johann/ Osanka 1989: 262), in its effects pornography is life relationship family human woman children sex ANTIsocial environment community culture conscience God The promotion of the dissociation of physical sex from its emotional and social context is something that probably not only Christian conservatives bemoan but which other groups - including women-oriented groups - also find problematic. So although the conservative position is often polemically described as extremist, there are elements that can be found elsewhere, too (and probably are relatively widespread). 2.1.3 Regulation As far as the legal side of the issue is concerned, conservatives are in favour of the state legally intervening, censoring material that constitutes a danger to public morality. The instruments of censorship are obscenity laws. These are criminal laws focusing on obscene material rather than on pornography per se. Obscenity laws were introduced to exempt material from the right to free speech on the grounds that it corrupted and depraved consumers, violated accepted community standards of decency, and just served prurient interests. These laws are still in power in both Britain and the United States, though in the 20 th century - the 1960 acquittal of Penguin Books for publishing a full version of D.H. Lawrence’s Lady Chatterley’s Lover playing an important role here (cf. Hunter/ Saunders/ Williamson 1993: 148ff.) - they were mitigated by amendments or interpretations to the effect that literary, artistic, educational or scientific values qualify an otherwise obscene work to go scathless (cf. Dunn 1987: 394f., Hunter/ Saunders/ Williamson 1993: ch. 5 and 7, Easton 1994: ch. 12). The fuzzy concepts contained in obscenity laws - prurient interests, accepted community standards of decency, etc. - have always left room for interpretation for executive organs and courts. It is my impression that conservatives assume that common sense ethics will fill the gaps left by this fuzziness and that obscenity laws thus indirectly serve their views on morality. They will therefore leave the laws in place. <?page no="28"?> Theoretical Background 28 2.1.4 Alternatives/ Positive effects/ Sex workers For hard-sot conservatives the problem is obscenity and they subsume all sexually explicit representations under it. Consequently, criticism does not only encompass pornography as the most extreme form of sexual explicitness but extends to erotica and also to information material distributed at schools and elsewhere on topics such as contraception, abortion, or homosexuality. There are no alternative routes designed: sexual expression should preferably be suppressed. It does not come as a surprise that Christian moralists do not deal with possible positive effects of pornography: there simply are none. They are not concerned about those working in the sex industry either. The latter are more or less assumed to have been seduced. 2.2 Liberals My personal sexual revolution will come when I do what I really want to do sexually, don’t do what I don’t want to do, let others do what they want to do, with a whole heart. Sallie Tisdale (1995: 247) Liberal positions on pornography have probably always existed alongside conservative ones, functioning as their antagonists. They have, however, come strongly to the fore in the sexual revolution of the 1960s. Contrary to Hunter, Saunders and Williamson’s claim (1993: 162), liberalism is a relatively heterogeneous position in the debate with views ranging from the libertine appreciation of pornography to a mere rejection of its legal regulation. I will include divergences where necessary. 2.2.1 Sexual morality Liberal morality is generally relativistic and individualistic. Relativistic means that no superordinate authority determines what is really good or bad. So values do not exist in the outside world but just for the individuals who have them. Individualistic means that the main concern is the freedom of individuals, who should be able to fulfil their potential and live according to their own needs and desires as long as their actions do not infringe the liberties of others (John Stuart Mill’s harm principle, cf. Easton 1994: 1ff.). For sexual morality this entails that everybody has the right to engage in any form of sexuality they like as long as no one is involved against their own free will. For some, sex is impossible without the <?page no="29"?> The issue and the camps 29 social context of a loving long-term relationship, for others it is a purely bodily pleasure best enjoyed without any commitments. We might prefer one lifestyle to the other but we are not in a position to judge whether one is better than the other. And neither the state nor any other institution is entitled to intervene in this private matter unless a sexual practice involves non-consenting partners or harm to other people. Liberals tend to have a very positive view of sexuality. This is partly a result of their biologistic conception of it. Sex is regarded as a natural, biologically-given drive. From the naturalness of sexuality liberals derive the maxim to view it as positive and fundamentally good. We should therefore rejoice in sexuality and its diversity instead of despising or even prosecuting anyone who has chosen a different set of sexual practices for her/ himself (liberals do not see a contradiction in sexuality’s biological foundation and the diversity of sexual lifestyles). 2.2.2 Language Like conservatives, liberals have a rather simplistic common sense view of the power of language. Fiction - and pornography obviously is subsumed under this category - is “non-propositional” (Soble 1991: 96), i.e. not asserting anything about reality. Since recipients are aware of its fictitious nature, porn is incapable of influencing consumers’ views of reality, let alone their behaviour. It is thus more a sort of daydreaming (cf. Gagnon 1977, Schmidt 1985: 20-22, both cit. in Ertel 1990: 86). For liberals, pornography is thus the representation of diverse sexual practices, reflecting individuals’ diverse sexual tastes. The sexual acts depicted in pornography are generally not seen as problematic: nonconsenting and extraordinary sex is rare and, if represented, it is usually part of a ritualized form of behaviour such as sado-masochism, which is violence and dominance played out and not enacted (cf. Michelson 1986: 168, Smith 1993: 82, Rubin 1995: 245f.). And consenting sex, in whatever form it is practised, can only be criticized by those believing in absolute sexual values. Pornography, furthermore, just evokes sexual fantasies, but as an apparently fictitious discourse it does not affect us in our sexual behaviour. In sum, pornography does not - or has yet to be proved to - cause harm. 2.2.3 Regulation Against this background, it does not come as a surprise that liberals favour legal deregulation of pornography. No official institution should decide what people are to watch or read but the latter should take the <?page no="30"?> Theoretical Background 30 responsibility themselves. In other words, only the dynamics of a free market should regulate what kind of - or whether at all - pornography is further produced. The most serious form of regulation, namely censorship, violates the freedom of speech, one of the most fundamentally liberal principles. Even if we find pornography to cause some kind of harm, e.g. it might be offensive or discriminating against women - and the research is at best inclusive with regard to this issue - this harm would have to be extremely serious to outweigh the freedom of speech (cf. Manning 1991: 156). And generally, it is particularly the speech we hate that the principle of free speech has to apply to (cf. Ronald Dworkin 1991). Censorship, apart from being morally and legally unjustified, is furthermore very unlikely to have the desired effects. On the contrary, it could have two negative consequences. Firstly, it might drive the porn industry underground and criminalize it (cf. Toolin 1983: 173), possibly making it more extreme - now it at least has to preserve some kind of moral face - and making the working conditions for female models and actors more difficult than they are now (cf. Smith 1993: 83, Rubin 1995: 249). Secondly, in the hands of conservatives, censorship will - and did so in the past - target not only pornography but also information on abortion, contraception, and any material - information or erotica - by sexual minorities (cf. Easton 1994: 65ff.). As pornography might be offensive to some people, the only regulative measures to be taken are to prohibit adult shops to publicly display their products (cf. Cornell 1995: 104f., cit. in Bristow 1997: 160) or to have zoning laws that either concentrate outlets in particular areas or that evenly disperse them across towns (cf. Goldman 1983: 125, Easton 1994: 28f.). This was exactly the measures propagated by the first US and UK commissions examining the issue of pornography, which were influenced by the liberal spirit of the 1960s (cf. Einsiedel 1988). 2.2.4 Alternatives Libertine liberals consider pornography a positive phenomenon worth promoting. They concede that the bulk of pornography available today is dull and boring but attribute this to the general negative attitude toward sexual explicitness and to its criminalization. If pornography were culturally appreciated, it would automatically improve in quality and <?page no="31"?> The issue and the camps 31 open up to a wider audience - in particular to women 2 - and it could thereby fulfil its positive functions (see below). There are, however, also liberals who, even though they do not want to see pornography criticized and regulated on the grounds that it does not correspond to conservative sexual morals, still think that it is fundamentally flawed. D.H. Lawrence, for instance, himself not reluctant to portray sexuality very explicitly, thought that pure pornography was seriously lacking in aesthetic values (1936: 175ff.; cit. in Hunter/ Saunders/ Williamson 1993: 97ff.; for a thorough examination of how some literature has been re-defined as non-pornographic on aesthetic grounds, cf. Pease 2000). And gay activist Antony Grey celebrates sexual freedom in his book Speaking of Sex (1993), but at the same time says that pornography strips sexuality of its necessary mysteries (a common claim by people from various camps, incidentally). 2.2.5 Positive effects Liberals most strongly support pornography. They therefore are also the only ones that have been presenting whole catalogues of positive effects. Supporters of this camp view the history of sexuality from the Puritan to the Victorian ages as one of repression, in particular through the agencies of Christianity. Some therefore think of the emergence of pornography as a revolutionary event questioning the churches’ dogmatic perspectives on sexuality. They still believe in its political potential and its providing a mouthpiece for oppressed sexual minorities (cf. Rosenfield 1973, cit. in Brigman 1983: 133). 3 Considering the apparently commercial interest of the sex industry and the content and general quality of the products, however, not many liberals defending pornography claim that it has maintained its political edge. It is probably wishful thinking that pornography as a whole has the same impact that Marquis de Sade’s work - rightfully or not - has enjoyed. A further positive feature attributed to pornography is that it enables consumers to explore their own sexuality so as to find the practices that they and their partners find pleasurable, and to learn to perceive, and act with, their bodies in a positive, affirmative way, which leads to more mutually satisfying relationships (cf. Gardiner 1993, cit. in Strossen 1995: 165f.). Henner Ertel (and colleagues), in a German 2 Liberals such as Christensen (1990: 3f.), however, in their biologistic conception of gender, see male - as opposed to female - sexuality as visually cued. They consequently even regard pornography’s androcentrism as justified. 3 Perhaps due to a misinterpretation of the word pornography. Writing of prostitutes does not mean that the prostitutes write but that they are written about. <?page no="32"?> Theoretical Background 32 survey of attitudes towards pornography (1990: 207), found that occasional users are the group that talk most about sexual preferences with their partners and spend most time in foreplay, which is some indication that this group - possibly through the influence of pornography - has indeed found a positive, honest and also gentle way to deal with sex. In sum, pornography can be sexually liberating. Pornography may also be applied in the therapy of persons with sexual problems or sexual dysfunctions (cf. Stoller 1976, Wilson 1978, both cit. in Malamuth/ Billings 1984: 119ff., Strossen 1995: 163). It is also said to have a cathartic effect on those likely to sexually aggress, providing a safety outlet (cf. McCormack 1985, cit. in Berger/ Searles/ Cottle 1992: 42, Green, in Carmen et al. 1986: 18), and to be practical in the therapy of sex offenders (cf. McKay/ Dolff 1984, cit. in Pally 1994: 51f.). Data from studies showing a correlation between availability of pornography and falling rates of sexual crimes (cf. section II.1.1) corroborates this view. Pornography is furthermore claimed to represent a form of sexual education, allowing consumers to scrutinize the bodily details of the opposite sex without embarrassment (cf. Malamuth/ Billings 1984: 118, Berger/ Searles/ Cottle 1992: 77) and to compare their genitals with those of the actors of the same sex (cf. MacDonald 1995). Finally, it shows men different sexual practices and, at the same time, teaches them how to perform them. Most of these practices are considered extreme and rare by non-consumers. Tests have shown exposure to pornography to raise the estimated values considerably (cf. Zillmann/ Bryant 1982 and 1984; cit. in Zillmann/ Bryant 1988a: 520). 2.2.6 Sex workers Liberals take a less critical view of the sex industry. For them, people working there are just employed in the service sector. It is argued that if it were not for its criminalization, the whole sex industry could provide safe, clean and well-paid jobs. These, in turn, could help women to make a step forward in their economic emancipation (cf. Tisdale 1995: ch. 8). Against the view that erotic models and actresses as well as prostitutes are forced into their jobs, examples of women who have reached top positions in the sex industry are cited, for instance, Candida Royalle, a former mainstream porn star who started Femme Productions, a company targeting the women and couples market for pornography (cf. Juffer 1998: 173, Dines 1998b: 164). <?page no="33"?> The issue and the camps 33 2.3 Anti-pornography feminists Getting pleasure is not my particular agenda; getting equality is. If sexual pleasure is in the way, we need to think about it. A small number of women with loud voices, magnified through every available means of publication, have argued that they love it the way it is. […] This is because it is essential to their position that inequality is not a problem for sex so long as it produces pleasure, that inequality in the pursuit of orgasm is bliss. Their basic query is not what about equality but what about orgasm. Catharine MacKinnon (1992b: 134) The debate over pornography has become a cornerstone of feminism in the last four decades, pinpointing the growing awareness that the sexual revolution of the 1960s, despite the loosening of sexual mores, did not bring gender equality. In the wake of the 1978 Feminist Perspectives on Pornography Conference in San Francisco, which ended in a protest march through the city’s pornography district under the banner “Take back the night” (cf. Laura Lederer’s (1980) conference volume of the same title) (cf. Hardy 1998: 13), feminists were beginning to see pornography as fundamentally linked to the crucial gender issues of male-female relations in general and of sexual aggression in particular. The most prominent women in the anti-pornography camp, which appears as a more coherent camp than liberals and conservatives, are Columbia law professor Catharine MacKinnon (1989, 1993) and writer Andrea Dworkin (1981, 1987). But Susan Brownmiller (1976), Susan Griffin (1981), Susan Kappeler (1986), Mary Caputi (1994), Kathleen Barry (1995), Susan Cole (1995) and Diane Russell (1998) have also gained some importance in the debate. Particularly in the USA, where political activism is more firmly rooted, many anti-pornography activist and/ or lobbying groups have emerged in the past 40 years, e.g. WAP (Women Against Pornography), OAP (Organizing Against Pornography) but, to be fair, also opponent groups such as FACT (Feminist Anti- Censorship Taskforce) (cf. Johann/ Osanka 1989: 267ff.). 2.3.1 Sexual morality Feminists depart from a social constructionist conception of the world (cf. section III.1.1 for an elaborate discussion of social constructionism). This means that attitudes and world views do not reflect external reality but are constructed by us and then projected onto (our conception/ perception of) the outside world. This stands in contrast to the conservative theory of values (and truth) as absolute and naturally given and consequently commits feminism to moral relativism. Constructing the <?page no="34"?> Theoretical Background 34 values and categories in which we think about and perceive reality, however, is a social process constantly taking place in all forms of interaction within society. Morality and world views are thus social and cannot simply be taken to be based on individual preferences and decisions, which challenges the moral individualism of liberals. The social nature of the construction process also means it is sensitive to relations of power: the more powerful are more likely to get their categories and values established in a society than the less powerful, which, as a consequence, will usually serve the interests of the ruling class. Although feminists would generally endorse the tolerant attitude inherent in moral relativism, their social constructionist perspective on the world entails that acceptance has to be preceded by a critical scrutiny of values and world views and, if necessary, socio-political intervention. Otherwise the concessive approach is prone to be instrumental in the perpetuation of a status quo unfavourable to the socially weak. Feminists also regard sexuality as socially constructed and conceive of it as a culturally acquired set of sexual behaviours and values and not simply as a natural and individualistic phenomenon (cf. especially Michel Foucault’s Histoire de la Sexualité, 1992 [1976], 1993a,b [1984]; 4 cf. also the references in section II.2). Sexuality, too, is thus affected by power relations (a point also most prominently made by Foucault). So although there are no absolute ethical principles, feminists cannot promote an ‘everybody should do as they please’-attitude because sexual values and categories are socially modelled to favour the powerful, albeit not necessarily in an overt manner. And being embedded in a patriarchal culture, these values and categories of sex correspond primarily to males’ desires and advantages and may consequently prevent women from finding their authentic pleasure/ s or, more seriously, could even hurt and harm them (cf. MacKinnon 1989, Barry 1995: 55f.). In other words, sexuality is never simply the expression of individual desires but is socially conditioned and socially influential. In a patriarchal culture, it will not work in the interests of women as a group. The social conditions of pleasure therefore have to be exposed to critique before being enjoyed. What feels good may only superficially and momentarily do so, but could be or even is bound to be sexist on the long run. 4 The idea has also been put forward by others at the same time or even before Foucault, for instance, by Mary McIntosh 1981 [originally 1968], Jeffrey Weeks 1977, or John Gagnon and William Simon 1973, all three cit. in Bristow 1997: 17 and 199. <?page no="35"?> The issue and the camps 35 The social constructedness of sexuality within a patriarchal social order can lead to pessimism about sex. Considering equality an essential prerequisite to pleasure (cf. MacKinnon’s statement above), some of the main anti-pornography theorists go so far as to question the very possibility of female (hetero)sexual pleasure in a patriarchal society. At least “we have to question the role of sexual intercourse in a sexist culture” (Cole 1995: 46). 2.3.2 Language Feminists, particularly feminist linguists, have variously discussed the relation between language and thinking/ behaviour. There is a growing body of very sophisticated accounts of how recipients in interpreting texts change, modify or support their knowledge of a particular topic. 5 Great power is thus assigned to language and texts: they are considered the main factor in the social construction of phenomena such as sexuality (cf. Cameron 1990a, 1992a). As pornography is one of the seminal discourses on sex, which, due to its tight link with sexual arousal and activity, does not invite critical intellectual questioning, its role in the social construction of sexuality must not be underestimated. The liberals’ argument that fiction has no impact on its consumers is refuted by anti-pornography feminists. They consider the fact-fiction - or reality-fantasy, for that matter - dichotomy treacherous in the discussion of the effectivity of representations in general and pornography in particular. Although most of the events in fiction have been invented, fictitious worlds still are sufficiently realistic, which means that in most of their features they correspond to the real world. Fictitious sex will consequently not be interpreted as completely distinct from real sex (it wouldn’t sell if it were). Sexual fiction may thus help to structure the perception and conception of sexuality in real life (cf. Hill 1991: 69). Besides, most people find it hard to draw the line between reality and imagination even in their own memories of sexual encounters (cf. Barbach 1985a: xiii-xiv) In conclusion, for anti-pornography feminism the problem of the pornographic message is not the promotion of sexual fun per se, but the incorporation of categories and values unfavourable to or discriminatory against women as a social group (often summed up in categories of 5 Feminist linguists have moved on from earlier wholehearted adoptions of the Sapir-Whorf-Hypothesis (e.g. Spender 1985 or Hamilton 1985, cit. in Werlen 1989: 210f.). For a critique of the approaches that assume a one-to-one mapping of language systems and reality, cf. Cameron 1992c: 146ff. <?page no="36"?> Theoretical Background 36 misogynisms such as objectification, fragmentation, subordination, degradation, etc., cf. Dworkin/ MacKinnon 1988 6 ): women are depicted as dehumanized, even fragmented, objects to be had or consumed. They are insatiable for sex and constantly willing to ‘do it’ with whomever (as long as he is well-endowed) and whatever kinky way the partner chooses. Women are presented as even enjoying the physical, verbal and psychological violence and pain inflicted upon them. The erotic charge of sexuality does not presuppose trust, love, sensuality and reciprocity but firmly rests on the power imbalance between the sexes with the male on top. Inequality is thus rendered not only acceptable but even desirable. Neither sexual pleasure nor sexual pain, which is notoriously absent from pornography anyway, are thus defined in female terms or from a woman’s perspective. This is regarded as highly negative and sexual arousal does not relativize this (it rather conceals it). Assuming that representations and discourses can influence people’s knowledge and behaviour, anti-pornography feminists conclude that pornography will have an effect on its consumers. And if it does indeed include the aspects mentioned above, male sexual behaviour informed by this discourse will be to women’s disadvantage. On the lighter side, men’s actions are unsatisfying to women, focusing exclusively on male pleasure. On the more serious side, however, pornography creates a cultural climate characterized by so-called rape myths, i.e. assumptions that convey a distorted and trivialized picture of sexual aggression, e.g. that women mean “Yes” even when saying “No” (cf. Plummer 1995: 67f., Johann/ Osanka 1989: 205ff. 7 ). This climate may make it easier for males to actually resort to sexual aggression and, by laying the foundation for it, pornography may thus be a causal factor in this kind of violence, concisely formulated by Robin Morgan in her famous sentence, “Pornography is the theory and rape is the practice” (1980: 139). It is, of course, widely acknowledged within this strand of feminism that the relation between pornography and sexual aggression/ rape is not an immediate, monocausal one and sophisticated models have therefore been proposed (cf. Russell 1998: part III). Pornography probably does not affect all consumers in the same way and does therefore not necessarily lead to harm. But if some of it does some of the time, this would be enough to take action against it as 6 Some of these will be re-defined in chapter IV (section 4), as they will form the backbone of the hypotheses to be tested in the analyses. 7 Some studies show that adolescent males consuming pornography are more likely to subscribe to rape myths (cf. Cowan/ Campbell 1994), which point to a possible link between the latter and pornography. <?page no="37"?> The issue and the camps 37 “[…] we have an obligation to those with the least power, to those who are hurt in this pornographic world” (Jensen 1998a: 161). Anti-pornography feminism rejects the equation of their cause with that of the conservatives (see also next section). The two major differences are that conservatives’ main concern seems to be the moral corruption of the male consumer himself, while feminists concentrate on the victims against whom the male consumer, incited by pornography, may exercise psychological, verbal and physical sexual violence. Secondly, Christian moralists oppose pornography because it first and foremost morally corrupts, while feminists fight it because it literally hurts (cf. Bristow 1997: 149). Anti-pornography feminism, furthermore, believes that the religious right have established the androcentric and misogynist world view inherent in pornography in the first place (cf. Weaver 1991) and are therefore the liberals’ allies. 2.3.3 Regulation Feminists have never pleaded for censorship and obscenity laws. As the harm is done to individual women, the latter should benefit from legal remedies (cf. Cole 1995: 86f.). They should have the chance to sue the producers, publishers and/ or distributors of certain materials. For this purpose, civil right ordinances were drafted by, among others, Dworkin and MacKinnon in the 1980s (Dworkin/ MacKinnon 1988: 36-41) and introduced in Minneapolis and Indianapolis. The Minneapolis ordinance was passed but was ruled down by the Supreme Court on the grounds that it violated the First Amendment, the Indianapolis ordinance was not passed in the first place (cf. Linz/ Malamuth 1993: 13ff., Hunter/ Saunders/ Williamson 1993: ch. 8, Easton 1994: ch. 11). Booksellers and publishers have been anxious not to let such ordinances be accepted by any legislative body since if they were liable to civil cases then they would have to tailor their products in a fashion that could amount to preconstraints more restrictive than the censorship imposed by obscenity laws (cf. Easton 1994: 109ff.). Civil laws, of course, also limit the free expression of pornographic content. Anti-pornography feminists have therefore repeatedly said why they think that pornography should be treated as an exception to free speech. First of all, the right not to be discriminated against (the Fourteenth Amendment of the American Constitution) could override the right to free speech, i.e. the right to consume and produce any material you like (cf. Longino 1991: 91, Hunter/ Saunders/ Williamson 1993: 233). Pornography could thus be treated under a law against incitement to sexist hatred, in analogy to the law against incitement to racial hatred (cf. Easton 1994: 158ff.; for a counterposition, cf. <?page no="38"?> Theoretical Background 38 Weinstein 1998). Secondly, it is doubtful whether pornography is to be treated under the heading of speech at all and whether it does not qualify more as sexual action (cf. Easton 1994: 91f.). Thirdly, the right to free speech was established to protect the free expression of political thought so that citizens had the widest choices of political ideas from which to choose. Pornography is commercial talk and is hardly valuable as a means in political education. It therefore does not constitute the prototypical example of a free speech case (cf. Easton 1994: 90). 2.3.4 Alternatives Anti-pornography feminists sneer upon the possibility of women exploring pornography to present their own fantasies because to them, the ‘sexploitation’ of women is inherent in the discourse itself. Women, furthermore, have to be cautious since what feels good is not necessarily (or may in fact undermine) what brings them equality (cf. Cole 1995: 39- 46, also MacKinnon’s statement above). Considering that the context of women-oriented pornography still is patriarchy, it is difficult to create really successful alternatives (cf. Myers 1995 8 ), which would quickly be adopted by mainstream pornography anyway (cf. Gottdiener 1985, cit. in Berger/ Searles/ Cottle 1992: 64). A gynocentric perspective is not created either by establishing a female editor as some magazines seem to have done (cf. Itzin 1992a: 42f.). Particularly if this editor poses in the nude on the first pages, this is quickly unmasked as a strategic move. Anti-pornography feminists do, however, not reject sexually explicit material in toto but endorse erotica as a valuable alternative of women-oriented sexual representations. Erotica feature explicit sexuality but put less emphasis on its quantity and more on its social and emotional context. They are also intended to go beyond mere physical gratification to a more intellectual-emotional sensuality. Gloria Steinem (1980: 37), for instance, says that in erotica there is usually sensuality and touch and warmth, an acceptance of bodies and nerve endings. There is always a spontaneous sense of people who are there because they want to be, out of shared pleasure. while pornography, as Audre Lorde (1980: 296) puts it, “emphasizes sensation without feeling.” Erotica are thus less likely to be immediately instrumental in sexual practices. Publishers are also cautious to market erotica as significantly distinct from pornography (cf. Juffer 1998: 26f.): they can be obtained from 8 She does not say this explicitly, but I conclude this from the questions proposed for feminist pornography. <?page no="39"?> The issue and the camps 39 major bookshops, their covers rarely show photographic nudity, and they usually just come in the most culturally accomplished medium, namely as books. The intended readership is primarily female, as can be seen from the fact that a lot of erotica include phrases such as for women in their titles. Contrary to some opinions (cf., e.g., O’Toole 1998: 7), experiments have in fact shown women to be able to clearly distinguish erotica from pornography. Erotica, for instance, do not cause the symptoms of distress or the mood changes that pornography evokes in female viewers/ readers (cf. Senn/ Radtke 1986, Check/ Guloien 1989, both cit. in Russell 1998: 4, Senn/ Radtke 1990, Saunders/ Naus 1993). While some feminists accept erotica as a legitimate form of explicit representation of sexuality, others are suspicious of them, considering them pornography for intellectuals. The latter do not want women to explore this genre at all. Andrea Dworkin is among them, according to Segal (1990: 39). 9 2.3.5 Positive effects Anti-pornography feminism does not see any benefits in pornography and strongly rejects the arguments in favour of it. Pornography is, for instance, not at all seen as a revolutionary medium lending a voice to sexual minorities. Anti-pornography feminists, on the contrary, argue that being deeply immersed in capitalism and possibly in organized crime pornography rather corresponds to (and creates) the tastes of the majority or at least to the tastes of a politically and financially powerful male minority (Russo 1998a: 17f.). It is, furthermore, ridiculous to consider the anonymous consumption of pornography in private homes, not bearing any risks, a political practice even if pornography may feature the violation of moral boundaries. I may add that the transgressing of taboos may also be ritualized, thus actually reconfirming them as taboos rather than undermining them (for the ritualized transgression of taboos, cf. Bataille 1962, and also the interpretations of his ideas by Harvey/ Shalom 1997a: 9 and Bristow 1997: 124f.). Another argument put forward in pornography’s defence, viz. that it represents a useful device in sex therapy, also has one serious drawback. People who do not enjoy pornography have been found to 9 As indicated above, there are also liberals that favour erotica over pornography. Their criteria of discerning the two are similar although mostly rooted in a literary aesthetics, e.g. the detail of description must always stand in some relation to the whole of a work, there must be some device creating distance to reality, it should not hurt feelings of morality, etc. (cf. Popp 1989). <?page no="40"?> Theoretical Background 40 have generally sex-negative attitudes (erotophobia) (cf. Becker/ Byrne 1985, cit. in George/ Lopez 1995: 277, Fisher et al. 1988), to be afflicted by sex-guilt (fear of anything sexual) (cf. Schill/ Chapin 1972, cit. in Morokoff 1985: 177), or to stick too tightly to gender-specific norms, e.g. that women are not allowed to like pornography. So it is exactly the group needing sexual ‘de-repression’ that is unlikely to open up to pornography at all. It is further also doubted that pornography can have a cathartic effect on those with socially unacceptable sexual preferences or on those likely to sexually aggress. No convincing model has been presented how catharsis should work psychologically (cf. Griffin 1981: 93-103), because most accounts drawing upon the concept imply a rather mechanistic conception of the human mind (cf. Turner 1999). A survey into the sex education of adolescents revealed that pornography is just a minor factor in their learning about sexuality, lagging far behind peers and school (cf. Trostle 1993). Claims about pornography’s value in this area are thus also exaggerated. Besides, if pornography is to be used as sex education it would have to be made available to minors, something that not even forefront liberals demand (cf. Easton 1994: ch. 1) There seems to be a contradiction in the liberal argumentation for positive effects anyway. While pornographic representations are considered to be without impact on consumers with regard to harm, great power is attributed to them if it comes to positive effects. The contradictions surface in other areas, too, because liberals, for instance, believe in representations in advertising being able to influence the minds of potential consumers (cf. Cole 1995: 56). 2.3.6 Sex workers Feminists have not only paid attention to the socio-psychological effects of pornography, but also to the situation of the women working in the sex industry as models and actors. For them pornography often actually is sex and violence and not just the representation of it. In the light of descriptions of the sex industry from the female perspective such as Linda Lovelace’s (1980) or Evelina Giobbe’s (1995), it is cynical to claim that this kind of business may contribute to female emancipation. As Russo (1998a: 22ff.) points out, it is not the aim of feminism to undermine the individual woman’s right to participate in the sex industry. But we have to examine the structural inequalities that force some women into this branch of capitalism. <?page no="41"?> The issue and the camps 41 2.4 Anti-censorship feminists As long as a long-suffering, victimized, and repressed natural female sexuality is viewed as the antithesis to a falsely ideological, constructed, sadistic male sexuality (or any other kind of “perversion”), practical resistance to what many women do find inimical in that sexuality is limited to the condemnation of unorthodoxies measured against an orthodox norm. Linda Williams (1999: 23) Anti-censorship feminism, which is also called pro-pornography or prosex (found, e.g., in Cameron 1990a, Senn 1993, Hardy 1998), emerged as a separate camp in the pornography debate in reaction and strong opposition to anti-pornography feminism. The pro-sex Barnard conference in New York in 1982 might be regarded as the first organized event of this strand of feminist thinking (cf. Hardy 1998: 15-18, cf. also the conference volume Vance 1992 [originally 1984]). Many of the anticensorship feminists work in academia (cf. Hardy 1998: 20) and/ or are active in the struggle for the rights of sexual minorities. The most important anti-censorship feminists are Ann Snitow (with Stansell and Thompson 1984b), Carole Vance (1992/ 1984), Alison Assiter (1989), Lynne Segal (1990, with Mary McIntosh 1993), Gayle Rubin (1995), Nadine Strossen (1995), and Laura Kipnis (1996). Interestingly enough, anti-pornography feminism is not really widely represented in the female population, which implies that it entertains views that not many women outside a mostly academic world hold (cf. Senn 1993: 337, Hardy 1998: 20). The anti-censorship feminists pose a problem to the anti-porn countercamp since they undermine the feminist consensus that the latter tries to conjure up. Andrea Dworkin, while sharing the panel with people from the right and from the liberal camp, was therefore more reluctant to discuss with anti-censorship feminists in public (cf. Strossen 1995: 85). This does, however, not come as a surprise given the fierceness with which the two camps fight each other (cf., for instance, the derogatory blend MacDworkenites, which Strossen (1995) uses). Anti-censorship feminists mainly define themselves in opposition to the anti-pornography camp, as will become clear from the account below. There is much less friction with liberal positions although there is a lot of women-oriented common ground between the two feminist camps, even if not stressed very often. <?page no="42"?> Theoretical Background 42 2.4.1 Sexual morality Anti-censorship feminism shares the anti-pornography feminists’ conception of morality in general and of sexual morality in particular as socially and discursively constructed. Granting that dominant sexual values and categories are not primarily determined by women in patriarchy does, however, not implicate a pessimistic and negative attitude towards sex. Being critical about sexual values and categories should not eventually prevent women from seeking their own sexual pleasure, which is possible even in a misogynist climate. And being critical about an individualistic ‘everybody should do as they please’view of morality should not lead to being prejudiced against certain sexual practices (cf. Michelson 1986, Segal 1990: 32, Strossen 1995, Rubin 1995). Anti-porn feminism is accused of being inconsistent in its constructionism and of falling prey to essentialism because it eventually resorts to the unquestioned categories of the powerful male with his fundamentally aggressive sexual nature and the passive and vulnerable female, who can only be a victim of pornography but would never appreciate it (cf. Snitow/ Stansell/ Thompson 1984a, cit. in Hardy 1998: 16). Such views a priori exclude the possibility of non-sexist (hetero)sexuality and, what is worse, are prescriptive with regard to women’s or at least feminists’ attitudes towards pornography (in the sense of ‘If you like pornography, you cannot be a feminist’) (cf. Heise 1995: 124, Bristow 1997: 157f.). A second consequence of this essentialism is that it implies a monolithic view of what constitutes female pleasure, focusing primarily on a romantic conception of sexuality as socially and emotionally embedded. As a result, everything that goes against this vein, including the playful enacting of inequality, is seen as a danger to women and thus to morality. This view ignores that there are different female sexualities, some of which stand in opposition to the romantic notion just mentioned. Not surprisingly, the first group to provide a hint at the corrosion of the feminist consensus about pornography was SAMOIS, a lesbian sadomaso group (cf. Michelson 1986: 167ff., cf. also SAMOIS’s (1987) collection Coming to Power). 2.4.2 Language Anti-censorship feminists assign a lot of power to language and discourses. In a postmodern vein, however, they stress the readers’/ viewers’ interpretative autonomy. Everybody can, if they choose to do so, reject the preferred readings of texts/ images and construct their own <?page no="43"?> The issue and the camps 43 alternative interpretations (cf. Kuhn 1995, McNair 1996, Williams 1999). This ability might, however, be unequally distributed, i.e. some groups - e.g. anti-censorship feminists - can come up with subversive readings while others cannot or do not want to (cf. Juffer 1998: 20). Some pro-sex feminists also re-iterate the liberal argument about the purely fictional nature of pornography, which locates sex in the realm of fantasy and has no effect on reality (cf. Assiter/ Carol 1993, Carol 1994, Strossen 1995, Kipnis 1996, all cit. in Dines 1998b: 163). These perspectives on sexual morality and representations imply that pornography as a very widely distributed sexual discourse could theoretically help to construct sexuality in a way unfavourable to women. In their cultural and sexual optimism, however, anti-censorship feminists do not consider it a real threat. Firstly, they believe that most consumers are not ill-meaning males with sinister sexual desires who passively adopt the sexism from the texts/ images but men who can actively and critically create their own meanings. And secondly, pornography is usually not consumed in an environment of acquiescent, compliant females but in a climate that allows active female counterculture, which should sit uncomfortable on the preferred readings of pornography. Anti-censorship feminists think that the hysteria created by antipornography theorists and activists goes beyond any justified critical doubts and indicates a deep alienation from sexuality and the possibilities of bodily pleasures. Female sexuality has so long been thwarted in its prospering and incarcerated, only beginning to be liberated by such efforts as Betty Dodson’s teaching about female masturbation and its emancipatory potential (first appearing in Dodson 1974, and summarized in Dodson 1996), the Shere Hite report on female sexuality (1976), Nancy Friday’s My Secret Garden. Women’s Sexual Fantasies (1973), or Our Bodies, Ourselves by the Boston Women’s Health Book Collective (1971), a guide to women’s sexuality and sexual health. In the light of such major steps forward, a minor problem such as pornography should not endanger the newly-won ground and should not prevent women from further striding ahead (cf. Michelson 1986, Segal 1990: 32, Strossen 1995, Rubin 1995). 10 10 Sexual knowledge is sexual power is the motto of one magazine to be analysed in Part 2. This sentence allows for two interpretations nicely pinpointing the difference between the two feminist camps: either women by gaining more knowledge about sexuality will become equal to men in sexual matters, or men by acquiring a certain kind of knowledge - however distorted - will exercise power over women in sexuality. <?page no="44"?> Theoretical Background 44 Besides, compared to the wide array of sexisms in our society that women find themselves confronted with - lower wages, single motherhood, too little quality child care, etc. - pornography still seems to be a minor problem (cf. Rosen 1994), which is the symptom of sexism rather than its cause anyway (cf. Dority 1991: 113). What anti-censorship feminists deem to be the major faux pas in the issue is that their anti-porn ‘sisters’ make common cause with conservatives. The latter, and not libertine pornography consumers, are the vanguard of patriarchy, oppressing women, keeping them at home, and not granting them the right to decide over sexual matters themselves. 11 2.4.3 Regulation As becomes clear from the arguments mentioned above as well as from their name, anti-censorship feminists hold the view that any strict form of legal regulation of pornography, which will always amount to censorship, whether in the form of civil or criminal laws, is unlikely to reduce sexism in our society. Supporters of the anti-censorship agenda point to the slippery slope of censorship: in the hands of patriarchal conservatives it will not only target violent hard-core pornography but also sexually explicit material for women and for sexual minorities, especially for lesbians, as has been the case in Canada, where a law modelled after the civil ordinances was passed and was subsequently used to raid lesbian bookshops (cf. O’Toole 1998: 30f.; Cole (1995: 27f.), however, says that the Canadian police has always focused on such groups and the law has in fact not aggravated the situation). So legal regulation will not liberate female sexuality but it will put it more firmly under the patriarchal yoke (cf. Soble 1985; cit. in Strossen 1995: 40). 2.4.4 Alternatives According to anti-censorship feminists, considering erotica a real alternative to pornography is a view based on a conservative and romantic concept of female desires: women are restricted to bodily pleasures that can be integrated into romance and a social and emotional context. Anti-censorship feminists, however, claim that the concept of 11 A study by Larry Baron (1990) (see also chapter II) shows that gender equality, as a matter of fact, tends to be positively correlated with pornography sales. On the other hand, a study by Lawrence and Herold (1988) has revealed that in individuals, feminist orientation and liberal sexual attitudes do not correlate positively. A paradox to be solved by feminism? <?page no="45"?> The issue and the camps 45 erotica is often just used as a sublimation and ‘cleaning’ of ‘dirty’ sexual desires, in the sense of “What turns me on is ‘erotica’, what turns you on is ‘pornography’” (Strossen 1995: 18). The anti-censorship camp therefore sees room for alternatives within the pornographic discourse itself. Women may find pornography boring or even offensive but the reason for this does not lie in the discourse per se but rather in the fact that at present there is mainly pornography that caters to male fantasies and in the strong social restrictions on women enjoying pornographic representations (including restrictions by anti-porn feminists themselves). If we take a more relaxed attitude towards pornography, it will eventually broaden to free itself from the patriarchal limitations. These assumptions are supported by research. There is, for instance, evidence (cf. Ertel 1990: 434) that women do not want to be restricted to erotica and sexually tainted romantic novels. They want sexual representations that are as explicit as males’ pornography but integrate a female perspective. And a study by Cowan and Dunn (1994) has shown that women do not find the explicit and unmediated representation of sex offensive in pornography but rather the concrete degradation and objectification of women and the penis worship, i.e. women’s going overboard once they catch a glimpse of a penis or of semen. The need for a more positive approach to the potential of the pornographic discourse becomes particularly clear if we appreciate the idea that female sexuality is not a monolithic phenomenon but that there are various different sexualities with their different modes of expression. The group that most prominently shows this are lesbians. Their sexuality has been denied them more fiercely than any other women. So it is not surprising that their need for explicit representation of sexuality is also different. Anti-pornography feminists find it difficult to deal with lesbian pornography. On the one hand, lesbians are a group that is wholeheartedly subsumed under feminism, being women that have completely rejected patriarchy (some feminists even find lesbianism the protofeminist sexual orientation, cf. Mills/ White 1997). On the other hand, many lesbians do not see themselves as too high-mindedly intellectual to admit their bodily desires and their need for raunchy sexually explicit material. On the contrary, in their sex lives as well as in their sexually explicit material many even experiment with and exploit typical gender roles from patriarchal pornography (butch and femme , denoting a mock heterosexual role division). Many lesbians think that in a postmodern age we should become aware that there are differences even among women, and that these differences may be essential to female eroticism (cf. Dunn 1990, Bensinger 1992, Henderson 1992). <?page no="46"?> Theoretical Background 46 2.4.5 Positive effects/ Sex workers Trying to exploit the pornographic discourse to promote the sexual liberation of women is the major positive aspect for anti-censorship feminists. They also see some potential of pornography in the aspects mentioned by liberals (cf. 2.2.5 above). It can also provide information for women, e.g., to compare their partners with the protagonists in the films, pictures or texts (cf. Gardiner 1993, cit. in Strossen 1995: 165f.). Otherwise they are not very explicit on this topic. With regard to workers in the sex industry, pro-sex feminists agree with liberals that there should be less ethical pre-judgement on the industry, which will immensely help to make it more respectable, thereby improving the conditions for the workers. A group that is trying to achieve this goal is COYOTE (Call Off You Old Tired Ethics), founded in 1973 (cf. Strossen 1995: 115). 2.5 My own position I have been strongly influenced by the depth and profundity of feminist theories on pornography. Although inclined towards the anti-pornography camp, I also appreciate many of the objections of the anticensorship opposition. This feminist background has informed the focus of my analysis as well as my methods, as will be shown in the chapters to come. Just to add, I have no personal history of pornography consumption. The material I collected marked my first real contact with pornography. I have not developed a liking for it. This is not supposed to morally excel myself, but only to provide some background. 3 Summing up Although Christian moralists, liberals, anti-pornography feminists, and anti-censorship feminists differ in the theoretical sophistication and socio-political sensibilities with which they approach the topic of pornography, their perspectives primarily can be distinguished on the basis of their answers to the fundamental questions of the debate: <?page no="47"?> Summing up 47 • What kind of ideas does pornography contain and are they harmful? • Can the ideas contained in pornography influence consumers’ thinking and behaviour? Although the four positions at times appear to be written in stone, being based on ideological views of morality and the power of language, and although they seem so much opposed to each other, trying to find systematic evidence in order to answer these questions still is a worthwhile enterprise. The next chapter will look into how this has been done, reviewing the most important research into pornography. <?page no="49"?> II Researching Pornography: Sexology and Hermeneutics One can consult all the experts he [sic! ] chooses, can write reports, make studies, etc., but the fact that obscenity corrupts lies within the common sense, the reason, and the logic of every man [sic! ]. Charles H. Keating, Jr. (1991: 30) This statement shows that to some - particularly to conservatives - there does not seem to be a need to carry out research on pornography, as its harmfulness is obvious. It is intuitively given and we therefore just need to appeal to common sense (and possibly the Bible). More commonly, however, theorists and activists want to base their views on evidence from systematic studies, seeking external support for their positions. Different types of studies have been carried out for this purpose, yielding different results and thus inviting cheers from different camps. Researchers have, roughly speaking, drawn upon two research paradigms, namely sexology and hermeneutics. The two sections of the current chapter will deal with the different methods within these two traditions. I will briefly sum up the main features of these methods, reviewing the most important studies and evaluating their merits and shortcomings. I will conclude by pointing to the basic demands that still remain unfulfilled, particularly with regard to the two questions raised at the end of the last chapter. This will connect the current chapter to the following one, in which I will present the basic tenets of my own approach, viz. Critical Discourse Analysis. 1 Sexology Sexology is the pronouncedly scientific study of sexuality. It takes the ‘thing itself’, i.e. the biologically given sex drive, as its object and examines its manifestations in behaviour and physiology with the empirical and quantitative methods of sociology, psychology and medicine to find the ‘laws’ governing it. Sexology thus applies the principles of positivist empiricism and (sophisticated) behaviourism to the field of sexuality (cf. Gagnon/ Parker 1995a: 7, Bristow 1997: ch.. 1), conceiving of humans primarily as objects conforming to natural laws (cf. Hampshire 1978: 67, cit. in Wood/ Kroger 2000: xi). <?page no="50"?> Theoretical Background 50 Although there may have been predecessors in history (cf. King 1994, Crawford 1994, Porter 1994), mainstream sexology started at the turn of the century with Richard Freiherr von Krafft-Ebing’s Psychopathia Sexualis (1984 [1894]), followed by Havelock Ellis’ Studies in the Psychology of Sex (1936 [1905-1928]), and peaked after World War II with the Kinsey Reports (Kinsey et al. 1948, 1953) and Masters and Johnson’s Human Sexual Response (1966). The advent of sexology constituted a revolution since for the first time people sought for reliable knowledge in a field so much covered by shame, guilt and repression, knowledge that promised remedies for sex-related problems ranging from psychopathology to infertility (cf. Porter/ Teich 1994a: 4) and consequently led to a proliferation of how-to guides (cf. Bristow 1997: 12). In the age of structuralism and neo-positivism, the positive empirical sciences were to provide the means to gather this knowledge. Although alternative approaches to the study of sexuality have been abounding since the 1960s, sexology is still holding its place as the standard research paradigm as major textbooks in the field bear evidence to (cf. King 1998, Byer/ Shainberg/ Galliano 1999, Hyde/ DeLamater 1999; even though some of the works include feminist/ ethnic perspectives, they still are clearly sexological). It therefore does not come as a surprise that the main bulk of studies into pornography have been sexological in orientation. Such studies systematically search for empirically observable effects of porn consumption, trying to prove or disprove a more or less immediate causal link between the latter and sexual behaviour and attitudes, particularly of a clearly negative nature, e.g. sex crimes, misogynist aggression, etc. The most important research types are correlational studies, clinical studies and laboratory experiments (these categories - together with those of section 2 - are extensions of those given by Nemes (1992), who does, however, not make the distinction between sexological and hermeneutic research types). Generally speaking, the tenets of sexological research neatly fit into the liberals’ view of sexuality. Liberally-oriented studies have consequently usually been within this tradition. Conservatives and feminists, however, though generally sceptic of this kind of science, still embrace it if it corroborates their views. 1.1 Correlational studies In correlational studies, researchers look for quantifiable connections between demographic variables and the amount of pornography consumption. If the factors show a significant positive or negative correlation, conclusions are drawn about possible causal connections. <?page no="51"?> Sexology 51 Most correlational studies focus on connections between sex crime rates and pornography consumption. The most famous one is by Berl Kutchinsky (1970), who found a drop in the number of sex crimes in Denmark during a period of increased availability of pornography after legal deregulation. Similar patterns have been observed for other cultures, e.g. Japan (cf. Abramson/ Hayashi 1984, cit. in Hardy 1998: 30), the United States (cf. Gentry 1991), or Denmark, Sweden, Germany and the United States (cf. Kutchinsky 1991). Some studies have also found the inverse correlation, i.e. a rise in sex crime rates after the introduction of stricter regulation of pornography, for instance, in Singapore (cf. Donnerstein/ Linz/ Penrod 1987, cit. in Pally 1994: 60). Baron and Straus (1984) and Scott and Schwalm (1988a,b, all three sources cit. in Pally 1994: 55), on the other hand, found positive correlations between pornography sales or adult theatre rates and the number of rapes in American states. But the authors reject a causal link on the grounds that there are various other factors that these simple correlations miss out on (cf. also Baron/ Straus 1987, 1989, cit. in Pally 1994: 55). Gentry (1991) also argues that possible connections between rape rates and pornography sales are misleading because they usually are based on a third variable often neglected, namely the number of men aged 18-34 living in the area. As the majority of both rapists and pornography consumers come from this group, the likelihood of rape and pornography use increases with the size of the group mentioned, without them being causally related. Only a few correlational surveys go beyond the two-factor studies mentioned above. The most interesting one is by Baron (1990). He tested the hypothesis that pornography leads to social conditions unfavourable to women in the USA. He found that, contrary to his feminist-oriented hypothesis, there was a strong positive correlation between the sales of major pornographic magazines and a Gender Equality Index comprising indicators of the status of women relative to men (e.g. percentage of women senators or women blue-collar workers, equal pay laws, etc.). In other words, women and men are more likely to be treated as equals in those American states with higher pornography consumption. Baron, however, thinks that this is not a causal relation but rather a result of a third, underlying factor, namely political tolerance. Correlational surveys are important, being indicators of largescale relations and developments. With their strong emphasis of quantitative sociological research they also rest on a solid scientific foundation. Many aspects of correlational research, however, remain problematic. The main objection is that mere correlations cannot prove causality. <?page no="52"?> Theoretical Background 52 Often, there are underlying factors, as in the Baron study described above, which blur the picture. In addition, the demographic data drawn upon is not always as straightforwardly reliable as suggested. Figures of pornography consumption and sex crime rates, for instance, are based on often implicit and/ or arbitrary definitions of pornography and sex crime, which also has an effect on accounting procedures (cf. Court 1984, cit. in Hardy 1998: 29). The actual numbers of sex crimes, furthermore, cannot take into account those assaults that are not reported. Pornography may be considered instrumental in creating a liberal climate where more such incidents are reported or a repressive climate where fewer incidents become known, the victims’ credibility being undermined by the dissemination of rape myths. Despite these shortcomings, the uniformity of the results of correlational studies leaves a lot to explain to the anti-pornography camp. 1.2 Clinical studies Research in this category aims to study the eminence of pornography use in special populations such as convicted sex offenders including rapists. Some studies corroborate the view that convicted criminals and especially rapists have a history of porn consumption (cf., for instance, Caputi 1992: 205, interpreting FBI figures, cit. in McNair 1996: 65). Other studies, however, reveal that this group sees less pornography than other groups (cf. Goldstein 1973, Savitz/ Johnson 1976, both cit. in Nemes 1992: 470f., Becker/ Stein 1991, cit. in Pally 1994: 45) and that other aspects such as sexual/ physical violence, alcohol abuse, and involvement in a delinquent peer group may weigh more heavily as causal factors (cf. Ageton 1983, Becker/ Stein 1991, both cit. in Pally 1994: 45). Clinical studies are interesting because they focus on those populations directly involved in sex crimes, and they also usually combine qualitative in-depth interviews with statistical evaluation. Both positive and negative results nevertheless have to be taken with a grain of salt. As for positive results, convicted rapists may - perhaps as a result of therapy - be eager to find a cause for their crimes that lies outside their own responsibilities and may therefore choose pornography as a scapegoat (cf. Cameron/ Frazer 1992: 377, cit. in McNair 1996: 72). It also remains unclear why pornography makes certain men want to rape women while it does not have this effect in the majority of its viewers/ readers (cf. Dority 1991: 114). As for negative results, convicted rapists might, on the other hand, form only a segment of those who already have raped or those who are likely to sexually aggress and the real role of pornography can thus not be accounted for in clinical <?page no="53"?> Sexology 53 studies. Furthermore it is generally difficult to separate the various factors involved (cf. Surgeon General’s Workshop 1986, cit in Pally 1994: 29f.). Although there are psychological indexes available that measure male subjects’ proclivity to rape or sexually aggress (cf. Abel et al. 1977, cit. in Malamuth/ Check 1983: 55f.), no systematic study has, to my best knowledge, been conducted to find out about the porn consumption of those who rate high on such scales. Besides, particularly Neil Malamuth and his colleagues (cf. Malamuth/ Check 1983, Malamuth 1986, 1989, Malamuth et al. 1991) have developed sophisticated models of sociopsychological and intra-psychological causal factors predicting sexual aggression. To do justice to such models, clinical research would have to be considerably updated and fine-tuned and the methodology would have to be adapted to the more complex situations. The major problem of clinical research, which probably is a result of metatheoretical and methodological inconsistencies, are the contradictory findings, which have not provided consistent reliable answers to the question of harm. 1.3 Experimental studies The main body of research into the effects of pornography consists of socio-psychological experiments in a laboratory setting. In the paradigmatic procedure, subjects are exposed to pornographic material, either on one occasion or regularly over a longer time span, and then physiological/ arousal-related, attitudinal, and/ or behavioural effects are assessed. The group of subjects is usually carefully sampled according to certain criteria, e.g. class, age, gender, sexual preferences, or certain sexual features such as likelihood to sexually aggress or degree of sex guilt. The pornographic material presented sometimes is also categorized according to content criteria such as derogatory but non-violent, violent, or including paraphernalia. The focus on arousal-related effects constitutes the basic level of experimental research. In the simplest set-up, subjects are exposed to specific types of pornography in order to see which kind of sexually explicit material triggers arousal in which kind of people. The degree of arousal is either assessed by the subjects themselves or is physiologically measured (tumescence of the penis, secretions of the vagina). Much of the data remains heterogeneous, often yielding conflicting results. While Malamuth and Check’s study (1983), for instance, shows that men who consider themselves more likely to rape and are more susceptible to rape myths are more aroused by material featuring <?page no="54"?> Theoretical Background 54 non-consenting sex (cf. also Heilbrun/ Seif 1988 for similar results), Barbaree and Marshall’s (1991, cit. in Pally 1994: 44f.) reveals that depictions of forced sex decreased arousal in all men by 50%, a percentage at which penetration becomes impossible. Some surprising results have also been gained. It has, for instance, been found that in general females are not necessarily less aroused by pornography than men (cf. Ertel et al. 1990: 58). Attitudinal effects are changes in evaluations of sexuality in general, of sexual roles, of genders, of social institutions, etc. They are directly assessed by interviewing or questionnairing subjects, or indirectly, by having subjects, for instance, watch a rape trial and then pass a sentence over the defendant. The results of experiments of this kind suggest that exposure to pornography may modify views towards a more libertine perspective. Particularly the results of studies by Dolf Zillmann and Jennings Bryant (1982, 1984, 1988a,b) have shown that there is a decrease in respect for social institutions such as the family but also for traditional gender roles, a greater acceptance of promiscuity and extramarital sex and of more uncommon sexual practices with generally higher estimates of the commonness of practices such as oral or anal sex (to more accurate numbers, as is suggested by the General Surgeon’s Workshop 1986: 17, cit. in Pally 1994: 35), and a less severe condemnation of rape. The two researchers have also found that through habituation some people feel a desire for more ‘extreme’ forms of pornography. None of these results have gone uncontested: Donnerstein (1984), Krafka (1985), Linz, Donnerstein and Penrod (1988, all three cit. in Pally 1994: 41), Donnerstein, Linz, and Penrod (1987: 176, cit. in Donnerstein/ Linz 1988: 180), and Ertel et al. (1990: 310-329), in their studies and in their reviews of other studies, do not corroborate Zillmann and Bryant’s findings (for their counterargumentation, cf. Zillmann/ Bryant 1988c). Ertel et al. (1990: 281f.) think that the latter’s experiments are also problematic due to a neglect of initial factors (affective stimulation at the beginning of the experiment, general attitude toward sexually explicit material, initial sexual arousal, etc.). Studying behavioural effects means looking for more direct connections between exposure to pornography and concrete acts, usually of a violent nature. Behavioural effects are tested by, for example, having subjects, after exposure to pornography, administer electric shocks to somebody, in the belief that this is part of a different experiment. Sometimes further emotions are triggered in the subjects, e.g., by angering them or making them laugh. <?page no="55"?> Sexology 55 The first experiments of the kind described suggested that pornography reduced the threshold level to use electric shocks as a punitive measure for subjects (e.g. Donnerstein/ Berkowitz 1981, cit. in Fisher/ Grenier 1994: 24). But further research showed that the same effect can be reached by ‘working people up’ through other means than sexual arousal, e.g. jogging, aerobics (cf. Zillmann 1984, Donnerstein 1989, both cit. in Pally 1994: 42f.), or that, if the subjects are pleasantly surprised, they do not become more violent but more friendly (cf. Zillmann 1984, cit. in Pally 1994: 43, Tannenbaum, 1 in Donnerstein/ Linz/ Penrod 1987, cit. in Segal 1990: 33). In sum, this means that emotions already activated can be rendered more acute by increasing the heart rate and the blood circulation. The 1990s saw an increase of studies on sexual behaviour combining cognitive models of processing, in particular schema theory (see section III.1.1.1), with empirical testing. Geer and McGlone (1990) and Kirsch-Rosenkrantz and Geer (1991), for instance, tested gender differences in memory for a sexual text. They found that males, stereotypically having more knowledge of sex, incorrectly recalled more content of a sexual nature than females. Pryor and Stoller (1994) explored the cognitive processes in men with sexual harassment proclivities. These men recalled word pairs that associated sexuality with dominance better than other word pairs, which shows that these two aspects are strongly correlated in their thinking. In the most interesting experiment of this kind, McKenzie-Mohr and Zanna (1990) explored cognitive sexism. Cognitive sexism means that women are perceived by males not in terms of their achievements but primarily in terms of physical attractivity and sexual attributes (called a heterosexual schema). It was found that after viewing pornography men corresponding to the cultural stereotypes of masculinity paid more attention to, and remembered more of, a woman’s physical details than to what she had actually said, even in situations such as a job interview. Though none of the experiments described in this paragraph concretely examined the pornography-violence-relation, they still point to highly relevant dimensions that need to be addressed in further research. The major advantage of laboratory experiments is the tight control of possible variables. As these variables are usually both empirically observable and quantitatively measurable, this kind of research adheres to strict scientific principles. 1 Tannenbaum is subsumed under Donnerstein/ Linz/ Penrod 1987 without separate reference in the citation from Segal 1990. <?page no="56"?> Theoretical Background 56 Some critics, however, say that the very idea of testing the effects of pornography in a laboratory setting is seriously flawed. Firstly, experiments do not take into account that people assign different meanings to the same ‘observable’ phenomena. Physiological arousal will, for instance, not necessarily be interpreted as, and can consequently not simply be equated with, sexual arousal (cf. Hardy 1998: 34, Tisdale 1995: 203). Secondly, violent behaviour within the laboratory may also bear different meanings than in the real world so that electric shocks may only be a weak indication of sexual aggression (cf. Morris 1985: 23, Linz/ Donnerstein 1990: 37f., both cit. in Pally 1994: 43). Thirdly, if effects are found primarily for violent pornography, we can assign them to the violence rather than to the combination of violence and sexual explicitness, which might be justified in the light of studies yielding similar results but using non-sexual violent material (cf. Donnerstein/ Linz 1986: 59, cit. in Pally 1994: 42). Finally, there are factors that are problematic for any social psychological experiments: college students as the main subject group (cf. Brannigan/ Goldenberg 1987, cit in Linz/ Malamuth 1993: 36; Lab 1987, cit. in Johann/ Osanka 1989: 130ff.), the limited time scale (cf. Howitt/ Cumberbatch 1990: 84, cit. in Pally 1994: 44), and the artificiality of exposure to the pornographic material (it is always other-directed, which means subjects do not see/ read what they themselves would choose to see/ read, cf. Ertel et al. 1990: 308f.). As this brief review already indicates, there is a great variety of often contradictory results. The heterogeneity of results, with one study claiming something and the other study claiming the exact opposite, again reflects the problematic status of laboratory experiments. The only thing following from the experimental studies is that nothing has been proved beyond a reasonable doubt, no matter whether the ball is actually in the court of those attacking pornography or those defending it. I nevertheless think that experiments in the cognitive vein, described last above, could be promising. 2 Hermeneutics Hermeneutics is philosophically based on social constructionism. Social constructionism, already schematically described in the context of feminist positions in the last chapter, assumes that sexuality is not primarily an innate drive but rather the product of socially negotiated/ shared interpretations, classifications and evaluations of the world, which means that sexuality is actually not a biological given but a social construct. This becomes manifest in the historical and cultural diversity <?page no="57"?> Hermeneutics 57 of sexual practices and expressions, which cannot be accounted for in terms of biology (for social constructionism, cf. Burr 1995, Gergen 1999; for social constructionism in the field of sexuality, particularly its relation to sexology, cf. Porter/ Mikuláš 1994a, Gagnon/ Parker 1995a, Laumann/ Gagnon 1995, Bristow 1997). 2 As a consequence, hermeneutics is not interested in stimulusresponse pairings but, in its anti-behaviourist vein, in the interpretations, classifications and evaluations which necessarily mediate between stimuli and responses. As these socio-psychological processes are not immediately open to empirical observation, hermeneutic research methodologies focus on understanding rather than on proving (cf. Salamun 1986: ch. 3), conceiving of humans as thinking and meaningmaking social beings (cf. Hampshire 1978: 67, cit. in Wood/ Kroger 2000: xi). Hermeneutics was developed as an alternative in sexual studies (and in other social sciences, for that matter), responding to the inability of sexology to incorporate constructionist aspects into its epistemological framework. Its emergence is also a result of the critique launched against sexology by feminism, gay/ lesbian studies, anthropology, postmodern theory and the sociological approaches of Symbolic Interactionism, Ethnomethodology and the Sociology of Knowledge (cf. Stanton 1992a: 16, Gagnon/ Parker 1995a: 7f.). Today hermeneutics has become established and stands alongside sexology, though perhaps lacking the latter’s prominence. As it is being more widely acknowledged that pornography is a discourse and thus is concerned with language and interpretations, an increasing number of studies on the effects of porn are hermeneutic in nature. The most important research methodologies used in these studies are evaluative surveys, testimonial studies (ethnographic interviews), and content analyses. With its social constructionist foundation hermeneutic research shows a connection to feminism. The main bulk of studies within this research paradigm have therefore been feminist in orientation, usually also with the anti-pornography slant. Few hermeneutic studies can clearly be assigned to the liberal or the conservative camp. 2 The study of sexuality from a constructionist angle has been taken up by linguistics/ literary studies, e.g. Cameron 1990a, 1992a, Plummer 1995, sociology, e.g. Gagnon/ Simon 1973, Gagnon/ Parker 1995a, Laumann/ Gagnon 1995, history/ philosophy, e.g. Foucault 1992 [1976], 1993a,b [both 1984], and anthropology, e.g. Caplan 1989a: 1ff. <?page no="58"?> Theoretical Background 58 2.1 Evaluative surveys Evaluative surveys are opinion polls. In personal interviews or questionnaires people answer questions about their own opinions on various aspects of pornography. The surveys usually combine questions concerning definitions (Is X pornographic? , Which of the following features are pornographic? ), evaluations of effects (Do you think that pornography does X? ), and opinions on regulations (Should pornography be banned? ). Some opinion polls have been conducted by women’s magazines: (female) readers completed questionnaires and sent them to the magazine. In the three surveys that Itzin and Sweet (1992) review (Women’s Day, Company, Cosmopolitan), the most striking results were that a large majority saw a connection between pornography and violence and, probably as a consequence, favoured strict restrictions on, if not a complete ban of, pornographic material. Although due to selfselection of participants these polls are far from representative, it is still interesting to see that women with anti-pornography attitudes felt most called upon to participate. Most evaluative surveys, however, aim at statistical representativeness and therefore use sampling in selecting subjects. Despite this more scientific approach, there are very few consistent findings. While in some surveys less than 50% of the people questioned (Germans) consider pornography to cause derogatory views of women or violence and a majority find pornography very interesting and attractive rather than disgusting (cf. Ertel et al. 1990: ch. 4.2), the majority of subjects in other studies (usually Americans) believe that pornography can lead men to commit rape, that it is at the root of a breakdown of public morality, that its presence harms children, and that it is not simply entertainment (cf. Toolin 1983, Johann/ Osanka 1989: 17f., reviewing an 1986 Washington Post poll, Russell 1998: 157-162, reviewing several polls). Russell (1998: 162-166) also found that the ‘haves’ (male, white, high income, good education, etc.) take a less critical stance towards pornography than the ‘have-nots’. All opinion polls that take the gender of the respondents into account have found that women are in general more negative about pornography, i.e. their dislike is greater, they assign more negative effects to it, and they are in favour of stricter regulations. Evaluative surveys, regarding pornography as a social phenomenon, study the social perception of it rather than possible mechanistic effects. They thus start exactly at the interpretative dimension and are therefore a highly relevant undertaking. <?page no="59"?> Hermeneutics 59 The major shortcoming of evaluative surveys, however, is that they cannot really reveal anything about the harmfulness of pornography. Strictly speaking, they do not examine pornography but people’s attitudes toward it and (possibly common sense) views on its effects. The results of evaluative surveys - even if flawed from a scientific point of view - still suggest that there is a critical attitude toward pornography in the population. 2.2 Testimonial studies Testimonial evidence comes from interviews with people involved in pornography, whether as consumers, relatives and/ or victims of consumers, or as workers in the sex industry. The focus in this qualitative research is not so much on proving that pornography causes harm but rather on understanding the individual’s lived experience with pornography in order to see whether certain patterns can be derived (cf. Jensen 1995, 1998b, Hardy 1998: 36,). Many feminists favour such studies as, in the tradition of ethnography, ‘listening to stories’ implies that I can aim at understanding the phenomenon as a whole without merely quantifying and thus somehow devaluing the individual’s experience, particularly the experience of women hurt in the real world (cf. Cole 1995: 75). The studies therefore often present the actual voices, including lengthy quotations, in addition to or instead of quantitative patterns. Many testimonial studies have been conducted including or exclusively focusing on female sexual assault survivors, e.g. Kelly (1988), Russell (1980) (both cit. in Jensen 1998b: 110-112). Many of the victims mentioned pornography as part of a continuum of violence, even if not particularly asked about it. An interview project with street prostitutes (cf. Silbert/ Pines 1984; cit in Jensen 1998b: 109f.) also showed that pornography had sometimes played a part in rapes, which many of them had experienced. Other testimonial studies concentrated on male consumers. Jensen (cf. 1998b: 119-134) conducted interviews with sexual perpetrators and ‘ordinary’ men. The study reveals how pornography was not only instrumental in shaping a stereotypically male sexuality but also how interviewees from both groups had used it to break down resistance, and how they sometimes had difficulties distinguishing between fiction and reality. Hardy (1998: ch. 5), in interviews with pornography users, found that they usually went from a more addictive phase of use to a reconciled stage, where they rarely used it at all because of the commitments in steady relationships. A further result of the study was that the <?page no="60"?> Theoretical Background 60 interviewees were aware of the sexist content of pornography but did not necessarily accept it. Some studies also concentrate on people directly involved in the pornography industry, e.g. Stoller (1991), who has found that most of these people have deliberately chosen this path, even if this might have been due to sex frenzy/ hysteria. Some autobiographical accounts of sex workers, 3 though not actually intended as research, can also be subsumed under this category here because even if they just represent one voice, they still provide deep insights as a result of the detailed depictions. Linda ‘Lovelace’ Marchiano, the main actress in the first porn cult film Deep Throat, in her autobiography Ordeal (1980) describes the hardships and tortures she had to go through when shooting the film. Evelina Giobbe (1995), a prostitution survivor and founding member of WHISPER (Women Hurt in Systems of Prostitution Engaged in Revolt), tells the story of how she was forced into prostitution and pornography: “Raised in New York City brothels by pimps and johns, pornography became my family photo album” (Giobbe 1995: 315). 4 On the other hand, there are also more favourable accounts by sex workers (e.g. Nina Hartley n.d.) and also by women who have become directors/ producers of pornography (e.g. Ann Sprinkle 1998 and Candida Royalle 2000). The main asset of testimonial studies is that they do not pretend that social processes can be reduced to a set of factors. The understanding is therefore more important than the explaining. This also implies a strong sensitivity to and a respect for the experience of those researched. Listening to stories further means that researchers focus on a group directly involved in or affected by pornography. One problem of testimonial evidence, however, is the non-representativeness of the groups of interviewees. This may be less important if the focus is really restricted to individuals. But there usually is an implicit appeal to common patterns, which can be seen from the very fact that some of the studies still give statistical relations. To find patterns also runs counter to the intention to concentrate on individuals. Another problem is that anecdotal studies help to understand the role of pornography in the lives of individuals but cannot scientifically 3 There are also such personal stories by pornography consumers, e.g. Jensen (1998b: 134-145), MacDonald (1995). 4 The personal accounts are not restricted to sex workers. Ann Russo (1998b), for instance, tells her own story, combining a personal narrative of sexual victimization and stigmatization and the attempt to come to terms with it by ‘digesting’ the theories on pornography and by entering the debate herself. <?page no="61"?> Hermeneutics 61 prove the harmfulness. This is also a result of not always focusing on the immediate effects of pornography but on processes that are only indirectly related to it (e.g. when interviewing victims of pornography users we, for instance, deal with their interpretations of the assaulters’ interpretations). To be fair, to understand the workings of pornography on a primary level is not always the main intention of these studies. Summarizing the results of testimonial studies is difficult as the very nature of these studies defies generalizations. But most of them rather support the view that pornography may be involved in sexual aggression and crime and thus should be taken seriously by the other side in the debate. 2.3 Content analysis In content analyses, researchers, roughly speaking, search a collection of material to see how frequently certain content categories occur. The categories that are normally of interest with respect to pornography are types of acts/ practices, e.g. sex acts generally, fellatio, masturbation, verbal or physical aggression, or rape, and types of participants, e.g. professional males, housewives, female students, or gays/ lesbians. A major focus in content analyses has been the amount of violence in pornography. No matter whether pictorials and/ or cartoons in major magazines such as Playboy and Penthouse (cf. Malamuth/ Spinner 1980, Scott/ Cuvelier 1987, 1993, all three cit. in Russell 1998: 13-15, Matacin/ Burger 1987), cover images of magazines (cf. Dietz/ Sears 1987/ 88, cit. in Russell 1998: 17-21), or sex scenes in videos (cf. Ertel et al. 1990, Duncan 1991, Cowan/ Lee/ Levy/ Snyder 1988, cit. in Russell 1998: 21-24) were analysed, most studies found violence not a dominant but still a fairly common subject of pornography (for an opposing view, cf. Thompson 1994, cit. in McNair 1996: 83). There are, however, contradictory results concerning pornography’s development with regard to violence. Brown and Bryant in their review of content analyses (1989: 18f., cit. in Jensen/ Dines 1998: 69) say that there has been an increase of violence, while Pally (1993: 5, cit. in McNair 1996: 67) in hers claims that, striving more and more to appeal to couples and women in particular, pornography has become increasingly non-violent (for a synopsis of contradicting results, cf. also Fisher/ Grenier 1994: 23ff.). Some content analyses further focus on stereotypes implicated in pornography. Despite claims that it presents a great variety of different and new sexual practices, Jensen and Dines (1998) have found that intercourse usually follows a rigid pattern in pornographic videos and novels: cunnilingus and fellatio quickly lead to vaginal penetration, the <?page no="62"?> Theoretical Background 62 end of which is marked by ejaculation (in filmic representations, this is called the money-shot; it always occurs outside the female body, cf. Linda Williams 1999). Intercourse is framed by the rise and fall of the penis, thus reflecting a male-dominated form of sexuality. Pornography also features relatively fixed gender roles in sexuality. Both Jensen and Dines’ (1998) and Hardy’s (1998) studies, for instance, reveal that women fall into one of two categories: nymphomaniacs, who go over the top for sex right from the beginning, or hesitant prudes/ uninitiated youths, whose initial resistance has to be overcome. The result is the same: both types of women end up being undiscerning in their sexual responses and men bring them pleasure and satisfaction, no matter what they do. Women are also portrayed as more submissive and less dominant than men. This is often also reflected in the different professional statuses of males and females: while the former appear as doctors, teachers, employers or supervisors, the latter are secretaries, students, or nurses (or housewives, I may add) (cf. Jensen/ Dines 1998). Pornography does not only stereotype women but also people of colour. Studies concentrating on racial stereotypes (cf. Cowan/ Campbell 1994, cit. in Russell 1998: 24-26) show that certain gender stereotypes are further enhanced if race comes into play. Black men are thus more dominant - the anatomical equivalents of their power are their particularly long penises - while black women are more submissive. This becomes particularly conspicuous as people of colour usually have sex with white people. So not only women are generally unfavourably represented but also non-dominant ethnicities. Content analyses are highly relevant because unlike the other research methods, they start at pornography itself, i.e. the representations, and do not treat it as a given as approaches focusing on possible effects often do. It should be noted in this context that most experimental studies draw upon content analysis if they focus on effects of particular features of content on viewers/ readers, e.g. homosexuality, bestiality, S/ M, degradation (for research involving reactions to a range of topics, cf., e.g., Zillmann/ Bryant/ Carveth 1981 or Cowan 1990, both cit. in Linz/ Malamuth 1993: 20 and 55,). But such studies again concentrate on reactions and the content analysis is just a necessary prerequisite. On the negative side, content analyses, however, focus too much on the representations per se, i.e. the pictures and/ or texts, and less on the interpretative processes. They hardly ever consider, for instance, that the researchers’ content categories might not feature in the consumers’ interpretations and thus might not have great effects. There may also be studies that combine content analyses with interviews with consumers <?page no="63"?> Hermeneutics 63 (cf. Hardy 1998), but even here no real attempt is made to match content categories with interpretative processes so that the two aspects remain largely unrelated. In addition, some categories are difficult to quantify, how, e.g., do we count violence (cf. Hardy 1998: 29)? Content analyses always start at the macro-level, namely content categories, and not at the micro-level of language (though the methodology is being sophisticated, cf. Riffe/ Lacy/ Fico 1998). Macrocategories necessarily involve an a priori interpretation by the analyst (cf. What constitutes a sex scene? What is foreplay, what masturbation? - such aspects do not automatically spring from the text), an aspect that reduces the relevance of such research. Despite these limitations, we can note a certain consistency in the findings of content analyses: women (together with other disadvantaged social groups) are presented in a more unfavourable light than men. Although with regard to the amount of violence there are no clear answers, content analyses thus usually support the anti-pornography feminist positions. 3 Summing up It seems obvious that the two questions that emerged from the discussion of the different positions in the pornography debate, namely whether pornography contains problematic ideas and whether these become part of consumers’ thinking and acting, have not really been satisfactorily tackled let alone answered by the types of studies under review. Let me briefly focus on these two questions again and assign them a place in the research designs. Schematically, the working of pornography in the context of sexology and hermeneutics could be represented by the figure below (see next page). The latter represents sexology as taking a behaviourist point of departure, concentrating only on the inputoutput relation, but neglecting the black/ grey box: it neither is (really) interested in pornography itself nor in the actual interpretative processes but just in its quasi immediately and directly produced and caused effects on consumers’ beliefs, behaviours and attitudes. Sexology, as a consequence, never actually gets to the two questions mentioned. It is, however, very systematic and consistent in the scientific procedures used to study input-output relations. Methodologically speaking, sexological studies of pornography thus score high on reliability but low on validity. Hermeneutics, on the other hand, regards the human practice of assigning meaning as its real object matter, thus focusing primarily on the grey area in the figure above. As this is at the core of pornography <?page no="64"?> Theoretical Background 64 consumption, it provides a promising foundation. Unfortunately, hermeneutic approaches have not made great progress in answering the two questions either, mainly for three reasons. Firstly, some of the hermeneutic approaches do not concentrate on the interpretations of porn itself but on secondary and tertiary aspects (e.g. the evaluations of it by non-consumers, or the lived experience of consumers’ relatives). These are interesting undertakings but do not really affect the core of the harm issue. Secondly, those approaches that focus on the grey area usually investigate one of the questions at the cost of neglecting the other: content analyses say very little about interpretations, interviews with users treat the texts as given. This may also be a result of the strict assignment of the two questions to either of the two dimensions (text or interpretation). Thirdly, and most seriously perhaps, hermeneutics has not managed to develop systematic empirical, intersubjective and explicit methodologies that would make its results more valid, reliable, and comparable to other results. Besides, studies in the hermeneutic vein have usually skipped the question of what constitutes problematic ideas, thus neglecting the socio-political dimension of the research. Another problem is that both paradigms, as shown, are ideologically biased, which severely complicates the communication and the exchange across theoretical and methodological borders. This, in turn, makes a real comparison between studies impossible. Sexology Input Pornography → What kind of ideas are contained in pornography and are they harmful? → Do the ideas contained in pornography influence consumers’ thinking and behaviour? → Output Sexual beliefs, attitudes, behaviour Text Interpretation Hermeneutics Figure 1: Schematic representation of the effects of pornography. To conclude, my review of studies into pornography points to the need of hermeneutically oriented research which focuses on pornographic texts and their interpretations and thus on the two core questions in the <?page no="65"?> Summing up 65 debate and which does not renounce the scientificness exemplified by sexology so that results can be communicated and thus also compared across the ‘great divide’. My version of Critical Discourse Analysis may fulfil these demands, as the introduction of its basic tenets in the following chapter is supposed to show. <?page no="67"?> III Researching Pornography: Critical Discourse Analysis Critical Discourse Analysis can serve as an alternative to the approaches in the study of pornography described in the preceding chapter. This objective requires concentrating on texts and people’s interpretations, 1 not neglecting the socio-political dimension of these, and at the same time adhering to empirical, intersubjective and systematic principles of research. In other words, it requires marrying hermeneutic theory with the scientific rigour of sexology. Outlining its basic theoretical and methodological tenets, I will show that Critical Discourse Analysis can meet both requirements and thus is not only a viable but also a highly valuable supplement to other approaches in pornography research. The structure of the present chapter corresponds to the two requirements. The first two subchapters are dedicated to the hermeneutic background, explaining in more detail the theoretical ideas that make CDA suitable for the study of pornography, with the first one taking a discourse-oriented and the second one a socio-political perspective. The third subchapter focuses on scientificness, delineating the methodological tenets of research in CDA. The concrete details of their implementation in the actual study to be undertaken will be discussed in the next chapter. As I do not wholeheartedly subscribe to any of the main frameworks available and as the study of pornography requires adaptations, I will be developing my own version of CDA in this book. I am nevertheless still heavily indebted to those who have been working in the field for years and whose writings have greatly influenced and inspired me. Critical Discourse Analysis emerged at the end of the 1970s as the study of the political and social messages implicated in the specific ways of using language in particular situations (cf. Fowler/ Trew/ Kress/ Hodge 1979, Hodge/ Kress 1979 (2 nd edition 1993)). It was, however, only at the beginning of the 1990s - marked by the publication of Norman Fairclough’s book Language and Power (1989) and the launching of the 1 I will use the term interpretation as covering both productive and receptive discourse processes as well as their results. As the main focus of my research is on the receptive dimension, this one-sided extension seems legitimate. <?page no="68"?> Theoretical Background 68 journal Discourse and Society (1990) - that Critical Discourse Analysis started booming. It has since then become a major linguistic discipline, with Norman Fairclough (1992a, 1995), Roger Fowler (1991, 1996), Gunter Kress, Bob Hodge (Kress/ Hodge 1993), Teun van Dijk (1987, 1988, 1993, 1997c,d, 1998), and Ruth Wodak (1996, et al. 1990, with Matouschek 1993, et al. 1998) as its main proponents. All these linguists represent a broad spectrum of perspectives: while the early proponents (then, and sometimes even today, called Critical Linguists) clearly start at language proper, Teun van Dijk takes a socio-cognitive point of departure, and Ruth Wodak’s and Norman Fairclough’s approaches are strongly based on social, historical, political and cultural theories. These differences on the theoretical plane often also entail differences in methodology. For an overview of the different approaches within CDA, see Fairclough and Wodak (1997). 2 1 Hermeneutic theory I: What is discourse analytic about Critical Discourse Analysis? Critical Discourse Analysis combines two research procedures, viz.: • Discourse analysis: A linguistic analysis of discourse with the basic research question: What kind of ideas influence people’s thinking and behaviour through the meaningful use of particular linguistic forms? • Critical evaluation: A socio-political contextualization of the linguistic analysis with the basic research question: Are these ideas problematic and harmful? The present subchapter and the next one are dedicated to these two areas, respectively. They both discuss theoretical backgrounds first and then examine how the latter suggest a certain approach to research. Although the focus will be general throughout, I will always conclude by pointing to the particular implications for a Critical Discourse Analysis of pornography. The research questions, though similar to the two that have emerged in the pornography debate and particularly in the hermeneutic 2 Some approaches are either similar in conception to CDA, e.g. Conversation Analysis (cf. Hutchby/ Wooffitt 1998) and Discourse Psychology (cf. Potter/ Wetherell 1987), or incorporate CDA as a methodological orientation, e.g. Ecolinguistics (cf. Fill 1996a). <?page no="69"?> What is discourse analytic about Critical Discourse Analysis? 69 research into pornography (whether pornography contains harmful beliefs and attitudes and whether they are conveyed through language), still divide the problem slightly differently. The reason for this is that CDA thinks that discourses cannot be said to contain ideas that they do not convey. The present subchapter will examine in some depth in what sense Critical Discourse Analysis is a form of linguistic analysis. Before dealing with more research-oriented issues I will discuss why the analysis of language could be of relevance at all and where it would have to start. The theory that provides convincing answers in this respect is social constructionism. 1.1 Social constructionism I have already mentioned social constructionism several times above. But here, I want to present an in-depth discussion of this position. Social constructionism theorizes the connections between language, thinking, acting, society, and reality. It holds that cognition, i.e. whatever is going on in our minds, does not reflect reality but, on the contrary, that reality is a projection or a construction of cognition. As cognition and particularly knowledge are not private but socially shared among members of a society, constructions also have a social side. Reality is thus a social and a cognitive phenomenon. I will be concerned with these two aspects separately in the following. Social constructionism has been developed in the sociological schools of Symbolic Interactionism (cf. Treibel 2000: 111ff.), Ethnomethodology (cf. Treibel 2000: 138ff.) and the Sociology of Knowledge (cf. Curtis 1972, Dant 1991) and it received its first book-length analysis in Berger and Luckman’s classic The Social Construction of Reality (1971 [1966]). It has also been treated from a more cognitive perspective in Social Representations Theory (cf. Farr/ Moscovici 1984, Augoustinos/ Walker 1995: ch. 6). Like these approaches, CDA treats the theory as its basis, arguing for its plausibility without, however, putting it to the test itself, at least not on a primary level. Social constructionism is nevertheless essential for hermeneutic research as it provides a starting point and in its broad outline of the theoretical foundation points to the areas where the actual analysis has to begin. As has been shown in chapter I, the anti-pornography feminist arguments, presupposing a tight relation between language, thinking, and acting, are hermeneutic in nature. It might therefore be objected that to start from hermeneutics means that I am deeply entrenched in antipornography feminism even at the outset and cannot help coming to the <?page no="70"?> Theoretical Background 70 same conclusions. This, however, is an ill-founded argument since assuming that there is a connection between language and constructions of reality does not automatically reveal what these versions of reality look like. A fine-grained and systematic analysis of real data may thus yield results that do not support assumptions about the harmfulness of pornography. To completely reject the idea of a link between language and beliefs about reality, however, precludes the possibility of research and not even sexological studies have therefore seriously denied it. 1.1.1 Cognition What does it mean for cognition to project or construct reality? This section will deal with this question, dissecting the concept of knowledge as the most important area of cognition involved and its function in the perception of reality and in acting, also touching upon the issue of the creation and maintenance of knowledge. Knowledge is used in a broad sense here and encompasses beliefs and ideas (= knowledge of the world; this also includes identities as knowledge about the self) and attitudes and values (= knowledge of evaluations or evaluative elements of knowledge). Knowledge does not entail epistemic certainty and is thus not distinct from believing or assuming. (Potential differences between ideas and beliefs and between attitudes and values will not be further explored in this study.) In cognitive psychology, knowledge is representations of information in long-term memory. It is divided into episodic knowledge (or episodic memory), which contains information about concrete events and situations in our lives (including the representations of the meanings of individual texts), and semantic knowledge (or semantic memory), which comprises general information about the world (cf. Stillings et al. 1995: 112f.). So while we store information about our ‘first time’ in the former, information about the general structure of sexual intercourse is captured by the latter. The knowledge that constitutes sexuality and similar phenomena is general in nature and I will therefore focus on semantic knowledge in the following. General knowledge per definition is not knowledge of individual items but about classes of these items. Semantic knowledge thus requires that we divide up the continuum of (our experience of) reality into a limited number of categories of persons, things, attributes, and events. 3 This process of categorization is based on the fundamental cognitive 3 To be precise, we divide reality into the global categories of persons, things, attributes and events in a first step. <?page no="71"?> What is discourse analytic about Critical Discourse Analysis? 71 ability to find similarities between aspects of the world and to assign them to classes mentally. These classes are often called concepts (cf. Van Dijk/ Kintsch 1983: 127, Schwarz 1992: 83ff., Smith 1995). Categorization is the mental operation of organizing data from our sense organs or from internal cognitive processes. Although there probably are universal cognitive constraints on categorization, we are still left with an almost infinite number of mental categories to which we could assign (aspects of) the same experience. This entails that the categories do not spring from whatever may be out there but that they are cognitive in a primary sense and are thus independent from outside reality. This becomes manifest in the cultural and historical variability of concepts: different cultures - and even different social groups within a culture - at different times divide up the continuum of reality quite differently (cf. Van Dijk/ Kintsch 1983: 127, Wierzbicka 1992, Ungerer/ Schmid 1996: 50). Categorizing processes usually run automatically, based on routine, so that we hardly ever are consciously aware of them. In everyday life, we therefore retain the belief in the givenness of reality, as this saves time and energy and normally does not conflict with our goals. Concepts are represented as schemata in the mind (cf. Bartlett 1932, Rumelhart 1980). 4,5 A schema is a network of information on types of persons, things, attributes, or events. Some of the information is fixed: sexual intercourse, for instance, obligatorily requires participants who engage in it. Much of the information, however, contains variables. Number and sex of participants in intercourse, for instance, are variables: there might be two or more participants, they might be of the opposite or the same sex. Most of these variables, however, have default values, which we assume in the absence of contradicting evidence. So we will think that, prototypically, the number of participants in intercourse is two, and they are male and female. The information contained in a schema also differs with respect to its centrality. Some aspects will be understood as more characteristic and distinctive, others 4 Bartlett and Rumelhart are the two classics - one old, one modern - of schema theory. The following, however, have more crucially informed my notion of schema: Van Dijk/ Kintsch 1983, Baddeley 1990, Schwarz 1992, Stillings et al. 1995, Semino 1997, Carrol 1999, Harley 2008. 5 The distinction between concept and schema is not consistently made (cf. Howard 1992 for an attempt). For me, there is a slight difference in focus: schema refers more to the organization of information in the mind, while concept refers more to the process of dividing up reality into categories. In many contexts, however, the two terms can be used interchangeably. <?page no="72"?> Theoretical Background 72 as more marginal. The function of a condom, for instance, will always be regarded as more essential than its colour (this has been captured in prototype theory, cf. Ungerer/ Schmid 1996: ch. 1). Most of the information in schemata is declarative in nature. Although there has been little research into this aspect, it is, however, assumed that schemata may also contain evaluative and attitudinal elements (cf. Augoustinos/ Walker 1995: 26, 48). Activating a schema in semantic memory may also be accompanied by affective responses so that there could be a tight connection between cognition and emotion (cf. Miall 1989, Semino 1997: ch. 6.6). Schemata never stand on their own. They are differently linked to other schemata and also become part of more inclusive ones. The schematic structure (or the conceptualization) of a field is how information concerning this field is organized into different schemata and how they are further integrated into larger patterns (cf. Denhière/ Baudet 1989, Baddely 1990: 337; cf. also the similar concept of semantic/ propositional networks, Anderson 1995: 148, Stillings et al. 1995: 28f.). The schematic structure of our memory also constrains our associations or our ability to retrieve information. If one unit in this network is stimulated, activation spreads along the connections to other units (spreading activation) (cf. Stillings et al. 1995: 30, Beaugrande/ Dressler 1981: 88f.). The activated elements are then more likely to be part of the associations of the element originally activated, and they are also easier to access and retrieve. Semantic knowledge is essential in our orientation in and towards the world as it is not only a passive reservoir of information but is used actively to make sense of reality. We usually attempt to match as much incoming data as possible to the schemata in our minds. In other words, we tend to perceive the world in terms of schemata, assigning concrete tokens (or part of tokens) to abstract concepts. Only by virtue of this process can these tokens and thus our experience of external reality acquire meaning - in a very general sense - for us. As we are able, for instance, to match the perception of a translucent latex object with our schema of a condom, we can assign the meaning ‘condom’ to the said object. The meanings thus do not reside in the outside world, but we assign them to it with the help of our mental schemata, thus actively constructing reality as composed of concrete realizations of abstract concepts or, put differently, perceiving reality as meaningful. Constructionism is thus a theory of how we perceive reality, i.e. an epistemological theory, and not a theory of what actually exists, i.e. an ontological theory (cf. van Dijk 1998: 25, Torfing 1999: 94). So if we do not have the concept of condom, we still see the object, it just is not a condom for us. <?page no="73"?> What is discourse analytic about Critical Discourse Analysis? 73 The main advantage of interpreting the world in terms of schemata is that they allow us to infer information that is not immediately accessible (cf. Stillings et al. 1995: 87). We may infer, for instance, the function from the appearance of a thing: a small translucent latex object, if matched with our schema of condom, will give rise to the assumption that it is instrumental in avoiding unwanted pregnancy and sexually transmitted diseases. Inferences help us to connect past with present events (= backward/ bridging inferences), e.g. we understand the connections between the wrong handling of condoms and pregnancy, and they enable us to make predictions about the future (= forward/ elaborative inferences), e.g. we might expect that the latex object will be used in a certain way if it is unpacked by two people who have just undressed (cf. Singer/ Ferreira 1983: 437, Singer 1990: 175). Inferences may not always be warranted by reality - after all, the object might have been a balloon and not a condom - and may lead to distorted interpretations (cf. Schacter 1989: 693, often as a result of different cultural backgrounds and concomitant differences in schemata, cf. Steffensen/ Joag-Dev 1984: 53-60). But without the ability to detect similarities and to categorize, we would have to treat any experience as new and, as a consequence, we could not make sense of the world. The active interpretation of the world with the help of schemata thus allows us to bring our past experience to bear on novel situations. The ability to use semantic knowledge is also an essential prerequisite to acting. Action is goal-oriented activity for the preservation, change or development of material or social reality. It is only possible if we are able to categorize situations (and their elements) and speculate on how we can manipulate them to achieve an intended goal. Action is impossible if we cannot draw inferences concerning the nature of the situation and concerning the nature of our manipulating activity on the basis of the schemata in our minds. If we see the small latex object without assigning the meaning of condom to it, we would probably not take chances to use it in sexual practices. In everyday life we, of course, can typify a situation at a glance to immediately infer what actions it requires or permits. But we have to bear in mind that this is only possible by virtue of the schematic structure of our mind and our ability to use it in our top-down interpretation of the world (cf. Stillings et al. 1995: 87). Schemata are not fixed entities but are in a continuous state of fluctuation. As they are constantly matched against new data, they and their internal structures, i.e. constants, variables, default values, central and marginal elements, as well as their external connections to other schemata may be subject to reconfirmation or change. The processes <?page no="74"?> Theoretical Background 74 probably run automatically and unconsciously. The extent to which new information can affect semantic knowledge depends on: • how often it occurs (frequently-encountered information > rarelyfound information); 6 • how much variation it displays (information that occurs in different shapes and contexts without essentially changing in content > information that always comes in the same form); • how densely it is interconnected with other elements (information featuring numerous and tight links with other information > isolated information); • how easily we can integrate it into existing knowledge (conforming information > contradicting information); • the status or credibility of the source (information from a trustworthy source > information from a doubtful source). (Enhanced version of principles presented in Marko 1995) The marginality of information may also be relevant in the transition, which means the less foregrounded data is, the more easily it can be accepted into semantic knowledge (cf. Marko 1995). But this feature may be more doubtful since it is far from clear what marginality actually means. Generally speaking, however, little is known about the emergence and change of schemata and the exact nature of the influence of confirming and disconfirming information on semantic memory (for various models, cf. Augoustinos/ Walker 1995: 52-55). To sum up, semantic knowledge consists of a network of schemata. We impose this structure on incoming data on all levels of perception and interpretation. As schemata do not reside in external reality but just in our heads, they are constitutive of reality, or, in other words, they construct (a meaningful) reality for us. 1.1.2 Society Knowledge in the shape of schematic structure divides up the world into categories and thus allows us to assign meaning to reality. In this sense, reality is cognitively constructed. In this section, I will now focus on the social side of the construction process, demonstrating that knowledge, though inhering in the minds of individuals, is still social in nature. Theoretically speaking, we are independent individuals and we could 6 > means ‘is more likely to be accepted into semantic memory’. <?page no="75"?> What is discourse analytic about Critical Discourse Analysis? 75 therefore categorize the world idiosyncratically as long as our own schemata allow us to act purposefully. But as a matter of fact, we do not develop our own schemata but instead strive to share them with other people. Why and how we do this will be discussed in the following. 7 There are two main reasons why we would share knowledge. Firstly, we are social beings, deriving our identities from membership in different social groups. These groups are mainly - though not exclusively - characterized by their shared beliefs and attitudes, i.e. by their shared knowledge. Belonging to a group hence means adapting our schemata to those of the group. Secondly, as social beings we also act upon, or together with, other humans, who act and behave back, so to speak. Social action is therefore often regarded as inter-action, requiring the co-ordination of individual acts and goals (cf. van Dijk 1997b: 10). In social action, we therefore not only have to predict reactions in/ by the material world but also reactions in/ by the social world. This means we have to assume that we interpret the situation similarly to our interactants, which, in turn, presupposes that we share - and know that we share - the schemata for the situation and for its component acts, roles, and entities, or, more generally, that we share knowledge of the (social) world. The sharing makes it easier to interact with each other without lengthy negotiations, thus also reducing insecurity. An aside: such schemata of situations and their concomitant social acts or practices (i.e. types of acts) - the paradigmatic case being the visit to a restaurant - are often called scripts (cf. Schank/ Abelson 1977). Sexuality primarily consists of social practices and it is therefore sometimes said to be scripted. Social constructionist approaches to sexuality in sociology accordingly often take scripts as their point of departure (cf. Gagnon/ Simon 1973, Gagnon/ Laumann 1995a). I will nevertheless stick to the concept of schema as it appears to be wider, not only covering aspects of social practices but also of objects, persons, attributes, etc., all of which are relevant in (sexual) knowledge. Sharing conceptualizations of the world is essential for social interaction to work. But how is this sharing achieved? We can only learn whether we share conceptualizations with other people by observing their behaviour in interaction and infer from it what the background knowledge they are drawing upon consists of. Social interactions 7 Strictly speaking, there are two social dimensions in the construction process, namely the social purpose (= Why? ) and the social nature of the genesis (= How? ) of knowledge. As they mutually implicate each other in parts, I will treat them together in this section. <?page no="76"?> Theoretical Background 76 therefore do not only serve their immediate purposes but also provide ample opportunity to compare other people’s interpretations/ categorizations of the world with our own. If the interaction runs smoothly, we can rightfully assume that we share the relevant portions of background knowledge. If one interactant, however, does something unexpected, something not in line with our background knowledge, we may, on the one hand, sanction this, e.g. by confusion, neglect, breaking up the contact, or we may, on the other hand, attempt to adapt our beliefs and attitudes. To give an example for this process, agreeing on the definition of a sexual situation, e.g. that a man and a woman are lovers and are about to make love, is only possible if the two participants share at least some background knowledge, e.g. what it means to be lovers, what the necessary prerequisites to making love are, etc. If the interaction runs without any problems, as is usually the case, their sexual schemata will be reconfirmed. If, however, the man violates against a schema of the woman - I take a heterosexual situation - by, for instance, remaining fully dressed, she might be confused by this apparent difference in their versions of reality and as a consequence refuse to have sex with him. She might, however, also modify her sexual schemata by, for instance, cancelling nakedness as a prerequisite to intercourse (or at least reducing its default value). Most social acts involve the use of language because language is a symbolic system capable of representing things dislocated, thus enabling people to achieve a wider range and a greater variety of goals. The act of representing the world, often subservient to the act of doing something to or with other people by means of language, 8 also presupposes a certain amount of shared knowledge: only consent on the meanings of linguistic items as well as on more general aspects of the world enables the (re)construction of fully-fledged mental pictures and the bridging of gaps left by the usually relatively meagre - compared to the mental pictures - textual input. Acts of representing the world thus rely as crucially on a common set of schemata as other aspects of interaction. Representing acts, like any other part of interaction, at the same time are a testing ground for these schemata. If understanding operates without any breakdowns, the schemata drawn upon by producers and receivers will be reinforced. If it does not, schemata may have to be revised and modified. To sum up, on the one hand, only socially shared knowledge makes smooth interaction possible. On the other hand, interaction can 8 The difference between representing and doing roughly corresponds to that between locutionary and illocutionary acts (cf. Levinson 1983: ch. 5). <?page no="77"?> What is discourse analytic about Critical Discourse Analysis? 77 reproduce knowledge - if nothing interferes in the interpretative processes, the confidence in the knowledge we have drawn upon is strengthened - or modify knowledge - if there are breakdowns, we have to reconsider our schemata and may adapt some aspects of them, as a consequence. So we must conceive of social interaction as a dialectic process in which knowledge influences action and action influences knowledge. As language allows representing the world in minute details, much more knowledge is at stake in forms of interaction involving language than in others. Social constructionism, in a nutshell, assumes that semantic knowledge in its schematic structure is the prime factor in our assigning meaning to the outside world and can thus be said to construct reality. The schemata are created, modified and reconfirmed in social interaction, primarily in language use, and the construction of reality is therefore also social (and linguistic/ discursive) in nature. From a social constructionist perspective, sexuality is founded on sexual knowledge, i.e. how we organize all information about the sexual into categories stored as schemata in our minds. This includes, for instance, a schema of heterosexual penile-vaginal intercourse: how it is structured, what its component acts are, in which order they occur, which roles it allocates to the participants, which evaluative moments it contains, how it is integrated into the more abstract schema of sexual practices, and how it is related to other schemata, e.g. that of emotions, relationships, or gender. Such a sexual schema is then imposed on the perception of the sexual in the external world, thus constructing the meaning of whatever is ‘out there’ biologically, and guides us in our acting. As sexuality involves interaction with other social beings, we have to share our sexual categories since we could otherwise never be sure that what we intend to do accords with what our partners intend to do. The sharing is achieved by taking other people’s interactive behaviour, whether in sexual events themselves or in, for instance, sexual talk, as indices of their background knowledge. This means that in interacting and particularly in talking with other social beings we constantly look for evidence of whether we share the categories in which we think and act sexually and may reconfirm, modify or change them accordingly. Social interaction is thus pivotal for the creation and maintenance of sexual knowledge. As sexual knowledge constitutes sexual reality, the latter can be said to be interactionally and thus socially constructed. Pornography is a form of talking about sexuality and it therefore must be assumed to be an important factor in the construction of <?page no="78"?> Theoretical Background 78 sexuality for its consumers, disseminating the schemata and scripts which will inform their sexual conduct. It is not difficult to imagine a situation where male and female sexual knowledge concerning, for example, pre-conditions for intercourse differ and lead the man to do things the woman would not like or would find nauseating or painful (there is, of course, more to sexual violence than mere misunderstandings). I would like to finish this section by pointing to what social constructionism means for research, establishing a bridge to the ensuing sections. The task for research is to retrace aspects of knowledge from the interactive processes - particularly from the use of language - that create and maintain it. As the analysis thus amounts to undigging knowledge from the socio-cognitive processes in which it is at stake, Michel Foucault has described it as the archaeology of knowledge (1995 [1969], cf. also Gutting 1994a). 1.2 Discourse Analysis CDA sees language as it is actually used in interaction as the main site of the construction of phenomena such as sexuality. Language in use therefore is the place to start CDA’s archaeological enterprise and is thus the prime object of analysis. To distinguish language in use from language as a system I will call the former discourse. This section will in some depth discuss what this focus on language as discourse and language analysis as discourse analysis entails, why it is particularly relevant for the study of pornography, and in what sense it distinguishes CDA from other linguistically-oriented approaches. We can look at discourse on three levels of abstraction - incidentally, they reflect three senses of the term. These three levels correspond to three dimensions of discourse analysis (cf. van Dijk 1997a: 3f.), which will be discussed separately below. a. Discourse as language in use (used as an uncountable noun): corresponds to the theoretical conception of language and thus to the general perspective of discourse analysis. b. Discourse as type of language event (used as a countable noun): corresponds to the referential frame of discourse analysis. c. Discourse as concrete language event (used as a countable noun): corresponds to the data-dimension of discourse analysis. <?page no="79"?> What is discourse analytic about Critical Discourse Analysis? 79 1.2.1 Discourse as language in use Discourse as language in use is conceived of as a dynamic event which features two main components, viz. the text, i.e. the semiotic output, and the context, i.e. the setting and the participants with their goals, their relationships, and their background knowledge (for the definition of discourse as event, cf. van Dijk 1997a: 2). The two major discursive processes of producing and interpreting meanings have to be seen as intricate dynamic forms of interaction between texts and contexts. The task of discourse analysis is now to examine these relationships between textual and contextual factors. Texts are a central element in this enterprise but in discourse analysis we look at the whole event and analyse the role of texts for the latter. The meaning of discourse I am using here covers language events with both written and/ or spoken texts (as, for instance, in the introductory textbooks on discourse analysis by Stubbs 1983, van Dijk 1997c,d, Renkema 1993, and Gee 1999) and does not want to restrict it to spoken conversation (as in the introductory textbooks by Brown/ Yule 1983 and Coulthard 1985). CDA, as a necessary consequence of its social constructionist background, sees discourse as a socio-cognitive event. It is a cognitive event because the interpretations are mental processes in which participants draw upon their background knowledge of the world, the current situation, themselves, etc. and in which this knowledge may be reconfirmed, modified, or changed. Discourse is a social event, too, because participants are not only individuals but are also members of social groups and the knowledge that is being drawn upon is a shared resource of such social groups and all modifications of this knowledge base may have social consequences. This entails that CDA looks at various aspects of discourse, particularly textual, situational and interpretative ones, to find patterns and to draw conclusions on the socio-cognitive resources standing behind these. Or, in other words, CDA analyses certain aspects of discourse to retrieve the knowledge that is being created and maintained. The special integrative view of language as a complex event is the reason why CDA has re-shuffled the content of the two original research questions of the pornography debate, which is based on a strict separation between textual analysis, which finds harmful or harmless ideas in the pornographic text, and discourse analysis, which examines whether the said ideas actually play a role in the consumption process of pornography. Discourse analysis, as outlined above, does not regard meanings as simply residing in the texts but as ‘happening’, as it were, in interaction between textual and contextual factors. The strict dichotomy <?page no="80"?> Theoretical Background 80 of textual and discursive is therefore rejected in CDA. Although we can make a distinction between the meaning potential of texts and those meanings realized (actual meaning) in discourse (cf. Widdowson 1979, cit. in Alerderson/ Urquhart 1984a: xix), it still seems absurd, from a discourse analytical perspective, to come to a conclusion that pornographic texts contain lots of problematic ideas but none of these feature in the consumers’ interpretations. 1.2.2 Discourse as type of language event Particular types of using and acting with language with regularities on the linguistic and the social sides develop and get conventionalized in societies. People with certain social backgrounds will talk to each other about certain topics in certain settings in certain interactive practices using particular linguistic items. These types of language use are also called discourses, but this time the word is used as a countable noun. Examples of discourses are business letters, advertising, political debates, or pornography. It should be noted that we can look at discourses at different levels of specificity. We can thus consider the discourse of the mass media (very general) or the discourse on the current (2008) political situation in Austria in quality papers (very specific). Although there are no clear-cut taxonomies of discourses, there are categories which people seem to accept - usually signified by the fact that they have been assigned a label - and which seem to play a role in their discourse practices. Analysis should start at such popular codified discourses. Discourse analysis is not simply the analysis of discourse but also the analysis of particular discourses. This approach is warranted by the social constructionist perspective. Using certain linguistic items rather than others leads to a regularized way of representing the world and thus, in turn, a conventionalized way of conceptualizing the world. This implies that the social construction of a field is implicated in the emergence of a discourse and CDA is more likely to find the patterns and regularities in such a specific discourse. The referential frame of research, i.e. the thing that we want to make statements about, thus normally is the particular discourse examined. This, of course, does not preclude the possibility of comparisons with other discourses or generalizations across a set of discourses. On the contrary, the position of the discourse under scrutiny in the order of discourse (or order of discourses) has always been regarded as very important in CDA (cf. Fairclough 1989: 28ff., 1992a: 68ff., also Foucault 1991 [1972], who coined the expression in the first place). <?page no="81"?> What is discourse analytic about Critical Discourse Analysis? 81 It has widely been acknowledged, particularly following Bakhtin (1994), that discourses are never completely homogeneous and that discourses mutually invade each other with strong effects (cf. Fairclough 1992a: ch. 4, Lee 1992: ch. 3). These effects are examined under the heading of intertextuality (or interdiscursivity) (cf. Fairclough 1992a: ch. 4, and also section 2.1.2 in chapter IV). There usually is, however, one discourse that can be seen as the dominant one and therefore remains the focus and the referential frame of the analysis. 1.2.3 Discourse as concrete language event On the most basic level, discourses are also concrete language events. Discourse analysis is consequently also the analysis of such events. This implies that the data for discourse analysis as implemented in CDA has to be authentic discourses in all their details and not introspective impressions of the analyst, even though the validity claims, of course, extend beyond the concrete events (to the type of events, as just suggested). The compactness of the event is, of course, not very strong with written monological texts in the mass media because it is usually not a homogeneous and concrete unit but participants may be temporally and locally distant, possibly not even knowing each other. The relationship between discourses as concrete events and discourses as types of events mirrors the one between the concrete interpretation of a text and semantic memory. As discourse analysis in a first instance is the analysis of the concrete event, yielding results about the concrete interpretations, the methodological question implied is: what warrants our generalizations? As I will explain in more detail in sections 3.1 and 3.2 below, the key is a high number of concrete discourses (and thus texts) and the inclusion of frequency analyses. The discussion of discourse/ s given above should make it easier to appreciate the conception of pornography as discourse with the three facets. Firstly, it constitutes language use and thus justifies analysing it as a socio-cognitive event with producers, consumers, semiotic elements and further situational factors interacting in particular ways. Secondly, it is a regularized way of using language or, more specifically, of using language to represent sexuality. Examining it will thus allow a description of the sexual knowledge it is likely to construct. And thirdly, pornography is also a concrete language event and the analysis must start at these concrete events and not at speculations/ generalizations (in the style of “Having looked at so much pornography, my conclusion is…”). <?page no="82"?> Theoretical Background 82 1.2.4 CDA versus other linguistically-oriented approaches In its discourse analytical orientation, CDA differs in crucial ways from other linguistic theories, in particular with regard to the conception of language and language studies. I will briefly comment on the most relevant ones. This is supposed to position CDA on the spectrum of linguistic schools, thus providing a broader perspective. System versus use Critical Discourse Analysis, though focusing on language in use, does not deny that on a first level, language is a sign system and that to study the properties of this idealized system in theoretical linguistics is relevant, not least as input to its own research. But it rejects the rigid dualism of system/ langue and use/ parole proposed by structuralism and generativism on two grounds. Firstly, parole is not simply the idiosyncratically messy application of language but it is itself highly structured, with its own social/ cognitive regularities, which are a worthwhile object of study (as has already been observed by Firth 1935: 66ff., cit. in Stubbs 1996: 41, cf. also Fairclough 1992a: 63). Secondly, parole is not simply deduced from langue but there is also a reverse, inductive relation, which means use influences the system (cf. Stubbs 1996: 45), which is an essential condition for language change. Formal versus functional For generativism and structuralism many aspects of language structure are arbitrary and are thus just formally analysed, i.e. without reference to meaning. By contrast, CDA sees language as functional on almost all levels. A fine-grained semantic theory that goes beyond the truth conditions of sentences shows that even distinctions such as passive and active can be explained in terms of meaning (e.g. regarding perspective or informativity). This explains CDA’s strong affiliations with Hallidayan Systemic Functional Linguistics and Functional Grammar. Autonomy versus integration Particularly generativism has always claimed that language is an autonomous cognitive faculty whose operations are independent of other cognitive faculties. Such a modular conception of language is incompatible with an approach taking a functional view of the language system and regarding discourse as a major site for the construction of knowledge. CDA therefore postulates tight connections between linguistic knowledge and other cognitive faculties, particularly with general world <?page no="83"?> What is discourse analytic about Critical Discourse Analysis? 83 knowledge. There is supposed to be a particularly close link between the meanings of words and grammatical constructions, on the one hand, and general knowledge, on the other (a view strongly endorsed by cognitive grammarians, cf. Langacker 1987). Meaning potential versus actual meaning The Sapir-Whorf-Hypothesis (cf. Whorf 1956, Werlen 1989) claims that the categories into which we divide the continuum of reality are likely to be those that our language system provides. Language thus functions as a kind of conceptual grid through which we perceive reality, thus strongly informing the way we think. Similar ideas have also been developed by Wilhelm v. Humboldt, Leo Weisgerber, and Ferdinand de Saussure, and more recently by Michael Halliday: “‘[C]ognition’ is just a way of talking about language” (Halliday/ Matthiessen 1999: x). In contrast to this perspective, CDA assumes that although the conceptual nature of the language system is instrumental in the construction of knowledge, it only becomes effective when it is actually used, that is, in discourse. The system provides the meaning potential which is realized in discourse because the latter is the site where linguistic items really occur and combine, rarely or frequently. So only in discourse can we actually speak of a conceivable influence of language on thinking. In this context, it is interesting that Critical Linguistics explicitly appeals to Whorf (cf. Fowler et al. 1979: 1, Fowler 1991: ch. 3), and has therefore been termed Whorfian stylistics (Dillon 1982). But Edwards (1996: 204ff.) maintains that Whorf’s approach is more discourseoriented anyway than commonly believed. The difference between system-oriented and discursive approaches to the study of social constructs can, incidentally, very well be illustrated with a comparison between the - quantitatively sparse - research on the connection between language and sexuality produced so far and my own study. Richter in his book on language and sexuality (1987) and in his Dictionary of Sexual Slang (1993) as well as feminist linguists like Dale Spender (1985) primarily look through the lexicon to find words denoting sexuality, drawing conclusions from the conceptual structure of these words (partly also from their etymologies) on cultural assumptions concerning sex. My discourse analytic approach would deny that the mere existence of words allows deep insights into culture and world views. This can only be achieved by a discourse analysis, which takes into account when these words are used, how often, by whom, in what kind of situations, and for what purposes. <?page no="84"?> Theoretical Background 84 Language versus society Stratificational sociolinguistics (most prominently represented by people like William Labov and Peter Trudgill) primarily tries to find (statistically valid) correlations between linguistic and social features. Underlying this approach is the assumption that language and society are strictly separate spheres and that language reflects social aspects (for a critique, cf. Cameron 1990b). In contrast to such a view, CDA assumes that as discourse language becomes itself a social feature and it can enter into complex relationships with other social aspects, which might be reflective as well as constructive (i.e. you create a social feature by using language in a certain way). The discourse analysts’ task is to examine these intricate interactions. Generalized context versus macrosociological context Unlike pragmatics, which - at least in its prototypical form (cf. Levinson 1983, Thomas 1995) - is mainly interested in a generalized notion of context and language use, CDA focuses on the specific features of situations, particularly also on those features of macrosociological relevance, e.g. whether the discourse is taking place in an institutional setting, to which social groups the participants belong, and tries to see the mutual influence of the macrosociological dimensions and the particular discourse analysed. CDA thus would not confine itself to saying that Fuck me! contains an imperative, and is consequently a direct and possibly impolite way of expressing a directive speech act. Instead, it would examine the consequences of a pornographic text representing a female using this utterance rather than, for instance, I was wondering whether you could satisfy my deepest bodily desires. 2 Hermeneutic theory II: What is critical about Critical Discourse Analysis? This subchapter examines in what sense Critical Discourse Analysis also involves the socio-political evaluation of potentially problematic or harmful ideas conveyed by the discourses analysed and what it takes for such an evaluation to be critical. Critical theory provides the background for such a discussion. Critical Discourse Analysis owes its political and particularly its critical edge mainly to Neo-Marxist social theories, which hold that hierarchical social structures are not only a result of economic forces, i.e. <?page no="85"?> What is critical about Critical Discourse Analysis? 85 differences in the positions towards the means of production, but also by the world view conveyed by capitalism. Rather than concentrating on purely materialist aspects as in classical Marxism, they thus focus their attention also on non-materialist cultural phenomena. The main influence of this kind comes from the Critical Theory (with capital C and T) of the Frankfurter School of Sociology, particularly from Jürgen Habermas and his Theory of Communicative Action (1995) and from neo-Marxist thinkers such as Antonio Gramsci and Stuart Hall. CDA has, however, also eclectically drawn upon social theories by Anthony Giddens, Pierre Bourdieu, Ernesto Laclau and the postmodernist thinking of Michel Foucault (for an overview of theorists influencing CDA, cf. Chouliaraki/ Fairclough 1999: ch. 5-7). 2.1 Critical theory Critical theory is a theoretical conception of social and political phenomena, with a particular focus on socio-politically problematic and harmful ideas, inviting socio-politically sensitive and committed critical research. This section examines the meaning of socio-politically problematic and harmful and what these attributes have to do with ideas in critical theory, with various practical implications for critical research being discussed in the following section. Critical theory holds that politics is concerned with the question of who has access to resources of both economic (e.g. capital, means of production) and cultural values (e.g. education, prestige, liberties) and who is assigned specific roles in social practices, i.e. what they are by convention allowed or required to do in certain types of interactions, and how and why this should be so (= distributive justice, cf. Jary/ Jary 1995: 345). Although these aspects may be a matter of individual circumstances, too, they depend to a larger extent on which social group we belong to and on the relations of this group to other groups. In other words, distributive justice is tightly linked to social orders. Grossly simplifying, social orders can be conceived of as systems of social groups and their relations to each other. Group membership entails a certain position towards resources and a certain set of conventional roles in social practices. Certain groups may have less access to capital and education and they may be required to do less profitable, in both an economic and a figurative sense, work than others (e.g. people of colour or women in many Western societies). Groups are thus not only different but they are unequal and social orders are, as a consequence, hierarchically structured, with those on the higher echelons having <?page no="86"?> Theoretical Background 86 better/ more/ easier access to resources and being assigned better - ranked on a cost-benefit scale - social roles. The hierarchical structures of most social orders translate into power relations between groups. These power relations become primarily manifest in the fact that it requires less effort for the higherranked groups to act and think as they like to, to receive positive judgements, to make lower-ranked groups do and think what might not be in their own best interest, and to pass negative judgements over the latter. For the groups positioned lower in the hierarchy all this is more troublesome and requires more effort. Though Foucault (particularly 1994 [1975], cf. also Rouse 1994) has shown that power is an intricate social phenomenon not absolutely determined by our positions in social orders, structural power, i.e. the one inherent in social structure, still is very important. This is just a rough outline of social orders and the power relations they implicate. Of course, social orders are infinitely more complicated. Firstly, they include several levels - individuals, families, institutions, society as a whole - which constitute orders in their own right with complex interactions between them. Secondly, multiple group membership is not only possible but is generally the rule. But the simplified version should suffice for my purpose here as I will primarily be concerned with the societal level, not going too much into institutions and small-scale social configurations. Socio-politically problematic now means ‘relating to power relations and inequalities between groups in hierarchical social orders’. Socio-politically harmful refers to the practical side of the coin, viz. to the fact that structural power relations can be manifested in concrete individual or collective acts ranging from insensitivity to sheer violence by the powerful to the non-powerful. What constitutes the link between social problems and social harm to ideas and values? From a constructionist perspective, social orders and the system of distributional justice that they imply are not a given but are themselves socially constructed. Social hierarchies and their consequences can thus be regarded as socially shared knowledge. This area of social knowledge is commonly called ideology (cf. van Dijk 1998 for a narrower - no value judgement! - conception of ideology). An ideology can thus be defined as a set of socially shared beliefs and attitudes concerning social groups and their relationships, social institutions, social roles, social practices, social hierarchies, social inequalities, and power relations. The idea of bias is inherent in the concept of ideology because social hierarchies, which constitute its (main) content, necessarily work in favour of groups on the higher echelons and against <?page no="87"?> What is critical about Critical Discourse Analysis? 87 those on the lower ones. The concept is often used to denigrate the ‘false knowledge’ of others, as opposed to our own ‘true knowledge’. For CDA, however, all knowledge of the social world is ideological. And we also have to bear in mind that despite their bias, ideologies also have an important function in regulating any kind of social encounter. Although the most pervasive ideologies are those that are shared - to a certain extent - by most people on a societal level, there are also ideologies on lower levels. Minority groups develop their own ideologies, usually in opposition to the dominant ones (cf. Eagleton 1991: 6). These, however, do not usually have the same kind of impact on a larger scale (but might become more prominent in times of social turmoil and revolution). We are, as has already been mentioned, neither consciously aware of the fact that our knowledge does not reflect reality but that it actually projects and constructs it, nor of the fact that we are constantly engaged in updating this knowledge in interpreting and producing texts and in any other form of interaction. Much of this socially shared knowledge thus to us is self-evident and natural. It represents common sense and we take it for granted. Ideologies, as constructions, thus work towards the naturalization of social hierarchies. The power relations inherent in these hierarchies, as a consequence, are not perceived as arbitrary but as inevitable. So even those on the lower ends of these relations thus unconsciously comply with their positions and roles. Ideologies therefore help to exercise power not by coercion, i.e. by some kind of force, but by consent, i.e. everybody implicitly agrees that the status quo is the natural state of affairs (for this distinction, cf. Fairclough 1989: 3f.). (Most) Ideologies thus support the interests of the powerful. Ideologies, again like all social constructions, are not fixed once and for all times but they are constantly subjected to potentially uprooting forces. Those favoured by ideologies therefore have to strive to maintain their positions by, for example, monitoring discourses and structuring them in certain ways. This condition of power relations being actively, though not necessarily consciously, 9 upheld is called hegemony. It introduces a dynamic element to the concept of structural power (cf. Gramsci 1971, cit. in Fairclough 1992a: 91ff., van Dijk 1997b: 19f.). 9 I will not speak of manipulation in this context because for me it presupposes that the attempt to maintain ideologies in discourse is conscious and that ideologies are false knowledge, which, in turn, implies that the analyst must have access to ‘true’ knowledge. <?page no="88"?> Theoretical Background 88 The concept of hegemony points to another aspect of power: power is not only important within interaction but also outside. The powerful have a greater impact on how different types of interaction can influence knowledge. Having access to the mass media, which are characterised by a great imbalance between the number of producers and recipients, is, for instance, of great importance in the hegemonic struggle for power (cf. van Dijk 1996). (This aspect of power will, incidentally, remain a marginal topic in this study and will not receive special attention.) From an analytical point of view, it is important that to examine ideologies means - among other things - to retrace them in discourses because using language helps to shape and maintain them. In conclusion, ideologies are socially shared knowledge of the social world, including social hierarchies and power relations. They help to make imbalances seem natural, thus sustaining the positions of the powerful. In the form of ideologies, social constructions are at the root of social inequalities and thus become socio-politically problematic and potentially harmful. Just as an aside comment: to treat ideologies as knowledge does not at all amount to cognitive reductionism (cf. van Dijk 1998: 27), an objection often raised against similar approaches (cf. Gee 1992, Edwards 1996). Ideologies have both a social and a cognitive side and in order to understand certain aspects about them it makes sense to talk about them in cognitive terms and for others - admittedly the more crucial ones - it is more useful to treat them in sociological terms. But for both practical and theoretical reasons, it is not advisable to leave out one of the two dimensions, which the critics mentioned do with the cognitive side (they thus engage in social reductionism). With respect to sexuality, socially problematic and harmful ideas are those that imply inequalities between different social groups within sexual practices and are not unlikely to result in actions bringing satisfying pleasure to only one group, leaving the other group in unsatisfied discomfort and/ or pain. This in particular applies to the ideas subsumed under objectification, which will therefore form the basic targets in my analyses (see section IV.4 on hypotheses). So there is a lot of room for ideologies in sexual knowledge. Pornography as a seminal form of interacting about sexuality helps to construct and sustain those. With respect to sexual morality, I thus follow the feminist path: sexual morality is not so much concerned with whether I reject or endorse certain practices but to fight social inequalities that might be inherent in them. <?page no="89"?> What is critical about Critical Discourse Analysis? 89 Let me conclude this section by pointing to the research-related dimension. Critical evaluation should provide the context to the undigging of knowledge, making sure that the archaeological project focuses on socially contested areas and integrating findings into a larger sociopolitical picture. The researcher should, if at all possible, also take into account what other groups may find ideologically problematic and not just stick to her or his own perspective. 2.2 Critical evaluation Critical theory with its insights into the social constructions of inequalities and power relations requires that the socio-political significances of the ideas recovered from discourses be examined. Discourse analysis should therefore be complemented by critical evaluation in CDA. This section will discuss critical evaluation, with a special subsection also focusing on the socio-political significance of the research process itself. The prime task of critical evaluation is to examine the sociopolitical context of the discourse under scrutiny, gathering information about which groups are involved in the discourse, what their stake is in it, and what the relationships of these groups are. This puts researchers in a position where they can evaluate which conceptualizations created in the discourse are in fact ideological and which effects they may have on social constellations. Considering, for instance, that male-female relationships figure prominently in the pornographic discourse and that male sexual insensitivity and violence are problematic because they constitute and enact inequalities, I will focus on particular aspects in pornography, e.g. the representation of women as body parts (= fragmentation) or as passive (= passivization), to assess their social significance. It is essential that researchers are aware of whether there has been a debate about the topic they are going to deal with because this debate will already have brought to the fore the most important potentially problematic areas. Researchers can and should then take a closer look at those. Besides, the fact that there has been a controversy indicates that the subject is a socially and politically contested one. Critical evaluation should not be understood as a sequential step in the research process, i.e. that first we do the discourse analysis and then the critical evaluation. Critical evaluation is rather a process (sometimes even a perspective or an attitude rather than a distinct process) that embeds discourse analysis (see also the graphic representation of CDA at the end of this chapter). It is relevant before the actual analysis of <?page no="90"?> Theoretical Background 90 discourse, e.g. in the selection of a socio-politically relevant subject area, during the analysis, e.g. at the hypotheses stage, and after the analysis, e.g. in the evaluation of the consequences of certain conceptualizations found. Discourse analysis and critical evaluation are thus tightly interwoven research steps, which justifies merging them into one approach, viz. Critical Discourse Analysis. As far as the topic of this book is concerned, the chapters on the pornography debate and on research into pornography have already provided a lot of background for a more fine-tuned focusing on specific aspects of sexual knowledge that feature potential inequalities between genders and for evaluating the implications of findings. This should show that critical evaluation has already begun. 2.2.1 Political dimension of the research process It is important to note that performing discourse analysis embedded in critical evaluation is not only socio-political due to the nature of the subject matter but that it has socio-political significance itself. CDA’s main activity of undigging ideologies from discourse means shedding light on what we normally do not pay attention to as we take it for granted. The analysis (cum evaluation) is thus also a form of denaturalizing, defamiliarizing or deconstructing ideologies, lifting the veil of naturalness and revealing their arbitrariness (cf. Fowler 1996: 5). This usually entails questioning the state of distributive justice inherent in the ideologies. The socio-political significance of a Critical Discourse Analysis of pornography becomes manifest in the deconstruction of the naturalness of certain aspect of gender relations in sexuality (inferior roles of women, etc.). This significance is further enhanced by the fact that there has been a debate about pornography with various camps voicing opposing views, so that research findings necessarily enter an already strongly politicized arena. There are three important corollaries of the fundamentally sociopolitical nature of research in CDA: a. Socio-political commitment: As CDA is aware of the fact that it is a priori socio-political, it does not aim at unbiasedness but is, right from the outset, a socio-politically committed form of science. Its main ethical principle is that researchers should actively seek to undermine the positions of the powerful, siding with the weak. CDA is therefore positioned on the left side of the political spectrum (cf. van Dijk 1990, Wodak 1996: 20). <?page no="91"?> What is critical about Critical Discourse Analysis? 91 b. Reflexivity: CDA emphasizes the fact that all discourses are to a certain degree ideologically informed. We are therefore required to constantly reflect on our own views and how they might influence us and to be as sincere and explicit about them as possible so that others see the socio-political commitment (cf. Fairclough 1989: 167, van Dijk 1993: 252, Wodak 1996: 20, Chouliaraki/ Fairclough 1999: 66, Pennycook 2001: 8). c. Emancipatory potential: Critical research is of an emancipatory nature (cf. Prechtl/ Burkard 1996: 278), i.e. its ultimate objective is to liberate people from believing in the givenness and inevitability of certain aspects of their social realities. CDA does this by showing the role that language and discourse play in this, which might, in turn, influence people’s decisions on which action to take. Ideally, CDA should teach people how to apply its deconstructing methods by enabling them to perform the analyses themselves in the future (cf. Fowler et al. 1979: 4, 196f.). To make this superior goal more prominent, Critical Language Awareness has been established as a branch of CDA whose main objective is that students on all levels of education should be taught to see and analyse the potential effects and implications of language use (cf. Fairclough 1992b). Readers will have noticed that I have already been trying to comply with these three principles: as the pornography debate centres around the issue of harm to women (and possibly also other groups), a project such as mine focuses on the less powerful group within society (primarily on women) and their roles in sexuality, starting the research from their perspective. I have not tried to conceal the generally feminist orientation of my study, thus making my socio-political views explicit. And finally, a study that features a discourse that is of common interest and has received so much controversial attention as pornography may be particularly valuable in the long-term pedagogical objective of CDA. 3 Scientific methodology and metatheory I have emphasized that an approach to the study of pornography promising to yield interesting results has to graft a scientific orientation onto the hermeneutic foundation. In this section, I will outline how these basic principles of scientific research are reconciled with a hermeneutic perspective in CDA, with a special emphasis on studying pornography. Concrete aspects of how these principles of methodology and <?page no="92"?> Theoretical Background 92 metatheory are implemented in the study at hand will be described in chapter IV, giving specific information about the research project. 10 Hermeneutics is an interpretative approach concentrating on what happens in the understanding of texts. Methodologically, this usually means that researchers take the position of the recipient in the discourse, go through the interpretation process themselves, and then extrapolate their own insights, or, to put it more succinctly, analysis is interpretation (cf. Fairclough 1992a: 35). 11 The quality of research therefore primarily depends on the researchers’ ingenious ways to use intuitions and introspective insights. This has made hermeneutic research, particularly studies into sexuality and pornography, speculative and unsystematic (or at least not systematic in an obvious way), even though the findings of such studies cannot simply be discarded as irrelevant. Being based on hermeneutic theory, Critical Discourse Analysis, too, is an interpretative approach. It, however, constrains its subjective element by a broad intersubjective and systematic orientation in its methodology. An intersubjective orientation means that the ultimate aim is that others can follow and understand all steps in the analysis, i.e. we are as explicit about procedures and background assumptions as possible and most data should be accessible to everyone. A systematic orientation means that we try to be consistent in our use of analytical tools and to avoid ad hoc solutions. Intersubjectivity and systematicity add a strong scientific dimension to studies in CDA, making them transparent (and, in principle, replicable, enhancing their reliability) and opening them up to communication - you can talk about results across disciplinary and ideological borders - and to critique. CDA’s scientificness rests on five pillars. The first three can be assigned to intersubjectivity, the last two to systematicity, although a clear separation of the two areas is not possible: a. Empirical data b. Quantitative data c. Comparative data 10 My analysis will be concerned with pornographic short stories, which, apart from some pictures, consist just of verbal written text. So the points made below mainly refer to this kind of data. I will, however, mention problems for other material. 11 There is an ongoing discussion about the warrantability of the readings/ interpretations of analysts. For CDA-critical arguments, cf. Widdowson 1995 (opposed by Fairclough 1996), 1996, 2004, and Stubbs 1997. <?page no="93"?> Scientific methodology and metatheory 93 d. Model-based e. Hypotheses-based What I am going to say about scientific methodology in the following is almost exclusively concerned with the discourse analytical dimension of research in CDA and much less so with the critical one. This is a consequence of the fact that CDA and its proponents come from linguistics rather than from sociology or political science. But it points to the need for interdisciplinary collaboration in the future. 3.1 Empirical data A major requirement of scientific research is starting at observable, i.e. empirical, and thus (potentially) intersubjectively given data. I will argue that this necessitates a corpus-based approach and a focus on linguistic details. I will also discuss two important caveats for the empirical basis of CDA research. Although neither knowledge nor discursive processes - and thus the aspects that CDA is primarily interested in - are immediately open to observation, there is one element of discourse that can be rightfully assumed to be empirical, namely the texts. An empirical orientation for CDA therefore means in the first place to take texts - together with, of course, any observable information available on the context - as its starting point. The actual analysis of texts and discourse can then be best described as reconstructing likely discursive processes - i.e. interpretations and productions - and, in a second step, semantic knowledge from the empirical data that texts constitute. Unlike other approaches in the hermeneutic vein, CDA starts at a well-defined, limited and (potentially) intersubjectively accessible collection of texts involved in specific discourses (such as pornography). Such a collection is called a corpus. CDA is thus not simply textoriented but it is corpus-based. The collection has to be well-defined as CDA’s focus is on the conceptualizations implied in specific discourses. The collection has to be limited as only this enables the analysts to make specific statements about it (particularly about frequencies, see also below in the section “Quantitative data”). The fact that the corpus is limited also allows making it available in some form (this unfortunately is only a theoretical condition due to the severe restrictions imposed by copyright laws). This guarantees that the empirical data which a study is based on is at least theoretically intersubjectively accessible, which, in <?page no="94"?> Theoretical Background 94 turn, makes the research more transparent and allows for better-aimed critique. An empirical orientation also requires concentrating on linguistic details rather than on large-size units. Even straightforwardly descriptive categories such as adjective, clause, or cohesion are to a certain extent the result of the researcher’s interpretations rather than empirical givens. Small elements of language, e.g. words or clauses, however, presuppose less interpretative work than large-size units, e.g. the parts of a story, argumentative moves, etc. Starting at the former thus will better preserve the empirical character of the text as a research object. In other words, an analysis is more data-oriented and empirical if it starts at the ‘nittygritty’ details of language rather than at large, ‘contenty’ chunks. In conclusion, to be empirically oriented in its methodology CDA starts with the close analysis of the linguistic details of a corpus of texts (together with some information from context), reconstructing interpretations and social constructs (as semantic knowledge) from it. There are two aspects that invite closer scrutiny in connection with CDA’s empirical orientation, viz. the analysis of semiotic forms that go beyond verbal language and the empirical nature of discursive processes themselves. 3.1.1 Beyond verbal language Today’s texts usually employ semiotic modes going beyond verbal language. I will just briefly comment on the fact that the question of intersubjective empirical data is much more intricate if we understand texts as multimodal phenomena, as is often required by CDA (cf. van Leeuwen 1993, Kress 1993, 1996, Fairclough 1995: 4). Verbal language consists - at least on lower levels - of clearly discrete elements such as phonemes, morphemes, words, sentences, etc. Other semiotic modes, particularly images, are analogical and thus continuous. In order to interpret pictures in the same or a similar way as verbal language, we have to find structure, which means we have to divide them up into components. Though there may be some legitimate ways to do that, based on insights from perceptual psychology, most attempts will be strongly informed by the analyst’s interpretations. So there is much more a priori analysis, which has to be considered in evaluating results. (For interesting alternative approaches to the semiotic analysis of visuals, cf. Kress/ van Leeuwen 1996, Rose 2001.) To illustrate this problem, let us take the example of a woman being represented as passive or active. In a - pornographic - picture we might be able to isolate aspects that seem to suggest that she is passive (averted downward gaze, body posture, etc.), but we can never be sure <?page no="95"?> Scientific methodology and metatheory 95 whether these aspects really feature as interpretative categories for viewers. We do know, however, that the distinctions in verbal language between the active and passive thematic roles of agent and patient will be a mandatory part in interpretative processes (cf. Kress/ van Leeuwen (1996), however, argue that visuals are as much structured as verbal language). 3.1.2 Interpretative practices As mentioned above, a Critical Discourse Analysis usually amounts to reconstructing interpretations and their effects or connections to knowledge from the empirical data of texts. It is often demanded that textual analysis should be complemented by a systematic and empirical study of interpretative practices, i.e. how people actually interpret texts in discursive events (Fairclough (1992a: 227f.) speaks of enhancing the corpus in this respect). This would mitigate the subjectivity necessarily involved in the reconstructive process and would challenge the results of the primarily text-based method of analysis. Common understanding/ producing practices deployed in discourses can be examined by means of ethnographic methods, where researchers get immersed in the discourse as participants themselves in order to study the discursive practices more immediately (for an attempt to combine CDA with ethnographic research, cf. Agar 1991), even if, strictly speaking, ethnographic research is not primarily empirical as it depends very much on the researcher’s interpretations of situations. Focus groups, i.e. groups of people discussing interpretations of texts (cf. Krueger/ Casey 2000), are a further method of studying discourse practices. Interviews, in which people are concretely asked about interpretations, are another viable form of research. All three methods are relevant in studying practices more generally, for example, if we want to find out how often people engage in a certain discourse, for what purposes, etc. This could constitute a highly important pillar in the necessary prerequisite steps to text-based analyses (cf. section 3.4 below). But when it comes to the analysis of discourse practices involving specific texts, the value of the said methods is clearly limited because they cannot focus on the minute details of interpretations of specific items. This makes a real comparison to textual analyses difficult. Besides, in focus groups and interviews the discursive practices are not studied immediately but are made explicit in what is, as a matter of fact, another discourse, which may distort results considerably. It might, for instance, draw the reader’s attention to aspects that normally work subliminally. <?page no="96"?> Theoretical Background 96 Discourse processes can, alternatively, also be studied with psycholinguistic methods. As with the sociological methods from above, they can reveal general features of cognitive processing, but they are not yet able to evaluate more than the most basic aspects of comprehension/ production of longer stretches of text. The problem of the lack of data on processes is more acute in written monologic discourses due to the temporal-local gap inherent in the discursive event and due to the often solitary nature of both production and reception. In interactive communication the participants themselves give clues as to how they have interpreted the preceding turn and what the background assumptions and intentions of their own utterances are (cf. Hutchby/ Wooffitt 1998). 3.2 Quantitative data Although reconstructing discursive processes and also mental representations from linguistic forms is primarily qualitative research, CDA - at least in my version - adds a strong quantitative element to this. As quantification provides more intersubjective transparency and comparability, it is a strong requirement of science. The quantitative orientation, however, is also a necessary consequence of the social constructionist basis of research in CDA. We concentrate on the effects of linguistic elements as used in discourse on semantic knowledge. This requires a qualitative analysis of the meanings and functions of linguistic elements and of how they may be represented in the concrete interpretations of texts, particularly focusing on aspects that go beyond truth-conditional and thus perhaps more immediately apparent meanings. But as it can be rightfully assumed that elements occurring more often, in greater variation and with more interconnections will exercise a stronger influence on general knowledge, as has been suggested in section 1.1.1 of this chapter, CDA does not confine itself to qualitative data but complements them with quantitative data, i.e. frequencies of linguistic items. A quantitative orientation means, for instance, that the difference between penis and love pole is first explained with respect to differences in meaning (metaphors, intertextual associations, etc.). But we need to know the difference in frequencies between these two lexemes to really draw any conclusions on effects on knowledge. Readers will have noticed that this quantitative orientation is one of the major differences between the analysis of social constructs in discourse analysis and in the more system-oriented approaches in linguistics as outlined in section 1.2. <?page no="97"?> Scientific methodology and metatheory 97 Only if we take a limited collection of texts - rather than an openended set such as the World Wide Web - does a quantitative analysis make sense. The quantitative orientation further requires - not necessarily but plausibly - that we work with relatively large corpora so that the object of analysis constitutes a (quasi)representative sample of the discourse to be analysed. Only with such a large corpus can the quantitative data show regularities in the linguistic structure of the discourse and thus of the conceptualizations probably being created and maintained. Modern methods of computer-assisted text analysis (see section 1 in chapter IV) help to handle such large bodies of data. 3.3 Comparative data Working with a corpus always bears the danger of seeing results that clearly correspond to our hypothesis as peculiar to the discourse under scrutiny. The same features may, however, occur in other discourses. It is therefore advisable to have some kind of comparison to a general corpus and/ or to a corpus of a similar discourse. The comparative aspect can be at the centre of the study but it can also function as an addendum. In connection with pornography there are a large number of possible comparative corpora. We could go to a general corpus of English, or corpora of other discourses featuring sexuality, e.g. sexological books, how-to guides, or erotica, or corpora of discourses accused of having similar effects, e.g. violent action novels and short stories. As will be described in the next chapter, I have chosen erotica for comparison in this book. 3.4 Model of the idealized discourse participant Reconstructing discursive processes and their potential impact on semantic knowledge from the linguistic details of texts does not mean that we engage in the processes ourselves but that we simulate what real discourse participants do. As mentioned above, however, we often take the position of the discourse participant and use our own interpretations in the simulation, generalizing them for the results. In order to reduce this subjective element and in order not to be exclusively dependent on our own intuitions we can draw upon a theoretical model in our analyses. A model in this context is a systematic framework representing invariant and general aspects of discourse processes. It thus enables researchers to bridge the gap between linguistic items in texts and interpretations. <?page no="98"?> Theoretical Background 98 A model actually postulates an idealized discourse participant (for this concept, cf. Marko 1997: 5, O’Halloran 2003: ch. 8 and 9), who can be described on two planes: • linguistically/ cognitively: in terms of the universal (cognitive) aspects of discourse comprehension/ production; • socially: in terms of the most likely discourse participants and their common social resources/ goals and the practices these participants are likely to engage in, also in relation to practices in other discourses. Idealized in this context means ‘abstracted from the individual’ and not ‘the best possible’. The linguistic-cognitive characterization of the idealized discourse participant constitutes the core of the model. It is more general in nature and should therefore be applicable in all analyses. The social side, on the other hand, depicts concrete aspects of the discourse under investigation and therefore has to be adapted to the latter. The social characterization of the idealized discourse participant could, as a matter of fact, be part of the research: interviewing or questionnairing people could, for instance, shed light on discourse practices (cf. 3.1.2 above). Working with a model has the following advantages: • Models reduce the amount of introspective interpretation. Representing the invariant aspects of discourse processes, they limit the space for purely intuitive assumptions. • Models can render research more systematic. The analysts are aware of the position of an element within the whole system and can thus better assess its relevance, contrasting it with and comparing it to other elements. • Models are explicit and as a result, the analysis becomes more transparent and open to critique by others. • Models are independent from, and precede, the actual analysis. This helps prevent circularity. Circularity means that we find what we expect in the texts, not because it is actually there but because we interpret whatever is there as an indication of our expectations. • Models integrate CDA more firmly into linguistic theory as they draw upon frameworks that have been presented elsewhere in linguistics (although they do not have to be adopted in an unmodified form). <?page no="99"?> Scientific methodology and metatheory 99 Practically speaking, a Critical Discourse Analytical project can only present the rough outlines of a model explicitly, with particular details being explained if they are of relevance in a specific area of analysis. This is also how I will proceed in my study. 3.5 Hypotheses-based It has been deplored that findings of hermeneutic and CDA research are often presented as though we could not have come to any other conclusions. Usually no connection between the results and the initial expectations are made. This sometimes conveys the impression that the research has found what it intended to find and thus has become circular. To avoid this impression and to be more systematic in its scientific procedure, my version of CDA follows the standard scientific procedure of forming hypotheses and then testing them against (textual) data. Hypotheses are important because they embrace the fact that analysis is always heavily informed by the analysts’ expectations and the consequential selective focus on certain aspects rather than on others. They additionally allow for the possibility of falsification: I may fail to find evidence for the existence/ occurrence of aspects of knowledge that I thought would be there. The hypotheses in CDA are, on a first plane, socio-political and then, on a second plane, are translated into cognitive and linguistic terms. In other words, I propose in a first step that certain beliefs and attitudes should have a prominence in the discourse under analysis (influenced by my analysis of social circumstances) and in a second step that certain linguistic items would have to occur in the corpus to realize the said ideas and attitudes. These two steps of hypothesis-formation actually represent a reversal of the simulation of discourse processes, starting at the sociopolitically significant knowledge and working its way down to language. This, however, just means that the simulation in a hypothetical form precedes the actual research, i.e. first step: this linguistic element would have this cognitive effect, second step: does it really occur? The procedure does not alter the general nature of the research outlined above. Hypotheses should concentrate on more general conceptual structures, i.e. on global conceptualizations (or superordinate schemata), organizing our specific beliefs and attitudes. Fragmenting women, for instance, i.e. thinking of women in terms of their body parts, would be such a global conceptualization. Despite the abstract nature of such hypotheses, they seem more interesting from a socio-political point of <?page no="100"?> Theoretical Background 100 view because ideologies can manifest themselves in abstract conceptualizations and they also appear to be the ideal starting point for largescale studies of particular discourses. Besides, the main contested areas emerging from the pornography debate, e.g. objectification - which I will adopt as the basis of my hypotheses - and subordination constitute such global conceptualizations. To give an example of how hypothesis-oriented research proceeds, my analysis of the socio-political context of pornography suggests that the question of female passivity will be a central one. If women are to be interpreted as more passive in sexuality, I expect them to occur more often in the patient role of verbs than men. This is an assumption that can easily - theoretically, at least - be tested against the data of a corpus of pornography. Hypothesis-based does not mean that the analysis narrow-mindedly postulates a one-to-one correspondence between linguistic elements and beliefs and attitudes, assuming that the hypothesis is verified if the linguistic elements are found and that it is falsified if they do not appear. As researchers we have to keep an open eye on other factors potentially influencing the linguistic elements. In conclusion, (my version of) CDA integrates a scientific perspective into its hermeneutic foundation by adopting an intersubjective and systematic orientation. It is corpus-based, i.e. it works with well-defined, limited and intersubjectively accessible collections of text/ s, focuses on linguistic details, analyses the data qualitatively and quantitatively, and compares its findings with data from other corpora. It furthermore makes its subjective interpretations more systematic and generalizable by basing them on models of discourse processes. Finally, it does not just interpret discourses but develops hypotheses that are tested against the data in the corpus and are thus falsifiable. 4 Summing up Discourse, i.e. language in use, is the most important form of social interaction. In discourse, we produce and interpret texts with different intentions and drawing on different types of background resources, checking whether we share our general semantic knowledge of the world with others. Discourse is thus essential for the creation, modification and reconfirmation of this knowledge. As we use the latter to actively assign meaning to reality, we can say that discourse also contributes to the <?page no="101"?> Summing up 101 construction of reality. So if we want to analyse reality constructions, we have to start at discourse. Critical theory holds that hierarchical social orders with their concomitant relationships of inequality and power between social groups are also a form of knowledge and are thus also socially and discursively constructed. To analyse discourse may thus reveal socio-politically problematic constructions of social reality working in favour of those in power. In discourse analysis, we examine what kind of semantic knowledge is created and maintained in the interpretations and productions of texts in context, i.e. in discourse. Methodologically this usually means taking the empirical data that texts constitute and simulating (reconstructing) interpretations from them with the help of models of discourse processing. Theoretically, this method could be complemented or replaced by empirical analysis of the processes themselves, for which, however, no convincing method has been proposed yet. As we are interested in specific types of using language, i.e. in specific discourses, we focus on a representative sample of texts - a corpus. The corpus is analysed qualitatively and quantitatively. The semantic knowledge thus ‘distilled’ from language in discourse analysis is critically evaluated, which means that its sociopolitical significance, relevance for social inequalities and power relations, is assessed. Critical evaluation is less a second step in the analysis than an accompanying practice in the analysis of discourse itself, in a way embedding the latter (as indicated in the graphics below). In a schematic form, CDA can be described as the description of forms with the aim of reconstructing the interpretations of meanings in the light of a critical evaluation of their socio-political significance. Critical evaluation of socio-political significance Interpretation of discourse meanings Description of textual forms Figure 1: Schematic representation of the methodological practice of CDA (adopted from Fairclough 1989: 25, but re-interpreted and adapted to my own approach). With respect to the two questions mentioned at the end of the last two chapters, CDA says that pornography as (a) discourse can influence its <?page no="102"?> Theoretical Background 102 consumers’ knowledge, explaining the connection in an intricate framework. It further assumes that a text-based discourse analysis, involving the simulation of interpretative process on the basis of textual material, can help to uncover beliefs likely to become part of typical pornography consumers, denaturalizing and making explicit what we normally take for granted. As some of these beliefs involve discriminations and inequalities between women and men, they can be argued to be potentially harmful. And it is these that CDA will focus on in its research. <?page no="103"?> IV Analysing Pornography: Concrete Features of the Study The preceding chapters have outlined the socio-political and metatheoretical contexts of a Critical Discourse Analysis of pornography, explaining why it is relevant and how it can be carried out. This chapter will now describe the concrete details of my research project. It will first focus on the corpus, then present the most important features of a model of discourse interpretation, and finally introduce the hypotheses to be tested in the second part of this book. 1 Corpus analysis: empirical data, comparative data, quantitative data This section will explain how my project meets the requirement that CDA should work on empirical data, comparative data and quantitative data. It will first introduce the corpus to be investigated in my project and the one I will be using for comparison. I will describe and explain their compositions and the technical details of the analysis, particularly the use of concordancing software. Besides, I will discuss the reasons why I have not included pictures in the corpus and why the project is primarily text-based and does not examine interpretative practices themselves. Even though some aspects of the magazines and stories mentioned in the following invite a more thorough critical examination in its own right, the account is limited to descriptions. 1.1 Empirical data: the corpus The empirical data of my project is a corpus of 486 pornographic short stories from 20 magazines specializing in this discourse, comprising approximately 680,000 (orthographic) words (orthographic words are defined as strings of letters bounded by spaces or punctuation marks). <?page no="104"?> Theoretical Background 104 Number of magazine (issues) 20 Number of stories 486 Number of words (tokens) 677,428 Number of words (types, i.e. different words) 22,547 Standardized type/ token ratio (in %) 1 42.66 Average number of words per story 1,394 Table 1: General statistical information on the pornography corpus. I have narrowed down pornography to pornographic short stories because they have obvious practical advantages over videos, largeformat magazines such as Penthouse, Hustler, or Playboy, and electronic media. As the former consist primarily of language, a corpus approach is easier to carry out. Besides, short stories are a more homogeneous and thus more easily definable discourse than what can be found in the above-mentioned magazines. Short stories are also to be preferred over pornographic novels. Being shorter they offer more possibility of variation, which benefits the representativeness of the collection of texts. And finally, although short stories stand for only a tiny segment of the pornography market, it is my impression that they contain all essential elements that also characterize more popular formats. The magazines in the corpus are mostly mainstream, which means that the stories do not cover any extreme sexually explicit material featuring rape or overt sexual violence and/ or fetishes such as sex with old, obese or disabled persons. Apart from the fact that mainstream pornography is easier to get hold of, 2 I also assume that if there are certain (negative) tendencies in my material, then these will aggravate in more hard-core sources. The only deviant magazine included is Family Secrets, partly featuring incestuous sexuality, which, however, was not classified as hard-core or special interest because it was sold together 1 Standardized type/ token ratios, which are measures of the lexical variation of a corpus, take into account corpus size since there normally is a negative correlation between corpus size and ordinary type-token ratios (cf. Mike Scott’s WordSmith 3.0/ 4.0 help manual). The respective value for the British National Corpus - a large general corpus of British English - is 43.5, just for comparison. 2 Considering the probably only limited market for textual pornography, I doubt that there are many fetishist magazines specializing on short stories at all. We, however, have to take into account that with the exception of child pornography there are hardly any restrictions on written pornography (cf. Ellis 1988, cit. in Hardy 1998: 50, McNair 1996: 119) so that it would be much easier to produce really hard-core fetishist textual porn. <?page no="105"?> Corpus analysis 105 with two other publications in one sealed package. The collection of magazines is homogeneous with regard to external factors (e.g. text type, i.e. short stories, appearance, price, availability), covering British and American sources only, though. The corpus’ representativeness is nevertheless questionable because the set of sources was not based on a careful selection process after a thorough analysis of what was available, but it was rather guided by intuitive criteria during two short research stays in England and the United States on a limited budget. I have included the full texts of all the stories contained in the 20 magazines and not just extracts of equal length. 1.1.1 The magazines The following magazines have been used (details in “References”): • Dear Diary. Confessions of Sex-Crazed Women • Mid-Night • Down ‘n’ Nasty • Night • Erotic Penpals • P.S. I Love You • Erotic Stories (3 issues) • Penthouse Forum • Experience • Silky Secrets • Experience Readers’ Erotic Letters • Turn-Ons • Family Secrets • Vibrations • Hustler Fantasies • Wet • Luv Letters • Wild & Nasty These magazines were acquired at pornographic outlets and at ordinary magazine shops (albeit on top corner shelves) in Washington, D.C., New York, and London. Six of them (always non-current issues) were contained in sealed plastic bags (at three magazines a bag) and sold at special prices. The only magazine not directly purchased is Erotic Stories, to which I subscribed for six months. To have more variation in the corpus, I have included three issues of Erotic Stories because its stories belong to a slightly different and more sophisticated kind of pornography (more plot, more variation, more intertextual allusions, generally greater length) and because some of the other magazines though bearing different names are still published by the same companies in the United States and partly resemble each other more strongly than the different issues of Erotic Stories. Since I did not want to spend too much time in the respective shops, paging through pornographic material, I had to make a quick selection based on external criteria. All magazines appear in small format (A5 or smaller) and cost 3.5-5 $ or 3-3.50 £ (in 1994/ 1995). They <?page no="106"?> Theoretical Background 106 all have glossy covers with colour photographs of nude or semi-nude women in seductive poses, all holding, touching, pointing to, or otherwise exposing primary or secondary sexual organs and nearly all of them catching the eye of the observer. None of the cover photographs features a second person or any sexual activity. The inside pages of the magazines are black and white and are printed on ordinary paper (exceptions: Vibrations, Experience, Readers’ Erotic Letters, which are printed on glossy paper and include some colour photographs, and Penthouse Forum, which has glossy paper and colour photographs exclusively). The titles of the magazines might be suggestive, but with the exception of Family Secrets, and, to a limited extent, Dear Diary, Experience Readers’ Erotic Letters, Luv Letters, and Erotic Penpals, they do not give any clues as to the content. On the cover page, the title is commonly framed by short headline-like phrases, usually but not necessarily indicating a section to be found inside and probably intended to enhance the title’s allusiveness, e.g. Easy Pickups & Easy Lays, No-name , Dirty Sex! , Vacation Fornication: Pussy in the Pines, or Just Friendly Fucking. Apart from these literally superficial features, there are also further apparent characteristics that the magazines share. Even though the main focus is supposed to be on text, the magazines also contain photos and other pictures (especially drawings). All in all, I have found 2,528 pictures, which means there is, on average, a picture every four pages. This number just includes editorial pictures (with cover photos) but not those in ads. Interestingly enough, the pictures are usually either completely unrelated or only vaguely connected to the stories. A further trait of the magazines is that they do not only contain stories (even though the corpus just consists of narrative discourses), but also feature (further) editorial discourses: introductions, disclaimers, and sometimes advice sections, reviews, cartoons, or photo-stories. Editorial material sometimes verges on narrative discourses with a female (pseudo? ) editor being presented as an erotic persona herself, saying, for instance: My name is PHOEBE GLANVILLE and I’m the new Managing Editor for Bellright Limited […] This is found in Vibrations, Experience, and Experience Readers’ Erotic Letters with a photo of a nude woman exposing her pubic area just above this text, making readers identify Phoebe Glanville, for example, with the picture. Erotic Stories, too, has a female editor who directly speaks to the readers and who appears in a photograph (with a seductive gaze though not in the nude). In the three magazines mentioned at the <?page no="107"?> Corpus analysis 107 beginning of this paragraph, there are also section or issue editors who feature in some photographs (most of them large-breasted blondes) and who address the readers directly (cf. also Itzin 1992b): I’m Bella Billingham and from this issue I’m taking over from Candy. Please keep sending in those horny letters so that I can make this magazine even more exciting than it has been. Remember it’s YOU who writes this magazine. Love, Bella They are also directly addressed in the letters/ stories (Dear Bella […]). The publishing company also boasts on the back cover: […] Bellright, with it’s [sic! ] posse of beautiful, sexy editors, will keep you coming! There is also a lot of promotional discourse. About 15% of the magazines’ space is on average devoted to advertising, most noticeably on the insides and outsides of the back covers and the inside of the front covers. Apart from personal ads, there are primarily advertisements for products belonging to the sex industry: subscription to pornographic magazines, videos, sex toys, lingerie and, most prominently, telephone sex. I assume that many of the companies advertising are related to the publishing companies so that the magazines themselves also serve to market other products, bringing more revenues. But this relay function is not as important as in large-format magazines (cf. Dines 1998a). 1.1.2 The stories The stories mostly follow the expected pattern of pornography: there are many sexual encounters, which are described in physiological/ visual detail. The typical plot features casual encounters between strangers which very quickly turn into sexual affairs, with fellatio and penilevaginal intercourse being the common practices. Differences in content between stories are primarily restricted to settings and props. Given that the average length of a story is about 1,400 words - even if some stories are much longer than this, particularly in the Erotic Stories issues - there is, judging superficially, little space for fully developed characters or detailed accounts of emotional and social motivations. Despite the predictability of storylines, there is relatively great variation in narratological devices. We find first person narrations (related by one character involved in the story) and third person narrations related (or focalized) through one character’s perspective (for this difference, cf. Toolan 1988: 67ff.). The former are often presented as letters - purely fictional or allegedly sent to the magazine by readers - or as diary entries. Narrative discourse occasionally is disguised as non- <?page no="108"?> Theoretical Background 108 narrative, e.g. in interviews with nightclub strippers or in argumentative essays about sexuality incorporating lengthy (narrative) quotes. Some of the texts in Erotic Stories are exceptional in that they operate with intertextual allusions to well-known sources, sexualizing books such as Charles Dickens Great Expectations (in “Small Expectations”), Agatha Christie’s Poirot series (in “Pierrot’s Strangest Case”), Enid Blyton’s Famous Five series (in “Famous Five Go Bonking Mad”), films such as Thelma and Louise (in “I’ve Seen that Movie, too”), Pretty in Pink (in “Pretty in Pink”) or the TV series Dr. Quinn (turned into “Dr Quim: Medicine Woman”), or pop songs such as Frank Sinatra’s “New York, New York” or Phil Collins’ “In the Air Tonight” (in stories of the same titles). The authors in this magazine are also more explorative of other discourses, e.g. science fiction, horror, fantasy, or romance. The writers are, judging by their names, both women and men. I have decided, however, not to take the gender of the writers into account, firstly because I doubt whether all writers with a female name are really women, secondly because, apart from certain changes in the gender of the narrator, there appears to be great uniformity of the stories of female and male writers, and finally because I think that the gender of the recipient is more crucial for the interpretation process and the possible effects that pornography has. Besides, authorship, i.e. the authentic identity of the person writing, is much less prominent in pornography. We thus learn nothing about the authors apart from their names, unlike in the erotica books (see below). 1.2 Comparative data: a corpus for comparison In order to relativize the results from the textual analysis of the pornography corpus, I have composed a comparative corpus of erotica, i.e. of a discourse that is often seen as indistinguishable from pornography. Although the comparison will not be the main objective of my study, it will still help to bring results into perspective. The comparative corpus is smaller in size (approx. 240,000 orthographic words) and also exclusively consists of short stories. They have been taken from three collections published as books. Table 2 contains the most interesting statistical features of the corpus. <?page no="109"?> Corpus analysis 109 Number of books 3 Number of stories 68 Number of word tokens 239,959 Number of word types (i.e. different words) 16,360 Type/ token ratio (standardized) (in %) 43.66 Average number of words per story 3,528 Table 2: General statistical information on the erotica corpus. Although there would have been a number of alternative comparative corpora - e.g. historical or gay/ lesbian pornography - the contrasts to erotica might make particular features of pornography more visible, features that seem more interesting for a feminist-oriented study. The comparison initially was only a side issue, becoming more important in the course of the research, i.e. at a time when I had already obtained my material. This is the reason for the relatively small size of the corpus, which of course also makes its representativeness problematic. 1.2.1 The books The erotica corpus consists of the 68 stories from the following three books, all acquired at a normal bookshop in Washington, D.C.: • Fever. Sensual Stories by Women Writers, edited by Michelle Slung, published in 1995 [originally in 1994] by Harper Perennial; • Pleasures. Women Write Erotica, edited by Lonnie Barbach, published in 1985 [originally in 1984] by Harper Perennial; • The Best American Erotica 1994, edited by Susie Bright, published in 1994 as Touchstone Book by Simon & Schuster. I have chosen to include two books clearly addressed to women readers because erotica seem to be primarily a female discourse (cf. Juffer 1998 for an elaborate argument why this is so), which often also implies female authorship. I further expected the contrast with the pornographic short stories to be more highlighted and also more revealing. Whether erotica differ from pornography in the language used and the conceptualizations of the world and particularly of sexuality conveyed remains to be analysed. There are, however, many apparent superficial differences that justify a distinction of these two discourses. What are these differences? <?page no="110"?> Theoretical Background 110 The very fact that the erotica short stories appear as books, are published by renowned publishers, and are therefore available at wellrespected bookshops culturally distinguishes them from their pornographic counterparts, inviting a different readership as well as a different kind of consumption. To be published as books entails difference in price - all three books cost 12$, which is, however, compensated by their greater lengths - and in semiotic design: the book covers feature painted instead of photographic nudity of women (on the cover of Fever, we even find the famous painting “Nude with Cat” by Balthus). None of the women’s poses could be interpreted as seductive: in two pictures, no sexual organs are visible and the women’s heads are averted, in one picture, there is no eye contact between observers and the woman either, and the exposing of chest and breasts seems to be directed at a schematic person standing in the background. It is worth mentioning that none of the books contains any further pictures. The titles Fever and Pleasures may not be that different from the pornographic magazines in their unspecific sexual suggestiveness (which The Best American Erotica 1994 lacks altogether). The subtitles Sensual Stories by Women Writers and Women Write Erotica, however, introduce a much more sober tone. The framing of the title by further sexual allusion is also missing. Non-narrative discourses in the erotica books are restricted to introductions by the editors (Pleasures also includes introductions to the individual sections) and the authors’ bios at the end. There are no advertisements. 1.2.2 The stories The erotica stories contain explicit sexual scenes though they cover much less space than in the pornographic stories. Being three times as long as the latter on average, the erotica can afford to develop the characters and the social and emotional backgrounds of their sexual desires and activities. This is also a reason why the plots are much less predictable. As far as narratological devices are concerned, there seems to be as much variation in the erotica corpus as in the pornography corpus. The use of other discourses and of intertextual allusions, however, is rarer, probably also as a result of the smaller number of stories (the main corpus contains seven times as many). A major difference in the presentation of erotica and pornography is that the stories in the erotica books are clearly assigned to an author, who is even introduced with a short biographical sketch at the end of the <?page no="111"?> Corpus analysis 111 respective book. This is relevant considering that the author-function is, according to Foucault (1984: 107, cit. in Juffer 1998: 116ff.), one of the most important features of (high? ) literature. 1.3 Limitations on empirical data: non-verbal signs I have already mentioned that the pornographic magazines of my main corpus contain pictures. These, however, will not feature in my analysis although I am aware that in multimodal discourses visuals and verbal language mutually influence each other in complex ways. There are, however, some reasons that justify my decision. Firstly, there is, as far as I can judge, still a lack of convincing models of how we proceed in our interpretations of multimodal discourses, let alone of texts such as those examined since here the pictures hardly ever have any direct connection to the content of the stories, as mentioned above. There might be coherence established on a deeper level with both images and words contributing to sexual arousal without being connected content-wise. But this would considerably complicate the picture (pun intended! ). Secondly, visuals cannot be integrated into the corpus (at least not in a way suitable to my approach) and can thus not be analysed in the same qualitative-quantitative ways as verbal language. Finally, I would be faced with all the problems of the interpretation of analogical signs described in section 3.1.1 of the last chapter. 1.4 Limitations on empirical data: interpretative practices Textual analysis should always be supplemented by a study of the interpretative practices. My study does not include such an investigation and I will briefly argue why not. Firstly, there are methodological reasons (cf. the discussion in 3.1.2 of the last chapter). But secondly, there are also very mundane practical reasons. Studies of interpretative practices require much more effort and are difficult to carry out by an individual researcher. Thirdly, pornography is probably a discourse where subjects are more selfconscious than in other discourses so that there might be a gap between how they say (and think) they act and how they actually act. The omission of this side of CDA research is certainly a serious shortcoming since it would definitely have been interesting to at least have some insights into what consumers generally do with pornography, where they read it, for what immediate purpose, etc. But perhaps further studies of pornography will remedy this problem. <?page no="112"?> Theoretical Background 112 1.5 Computer-assisted analysis of the empirical data As the quantitative analysis of a large corpus is impossible to carry out by hand, I typed all material on a word processor (due to the poor quality of printing, I could not employ OCR - optical character recognition - software) to make the corpus accessible to concordancing (i.e. text analysis) software. I then used the programme WordSmith (version 3.0) - created by Mike Scott and published by Oxford University Press - for the analysis. Concordancing software can basically do two things: firstly, it can produce word lists, i.e. lists - ordered either alphabetically or according to frequency - of all orthographic words in a corpus (cf. Table 3). highlights 4 the 25,847 (3.79%) highly 16 I 23,693 (3.47%) highly-polished 1 and 22,484 (3.29%) highs 3 my 17,307 (2.54%) highschool 2 to 16,795 (2.46%) hightech 1 her 16,221 (2.38%) highway 6 of 13,103 (1.92%) hike 4 a 13,004 (1.91%) hiked 8 was 10,969 (1.61%) hikers 1 she 10,067 (1.47%) Table 3: Word lists of the pornography corpus (extracts) (left: alphabeticallyordered; right: ordered according to frequency). Secondly, the software can list all occurrences of a word in its immediate co-text (the co-text is the verbal context). Concordance 1 presents an extract of the concordances of fuck. nfusion . ^ After a long fuck session , we lay exhausted s she started to fingerfuck her pussy . ^ I felt compl round 2: 30 with our new fuck -friends Frank and Gina bu s a tangled heap of four fuck -hounds locked in 69 . Lic h him , or lick him , or fuck him . ^ He was so turned-o e more . I want it all . fuck me ... fuck me hard , hard ant it all . fuck me ... fuck me hard , harder . Make me p . We ‘ve been like two fuck -machines . But that ‘s no hem . Andrew is the best fuck I ‘ve ever had ; any and e longer , I needed him to fuck me and fuck me right then Concordance 1: Search word fuck in the pornography corpus (extract). The heuristic power of concordances can be enhanced by an ordering function, enabling us to alphabetically order the lines according, for <?page no="113"?> Corpus analysis 113 instance, the first element following or the first element preceding the search word. This entails that all instances of, for example, fuck followed by her or him are displayed together. WordSmith provides for a wildcard option to widen the scope of the search. With the help of wildcards, we can, for instance, look for all words starting with fuck* (e.g. fuck, fucks, fucking, fucked). On the other hand, the programme also allows us to limit the output of a search by defining a context word, which has to occur within a certain span before or after the search word. By means of context words, we could, for instance, look for all occurrences of fuck followed by the personal pronoun him within two words. Finally, the concordancing function of WordSmith can also calculate the most frequent words (collocations) in the co-text. Table 4 lists, for instance, the most common collocations of fuck. Freq. 2L 1L C 1R 2R me 197 31 6 - 160 0 to 193 12 178 - 0 3 her 107 9 6 - 92 0 and 61 9 30 - 8 14 you 61 16 11 - 28 6 the 56 15 22 - 11 8 my 40 6 6 - 26 2 I 39 9 9 - 4 17 a 38 19 11 - 3 5 him 25 16 4 - 5 0 Table 4: Collocations of fuck; horizon 2/ 2 (i.e. two words before and two words after the search word); first column: total frequency, other columns: frequencies in the respective positions (e.g. 2L = two words to the left of fuck). In addition, WordSmith produces all kinds of statistical information about a corpus and also about its component sub-corpora, e.g. type/ token ratios, average wordand sentence-lengths, standard deviations of wordand sentence-lengths, etc. In a nutshell, a concordancing software of the quality of Word- Smith can help to quantify aspects in the language of texts because it has immediate access to elements scattered across the whole corpus and can display them together. One problem of concordancing programmes is that they define words as strings of letters. Sometimes, however, the same string of letters can represent more than one word class, e.g. fucking can be a verb in the present participle, e.g. they were fucking, a noun (gerund), e.g. the <?page no="114"?> Theoretical Background 114 fucking was great, an adjective those fucking exams, and an adverb/ intensifier, e.g. they were fucking gorgeous. The software cannot distinguish between these different versions of fucking because it is the same orthographic word, i.e. the word always consists of the same combination of letters. Tagging can solve this problem. Tagging means adding information on particular aspects of the corpus by inserting so-called tags. The most common form of tags are part of speech tags, added to words for marking their word categories. Tags consist of combinations of letters attached to the word separated by triangular brackets (or an underscore). The latter help to set them off from the rest of the text. Fucking as a noun could thus become fucking<NN1>, as a verb in the -ing-form fucking<VVG>, as an adverb fucking<RR>. A tagged passage of the corpus is given in Table 5. ^> Still<RR> dressed<VVN> in<II> my<APPGE> business<NN1> suit<NN1> and<CC> high<JJ> heels<NN2> ,<,> I<PPIS1> was<VBDZ> moving<VVG> props<NN2> around<RP> ,<,> not<XX> watching<VVG> where<RRQ> I<PPIS1> was<VBDZ> stepping<VVG> .<.> I<PPIS1> tripped<VVD> on<II> a<AT1> tangled<JJ> cord<NN1> and<CC> would<VM> have<VHI> really<RR> been<VBN> huff<NN1> if<CS> Peter<NP1> had<VHD> n’t<XX> caught<VVN> me<PPIO1> .<.> I<PPIS1> fell<VVD> against<II> him<PPHO1> ,<,> and<CC> when<RRQ> .<.> I<PPIS1> did<VDD> ,<,> a<AT1> bolt<NN1> of<IO> electricity<NN1> ran<VVD> through<II> my<APPGE> body<NN1> .<.> Judging<VVG> by<II> Peter<NP1> ‘s<GE> reaction<NN1> ,<,> he<PPHS1> felt<VVD> it<PPH1> too<RR> .<.> I<PPIS1> ‘ve<VH0> never<RR> felt<VVN> such<DA> intense<JJ> chemistry<NN1> with<IW> anyone<PN1> .<.> Table 5: Extract from the pornography corpus with part-of-speech tags. As it would be impossible to insert all tags by hand in a corpus like mine, I have used an automatic tagging programme called CLAWS. 3 This is a statistically-driven corpus-based programme that assigns the most likely word class to each word, basing its decision on evidence from other tagged corpora (cf. Garside 1987, Leech/ Garside/ Bryant 1994, Oakes 1998: 80-84). I have post-edited the result manually, correcting consistently-occurring and obvious mistakes. Tags can, as already mentioned, help to disambiguate the grammatical status of certain words. But they also enable us to carry out more general searches based on word-classes. We can, for instance, look for all occurrences of fucking as a noun that are preceded by an adjective. 3 I would like to thank Nick Smith from the Department of Linguistics and English Language at Lancaster University for helping me with the tagging. <?page no="115"?> Corpus analysis 115 VVI> such<DA> a<AT1 great<JJ> fucking<NN1> much<RR> long began<VVD> a<AT1 classic<JJ> fucking<NN1> action<NN1 HO1> a<AT1> damn<RG> good<JJ> fucking<NN1> now<RT> ,<,> " II> his<APPGE> passionate<JJ> fucking<NN1> ,<,> his<APPGE VHG> a<AT1> "<"> fabulous<JJ> fucking<NN1> time<NNT1> "<" > from<II> the<AT serious<JJ> fucking<NN1> I<PPIS1> ! <! > ^> Just<RR Friendly<JJ> Fucking<NN1> ^> Just<RR eating<VVG> this<DD1 wild<JJ> fucking<NN1> over<RP> and< D> and<CC> the<AT serious<JJ> fucking<NN1> began<VVD> Concordance 2: Search string ‘*<JJ> fucking<NN1> (= adjective + noun fucking)’ in the pornography corpus (extract). As becomes clear from the example of fuck, tagging can resolve certain grammatical ambiguities, but semantic ambiguities remain. It is, for instance, impossible to distinguish between the sexual meanings of fuck and its mundane meanings (‘cheat’, ‘make a mess of’, etc.) with a concordancing programme. I therefore almost always have to go through the results and ‘weed out’ cases semantically not fitting my purpose. 1.6 Quantification I will present data in quantitative terms. This primarily amounts to giving frequency figures for the occurrences of particular linguistic items or numbers of different items as a value of lexical variation. There will be no attempts at large-scale quantification, e.g. devising a quantitative objectification index. My study only uses basic tools of descriptive statistics (mostly absolute numbers and percentages). Only occasionally will I draw upon more complex mathematical procedures (e.g. calculating ratios between values or average percentages). I will, however, not employ more sophisticated statistical methods, which admittedly may be one of the major shortcomings of my study. Future research in CDA will obviously also have to explore which of these methods can substantially contribute to a more thorough understanding of socio-politically significant discourse processes. To put it in a nutshell, the main body of research in this study is the computer-assisted qualitative and quantitative textual analysis of a 680,000-word-corpus of pornographic short stories. Additionally, comparisons will be drawn between this and a 240,000-word-corpus of erotica. The project will focus on verbal language and pay less attention to visuals. It will look at the texts, trying to reconstruct discourse processes but will not examine the latter separately. <?page no="116"?> Theoretical Background 116 2 A model of the idealized discourse participant (pornography and erotica consumers) As detailed in 3.4 in the last chapter, if we want to know what happens to consumers of pornography without immediately examining them we have to simulate the consumption process with the help of a model of discourse processing, with a linguistic side and a social side. 2.1 Linguistic model The linguistic model is supposed to represent the invariant aspects of what happens when we interpret or produce texts, i.e. the invariant aspects of discourse processes. Such a model should eventually enable us to assign a certain (preliminary - of course, this might be subject to contextual variation) semantic value to all occurring linguistic items. As no linguistic model is as yet able to achieve this, I will just provide a framework that integrates eclectically selected semantic conceptions of various elements into a whole. I will present the skeleton of the model, leaving the introduction of individual linguistic items to the places where I actually apply them. Two loud words of warning are in place here: even though the ideas for the framework to be presented in the following have emerged in my own work with CDA as well as through dealing with other people’s studies, it still is much too abstract to be of immediate practical relevance. Its sole function probably lies in providing a system for orientation. It is also worth noting that I will present the model in much finer detail than is justified by the ensuing discussion, which is to say, many of the aspects to be mentioned will not play an important role in the analyses proper. I have decided to include the full description, however, to fulfil the subservient function of this study mentioned in the introduction, namely to contribute to the refinement of the theoretical background of CDA, however modest this contribution may be. The model of discourse processes is cognitive in nature: it stipulates what the idealized discourse participant does in her or his mind when assigning meanings to linguistic elements and what the relations between these meanings and semantic knowledge are. The cognitive perspective is a necessary requirement of social constructionism, on which the research is based. Although the model is generally not discourse-specific, i.e. it should apply to all language use, I still have to take into account that the <?page no="117"?> A model of the idealized discourse participant 117 focus in the research is on the effects of pornography on its consumers and that the model must thus describe discourse comprehension rather than discourse production. Cognitive theories of discourse are usually either result-oriented or process-oriented (for a general discussion of the distinction, cf. Alderson/ Urquhart 1984a). Result-oriented theories examine what mental representations the processing of linguistic elements result in. They are thus concerned with representations of meaning. Processoriented theories look into the cognitive processes that occur in discourse comprehension, e.g. which inferences are made under certain conditions and how they are made. In other words, a process-orientation means looking at how representations of meanings are arrived at. 4 My model wants to straddle theories of both mental representation of meaning and theories of discourse processes and present them in a unified framework. The focus is, of course, on the result, but always in relation to the process. The model has two intersecting planes. The first is concerned with different types of meanings or functions that elements in a text can have. The ideational meaning (I will not include references here as section 2.1.4 will be dedicated to the background of the model) of a text refers to the ideas of the world that are conveyed. Under the contextual meaning of a text I subsume all meanings that make reference to the situation of the discourse, i.e. to the context, rather than to the world that is represented. This includes the interactive meaning, i.e. the use of texts in acting with and upon other people and establishing relationships with them, and the evaluative meaning, i.e. the evaluations of ideational meanings according to different criteria such as likelihood and desirability. The information structure of the text, finally, organizes the information in the other two realms, arranging it in certain ways and according to criteria such as importance, newness, or perspective and relating it to other information. The second plane posits two representational levels of discourse with two equivalent processing stages. In the first stage, each linguistic item is directly assigned a semantic value (of one of the three types), resulting in a schematic representation, the semantic grid. In a second stage, this grid is enriched by different kinds of inferences, forming a fully-fledged representation, the full scenario. The first stage is thus characterized by a correspondence between meaning and linguistic 4 The difference becomes evident in the different conceptions of cognitive linguistics presented, for instance, in the textbooks by Ungerer/ Schmid 1996 and Rickheit/ Strohner 1993. <?page no="118"?> Theoretical Background 118 structure, the second stage by a constructionist relation as the semantic representation is actively constructed by the hearer 5 in interaction between the linguistic structure of the text and her or his background knowledge. Table 6 represents the model. As the two planes are intersecting I am left with a matrix with six fields. As we can divide the contextual world into an interactive and an evaluative part, the relevant box features two categories (cf. Fairclough’s (1992a: 64) distinction between the identity and the relational function of language). Semantic grid Full scenario Ideational meaning Conceptual structure Textual world Contextual meaning Interactive structure and evaluative structure Contextual world Information structure Local information structure Global information structure Table 6: Schematic representation of the linguistic model of discourse processing. All six categories contribute to the overall interpretations of texts in discourse and are thus worth analysing. I will describe them in the following sections. 2.1.1 Semantic grid The semantic grid consists of schematic representations of ideational meaning, contextual meaning and the information structure. Every aspect of the linguistic structure of the text is assigned an element in one of the three areas. I will briefly describe what kind of representing elements exist and to which linguistic items they are related. Conceptual structure The conceptual structure is a hierarchical network of concepts and propositions. Concepts with their different internal structures enter complex relationships with other concepts to form propositions. Propositions are units of meaning that, unlike concepts on their own, can give information about what is the case in the world or, in terms of logic, can be true or false (cf. Saeed 1997: 14f.). Propositions themselves form 5 I use the general terms speaker/ hearer in connection with the model. They, of course, include writers and readers of all kinds. <?page no="119"?> A model of the idealized discourse participant 119 links to other propositions by virtue of sharing elements or being causally, temporally, contrastively or otherwise connected. 6 The conceptual structure corresponds to lexemes and their grammatical connections, by way of which they form larger complexes (particularly phrases, clauses, and sentences), and to the grammatical morphemes that attach to lexemes. He leaned toward me and slowly massaged my large breasts. In a sentence such as this one, analysing the conceptual structure means dissecting the meanings of the lexemes (e.g. slowly, to massage, large, breast), including their word type status (e.g. the fact that massage is a verb here and not a noun), the relationship of the noun breast with the plural and that of the verbs lean and massage with past tense and simple aspect, the connection of the adjective large and the noun breasts, the type of processes the two verbs describe and the semantic roles the noun phrases he, my large breasts, the prepositional phrase toward me, and the adverb phrase slowly play in the two propositions, and the relationships between the two propositions signalled by and. Interactive structure The interactive structure features aspects that provide the basic foundation of the interactive functions of discourse (getting into contact, establishing and maintaining relations, acting together with or upon each other), particularly speech acts. The linguistic items realizing the interactive structure are sentence types, performative verbs and modal verbs (although they also play a role in the evaluative structure; see below). We might also include certain aspects of conversational behaviour, e.g. turn-taking, although it is not clear whether we can say that these aspects are directly signalled by linguistic items. Evaluative structure The evaluative structure consists of the elements that modify the conceptual structure with reference to the speaker’s and/ or the hearer’s judgement about likelihood, desirability, necessity, obligation, source of information, value, etc. The evaluative structure is created in analogy to special verbs (modal verbs such as should or must), certain adverbs (such as hopefully or presumably), disjunctive elements (e.g. There are, I 6 Many researchers posit such propositional networks as the basis of the representation of the meanings of texts (cf. van Dijk/ Kintsch 1983, Harley 2008: 379f., Fletcher/ van den Broek/ Arthur 1996: 142). <?page no="120"?> Theoretical Background 120 guess, many…) (these aspects are often subsumed under the grammatical system of modality, although the focus is usually on modal auxiliaries, cf. Lyons 1977: ch. 17, Palmer 1986), hedges (e.g. this sort of surprised me), and also to the non-descriptive aspects of lexeme meanings (their affective connotations, in traditional terminology, e.g. when ass is used instead of anus). We may also include prosodic patterns and typographical elements such as quotation marks or inverted commas. Local information structure The local information structure organizes information according to criteria of newness, givenness, importance, perspective, and connections to information included (or assumed to be included) somewhere else in the text or otherwise available to the hearer. This involves the speaker’s judgement about the hearer’s state of mind and a judgement of the hearer about her or his own assumed state of mind. The local information structure features aspects such as what we take to be the topic of clauses and what the comment on the topic (thematic structure or theme/ rheme structure), which information is explicitly stated and which is treated as given (assertion vs. presuppositions), and which referents are treated as new and which as identifiable and easily accessible (identifiability and activation). The local information structure corresponds to syntactic patterning, grammatical items (e.g. pronouns, articles, etc.) and certain adverbs/ adverbials (e.g. first, then, etc.). He leaned toward me and slowly massaged my large breasts. In this sentence, for example, looking at the local information structure means analysing the information into theme (he) and rheme (leaned toward me and slowly massaged my large breasts), examining the use of the personal pronouns he and me and the presupposed existence of breasts and their presupposed size (this information is not explicitly asserted but backgrounded in an attributive adjective phrase). 2.1.2 Full scenario The full scenario is the set of fully-fledged and coherent representations of ideational meaning, contextual meaning, and the information structure. To arrive at such representations the semantic grid is augmented by inferences, so that we get the formula: semantic grid + inferences = full scenario <?page no="121"?> A model of the idealized discourse participant 121 Inferencing is triggered by text and draws upon representations in long term memory, stored in the form of schemata. Inferences allow us to add or complement information, partly in order to fill the gaps on the grid level, to make predictions, organize information, and dissolve ambiguity. Following the three-part distinction of meanings, the full scenario consists, as a matter of fact, of three scenarios. Textual world, contextual world, and global information structure The textual world is the full representation of the states of affairs that we construct in our minds as hearers with the help of inferences (and the main bulk of inferences are concerned with the textual world). In other words, it is the full representation of ideational meaning. But we also build up a full picture of the current discourse situation, of the contextual world. It includes information about the setting, about the participants, with their social roles, their social relationships, their background knowledge and attitudes, their goals in the current discourse, and about the ongoing discursive action. To construct the contextual world, we need to infer aspects such as indirect speaker intention - this would include things such as indirect speech acts, irony, or politeness, which never directly spring from the wordings themselves - interactive patterns, and not immediately visible attitudes and judgements. The construction of the contextual world does not solely rely on textual data but we also draw upon extratextual information (anything that is situationally given but is not contained in the text - this is of more relevance in face-to-face interaction simply because more immediate context is shared by participants), which gives rise to further inferencing. The global information structure, finally, organizes information of the discourse as a whole. There are also inferences on this level, concerning the general and global organization of information in the textual and the contextual worlds with aspects such as overall topic, general focus, global structure, general perspective, etc. The most important aspect of the global information structure is salience. Salient information or salient elements of information take pivotal positions in the interpretation, frequently recurring in many different shapes and being linked to various other information. Such items of information are therefore more likely to enter semantic memory than non-salient ones, encompassing three of the five conditions for transition in chapter III, section 1.1.1 (frequency, variation, interconnectedness). Despite a certain conceptual affinity, salience does not necessarily correlate with conscious attention, often being taken for granted rather than being focused on. <?page no="122"?> Theoretical Background 122 I stress salience here because it is a key concept in corpus-based CDA. Only if an element is shown to gain some level of salience can it be argued to become established in discourse participants’ semantic memory beyond temporary significance (see also 2.1.3 below). Although salience, depending to a certain extent on individual background knowledge, does not correspond to specific linguistic structures, there are, as mentioned, three textual factors that contribute to its creation, viz. frequency, variation, and interconnectedness of certain elements or structures, which ensure that information associated with these is maintained in the mind for longer or more frequent periods of time, from which we will then infer that the information must be central. There are different aspects that play a role in the distinction of different types of inferences (e.g. conversational and conventional implicatures, forward and backward inferences, elaborate and minimal inferences, etc.). These categories, the majority of which is primarily of psycholinguistic relevance, point to the diversity of aspects that have to be taken into account when looking at inferencing. This diversity also suggests that the inferences will not be as quantifiable and systematically researchable as the elements of the semantic grid and they will therefore not feature as prominently in my analyses. But I will discuss one particular type, viz. intertextual inferences, because of their special status in CDA. Intertextual (or interdiscursive) inferences In intertextual inferences, the basic knowledge structures are representations of discourses (in the sense of different types of discourse). Discourses can be distinguished on three levels. a. Register: according to the subject-matter and the perspective on it (including such aspects as attitude and expertise), e.g. traditional medical register versus homeopathic register. Mind that this is what Fairclough (1992a: 127f.) calls discourse. Having defined discourse as the superordinate concept to genres, registers, and styles, I use the term register, in a meaning which is close enough to the one Coulthard (1985: 40) assigns to it. b. Genre: according to the type of activity performed with/ in the discourse and the general structure it entails, e.g. interview versus news report (cf. Fairclough 1992a: 126, see also Kintsch and van Dijk’s (1983) concept of superstructure). c. Style: according to the social relationship between participants, e.g. formal versus informal style (cf. Trudgill 1974: 106f.) <?page no="123"?> A model of the idealized discourse participant 123 As the definitions show, the three concepts can be assigned to the functions posited: registers to ideational meaning and the evaluative structure, genres to the information structure, and styles to the interactive structure. Although registers, genres and styles tend to occur in fixed patterns together - these patterns are what I have called discourses - there is still the possibility of independent variation. We could, for instance, write a linguistic article in a more or less formal style, we can write a formal scientific article on linguistics or psychology, or we can give a formal linguistic talk or write a formal linguistic review. Once any of these patterns has been activated in the hearers, it allows them to draw inferences about structure, the suggested attitude towards the subject matter, about the identity of the speaker, etc., all elements that will lead them to interpret the current text/ discourse in a particular way. Critical Discourse Analysis, following particularly the ideas of Bakhtin (1994), assumes that when using language people often draw upon different genres, registers, and styles so that most discourses actually represent a heterogeneous mix of voices, sometimes leading to contradictory interpretations and generally indicating an ideological problem in a certain area. The heterogeneity also points to ongoing discursive and thus also social changes (cf. Fairclough 1992a). There are, not surprisingly, a host of unresolved issues with regard to a model such as the one presented above. The major problem perhaps is that the model is not able to deal with more delicate aspects of the resultprocess interface. Critical Discourse Analysis assigns semantic values to all linguistic items, assuming that they might all contribute to the way the information is represented in the hearers, particularly in their semantic memories. On-line processing, however, is severely limited: our memory buffers cannot hold an infinite number of elements, which means that representations are sometimes just transient states of working memory: they are created, are used in further interpretation processes, and then they are discarded with only some aspects being retained. The meanings of certain linguistic elements might consequently be just temporarily represented but might be lost as the comprehension process progresses. My model postulates a few principles supposed to facilitate transition from discourse meaning to semantic knowledge, but to be honest, it does not have to say a lot about how we can deal with these on-line constraints. On the other hand, we also have to bear in mind that it is far from clear what it means for information to get lost. It may, for instance, be backgrounded, still having an impact on long-term memory representations <?page no="124"?> Theoretical Background 124 My model does not offer a solution to the problem of the depth of interpretation, i.e. how much of the internal meaning of lexemes is really represented. As linguists we, for example, easily find a plethora of conceptual differences between the meanings of make love and fuck. But it is not clear whether pornography consumers in the ‘heat’ of the comprehension process really distinguish that profoundly between the two meanings. A further intricacy is that there are phenomena that defy a clear category membership. Deictic elements, for instance, express content belonging to the textual world but also content that belongs to the contextual world. Metaphors, on the other hand, are part of the conceptual structure and thus of the semantic grid but probably also require inferencing. These two examples already show that we cannot assume the meaning categories posited to be fully distinct spheres without interfaces. There will be further research in psycholinguistics and cognitive semantics that may resolve some of the issues so that a model such as mine can be further enhanced. For the time being, however, I have to cope with a less than perfect version of it. 2.1.3 From interpretation to semantic memory I have already mentioned salience, which is tightly linked to frequency, variation and interconnectedness, as an important factor enabling the transition from interpretation to semantic memory. Further principles potentially instrumental in this, as has already been pointed out in chapter III (section 1.1.1), are integratability of information and credibility of the source. The former depends on the socially-shared resources available to the discourse participants, the second one on their assessment of a person or a medium. Both therefore have to be dealt with in the social model of the idealized discourse participant. 7 2.1.4 Background My framework is axiomatic in nature, which means the assumptions inherent in its conception will not be put to test in the project. It is, however, based on a review of the literature on approaches to semanti- 7 It could be argued that whether information is conforming or contradicting depends on the individual knowledge of the discourse participants. As I am dealing with the idealized discourse participant, I, however, abstract from concrete individual knowledge and see it rather as a socially shared resource. <?page no="125"?> A model of the idealized discourse participant 125 cally-oriented grammar and cognitive discourse comprehension. I will briefly mention the two most prominent sources that have influenced me. The idea of the three types of meaning is borrowed from Hallidayan Systemic Functional Linguistics (cf. Halliday 1994, Eggins 1994) with its ideational, interpersonal (~ contextual meanings) and textual (~ information structure) metafunctions. Language users make choices in the three areas which eventually result in the surface form that they produce. The main difference to my framework is that my metafunctions are understood more in a discourse-related than in a system-inherent sense (and that they are slightly more receptionthan production-oriented). Even if Halliday’s framework is the most elaborate spelling out of different types of meaning, there are other models which also incorporate at least the fundamental opposition between ideational and contextual meanings. The distinction corresponds, e.g., to Lambrecht’s (1994) between text-internal world and text-external world, or to Werth’s (1999) between text world and discourse world. The distinction between a corresponding level and a constructed level in discourse processing draws upon a similar multi-level approach most prominently proposed by van Dijk and Kintsch in Strategies of Discourse Comprehension (1983). For them, the meaning of discourse is represented in two forms: the first one is the text base and the second one the situational model (or mental model, as it is more commonly called by other researchers, cf. Johnson-Laird 1983, Rayner/ Pollatsek 1989: 313ff., Garnham/ Oakhill 1996, Carroll 1999: 170ff.). While the former represents the explicit information of the text in propositional form (similar to the semantic grid), the mental model is an analogical, picture-like representation of the text base enriched by inferences from background knowledge (similar to the full scenario). My framework primarily inherits the multi-representation dimension. It does, however, not make specific claims about the nature of the second level, which is why I have not termed it mental model. It furthermore fully integrates level 1 into level 2, while real mental models just adopt selected information from the text base. 2.2 Social model of the idealized discourse participant Many aspects of discourse interpretation depend on background knowledge, purpose of participation, attitudes, etc. These aspects cannot be described generally but have to be specified for the particular discourse under investigation. This means that I will characterize the idealized discourse participant as the prototypical pornography consumer (and <?page no="126"?> Theoretical Background 126 then also the prototypical erotica consumer), describing him in social terms. As I have said, this aspect would deserve a separate stage in the research, which, however, could not be realized in the present study. So the following description is rather a common sense picture which is based on the important assumption that there are strong parallels between the implied reader, i.e. the hearer the whole discourse creates, and the actual readers. The typical pornography consumer is male. Despite all claims to the contrary (cf. Strossen 1995: 144), I have not gained the impression that sex shops or magazines make a strong effort to address females and couples in the same way as men (this is the reason why the pornography corpus just contains what can be classified as male-oriented material at face value). As far as age and marital (relationship) status is concerned, there is some evidence (cf. Hardy 1998: 102ff.) that younger men not yet in a stable long-term relationship tend to consume more pornography than others. With regard to social class and profession, the picture is heterogeneous, though the whole presentation of some of the magazines conveys the impression that they are not targeted at a middle-class readership. But this might make it easier for middle-class men to buy it, as is the case with Hustler (the most notorious of the large-format pornography magazines), which many middle-class men read because they feel they are not meant anyway (cf. Dines 1998a). And it is sometimes claimed that pornography is a middle-class phenomenon with middle-class consumership (cf. Tisdale 1995: 96), which might be particularly true for pornographic short stories. Unlike other pornographic material, especially videos, which can be consumed with a partner or in a group of males, reading short stories is a solitary practice. It occurs in autoerotic activity - the term onehanded reading found in some magazines in their self-praise alludes to this - or serves as stimulation to the latter. Consumption may partly be excessive. There might be some readers who hold different views from the ones inherent in the stories and whose consumption of porn will lead to an internal conflict. But I suspect that generally the material will be received in an affirmative and non-critical way, with many aspects of the conceptualization of sexuality and society corresponding to the consumer’s own assumptions. Although it is often claimed by pro-pornography camps that pornography users are not passive victims, but have the chance to critically reflect upon the ideas imposed on them, I do not believe that even a large minority subject what they read to profound and <?page no="127"?> A model of the idealized discourse participant 127 lengthy critical reviewing (otherwise they would be unlikely to go on buying the magazines). As far as the relation of pornographic short stories to other discourses is concerned, I think that the typical reader will also turn to other pornographic material, particularly to large-format magazines such as Penthouse, Hustler, or Playboy, to videos and to the Internet, i.e. to discourses that will basically enhance the effects of the stories. I assume that he will not read both pornography and erotica. Other primarily sexual discourses such as sexological, hermeneutic, anthropological or psychoanalytical studies have an audience extremely limited in size and can thus be neglected. To what extent sexuality features as a topic in other discourses, e.g. personal conversations with partners or friends, medical interviews, TV talk shows, films, fiction, etc., is very difficult to judge. I think, however, that for the typical reader of pornographic short stories, pornography in general constitutes the quantitatively most prominent discourse on sexuality. I further assume that what is read in the short stories will rarely feature in other discourses. Readers will not normally talk about their experience with pornography to other people, unless in an apparently approving group (of other males). Pornography is thus not exposed to social and open critique either. Although these assumptions will not help in specifying the typical interpretative processes, what they suggest is that reading pornography is actually an act of consumption primarily intended for immediate pleasure and unlikely to critically challenge or question the beliefs and values suggested by the text, which points to a high integratability of most information. As for the credibility of the source, I assume that although most pornography users will be aware of the fictional nature of what they read, they will still consider it to relate to their own experience and thus to be realistic to some extent, otherwise a function in stimulation is hardly conceivable. The source of information, i.e. pornography, thus enjoys a certain level of credibility, which is probably enhanced by the encapsulated and isolated nature of its consumption, preventing it from being exposed to critique. As far as the typical erotica consumer is concerned, I assume her to be female and probably middle class. Her consumption of erotica will never be excessive and it will not be immediately related to sexual activity. Like the pornography consumer, she will accept much of the content in an affirmative and not too critical way, again because many of the assumptions will correspond to her world views anyway. Erotica might very well represent the only form of sexual entertainment she consumes, which means erotica readers do not necessarily watch films <?page no="128"?> Theoretical Background 128 with sexual content and will definitely not consume pornography. Finally, I expect the typical erotica consumer to be more open about her consumption, talking about it with partners and friends. This is, of course, easier as there are fewer taboos on erotica than on pornography. In sum, interpreting erotica is probably less an act of consumption and readers might therefore approach the beliefs and values more critically. As the latter, however, probably fit into their views of the world, such critique will not lead to a strong re-evaluation of the interpretations suggested by the texts. The information will thus also reach a high level of integratability, which will facilitate its access to semantic knowledge. As regards erotica’s credibility as a source, I assume that despite their fictional nature, readers will consider it to be a more authentic representation of sexuality than other, purely commercial sexually explicit material. This will also lower the threshold level for information to pass on. 3 Combining corpus analysis and the idealized discourse participant model I will now briefly describe which aspects of this model can play a part in a corpus linguistic approach to the analysis of texts and will consequently feature in the chapters of Part 2 of this book. To appreciate the selective focus that corpus research places on the use of the model, it is necessary to make a crucial distinction in textual analysis. Textual analysis can basically take two routes. In transtextual analysis we examine individual elements across long stretches of text. It thus is the selective analysis of a large set of data. Co-textual analysis, on the other hand, means that we analyse how elements interact with each other within a usually short passage of continuous text. It thus is the exhaustive and in-depth analysis of a selective set of data. These definitions indicate that the two types of analyses are suited for different tasks, transtextual analysis for quantitative corpus research, laying more emphasis on the amount of data, co-textual analysis for the qualitative analysis of individual texts (such as speeches, etc.), laying more emphasis on the amount of elements analysed. The two types, however, merely represent two poles of a scale with intermediate forms being possible. Co-textual analysis may be considered to simulate real discourse processes more naturally, being concerned with the whole of a text. It is, however, also more prone to subjective readings, as the simulated inter- <?page no="129"?> Corpus analysis and the idealized discourse participant model 129 pretative processes will be influenced by the researcher’s own interpretations. In transtextual analysis, such subjective interference is more likely to be suspended because it does not look at texts the way we normally encounter them, i.e. as linear. Besides, my interest in the category of pornography (and that of erotica) rather than in the individual text also prevents me from tackling the problem from a co-textual starting point. As my project is corpus-based, the analysis is generally transtextual in nature. But I will also look for combinations and patterns of elements across the whole corpus, an aspect introducing a co-textual perspective. Having the co-texts of troublesome elements available in concordances will also help me in the analysis, in particular in decisions about classifications (e.g. whether to count an adjective such as fair as a colour or a psychology descriptor). Opting for a corpus analysis and its transtextual orientation implies that it is not possible to examine all the levels included in the linguistic model of the idealized discourse participant. The latter thus does not automatically translate into a model of analysis, with all its components corresponding to separate steps in the investigation, as would be possible in an exhaustive co-textual approach. As a consequence, the model of the idealized discourse participant is rather a table with analytical options from which I have to select those that appear to be most interesting and viable with respect to the hypotheses. The transtextual orientation also entails that I will focus primarily on the semantic grid. With its strict association of semantic aspects and linguistic forms, it is easiest to research with corpus methods. The conceptual structure probably is the primary object of analysis as it contains most information and as any aspect might be of interest. But aspects of the evaluative structure, e.g. elements of modality, and of the local information structure, e.g. presuppositions, also lend themselves to a transtextual analysis with the help of corpus methods, allowing insights into proposed attitudes towards certain subjects or into pieces of information presented as taken for granted, respectively. Aspects of the full scenario, on the other hand, will not feature prominently in my analysis because due to the independence of the majority of inferences from formal aspects, they are very difficult to track down in a corpus. I am aware that inferences have a high priority in the actual interpretation in discourse and that, as a result, the analysis will remain ‘skeletal’, characterizing the grid but not the full picture. There is, however, one aspect concerned with inferences that will play a pivotal part, namely salience, which due to the role that the observable aspects of frequency, variation and interconnectedness play in <?page no="130"?> Theoretical Background 130 its creation, can be traced in a transtextual corpus analysis. Salience is so important because only meanings that are shown to be assigned a salient status can be argued to have a more permanent relevance for pornography readers’ knowledge. I will therefore always look at the frequencies, variation and - to a lesser extent - interconnectedness of the linguistic elements under examination to see whether the meanings they create will also gain salience. To conclude, the analyses will primarily start out with the semantic grid of the interpretations - i.e. the conceptual structure and, to a lesser extent, the evaluative structure and the local information structure - speculating how elements there contribute to particular conceptualizations. I will then check frequencies, variation and interconnectedness to see how salient the said elements are, interpreting their impact on the semantic knowledge of readers. 4 The hypotheses The hypotheses to be tested against the data from the corpus are assumptions about the beliefs and attitudes likely to be created or supported in the typical consumer through exposure to pornography or erotica. As mentioned in the previous chapter, I will focus on global conceptualizations, i.e. general beliefs and attitudes organizing more specific ones. They usually do not emerge through individual details but only come to the fore as trends over larger stretches of text, which is another reason for transtextual analysis. Taking a feminist perspective, my hypotheses will focus on conceptualizations supposed to be socially most problematic for the relations between women and men, giving rise to inequality and discrimination and, by blurring the preferences and limits of sexual partners, to harm and violence. The idea most often named in this context is objectification. Subordination is another one. But as its aspects are often subsumed under objectification, I will not treat it separately in this book. Despite its pivotal role in the discussion, the meaning of objectification has remained blurred and fuzzy. Andrea Dworkin and Catharine MacKinnon (1988: 36, cit. in Russell 1998: 6) say that it is created if “[…] women are presented dehumanized as sexual objects, things, or commodities.” Russell (1998: 7) defines it as “[…] the portrayal of human beings - usually women - as depersonalized sexual things, such as ‘tits, cunt, and ass,’ not as multifaceted human beings deserving equal rights with men.” And Jensen and Dines (1998: 93) describe it when saying that “women are routinely presented as objects for the visual or <?page no="131"?> The hypotheses 131 sexual consumption of men,” while men are “the sexual subjects, the beings with agency who [control] and [direct] activity.” These definitions show that objectification is not a unified, homogeneous conceptualization itself, but rather serves as an umbrella term, covering a number of different, though related ideas which all centre on different aspects of the concept of object. Firstly, an object is conceived of as not on the same level as a whole human being. Objectification here means not perceiving humans as wholes but segmenting them into parts, a process that can be called fragmentation. Secondly, an object is taken to represent physical - and primarily visual - matter. Objectification in this sense is focusing on the physical and visual properties of humans, for which I use the terms physicalization and visualization. Thirdly, objects are also thought of as being opposed to subjects and objectification consequently means reducing subjectivity, which includes not adopting someone’s perspective and backgrounding their inner lives. This form of objectification can be termed desubjectification. Fourthly, objecthood is also often associated with being at the receiving end of acts. Objectification thus also amounts to conceptualizing humans as passive beings (reflected in lexemes such as plaything, toy, commodity - as opposed to control and direct - which can regularly be found in accounts of women’s roles in pornography, see, for example, the definitions above), a dimension which I call passivization. Although other aspects also qualify as objectification, e.g. reification, i.e. the conceptualizing of humans in terms of things, plants, or animals, or deindividualization, i.e. the conceptualizing of human individuals in terms of non-individual classes, it is the four aspects mentioned above which do not only seem to emerge as the most central forms of objectification, but which also can be researched most thoroughly. The four global conceptualizations to be examined in the four chapters of Part 2 thus are: 1. Fragmentation (chapter V) = the tendency to conceptualize humans in terms of their bodies and their body parts. 2. Physicalization and visualization (chapter VI) = the tendency to conceptualize humans in terms of their physical and in particular their visual features. 3. Desubjectification (chapter VII) = the tendency not to conceptualize humans as thinking, feeling and perceiving subjects. <?page no="132"?> Theoretical Background 132 4. Passivization (chapter VIII) = the tendency to conceptualize humans as passive recipients of events rather than as active participants in them. The four conceptualizations on their own do not constitute testable hypotheses but just form the areas for which we can formulate assumptions. These assumptions are concerned with the questions whether the conceptualizing trend applies more to women than to men and how pornography compares to erotica with respect to gender differences. In order to see whether a conceptualization generally figures in a discourse - if it hardly figures at all, finding differences might be a futile undertaking - and in order not to focus exclusively on the female-male opposition, I will examine whether the discriminatory conceptualizations operate on a general level, too (e.g. whether there is a substantial amount of fragmentation of humans) and whether there are differences between pornography and erotica with respect to this (mind, however, that this is not relevant for all conceptualizations). As the questions are similar for all four aspects to be studied, the assumptions can be formulated as a set of schematic hypotheses. These are: • with regard to pornography only: a. The global conceptualization will figure prominently in pornography. 8 b. The global conceptualization will apply to women more than to men in pornography. • with regard to the comparison between pornography and erotica: c. The global conceptualization will generally figure less prominently in erotica. d. Women will not be subjected to the global conceptualization more strongly than men in the erotica or at least the gap between the genders will not be as wide as in pornography. These schematic assumptions have to be translated into more specific predictions for the respective aspect examined. 8 If I use in pornography (or in erotica), I usually mean ‘created in the idealized pornography readers’ minds or in pornography consumption’. So it is supposed to point to the receptive/ interpretative side of the discourse of pornography and not simply to the text. I will sometimes use this formulation in this book in order to avoid the clumsier one. <?page no="133"?> Summing up 133 The aforementioned feminist perspective has influenced me in my choice of hypotheses, but it does not predetermine the outcome of the research, because it lies in the nature of this investigation that its assumption may be refuted or may at least not be supported. 5 Summing up The main questions of the research project are whether women and men are represented and will consequently be interpreted as fragmented entities rather than as whole persons (fragmentation), as physical and visual (physicalization and visualization) rather than as sentient and social (desubjectification), and as passive recipients rather than as active agents of sexuality (passivization). These questions have been distilled as central problematic issues from the pornography debate about women’s objectification and the answers may thus be of interest and relevance for the ongoing discussion. I will test whether these semantic tendencies are likely to be included in the interpretations of pornography consumers. I will not examine the consumers themselves but will simulate their comprehension process of concrete text. The texts examined are a 680,000-word corpus of pornographic short stories, analysed with concordancing software. The analysis of the pornography corpus will be compared to those of a similar corpus of erotica (240,000 words). Qualitative ideas will be combined with quantitative methods, which primarily draw upon basic tools of descriptive statistics. The simulation of the comprehension will be based on theoretical assumptions about the interpretation of certain linguistic items that can be integrated into a large model of cognitive discourse comprehension and on general assumptions concerning social features of pornography consumers. <?page no="135"?> P ART 2: A NALYSES <?page no="137"?> Introduction to Part 2 Each of the following four chapters will take up one of the hypotheses centring on a global conceptualization mentioned at the end of Part 1 and test it against linguistic data from the corpora. I would like to shortly outline - and explain, wherever necessary - the basic structure of the chapters in addition to pointing to a few peculiarities. At the beginning of each chapter, I will introduce the global conceptualization under examination in all its facets. In addition to in-depth definitions of the conceptualization and its subaspects, I will explain its wider cognitive, social and political implications. I will also trace its roots in the pornography debate, providing background references wherever possible. The introduction will end in a preview of the linguistic elements and structures to be dealt with and in a general formulation of the hypotheses for the respective chapter, following the schematic hypotheses formulated above. All chapters consist of two subchapters each focusing on a particular linguistic element or structure. These subchapters will first provide a thorough discussion of the relevant linguistic background and of the importance of the elements or structures for the global conceptualization. This will then lead me to translating the general hypotheses into concrete predictions about the linguistic data (to avoid confusion, I will not speak of hypotheses in connection with the latter but rather of predictions or expectations although, to be strict, they are simply more concrete hypotheses). Following this theoretical account of the language aspect to be examined, a special section named “Technical and practical aspects of the analysis” will provide background information on the progress from the raw data to the processed data. This includes precise accounts of the actual corpus searches, how quantitative relations are derived from the concordances and lists that the latter yield, and which analytical categories are used. Although possibly boring to the non-specialist, such a section, in my view, is necessary to document that the data to be finally analysed has not been conjured up by some incomprehensible and nonintersubjective trickery. A “Presenting and discussing results” section will then present, describe and interpret the relevant data from the pornography corpus. The one following, “Comparing and discussing results”, will do the same <?page no="138"?> Analyses 138 with the data from the erotica corpus, always from a contrastive perspective. Data interpretation will not be limited to examining whether my expectations have been fulfilled or not but will also - if necessary - involve considering other factors possibly influencing the results. This may include: • Conceptual factors: The conceptualization under examination may be a multi-faceted phenomenon interacting with other conceptualizations. These subaspects and interactions may at times distort data. If fragmentation and passivization, for example, play a role in the use of a certain linguistic element, then this may lead to contradictory results. Or if the majority of the stories feature a female first person narrator, this may also have an impact on who is being passivized or fragmented. • Linguistic factors: There may be language-internal (the borderline to conceptual factors is not clearly definable) aspects influencing the linguistic structures under examination. If I, for instance, examine the differences in the semantic roles that fuck and make love take, I of course have to take into account that the most idiomatic role for the latter is a directive (due to the colligation with the preposition to). • Methodological factors: The results may be influenced by my way of collecting, categorizing or quantifying data. Analysing data sets of grossly different sizes (e.g. the words used for female and male breasts) may, for instance, produce inexplicably diverging results. As will become clear very soon, the analytical part of my study is dataheavy. This means that despite my restricted focus on particular linguistic elements and structures, the corpus searches will provide such a wealth of data - both qualitatively and quantitatively - that any attempt at an exhaustive examination will prove futile. The actual analyses will thus only concentrate on a very limited set of aspects. To do full justice to the richness of the data, I have decided to indirectly include it in my presentation. Indirectly means that readers can find all relevant data on a special website, complete with frequencies and discussions of problematic cases (at http: / / www.uni-graz.at/ georg.marko). Making the data accessible will allow readers to have a thorough look and draw their own conclusions on aspects not touched upon. It will also provide a more comprehensive description of the language of pornography (and of erotica), beyond what is actually analysed. And finally, in agreement <?page no="139"?> Introduction to Part 2 139 with what I proposed in chapter III, including the data will also open up my approach to critique, allowing readers to immediately take issue with, for instance, particular categorizations. To present the data in tables and diagrams is a demanding if not an impossible task given the former’s richness and complexity. Although I will try to be as consistent as possible, it still is often necessary to adapt font sizes, to use abbreviations, and to omit percentage signs to find a balance between layout-related restrictions and a synoptical form of presentation. The complexity of the data also makes it necessary to have clumsy captions. The penultimate section of each chapter, termed “Casting a side glance: phenomena not considered”, functions as a kind of dustbin, containing snippets of ideas and data that invites - even if just at first sideway sight - further research. All chapters end in a “Conclusion” section, reviewing the hypotheses in the light of the evidence yielded by the individual analyses. The vast majority of the examples cited to illustrate a point in the introduction to the linguistic structures and elements have been taken from the corpora. This is supposed to allow further glimpses into the language of pornography (and erotica) and to show that the aspects discussed can be exemplified by authentic data. All chapters of Part 2 of this book will contain analyses of many different but similar sets of data with respect to the same questions. Despite my attempt to achieve some ‘elegant variation’ by using different expressions and constructions in my descriptions and discussions (which, in turn, may earn me some criticism for not establishing unambiguous technical vocabulary), this will convey the impression that I am repeating myself. This effect on readers is, however, almost inevitable and seems to be part of the costs I have to pay for opting for the stony path of scientific methodology. <?page no="141"?> V Fragmentation She’s the hottest piece of ass a man could ever wish for She was a hell of a piece If the whole is more than its parts, as a well-known saying has it, then the parts are necessarily less than the whole. Transferred to human beings, this means that only the whole person can be considered a complete human subject while her or his material body and in particular her or his body parts have the status of objects. Thinking of persons in terms of their physical components, i.e. fragmentation (understood in this restricted material sense here), thus segments the whole into objects and if pornography manages to create this in its consumers then it is justly regarded as objectifying. The objectifying potential - and its discriminatory edge - can be assumed to be further increased if fragmentation involves a consistent focus on sexual organs (in a broad sense - this includes all parts that are presented as relevant for sexual activity, i.e. the genitals, the anus and the breasts). It is thus not merely fragmentation but also sexual fragmentation that is at issue in this chapter. What are the negative social consequences of thinking of humans in terms of their body parts rather than as wholes? Fragmenting humans means seeing the parts as equal to but also as independent from the whole, which will have implications for interpersonal relationships. Body parts themselves have no emotions and preferences. We can consequently treat them as objects serving the purposes that we attribute to them, and we just have to consider their functionality but not their thinking and feeling. And with regard to women’s fragmentation in sexuality, this functionality lies in the satisfying potential of their bodies to their lovers with the woman as a whole individual not being important. If this is part of the male consumers’ conception of sex, it can at least affect the level of satisfaction that women can achieve in their sex lives with men, but it might also influence the latter’s general behaviour towards women, reducing their ability to empathize with them. In addition to merging fragmentation and sexuality, sexual fragmentation may also be considered to reorganize our thinking of sexuality in general, pointing to a view that overemphasizes the role of the genitals at the cost of other sensitive body regions, which does not correspond to - perhaps stereotypically - female conceptions of sex. <?page no="142"?> Analyses 142 The importance of fragmentation, with special emphasis on women’s fragmentation, in pornography is acknowledged by Dworkin and MacKinnon (1988: 36; cit. in Russell 1998: 6), who mention the following aspect in their definition of pornography, “[…] women’s body parts - including but not limited to vaginas, breasts, or buttocks - are exhibited such that women are reduced to those parts […].” The fact that other feminists such Kuhn (1995) and Caputi (1994: 18) also see it as an essential feature of the pornographic code of representation shows the relevance assigned to fragmentation in the feminists’ arguments in the pornography debate. This central role is also the main reason why it is included as an object of analysis in this study. The current chapter will examine linguistic elements assumed to contribute to fragmentation and sexual fragmentation in the readers’ minds, thus analysing whether and to what extent these phenomena are part of the pornographic discourse and whether gender differences can be observed. I will first study the overall salience of bodies by looking at the general frequencies and variation of lexemes denoting body parts in the first subchapter and will then examine how different types of possessives can enhance this salience in the second subchapter. Following the schematic hypotheses outlined at the end of chapter IV, my superordinate assumptions are: • There will be a tendency towards fragmentation in general and sexual fragmentation in particular in pornography. • This tendency will be stronger for women than for men in pornography • In erotica, this tendency will not be found or will be less pronounced than in pornography. • There will not be a clear gender bias with respect to fragmentation and sexual fragmentation or at least the gap from women to men will not be as wide in erotica as in pornography. 1 (Over)Representing body parts: lexeme frequency and variation Fragmented views of humans or of particular social groups develop in readers if body parts in general take salient positions in interpretation. As mentioned in chapter IV, salience will be inferred by readers from a number of different linguistic factors. The mere quantity of expressions <?page no="143"?> (Over)Representing body parts 143 for a particular area is one of them, variation in the expressions another one, as is their interconnectedness, i.e. to how many and to which other elements of information they are linked. While the next subchapter will be concerned with one aspect of interconnectedness, the present one will deal with the quantity and diversity of lexemes denoting body parts. Banally speaking, elements occurring very often can be assumed to gain greater salience than less frequent ones because readers are more often concerned with them, inferring a connection between quantity and centrality. Let me take an example from a pornographic short story to consider the effects of linguistic elements. Lifting her billowing skirt, I parted her cunt lips and plunged my cock into the tight wetness of her pussy, as the dark shadows thickened and moved in on us. The sentence contains three references to body parts (shaded: her cunt lips, my cock, her pussy) as opposed to only two references to whole persons (bold: I and us). So on a small scale, body parts proliferate and will, as a consequence, be interpreted as salient. By foregrounding 1 body parts, the example above can thus be argued to create some degree of fragmentation. This implies that simply mentioning body parts frequently in a text - or overrepresenting them, if you like - will gain them a more salient status in the information structure of the readers’ interpretation of texts, which in turn will create - or at least contribute to - a fragmented conception of people or groups of people. The effect will of course increase if this trend continues over large stretches of text (as represented by corpora). The first thing to examine in the corpora is, as a consequence, how often female and male body parts occur. Salience is not only a product of quantity and thus of token frequencies, i.e. how often words from one category occur, but also one of the variation in the expressions used and thus of type frequencies, i.e. how many different words there are in a category. For illustrating purposes, I will take a short look at two sets of words, one for the (female) hand, the other one for the clitoris. 1 I use this term in the sense of ‘making salient’. It is not supposed to indicate that there will automatically be more conscious attention. <?page no="144"?> Analyses 144 hand bliss button clitty love bud nugget bud clitty button love nub passion point bullet clitty knob love pearl pleasure button button cum button magical button pleasure-knot cherry clit hidden button nodule rosebud clit joy button nub sex button clitoris knob Table 1: Lexemes denoting the female hand (left) and the clitoris (right) in the pornography corpus. If we take the frequencies of reference to either category in the porn corpus, we see a marginal preference for the hand, which is mentioned 619 times as opposed to 616 references to the clitoris. It is, however, striking that the corpus features 26 different expressions (at least - there are also some additional borderline cases) for the clitoris, but only one for the hand. The clitoris is thus overworded (or overlexicalized; for this concept, cf. Fairclough 1989: 115, Goatly 2000: 64). The expressions given differ only in fine semantic nuances, nuances which go beyond simple ideational aspects to include metaphorical but also evaluative meanings and to evoke further, e.g. intertextual, meanings by inferences. The overwording thus suggests that there is a need to make these subtle distinctions, indicating a strong pre-occupation with this area of the female body. This will foreground the clitoris in readers’ interpretations, thus also contributing to its salience in the information structure. Only a combination of token and type frequencies can therefore reveal the full picture. The analyses will start with the question how often body parts in general are referred to and by which variety of expressions. Additionally focusing on the question of sexual fragmentation, I will also be interested in frequencies and variation in the sexually relevant areas of the body. For this purpose, I will use categories of body part lexemes to be introduced in the next section. As regards the linguistic elements instrumental in the creation of fragmentation, I make the following predictions: • The type and token frequencies of body part lexemes will be high in the pornography corpus, the majority denoting sexual organs. • The type and token frequencies of lexemes denoting female body parts will be higher than those denoting male body parts in the pornography corpus, with the gap widening in the sexually relevant categories. <?page no="145"?> (Over)Representing body parts 145 • In the erotica corpus, the type and token frequencies of body part lexemes will be lower than in the pornography corpus, particularly those of sexual organs. • The type and token frequencies of female body parts will not exceed those of male body parts in the erotica corpus or at least the gap will not be as wide as in the pornography corpus, particularly in the area of sexual organs. 1.1 Technical and practical aspects of the analysis Searching a corpus for lexemes for body parts is an intricate task because the latter are a category defined by semantic criteria only. As there are no formal features that they share apart from their being nouns, they cannot be retraced easily. […] sticking her bouncing bosom in my face and groping my growing dick. We watched the girl’s hungry mouth devour the guy’s big cock. Hal could only see Marian’s ears, small and pink as shells […] The examples above, however, show that nouns denoting body parts (boldface) almost always occur with a possessive (shaded) in English, this is, with possessive pronouns, as in the first example, or a genitive NP, 2 as in the second and third examples. This implies that the starting point for my corpus search are such possessive expressions. With the focus being primarily on gender differences, I will select those that are gender-specific and can easily be recognized as such. This includes firstly, of course, the possessive pronouns his and her (distinguished from the personal pronoun her with the help of the tags). As the examples above show, however, an exclusive concentration on these two will not suffice and instances of my will have to be included. To make the first person possessive pronoun gender-specific, I have divided the corpora into subcorpora according to the gender of the first person narrator (if there is one). Presupposing that first person pronouns are used for the narrator, this allows me also to look for the body part nouns following my in the male and the female subcorpora, applying the same procedures as with his and her with the notable 2 I will use the abbreviations NP for noun phrase, AP for adjective phrase, AdvP for adverb phrase, VP for verb phrase, and PP for prepositional phrase. <?page no="146"?> Analyses 146 exception that I ignore my if it occurs in direct speech as it does not necessarily agree with the gender of the narrator in these cases. Other possessive pronouns (including the occurrences of my in direct speech) have not been considered for practical purposes: to determine the possessor’s gender in addition to categorizing the body part would simply complicate the search beyond feasibility (in contrast to searches in later chapters, where it is possible to determine the gender from the co-text because the search phrase itself is fixed). Since body parts can also be preceded by a genitive NP (either a common noun such as the guy or a proper name such as Marian), I also include these in my search. For this purpose, I have to look for all genitives, which are tagged as separate units in the corpora, e.g. Marian<NP1> ’s<GE> ears<NN2>. I then have to divide the resulting concordance into two parts, one with the genitive NPs denoting females and one with the genitive NPs denoting males (ignoring genderunspecific cases). (The very few) examples of the construction ‘body part + of + nominal possessive (his/ hers/ mine/ Julia’s…)’ as in I was anxious to slide into that creamy snatch of hers, are also included. I am aware that body part lexemes do occur in positions not following a possessive construction, too, for instance in: A tiny waist, large breasts, outstanding bottom and long legs - as good as any beauty queen. However, as I am interested in female and male body parts I focus on those that are explicitly assigned to their ‘owner’ by a possessive construction. This will provide me with an approximate rather than with a complete list, but one that allows conclusions to be drawn on the amount of fragmentation at work. Once I have selected the possessives, I use them as search words for different concordances such as the following one with his. re<VBDR> inside<II> his<APPGE> apartment<NN1> ,<,> I<PPI VD> over<RP> to<II> his<APPGE> place<NN1> .<.> that<DD1> ambling<VVG> at<II> his<APPGE> loins<NN2> as<CSA> he<PPH PPHS1> slipped<VVD> his<APPGE> hand<NN1> round<RP> and<C <PPIS1> sucked<VVD> his<APPGE> cock<NN1> .<.> But<CCB> I <PPHS1> opened<VVD> his<APPGE> pants<NN2> with<IW> one<M e<CS> stroking<VVG> his<APPGE> juicy<JJ> cock<NN1> until PHS1> unzipped<VVD> his<APPGE> fly<NN1> and<CC> hauled<V <CS> two<MC> of<IO> his<APPGE> fingers<NN2> churned<VVD> J> grin<NN1> on<II> his<APPGE> face<NN1> .<.> <^> Be Concordance 1: Search word his in the pornography corpus (extract). <?page no="147"?> (Over)Representing body parts 147 In order to facilitate the retrieval of all body part nouns from these concordance lines (a demanding task with more than 30,000 concordance lines for the pornography corpus alone), I use WordSmith’s Collocation function, which can give me all words occurring within the co-text of the possessor, with all body part lexemes necessarily being among them. I have defined the horizon as 0/ 4: this means that all words occurring within four words following the keyword (and zero words preceding it) will be considered by the programme. This should ensure that the list contains the most important nouns, even if there are adjectives or numerals in between the search word and the collocate. Table 2 lists, as an example, the top collocations of the possessive pronoun his. and 1,051 I 418 in 325 as 254 cock 785 to 390 her 302 hands 211 my 676 he 375 into 268 hand 196 the 507 was 357 tongue 261 mouth 184 Table 2: Collocations of his in the pornography corpus (extract). As can be seen from the collocations above, I have to ‘weed out’ words irrelevant for my purpose (e.g. and, my, the, I, to, he, was, in, her, into, as) and combine singular and plural forms (the lexeme or lemma hand, e.g., covers both singular and plural and thus occurs 211 + 196 = 407 times). By doing this, I arrive at a preliminary list of lexemes for body parts following the respective possessive (his in the example above). Expressions such as mountains of flesh (‘breasts’), back door (‘anus’), feminine mystery (‘vagina’), or most secret place (‘vagina’) show that there are also multi-word lexemes for body parts. Since they are not contained in the collocation list, spotting them requires checking the concordances manually, too, modifying the collocation list if necessary (e.g. reduce the frequency of the single word flesh if there is mountain of flesh). A question raised by this procedure is what constitutes a multiword lexeme and what distinguishes it from mere combinations of lexemes. Although there are no foolproof criteria, I include most N + N combinations (e.g. love pole or passion button) and all other combinations where the modifying element is essential for the meaning of the whole (e.g. place on its own does not denote the vagina, only in combination with most secret, as in juice that flowed from her most secret place, it does). The procedures described above will provide me with separate lists for her and his, my with female first person and my with male first person, and female and male genitive NPs. These lists eventually are <?page no="148"?> Analyses 148 combined so that I end up with two lists for expressions for female and male body parts, respectively. All figures comparing women and men in the two corpora have to be treated with some caution since it is difficult to judge to what extent they are dependent on the overall occurrences of the two genders. If women, for instance, feature quantitatively more prominently in the corpus, then authors will also refer to their body parts more frequently. I therefore need a value that gives me some indication of the women/ men ratios in the corpora. This value is based on the frequencies of references to whole persons (= holistic references). For reasons of practicality, I confine myself to gender-specific words, which means the value covers figures for personal and reflexive pronouns (he, him, himself, she, her, herself, I, me, myself), proper names (first names only) and the most prominent common nouns marked for gender (man, woman, girl, boy, lady, guy, Mr., Ms./ Miss./ Mrs., Sir, Madam, wife, husband, mother, father, daughter, son, sister, brother, grandmother, grandfather, uncle, aunt, nephew, niece, girlfriend, boyfriend). 3 Even though these items do not comprehend the whole set of expressions referring to women and men as whole persons, they firstly should represent the vast majority of these and secondly they provide an approximation to the real values, thus allowing a good estimate of the quantitative relation between females and males. I compute the frequencies of these words individually and then add them up to a female and a male value. Third person pronouns and common nouns are unproblematic since their frequencies can be read off the frequency lists that WordSmith can produce (separating homonyms with the help of tags). For all first person pronouns I use the genderspecific subcorpora and their word-lists, manually excluding occurrences in direct speech. Establishing the frequencies of proper names requires searches with the help of the respective tags. The Collocation function can also produce a list of all words assigned the respective tag, in this case of all proper names. The words then have to be divided into a female and a male list revealing how often women and men are referred to by proper names. In order to gain insights into which areas of the body are represented more extensively, I divide the body and thus also the sets of lexemes into the following eight categories, based on traditional classifications of the body and on knowledge of the corpora. I will also use 3 I have omitted common nouns and proper names in the genitive because this form corresponds to the possessive pronouns, which have not been included with personal pronouns either. <?page no="149"?> (Over)Representing body parts 149 subcategories, e.g. the limb category encompasses the subclasses of arm, hand, leg, and foot. These subcategories help me to organize the data in the presentation (see my homepage), but will not feature that importantly in the analyses themselves. General: elements covering the whole body, e.g. body, form, contours, weight (they are thus not ‘parts’ in the strict sense). Transcending: covering parts of the body that cannot be defined topographically, i.e. especially tissues and body fluids, e.g. blood, veins, hormones, muscles, bones. Head: covering the head and the neck, including internal and external parts (this also applies to the following categories.). I will distinguish between parts of the face (e.g. eye, nose), the head (e.g. brain, hair) and the mouth (e.g. lips, teeth) as the three most relevant subareas of the head. Torso: covering the trunk, ranging from the shoulders to the hips (e.g. shoulders, hips, navel and heart, lungs, intestines). Limbs: covering the arm (e.g. elbow, biceps), the hand (e.g. palm, thumb), the leg (e.g. knee, shin) and the foot (e.g. toe, heel). The four parts mentioned also constitute the subcategories of limbs. Breasts: covering the breast/ s, e.g. bosom, nipple, boobs. Anus: covering the anus and the surrounding areas, e.g. ass, bottom. Genitals: covering the genitals and the surrounding areas, at least as far as they are lexically and thus conceptually distinguished from the torso and the legs, e.g. pussy, clit, dick, pubic hair, ballsack. I will discuss the overall frequencies of body part nouns in the pornography corpus and the erotica corpus first and will then concentrate on the quantitative relations between the different categories. 1.2 Presenting and discussing overall results This section will first examine the numbers of different lexemes used for body parts to see how much variation they exhibit and what conclusions the data allows with respect to salience and fragmentation. Table 3 gives the relevant figures. <?page no="150"?> Analyses 150 Types Overall Absolute number 615 399 1,014 Table 3: Numbers of different lexemes for body parts in the pornography corpus. The fact that overall I have found more than a thousand different expressions 4 for the human anatomy and its components supports the expectation of high lexical variation in this area. It suggests that the body parts will indeed become salient and that fragmentation consequently will play a part in the discourse of pornography. I have to concede, however, that without comparison (see the following section) this conclusion remains tentative. Unfortunately, I cannot relativize these numbers by comparing them to overall numbers of lexemes (concluding that, e.g., two out of a hundred lexemes are body part nouns) as the latter cannot easily be calculated without a lemmatized corpus, i.e. a corpus where all forms of a word (e.g. hand, hands) are automatically counted as one type. Shifting our attention to the numbers for the two genders, we see a gap of more than 200 lexemes separating women from men. However, this difference has to be taken with a pinch (or even a lump) of salt because the figures are not easy to compare since they depend on token frequencies. Women occur more often in the pornography corpus than men (exact figures will be given below). We might consequently expect their body parts to be represented by more different lexemes. There is, however, no linear relationship, i.e. we cannot assume that if women occur twice as often as men, then the difference in the number of different expressions should also be twice as high (as a matter of fact, it can be expected to be far lower than that). With the aforementioned pinch of salt, I still conclude from the fact that female types exceed male ones by more than 50% that the overall variation in body part representation differs significantly, 5 as predicted, which increases the plausibility of the hypothesis that the salience of women’s body parts will be higher and that consequently the fragmentation of women is likely to be stronger in readers’ minds. 4 It has to be noted that different is relative since many of the words actually occur for both genders. Even if it may be partly justified to count these twice - after all, the same word can denote a different body part if used for males or females - the value for overall lexical variation has to be treated with caution, here and also below. 5 Significant(ly) will be used as an intensifying adjective and adverb in this book and not as indicating statistical significance. <?page no="151"?> (Over)Representing body parts 151 Moving from types to tokens, we get the following frequencies of lexemes for body parts in the pornography corpus in absolute and relative numbers. The latter are computed by dividing absolute frequencies by the number of words in the whole corpus. A value of 1% thus means that every hundredth word in the corpus is a body part lexeme (of the respective category). Tokens Overall Absolute frequency 16,687 9,560 26,247 Relative frequency 2.5% 1.4% 3.9% Table 4: Absolute and relative frequencies of lexemes for body parts in the pornography corpus. As the table reveals, the corpus contains a massive 26,247 references to body parts, which amounts to an average of 3.9 occurrences of a body part lexeme in 100 words. Readers thus encounter a body part approximately every 26 words. Although we have to compare this number to that of the erotica corpus later in order to put it into perspective, it seems very high, as I predicted above. The data thus suggests that body parts are likely to enjoy a certain salience in readers’ minds and consequently that fragmentation will be at work in pornography reading. On first view, the margin between the frequencies for female and male body parts appears to support my initial prediction of gender differences. This gap, however, cannot be treated as absolute because it depends on the general quantitative presence of women and men in the corpus, as already discussed above. I thus need to find out about the frequencies of references to women and men as whole persons for comparison. I have already indicated that I can calculate a value approximating the real quantitative relations of holistic references to women and men, based on the frequencies of personal and reflexive pronouns, proper names and a select set of gender-specific common nouns. The most important figures are contained in Table 5. 1 st ps. sg. personal pronouns 13,965 13,304 3 rd ps. sg. personal pronouns 13,884 8,595 Proper names 4,866 3,703 Common nouns 2,526 2,145 Totals 35,241 27,747 Table 5: Absolute frequencies of gender-specific words in the pornography corpus. <?page no="152"?> Analyses 152 As the total numbers in Table 5 show, women also occur more often than men if we consider references to whole persons. The crucial question, however, is whether the relation between female and male frequencies corresponds to that between female and male body parts. The ratio female/ male - the frequencies for women divided by the frequencies for men - for holistic references is 1.27, for body parts it is 1.75. This means that women as whole persons occur 27% more often than men, while female body parts are mentioned 75% more often than male body parts. In other and simpler words, there is a difference between the frequencies of female and male body parts that cannot be accounted for by the difference in the occurrences of women and men in general, which corroborates one of my expectations for this chapter. Another figure might show this more impressively. The ratio between the frequencies of women as whole persons (at least as far as they are covered by the selection made above) and that of female body parts is 2.11. This means for every two holistic references to women there is approximately one reference to one of their body parts. The respective ratio for men is 2.90. Like type frequencies, token frequencies thus point to an advantage of female body parts with respect to salience and thus to a slightly stronger fragmentation of women as opposed to men created in and through pornography. 1.3 Comparing and discussing overall results Let us now look at the corresponding results for the erotica corpus, starting once again with type frequencies. Types Overall Absolute number of types 228 175 403 Table 6: Numbers of different lexemes for body parts in the erotica corpus. Given that lexeme variation decreases with the size of the corpus, with more and more lexemes being repeated, the fact that the pornography corpus, which is three times as large as the erotica corpus with respect to tokens, contains 2.5 times as many different expressions for body parts (1,014 vs. 403) as its erotica counterpart indicates that the body and its components are more persistently overworded in the former, which agrees with my expectation. This, in turn, suggests that, although not absent from erotica, the salience of body parts and thus of fragmentation will, as hypothesized, be a more powerful factor in pornography. <?page no="153"?> (Over)Representing body parts 153 Depending in non-linear ways on the frequencies of the two genders occurring in the corpus generally, the figures for women and men in the table above pose the same difficulties for comparison as those from the pornography corpus. The quantitative advantage of female types is neither large nor small enough to allow a definite judgement with regard to the assumptions concerning gender differences. Having examined the variation in body part naming, I will now look at how often the different lexemes for body parts are used. The token frequencies of body part words in the erotica corpus are presented in Table 7. Tokens Overall Absolute frequency 3,595 2,236 5,831 Relative frequency 1.5% 0.9% 2.4% Table 7: Absolute and relative frequencies of lexemes for body parts in the erotica corpus. As expected, with 2.4 occurrences per 100 words, body part lexemes are less frequently used in the erotica than in the pornography corpus (3.9 per 100 words). This may lend some weight to the hypothesis about a higher amount of fragmentation in pornography, but considering that readers encounter a body part every 40 words, fragmentation cannot be claimed to be absent from the erotica. The predominance of female body parts again has to be compared to the frequencies of holistic references to women and men. Table 8 contains the respective figures. 1 st ps. sg. personal pronouns 7,184 1,013 3 rd ps. sg. personal pronouns 3,802 4,659 Proper names 1,150 1,217 Common nouns 739 840 Totals 12,875 7,729 Table 8: Absolute frequencies of gender-specific words in the erotica corpus. The ratio of female and male holistic references is 1.67, the ratio of female and male body parts is 1.61. This means that while women occur 67% more often than men in the erotica corpus, their body parts are mentioned 61% more often. The closeness of these numbers corresponds to the whole person/ body part-ratios, which are 3.6 for women and 3.5 for men. Women as whole persons, as a matter of fact, thus occur <?page no="154"?> Analyses 154 marginally more frequently relative to their body parts than men do relative to theirs. More importantly, this means that in accordance with my expectations, the difference between the genders is significantly smaller than in the porn corpus. This, in turn, can be interpreted to mean that female and male body parts will be interpreted as equally salient and, as hypothesized, female fragmentation can therefore not be said to be stronger than male fragmentation in erotica readers. A word of caution concerning the quantitative results for male body parts in the erotica corpus is in order here: the vast majority of the stories are written by women (a consequence of the fact that erotica are primarily a women-oriented discourse) and also have a female first person narrator. This means that women often occur in the first and men almost always in the third person. Now it seems that first person narrators are more likely to refer to themselves as whole persons than to other people and they are less likely to mention their own body parts than the body parts of others. That this is indeed a relevant factor is corroborated by the figures in the following table representing ratios of the frequencies of third person pronouns (she/ her/ herself and he/ him/ himself) to body part nouns preceded by her or his and the ratios of the frequencies of first person pronouns (I/ me/ myself) to body part nouns preceded by my in the two corpora. The ratios are calculated by dividing the absolute frequencies of the pronouns by the absolute frequencies of the nouns, which means that the higher the value, the more often there is reference to the whole persons compared to reference to their body parts. Porn: Porn: Erotica: Erotica: 1 st person 2.21 2.83 3.45 4.17 3 rd person 1.50 2.03 2.72 2.55 Table 9: Whole person/ body part ratios for first and third person pronouns. For females and males in both the pornography corpus and the erotica corpus the whole person/ body part-ratios are higher for first person pronouns than for third person pronouns. I, me and myself are thus used more often as compared to expressions for body parts introduced by my, than he, him, himself and she, her, herself compared to body parts introduced by her or his. If the assumption described above holds, then it entails that first person narrators are less prone to fragmentation than other characters. With the numbers of women and men as narrators being almost the same, this does not affect the pornography corpus. The partly higher frequencies of male body parts in the erotica corpus can, however, be attributed to the scarcity of male first person narrators in the erotica <?page no="155"?> (Over)Representing body parts 155 corpus. But this begs the more general question whether the fact that one aspect depends on another one - as the frequency of body part nouns on the genders of narrators - means that the former can be reduced to the latter, in particular if the two aspects are not equivalent and if the second one is not an accidental feature but represents the nature of a discourse,. In my view, the conclusion that there is no significant difference in frequencies and, correspondingly, in salience of female and male body parts and thus in the extent of female and male fragmentation holds, even though we have to bear in mind that fragmentation does not seem to be a completely independent conceptualization process. As this, however, is not only a contingent feature of the corpus but represents the essence of erotica, fragmentation cannot simply be reduced to a narratological aspect. 1.4 Presenting and discussing specific results Readers can find lexeme lists, complete with (sub)category membership and frequencies at my homepage, allowing them to appreciate the wealth of this data. But at this point I want to concentrate primarily on the overall quantitative relations between the body part categories. Table 10 lists all type frequencies of female and male body parts in the eight categories. As absolute frequencies are difficult to compare across data sets of different sizes, I have added relative frequencies. They have been calculated by dividing the number of lexemes for a particular category by the overall number of different lexemes for one gender (or by the total number of all different lexemes for the overall column). A value of 50% for men in a particular category thus means that half of all different lexemes used for men fall into the respective category. The percentages therefore represent the size and thus importance of the respective category in the description of female or male body parts (or body parts in general) in relation to the other categories (I will apply the same procedure to calculate the relative frequencies of elements of particular categories throughout this book). <?page no="156"?> Analyses 156 Types Overall % % % General 28 4.6% 13 3.3% 41 4.0% Transcend. 19 3.1% 13 3.3% 32 3.2% Head 56 9.1% 47 11.8% 103 10.2% Torso 40 6.5% 30 7.5% 70 6.9% Limbs 41 6.7% 38 9.5% 79 7.8% Breasts 58 9.4% 5 1.3% 63 6.2% Anus 63 10.2% 44 11.0% 107 10.6% Genitals 310 50.4% 209 52.4% 519 51.2% Totals 615 100% 399 100.1% 1,014 100.1% Table 10: Absolute and relative numbers of different lexemes for body parts in the individual categories in the pornography corpus. 6 0 % 1 0 % 2 0 % 3 0 % 4 0 % 5 0 % 6 0 % General Transcending Head, face, neck Torso Limbs Breasts Anus Genitals F e m a le M a le Figure 1: Relative sizes of the individual body part categories with respect to lexical variation in the pornography corpus. 7 As can be seen from the data in Table 10 and as the diagram in Figure 1 graphically displays, the preponderance of the genital category is striking. More than half of the lexemes used for body parts in general - for either females or males - denote the genitals or the genital area. This category can therefore be justly said to be overlexicalized, confirming what I expected. This tendency is further underlined by the fact that the other two categories which bear a strong sexual significance, namely the 6 If the totals do not equal 100% in the percentage columns, then this is a result of the rounding of individual percentages. 7 To undermine stereotypical associations of colours with gender, I am using darker shades of grey for females and lighter shades for males throughout this book. <?page no="157"?> (Over)Representing body parts 157 anus and the breasts, rank also very high in the type frequency lists (with women, the anus is second, the breasts fourth, with men, the anus comes third). The sexual organs will thus gain primary salience in readers’ interpretation, which lends support to the assumption that people will not just be fragmented in pornography but also sexually fragmented. A closer look at the lists of words reveals that the differences between the lexemes in the individual categories can be of a fundamentally distinct nature. Head, face, ear, nose, hair, neck, throat, tongue, and lip - to list just a few of the items in the head category - differ in the anatomically and also conceptually separate parts that they denote. There may be similar differences in the genital, anus and breast categories, e.g. between pubic hair, the pubic mound, the inner and the outer labia, the clitoris, etc., for the female genitals. Distinguishing the meanings of fuckhole, honeypot, cunt, pussy, snatch, twat, hole, entrance, or crack is, however, a less straightforward task. The conceptual differences may be more subtle, non-literal meanings (metaphors/ metonymies) come into play, as do evaluative and intertextual elements. Additionally, there is also a moment of linguistic playfulness and aesthetics in the naming of genitals, anus and breasts completely absent from the other categories, e.g. in the rhyming of pocket rocket or anal canal, the alliterative and pleonastic labial lips, or the widespread use of tropes (for a thorough analysis of metaphors for genitals, cf. section VIII.2). In my view, only in the latter situation can we really speak of overlexicalization since the wealth of subtle meaning differences and of playfulness indicates the preoccupation with a particular area and can create salience in the readers. To put it more concretely, the impact of using snatch, twat and pussy is greater than that of using ear, nose and hair. As such differences can only be found in the genital, the anus and the breast (in women) categories, their quantitative predominance becomes all the more impressive and possibly also more effective. In contradiction to what was predicted, no dramatic differences have been found between women and men, apart from the breast category (the number of words referring to male breasts is, of course, negligible). The hypothesis about a more consistent sexual fragmentation of females does thus not receive support by this data, probably because it has underestimated the dominant presence of the male penis - equally overworded as the female genitals - which may serve all kinds of symbolic functions (primarily probably male strength and power) and which pornography can therefore not afford to omit. If we move to token frequencies, we will note some deviations from type frequencies. The figures are included in Table 11 and are graphically displayed in the diagram below. <?page no="158"?> Analyses 158 Tokens Overall % % % General 901 5.4% 262 2.7% 1,163 4.4% Transcend. 170 1.0% 66 0.7% 236 0.9% Head 4,256 25.5% 2,236 23.4% 6,492 24.7% Torso 1,095 6.6% 499 5.2% 1,594 6.1% Limbs 2,920 17.5% 1,692 17.7% 4,612 17.6% Breasts 1,632 9.8% 85 0.9% 1,717 6.5% Anus 1,018 6.1% 273 2.9% 1,291 4.9% Genitals 4,695 28.1% 4,447 46.5% 9,142 34.8% Totals 16,687 100% 9,560 100% 26,247 99.9% Table 11: Absolute and relative frequencies of lexemes for body parts in the individual categories in the pornography corpus. 0% 10% 20% 30% 40% 50% General Transcending Head, face, neck Torso Limbs Breasts Anus Genitals Fem ale M ale Figure 2: Relative sizes of the individual body part categories with respect to token frequencies in the pornography corpus. While genitals again constitute the largest category, its lead over the other categories is not as dramatic as with types, and the breasts and the anus rank lower than with types. The reason for all this might be that we have a very large number of spontaneous lexeme formations in these categories with a single occurrence only. They thus have a large effect on type frequencies but not on token frequencies. Genitals are nevertheless the body parts that are mentioned most often in the corpus, and the breasts and anus categories are still sizeable, facts that are in line with my assumptions and hence add credence to the hypothesis that sexual fragmentation will be most salient in interpreting pornography and that <?page no="159"?> (Over)Representing body parts 159 there will be a strong element of sexual fragmentation in the comprehension process of this discourse. As far as gender differences are concerned, male genitals are - more clearly than with types - quantitatively more predominant than female ones, again a result of the overpowering presence of the penis. The gender gap narrows down if we take the results for breasts and the anus, where women have a huge preponderance, into account. But it does not correspond to my expectation, which saw a clear female advantage in all sexual categories. This cannot be interpreted as corroboration of the hypothesis of a more consistent sexual fragmentation of women in and through pornography. The data additionally points to the fact that sexual fragmentation in women extends to ‘tits and ass’, which was to be expected given the prominence of the female anus and breasts in pictorial pornography, while in men it clearly centres on the penis (as sign of penis worship). As a last interesting set of data, Table 12 lists the 10 most frequent nouns used for female or male body parts in the pornography corpus. Women Men pussy 1,246 cock 1,810 cunt 917 hand 647 mouth 902 tongue 507 body 788 dick 442 leg 640 finger 366 hand 619 face 330 lip 608 prick 330 tongue 596 mouth 313 face 522 ball 308 breast 516 eye 277 Table 12: The ten most frequent lexemes for body parts in the pornography corpus (sexual organs in bold). The fact that both lists are topped by lexemes for genitals and that among the 10 expressions, there are 3 (for women) and 4 (for men) lexemes denoting sexual organs underlines the quantitative dominance of these categories, further confirming my expectations and lending support to the assumption that there is a strong moment of sexual fragmentation - of both genders - in the interpretation of pornography. The tables, however, do not contain any evidence in favour of the assumption that there could be an advantage of female sexual organs. So <?page no="160"?> Analyses 160 there is once again no support for the hypothesis that the sexual fragmentation created by pornography will be stronger for women. 1.5 Comparing and discussing specific results Lists containing the lexemes found in the erotica corpus can again be found at my homepage. These, too, reveal interesting details, which readers are invited to review for themselves. This section, however, focuses solely on the quantitative relations. Let me first concentrate on the numbers of types in the different categories in the erotica, which can be found in Table 13 and - visualized - in Figure 3. To make the comparison between the two corpora easier, I have added - here and elsewhere in the comparative sections - a table synoptically contrasting the pornography and the erotica results. Types Overall % % % General 17 7.5% 13 7.4% 30 7.4% Transcend. 11 4.8% 9 5.1% 20 5.0% Head 39 17.1% 46 26.3% 85 21.1% Torso 35 15.4% 26 14.9% 61 15.1% Limbs 30 13.2% 30 17.1% 60 14.9% Breasts 14 6.1% 5 2.9% 19 4.7% Anus 11 4.8% 11 6.3% 22 5.5% Genitals 71 31.1% 35 20.0% 106 26.3% Totals 228 100% 175 100% 403 100% Table 13: Absolute and relative numbers of different lexemes for body parts in the individual categories in the erotica corpus. <?page no="161"?> (Over)Representing body parts 161 0% 10% 20% 30% 40% General Transcending Head, face, neck Torso Limbs Breasts Anus Genitals Fem ale M ale Figure 3: Relative sizes of the individual body part categories with respect to lexical variation in the erotica corpus. Types Overall Porn Erotica Porn Erotica Porn Erotica General 4.6% 7.5% 3.3% 7.4% 4.0% 7.4% Transcend. 3.1% 4.8% 3.3% 5.1% 3.2% 5.0% Head 9.1% 17.1% 11.8% 26.3% 10.2% 21.1% Torso 6.5% 15.4% 7.5% 14.9% 6.9% 15.1% Limbs 6.7% 13.2% 9.5% 17.1% 7.8% 14.9% Breasts 9.4% 6.1% 1.3% 2.9% 6.2% 4.7% Anus 10.2% 4.8% 11.0% 6.3% 10.6% 5.5% Genitals 50.4% 31.1% 52.4% 20.0% 51.2% 26.3% Table 14: Comparison of body part categories with respect to lexical variation between the pornography corpus and the erotica corpus. In the erotica corpus, the genitals receive the most intensive lexical treatment overall, with the highest number of different lexemes belonging to this class. Considering the additional fact that, like in the pornography corpus, there is more variation in subtle areas of meaning in this category, the genitals also clearly exceed the rest. In comparison to the pornography corpus, however, the data features lower percentages in the sexually significant categories, but higher ones in the five other classes. The degree to which sexual fragmentation will be effective in the erotica is thus certainly smaller, with the focus spreading from sexual organs to the more general categories of head, limbs and torso. This is in agreement with my hypotheses. <?page no="162"?> Analyses 162 With respect to gender differences, the data from the erotica corpus sees female genitals clearly ahead of male genitals, which only come second after head and are closely followed by limbs and torso. So in complete contradiction to my assumption, the gap in variation in naming sexual body parts is higher in the erotica than in the pornography short stories. These figures raise some doubts about the hypothesis of female sexual fragmentation constituting a more important force in pornography than in erotica consumption. If we go on to token frequencies, we witness some shifts in the quantitative relations, as can be seen from the tables and the diagram. Tokens Overall % % % General 215 6.0% 87 3.9% 302 5.2% Transcend. 114 3.2% 41 1.8% 155 2.7% Head 1,177 32.7% 816 36.5% 1,993 34.2% Torso 459 12.8% 176 7.9% 635 10.9% Limbs 958 26.6% 649 29.0% 1,607 27.6% Breasts 304 8.5% 77 3.4% 381 6.5% Anus 76 2.1% 51 2.3% 127 2.2% Genitals 292 8.1% 339 15.2% 631 10.8% Totals 3,595 100% 2,236 100% 5,831 100.1% Table 15: Absolute and relative frequencies of lexemes for body parts in the individual categories in the erotica corpus. 0 % 1 0 % 2 0 % 3 0 % 4 0 % General Transcending Head, face, neck Torso Limbs Breasts Anus Genitals F emale M ale Figure 4: Relative sizes of the individual body part categories with respect to token frequencies in the erotica corpus. <?page no="163"?> (Over)Representing body parts 163 Tokens Overall Porn Erotica Porn Erotica Porn Erotica General 5.4% 6.0% 2.7% 3.9% 4.4% 5.2% Transcend. 1.0% 3.2% 0.7% 1.8% 0.9% 2.7% Head 25.5% 32.7% 23.4% 36.5% 24.7% 34.2% Torso 6.6% 12.8% 5.2% 7.9% 6.1% 10.9% Limbs 17.5% 26.6% 17.7% 29.0% 17.6% 27.6% Breasts 9.8% 8.5% 0.9% 3.4% 6.5% 6.5% Anus 6.1% 2.1% 2.9% 2.3% 4.9% 2.2% Genitals 28.1% 8.1% 46.5% 15.2% 34.8% 10.8% Table 16: Comparison of body part categories with respect to token frequencies between the pornography corpus and the erotica corpus. Vagina, clitoris, penis, testicles, etc. are mentioned significantly less often than other areas of the body in the erotica corpus. If the body is referred to, the erotica short stories are much more likely to mention the limbs or the head. A comparison to the pornography figures underscores the missing predominance of the genital category even more clearly. The percentages are more than three times as high in the pornography corpus as in the erotica corpus. To underline this, I additionally point to the relative frequencies of genital lexeme in the whole corpus (if we divide the absolute numbers of references by the number of words of the whole corpus): here the results are 1.35% versus 0.26%, suggesting that porn readers encounter genital lexemes more than five times as often as erotica readers. All these numbers fully support my assumptions, thus adding to the plausibility of the hypothesis of sexual fragmentation being a more influential factor in the discourse of pornography than in that of erotica. The figures for genitals have men clearly ahead of women, mainly due to high frequencies of cock and penis, which are mentioned considerably more often than any lexeme for female genitals, which also explains the reversal of the relations found for lexical variation. The gender difference becomes smaller if we include the breast category, which is more frequently represented in females. This absence of a female predominance in the data corresponds to my predictions. A comparison to the pornography data is difficult because due to the unexpected token results in the pornography corpus, which also saw male genitals surpassing female ones - albeit all on a quantitatively higher level - the question where the predominance of female genitals is greater becomes irrelevant. The results thus may provide evidence for the hypothesis that there will be no clear gender bias with regard to <?page no="164"?> Analyses 164 sexual fragmentation in erotica, but whether it is indeed weaker than in pornography cannot be answered here. I will finally take a look at the 10 most frequent expressions for body parts in the erotica corpus. Women Men hand 318 hand 246 body 190 eye 144 leg 178 cock 125 breast 172 mouth 97 eye 159 arm 95 head 147 face 94 mouth 141 finger 93 face 130 head 85 arm 107 tongue 85 hair 106 lip 72 Table 17: The ten most frequent lexemes for body parts in the erotica corpus (sexual organs in bold). There is one lexeme for sexually relevant body parts for males and one for women in the list above. It is remarkable, but unsurprising according to my predictions, that both numbers lag far behind those for the pornography corpus (3 for women and 4 for males) and that the lists are not topped by sexually relevant body parts. This adds to the plausibility of the hypothesis of a difference between the two corpora with respect to sexual fragmentation likely to be created in readers’ minds. As regards gender differences, there is not enough data to allow a valuable judgement on the prediction about the relation between pornography and erotica. All data on the frequencies of body part lexemes considered, I conclude that many but not all of the initial assumptions have been confirmed. The lexical variation and the frequency of references to body parts are indeed very high in the pornography corpus, in particular in comparison to the erotica corpus. I have also found more different body part lexemes for women than for men in the pornography corpus, and these lexemes are also used more often. This gender difference cannot be observed for the erotica corpus. Following my argumentation that variation and frequency will assign body parts a more salient position in readers’ minds, these results show that the chapter’s hypotheses concerning fragmentation are plausible. <?page no="165"?> (Over)Representing body parts 165 With regard to sexual body parts, the data is not as homogeneous. On the one hand, we see a predominance of nouns denoting sexual organs over words for other body areas in the pornography corpus, particularly with regard to variation - genitals, anus and breasts are the parts of the anatomy that are heavily overlexicalized. This predominance is weaker in the erotica corpus, although overwording of sexual organs is not a phenomenon restricted to pornography. On the other hand, the data does not reveal a clear female preponderance in the variation or frequencies of lexemes for sexual organs, a result which was to be expected for erotica but not for pornography. We may thus conclude that sexual fragmentation will be a strong force in pornography consumption, much stronger than in erotica consumption. But it will equally apply to both genders. This is probably due to the omnipresence of the penis - or symbolically, of the phallus - whose power I underestimated in the initial formulation of the hypotheses. 2 Immediately attending to body parts: possessives As detailed in section 1.1, this study only considers nouns for the human anatomy explicitly assigned to their ‘owners’ by a possessive element introducing (or premodifying) the noun phrase, as in the example below. Ted then handcuffed Patricia’s left wrist to her left ankle and her right wrist to her right ankle, keeping her arms inside her thighs, which forced her legs wide apart and brought her knees up to her shoulders. The sentence also illustrates that possessives can, as already described, be realized by possessive pronouns such as her or by genitive NPs such as Patricia’s. It is this opposition between possessive pronoun versus genitive NP which I will be concerned with in this subchapter because I consider it to affect fragmentation. To understand the role of the opposition mentioned for salience and fragmentation, it is necessary to discuss the status of possessives in the information structure. In the example above, the possessive pronouns and the genitive NP are co-referential, which means that they all refer to the same person, viz. Patricia, as the owner of all the body parts mentioned. With regard to the ideational level, there is thus no difference between them. Possessive pronouns and genitive NPs - or pronouns and NPs in general, for that matter - differ, however, in their status in the local information structure, particularly in how they are linked to other information inside or outside the text. (Personal and possessive) pro- <?page no="166"?> Analyses 166 nouns are used if their referent is assumed to be activated and thus at the centre of attention in the recipients’ minds at the time of reading or hearing (cf. Lambrecht 1994: 93f.). If too much time has elapsed or too many pieces of information have gone in between since the last mentioning or if someone has not been mentioned before, the referent stands at or has been moved to the background and in order to draw her or him back to the centre an NP with a common or proper noun as head is used. The occurrence of either pronouns or full noun phrases is thus a sign of - and may contribute to creating - the state of activation a referent has or is assumed to have in the recipients’ minds. As speakers/ writers can refer to themselves and to their addressees only by means of first and second person pronouns, respectively, but not with a full noun phrase, the differentiation in the information structure can just be made for third persons. What are now the consequences of these differences in the state of activation for body parts, salience and fragmentation? If I turn (back) to someone then I will normally first perceive her or him as a whole person. After some time has elapsed with continuing attention, I might focus on her or his body parts, too. Linguistically, this entails that body part nouns will normally be preceded by possessive pronouns, which indicate that the mind has already been focused on the owner. A fragmented view of a person, however, can also mean that I immediately perceive a person’s body (parts) even when first turning my attention to her or him and possibly before acknowledging her or him as a whole. This will put bodily elements into a salient position. Linguistically, this process is reflected by the use of genitive NPs, which, as argued, indicate that I have just directed my attention to the body’s owner. In sum, fragmentation will play a role in readers’ interpretation if they encounter a large number of body part lexemes modified by a genitive NP. We may further assume that only in the course of perceiving a person will I turn my attention to her or his genitals, so that they are even more likely to be introduced by a possessive pronoun than a genitive NP. When sexual fragmentation comes into play, however, this trend will be reversed. Taking all this into consideration, my predictions are: • There will be a high proportion of lexemes for body parts and particularly sexual organs introduced by genitive NPs in the pornography corpus. • The proportion of genitive NPs will be higher for lexemes for female body parts and particularly female sexual organs than for those for male ones. <?page no="167"?> Immediately attending to body parts 167 • The erotica corpus will not feature as high a proportion of genitive NPs introducing lexemes for body parts and particularly for sexual organs as the pornography corpus. • The erotica corpus will not see the same female predominance with respect to genitive NP modification or at least the gap will not be as wide as in the pornography corpus. 2.1 Technical and practical aspects of the analysis I had to search for body parts introduced by genitive NPs or her and his separately to gain information about the occurrences of body parts. Hence the relevant aspects for this subchapter are contained in the data for the “(Over)Representing body parts”-chapter and no additional searches have to be carried out. I will not examine types in this subchapter because the opposition between possessive pronouns and genitive NPs only applies to individual occurrences of body part nouns and not to their general existence in the discourse. 2.2 Presenting and discussing overall results To obtain a value for comparison between the two possessive structures under examination in this section, I use ratios which are calculated by dividing the number of body part nouns modified by her or his by the number of body part nouns modified by a genitive NP. A ratio of 10 thus, for instance, means that for the set of lexemes mentioned possessive pronoun modification is ten times more common than genitive NP modification. The lower the ratio, the higher the proportion of genitive NPs, which, in turn, amounts to a higher degree of fragmentation, following my argumentation from above. I am aware that ratios are not absolutely reliable values due to the susceptibility to variation depending on the size of a category. If a category is small, then a single addition to either side will weigh much more heavily than in larger categories. I nevertheless think that ratios provide some point of orientation. The overall results for the pornography corpus can be found in Table 18 below. <?page no="168"?> Analyses 168 Overall her Gen. Ratio his Gen. Ratio her + his Gen. Ratio 9,273 1,096 8.5 4,236 620 6.8 13,509 1,716 7.9 Table 18: Absolute frequencies of possessive pronouns and genitive NPs (= Gen.) modifying lexemes for body parts plus their ratios in the pornography corpus. Without a comparative value, it is impossible to assess the overall ratio of 7.9 found for the pornography corpus. I will therefore not comment on it any further here. The ratios for women and men come as a surprise since men’s and not women’s physical components are introduced with a genitive NP more often (relative to the number of those introduced by a possessive pronoun). This indicates that turning to men and perceiving their body parts first may be taken to be as equally common or even as more common than doing the same with women and their bodies. The data thus does not support the assumption that a fragmented view of women is more likely to be created in pornography consumers. 2.3 Comparing and discussing overall results This section will examine the data on possessives and body part nouns in the erotica corpus. The absolute frequencies and ratios can be found in the table below. Overall her Gen. Ratio his Gen. Ratio her + his Gen. Ratio 1,396 119 11.7 1,830 163 11.2 3,226 282 11.4 Table 19: Absolute frequencies of possessive pronouns and genitive NPs modifying lexemes for body parts plus their ratios in the erotica corpus. A comparison between the overall ratios of body part lexemes introduced by possessive pronouns and by genitive NPs reveals a clear advantage for the pornography corpus. This suggests that it is, as assumed, more common to use a genitive NP for body part nouns in the pornography corpus, which should, according to my account of the structures above, lead to more fragmentation in the reading process of pornography than in that of erotica. The gender ratios are very similar, which agrees with the prediction that women’s body parts would not be introduced by more genitive NPs than men’s. Due to the unexpected gender ratios in the pornography <?page no="169"?> Immediately attending to body parts 169 corpus, however, the data cannot be taken to reconfirm the assumption that women’s fragmentation is weaker in erotica than in pornography. 2.4 Presenting and discussing specific results I will be concerned with the ratios of body part lexemes with possessive pronouns and those with genitive NPs in this section to examine whether there is any indication of sexual fragmentation in pornography. All relevant figures can be found in Table 20. Overall her Gen. Ratio his Gen. Ratio her + his Gen. Ratio General 505 56 9.0 134 18 7.4 639 74 8.6 Transcend. 95 10 9.5 32 2 16.0 127 12 10.6 Head 2,416 263 9.2 1,007 121 8.3 3,423 384 8.9 Torso 649 42 15.5 203 16 12.7 852 58 14.7 Limbs 1,809 139 13.0 912 77 11.8 2,721 216 12.6 Breasts 902 87 10.4 55 4 13.8 957 91 10.5 Anus 544 88 6.2 89 27 3.3 633 115 5.5 Genitals 2,353 411 5.7 1,804 355 5.1 4,157 766 5.4 Totals 9,273 1,096 8.5 4,236 620 6.8 13,509 1,716 7.9 Table 20: Absolute frequencies of possessive pronouns and genitive NPs modifying lexemes for body parts in the individual categories plus their ratios in the pornography corpus. We find the construction with genitive NPs to be significantly more common for the anus and the genitals, resulting in the - predicted - lower ratios for these two categories. This data suggests that, in agreement with my hypothesis, these two sexually relevant areas will play a more prominent role in the fragmented views of people in readers’ interpretations. It may thus contribute to a moment of sexual fragmentation of women and men. Interestingly enough, the breast category with its substantially higher ratios does not follow this trend, perhaps because the perception of breasts is more tightly associated with the perception of the whole person than that of genitalia and the anus. In contrast to my expectations, I do not see a clear gender bias concerning genitive NP use for sexual body part nouns in the ratios, even though there are differences between female and male results for the anus (here males have a lower ratio) and the breasts (here females have a lower ratio). These differences, however, have to be taken with some caution since males have very low token values in both categories, which may distort ratios. The data does not lend support to the <?page no="170"?> Analyses 170 assumption that women will be more consistently sexually fragmented in the discourse of pornography. 2.5 Comparing and discussing specific results To compare the extent of sexual fragmentation at work in the two discourses, I will present the relevant figures from the erotica corpus in Table 21, contrasting it with the pornography data in Table 22. Overall her Gen. Ratio his Gen. Ratio her + his Gen. Ratio General 65 14 4.6 59 14 4.2 124 28 4.4 Transcend. 49 7 7.0 39 0 - 88 7 12.6 Head 515 36 14.3 693 44 15.8 1,208 80 15.1 Torso 142 10 14.2 145 15 9.7 287 25 11.5 Limbs 391 35 11.2 554 38 14.6 945 73 12.9 Breasts 123 7 17.6 66 9 7.3 189 16 11.8 Anus 26 3 8.7 31 6 5.2 57 9 6.3 Genitals 85 7 12.1 243 37 6.6 328 44 7.5 Totals 1,396 119 11.7 1,830 163 11.2 3,226 282 11.4 Table 21: Absolute frequencies of possessive pronouns and genitive NPs modifying lexemes for body parts in the individual categories plus their ratios in the erotica corpus. Overall Porn Erotica Porn Erotica Porn Erotica General 9.0 4.6 7.4 4.2 8.6 4.4 Transcend. 9.5 7.0 16.0 - 10.6 12.6 Head 9.2 14.3 8.3 15.8 8.9 15.1 Torso 15.5 14.2 12.7 9.7 14.7 11.5 Limbs 13.0 11.2 11.8 14.6 12.6 12.9 Breasts 10.4 17.6 13.8 7.3 10.5 11.8 Anus 6.2 8.7 3.3 5.2 5.5 6.3 Genitals 5.7 12.1 5.1 6.6 5.4 7.5 Table 22: Comparison of ratios of possessive pronouns and genitive NPs modifying lexemes for body parts between the pornography corpus and the erotica corpus. The data reveals the same pattern that we saw in the pornography corpus, with the anus and the genitals taking genitive NPs more frequently than the other categories (apart from the general class). The proportions of genitive NPs in these two categories are, however, lower <?page no="171"?> Immediately attending to body parts 171 (and thus the ratios higher) than in the pornography corpus (even though the gap is even wider in the head category), thus being in line with my expectations and suggesting that sexual fragmentation will be a weaker force in erotica consumption. The ratios, however, also indicate that sexual fragmentation will be nothing alien to erotica readers. Oddly enough, the breast category differs from the other two sexually relevant body areas, as in the pornography corpus, indicating that breasts are generally more often mentioned in connection with a whole person than genitals and anus. Turning to the results for women and men, we see that women’s sexual organs more rarely occur with a genitive NP than men’s, suggesting that immediate attention to the sexually relevant parts of the anatomy is less common for females than for males in erotica. Although the pornography data also showed this trend, it is more perspicuous here, as predicted. These results taken into consideration, I come to the conclusion that possessives and the different degrees of salience they create for body parts will lead to less sexual fragmentation of women in erotica than in porn readers. Summing up, body part nouns in the pornography corpus are more often introduced by genitive NPs than those in the erotica corpus, a result in line with my hypotheses. But while there is not much difference between the genders in the erotica, as expected, the genitive NP is less common for women than for men in pornography, which does not correspond to my predictions. This suggests that immediate attention to body parts is not a phenomenon more commonly associated with females, which I find difficult to explain. The results thus support my hypotheses except the one concerned with women’s fragmentation in the pornography corpus. As regards the results for the individual body part categories, I have, as hypothesized, found high frequencies of genitive NPs for sexual body part lexemes in the pornography corpus, particularly in comparison to the corresponding frequencies in the erotica corpus. Male sexual organs are more likely to be introduced by a genitive NP than female ones in both corpora, which is line with my expectations concerning erotica but not concerning pornography. This suggests that there is support for the hypotheses that sexual fragmentation will indeed play an important role in pornography, that this role is definitely more significant than in erotica, and that women will not be more strongly sexually fragmented than men in the latter. The data, however, fails to produce evidence in favour of the assumption that women’s sexual fragmentation should be more pervasive in pornography than men’s. <?page no="172"?> Analyses 172 3 Casting a side glance: interesting phenomena not examined Fragmentation is created through a manifold set of linguistic structures. This chapter has examined only two of them in depth, but there still remain phenomena that have not been considered but seem to justify further research (not to mention the numerous details that, as noted in my discussions, invite more profound investigation). This section is intended to just touch upon some of these in passing. I might, for instance, look at how frequently activities are represented as attributes of body parts rather than as events in their own right (e.g., the throbbing penis instead of the penis was throbbing). For this purpose, I can examine activity-related adjectives or participles that are used to modify body part lexemes, a combination that is very frequent, as has been found in an informal analysis of such modifiers. Many words denote movement - in particular a form of iterative movement, e.g. quivering, spasming, throbbing, twitching, trembling, churning, bouncing, convulsing, pulsing, jiggling, lazily swirling, shuddering, shuttling, twisting, vibrating - and the majority is highly kinetic, which means a lot of energy is involved, as can be seen from words such as clenching, clutching, darting, hard-driving, invading, kicking, piercing, pistoning, pushing, racing, rampaging, stabbing, stinging, thrashing, throttling, thrusting, thumping, tight clenching, tight-gripping. These descriptors fulfil all requirements of stereotypical verbs. It is therefore interesting to see that they occur as attributive descriptors of body parts (primarily of the genitals) rather than as verbs. The activities are thus represented as properties of the body and its parts. The emphasis is consequently on the body rather than on the actions or events in which it partakes. This is another aspect enhancing fragmentation. Another aspect worth examining is to what extent the salience of body parts (and whose body parts) is further enhanced by adding information (thus increasing the interconnectedness), mainly as adjective phrases, e.g. her tight little pussy or his big and fat cock. The issue of APs modifying body lexemes will be taken up in the next chapter, even though the focus there will be on the types of descriptors rather than on the amount of modification. We might also examine whose body parts and which body parts also appear in semantic roles (for the concept of semantic role, see section VII.1) normally associated with whole persons, especially as agents (the person carrying out an action) or patients (the person to whom something is done), as this blurs the distinction between whole and part. Especially verbs denoting sexual intercourse occur with either <?page no="173"?> Casting a side glance 173 whole persons or body parts as agents or patients, as the two examples with the verb fuck demonstrate: […] his cock was fucking me in a steady rhythm. He fucked my mouth […] The semantic roles of body parts will be included in chapter VIII, even though with a different focus, viz. passivization. Finally, I would like to quote a lengthy example to prove that fragmentation can also be found if the corpus is examined co-textually (with all the limitations that apply to this form of analysis), demonstrating in the process what fragmentation ‘reads like’ in a coherent piece of discourse. In the passage from the pornography corpus, the high density of body part lexemes (shaded) - genitals being overrepresented; note also the many AP-modifiers - shows the overemphasis on body parts (although the cock is actually a plastic tool rather than a male genital here), with references to whole persons (in bold) being almost outnumbered. […] the head of the monster dick bumping against her thighs. "What you wanna hear? How much I missed this pussy? How good I’m going to fuck you? ” “Yeah, tell me about it.” Candy raised her knees a bit higher, causing her meaty shaved cunt lips to part and flash her pink gash. Her clit was already hard. “Tell me how you felt when you first saw my twat,” Candy added. She watched Kim rub the huge cock head up down her dripping cunt flaps. “You have an unbelievable pussy. I saw it when you first climbed up on that table at the hotel but not good. Then, when I moved your legs apart, I was thinking, even then, that I needed a cock - a nice thick cock.” In a comparative example from the erotica corpus, we find body part nouns occurring frequently, too, but not as often in comparison to holistic references as in the pornography corpus, and with a focus not limited to the genital area. I looked at him for a moment, then bent my head to his leg and kissed him gently just above the newly dressed wound. When I raised my head to look at him again, he took his glasses off with a quick, impatient tug that made me want him. Then he reached for me and pulled me toward him on the bed, up between his spread legs. He kissed my hands first, both of them. "How can I thank you? ” he asked. <?page no="174"?> Analyses 174 “You have,” I told him, “you are.” He touched my hair and my face, tracing the line of my cheekbone, and I put my hand on the back of his neck, at his hairline. His hair was thick and a little wiry. I could push my fingers into it. Then we kissed-a long, slow, exploratory kiss that relaxed us both. “Where's your husband? ” he asked. “In Lucerne for the night.” “Good,” he said and kissed me again. 4 Conclusion Conceiving of people in terms of their bodies and their body parts rather than as whole human beings means seeing them as objects. Fragmentation is thus a form of objectification. This applies even more to sexual fragmentation, where people are reduced not just to body parts but to those relevant in sexual activity. This chapter has dealt with the question to what extent and whether at all pornography can create fragmentation and sexual fragmentation in its readers’ minds and whether differences between women and men and between pornography and erotica can be found. For this purpose, it has linguistically focused on the frequencies and the variation of words for body parts and how often they are introduced by genitive NPs. My first hypothesis was that the linguistic structures analysed would suggest that there should be a high degree of fragmentation and sexual fragmentation in pornography readers’ minds. All of the data provides evidence that this claim is right. My second assumption was that the discourse of pornography would reduce women more consistently to their body parts and particularly their sexual organs in the consumers’ eyes. The clear quantitative advantage of female body part nouns with respect to both type and token frequencies suggests that there is indeed a stronger moment of female fragmentation created by pornography. The examination of frequencies is the more extensive and methodologically less problematic one in this chapter. The fact that the analysis of possessives introducing body part lexemes has not produced the results expected may cast a shadow of a doubt on the conclusion, but cannot really undermine it. With respect to sexual fragmentation, the results do not lend support to the claim that there is a difference between women and men here, a claim which has obviously underestimated the omnipresence of the penis, which in the state of erection seems to almost symbolize male power in pornography. <?page no="175"?> Conclusion 175 As regards my third hypothesis, concerned with the difference between the two discourses studied, the results of the analyses fully reconfirm the expectation that both fragmentation and sexual fragmentation feature more saliently in pornography than in erotica. The fourth hypothesis, saying that female fragmentation would be weaker in erotica than in pornography, is supported by the fact that the data from the erotica corpus does not reveal a persistent advantage of women (as a matter of fact, it does not reveal any clear gender bias at all). This also applies to the data on sexual fragmentation, even though a definite judgement on the comparison with pornography is difficult due to the surprising results from the porn corpus regarding this question. All in all, objectification in the shape of fragmentation in general and sexual fragmentation in particular is a key element of the pornographic discourse, much more so than of erotica. Its discriminatory effects, however, will not only apply to women but also to men, even though they may be a bit stronger for women. Although the data reviewed allows some conclusions concerning my hypothesis, I also have to admit that problems have also emerged in the analyses, particularly in the basic methodological design, the statistical evaluations (some values, for instance, depend too much on initial sizes of categories to be appropriate measures for comparison), and the conceptual interpretations (the preponderance of first person narrators in the erotica corpus may exert an influence on data). <?page no="177"?> VI Physicalization and Visualization Samantha is a short, very slender blonde… With his long hard fat dick up my cunt… We have no direct access to the inner cognitive lives of other humans. All the information about them available to us is what we perceive as their material physical existence. But by interpreting this information as a sign of cognition, we can experience other humans as endowed with inner lives and as capable of thinking, feeling and perceiving like ourselves, as other subjects, as it were. Suspending these interpretative processes, staying on the surface and perceiving and thinking about humans only or mainly in terms of their physicality (= physicalization) means seeing them as objects. If pornography manages to create this in its consumers, then it can be regarded as objectifying. The concept of physicalization is one-sided as it only concentrates on the properties of the perceived and must be complemented by a focus on the perceptive process. Although physicality - at least spatial physicality - can also be perceived via the tactile and the auditory sensory channels, the sense of vision is certainly dominant. I will therefore mainly be concerned with physicalization in the sense of visualization. Physicalization and visualization are two sides of the same coin, of the same phenomenon, and stand in a complementary relationship, i.e. what physically exists can normally (or theoretically) be visually perceived. In their effects, physicalization and visualization are in line with other forms of objectification. They are discriminatory processes as they restrict the perception of human beings to physical and thus material attributes, usually those that might be sexually interesting for the viewer, thus minimizing the importance of mental and other (e.g. spiritual) dimensions. In interpersonal relationships, this might also include ignoring the sexual partner’s thoughts, feelings, and preferences, which can lead to a less than perfect sexuality for the partner and, on the most serious side, to a proclivity for sexual violence. It goes without saying that, given the sexual climate of today’s societies, the effect will be more severe if physicalization and visualization become part of men’s conceptions of women’s role in sexuality than vice versa. The pivotal status of vision and thus implicitly of physicality has been widely acknowledged in analyses of pornography. This hardly comes as a surprise considering the predominance of visual forms of <?page no="178"?> Analyses 178 pornography, e.g. videos, films, magazines, etc. - Linda Williams (1999: 7) in this context even speaks of a frenzy of the visible (the term is borrowed from Comolli 1980: 121f.). Interestingly enough, it has not entered the discussions of objectification even though the link seems so apparent. But despite this lack of support from the literature, I still think that physicalization and visualization are core aspects of objectification. In the current chapter, I will look at linguistic items supposed to highlight people’s visual physicality, trying to find out whether these are phenomena integral to the pornographic discourse and whether differences between women and men can be found. In the first subchapter, I will examine descriptors for people, concentrating on attributive adjective phrases. The second subchapter will deal with descriptors of bodies and body parts, again confining itself to attributive adjective phrases. Although the main focus is on physicalization and visualization, I will also briefly comment on other categories of descriptors and their prominence in the corpus in order to bring the results into perspective. This comparative look is also important as non-visual attributes may - in addition to introducing new elements - enhance (e.g. tactile ones) or reduce (e.g. psychological ones) the objectifying potential of physicalization and visualization. I will additionally examine visual subaspects in the second subchapter, some of which are also believed to add to (e.g. shape, colour, beauty) or mitigate (e.g. size) the overall effects of visualization. I expect the linguistic patterns analysed to support the four hypotheses below, following the schematic hypotheses of chapter IV: • Physicalization and visualization will be dominant global conceptualizations of humans in pornography. • They will be stronger for women than for men in pornography. • In erotica, physicalization and visualization will on the whole play a less dominant role than in pornography. • There will be no apparent gender imbalance in erotica or at least the gap from women to men will not be as wide as in pornography. These hypotheses imply that aspects and/ or subaspects increasing the objectifying impact of physicalization and visualization will be in line with the trends mentioned. <?page no="179"?> Describing people 179 1 Describing people: attributive descriptors Theoretically, a lot of linguistic elements can create physicalization and visualization. If people are, for instance, often represented as participants in material processes expressed by certain verbs, e.g. walk, hit, or stumble, or in local relations expressed by certain prepositional phrases, e.g. in the garden, over the fence or on the hill, then this will contribute to their perception as physical. Describing them, i.e. assigning properties to them, however, is a more immediately relevant factor. I will therefore concentrate on linguistic elements - primarily adjectives - that play a prominent role in description and I will call these descriptors. The current subchapter will examine descriptors for people, descriptors for body parts will be the topic of the second one. There are a number of linguistic factors that need to be considered when analysing descriptors. Most of these will be relevant to both subchapters, which is why I will discuss them at the beginning of this first one. The main thing to look at in analysing descriptions of humans is, of course, which aspect of their being is emphasized. This requires categorizing the descriptors semantically. There are numerous candidates for semantic categories, as can be seen below: Sylvia is tall, with a ravishing figure […] […] spying on my crazy red-haired girlfriend Rachel Simpson was my Abnormal Psych professor during junior year in college. […] she was forty-two, but full-figured, huge-breasted […] […] we both are pretty open-minded. […] Aunt Mary was not pretty in the conventional way, but she was extremely attractive in a sexy, mature fashion. […] I meet a variety of young female students […] Descriptors might highlight physical and visual aspects such as shape or size (full-figured, huge-breasted), colour (red-haired), or attractivity (attractive, pretty), but they might also specify psychological (crazy, openminded) or social properties (Abnormal Psych professor, student, young, forty-two, female - if we classify age and gender as social, this is). What is now the link of such categories to the conceptualization of physicalization and visualization? I assume that, in analogy to what has been said about body part lexemes, readers will interpret the visual <?page no="180"?> Analyses 180 physical dimension as more salient and thus as more important to a person or a group, if they encounter particular types of descriptors frequently and in great variety. Examining frequencies and variations of visual physical descriptors is thus the prime object of this and the next subchapter. Although the visual physical class will be of paramount importance, it will not suffice to confine myself to distinguishing it from other descriptors. A fine-tuned analysis of the semantic classes of descriptors reveals that there are some that, though themselves not visualizing, may still add to the overall objectifying effects of physicalization and visualization. Tactile or olfactory properties, for instance, may have a similar physicalizing impact as visual ones. Evaluative descriptors may not be physicalizing but they also work towards objectification, being assigned to people who thereby become objects of judgement as opposed to the persons passing the judgement. Other descriptors may undermine or mitigate the effects of physicalization and visualization by increasing the salience of opposite features. Psychological and, to a lesser extent, physiological descriptors, being more concerned with people’s minds and inner selves, run counter to objectification. Social descriptors, which conceptualize people as embedded in a context of relationships to other people, take a middle position, although some of its subaspects, in particular age, gender and ethnicity, may be visualizing as a result of their perceptibility. Additionally, there are also visual subcategories that may affect the overall objectifying power of physicalization and visualization, e.g. size, in particular superlative size, may reduce objectification as a result of its association with power (for more on this link, see subchapter 2). The set of descriptor classes used will be introduced in the “Technical and practical aspect”-section below. While the categorial membership of descriptors is definitely most important, there additionally are subtle semantic differences created by word classes and syntactic positions worth examining. If we, for instance, take the following sentences, we see that (roughly) the same attribute can be expressed quite differently. the guy is fat the guy is a fatty John is a fat guy the fat guy came in I will argue that these differences, while not affecting the category of the attribute (we are always talking of physical size), still have semantic functions. On the one hand, they denote the attribute’s identity status, <?page no="181"?> Describing people 181 indicating whether fatness, for instance, is more or less essential for our conception of the guy’s identity. This is important since presenting physical features as identifying the very core of someone or as only describing a marginal aspect will further enhance or mitigate physicalization. On the other hand, the linguistic form of descriptors also has consequences for the information structure, contributing to the highlighting, the backgrounding, the subliminal conveying, etc. of features. But let me discuss these two aspects and the linguistic means by which they are realized individually. The identity status of attributes is represented by a scale with two poles. On the one hand, we have essential features, which define what a person (or a thing, in case we are dealing with non-personal descriptions) is like, on the other hand, we have accidental features, which are imposed from outside. The two poles are closely associated with a bundle of other oppositions such as core versus periphery, permanent versus temporary, situation-independent versus situationdependent, etc., which may cumulatively or separately be evoked in individual cases. Now which linguistic elements may contribute to the identity status of attributes? Generally speaking, descriptors can come as nouns or adjectives or - syntactically more precisely - as noun phrases or adjective phrases. The following two sentences illustrate this. Petra is an English teacher Petra is nice Although both descriptors assign a property to Petra, there still is a conceptual difference between them. Nouns do not only prototypically denote more permanent and fixed properties, they are also used for features that are closer to the very identity of a person. Adjectives, too, may refer to people’s permanent attributes, but they do not suggest that these attributes are essential for people’s personalities, i.e. they represent more accidental or peripheral features. As we tend to see professions or family roles as more central than physiological or psychological features, we, as a consequence, are wont to use nouns for the former (she is a teacher and she is my mother), but adjectives for the latter (he is hungry and he is aggressive). Theoretically, prepositional phrases, e.g. the guy was on top of me, can also function as descriptors. I will nevertheless not consider them in my study due to the more limited range of their semantic functions. A further important distinction for adjective descriptors is where they occur in the structure of a sentence, as exemplified by the following two sentences. <?page no="182"?> Analyses 182 She is red-haired. He watched the red-haired woman If the adjective phrase is assigned to a person via a copular verb (the copula be or verbs with similar meanings, e.g. become, cf. Bussmann 1996: 105), as in the first sentence, then the resulting clause denotes a description in the narrow sense of the word. 1 The function of the adjective phrase in this structure is called predicative. If the AP is contained in an NP, as in the red-haired woman, then this is called the attributive function (cf. Quirk et al. 1985: 402f., Biber et al. 1999: 505f.) There is a difference, however subtle, between the predicative and the attributive positions with respect to the identity status: attributive APs, by virtue of their being semantically combined with the head noun, are still interpreted to be part of the person and thus as more essential, while predicative APs are seen as further distant to the core, sometimes even as situational, i.e. as bound to the situation and thus describing behaviours and states. This becomes manifest in the preference for predicative APs to express temporary states: angry, hungry or frightened will, for example, occur more often in this position, expressed by predicative APs rather than by attributive ones. The difference also becomes evident if we look at sentence pairs such as she was aggressive vs. she was an aggressive person, where only the first sentence can denote a situational attribute. In brief, all things being equal, nouns tend towards the essential pole, predicative adjectives towards the accidental pole, with attributive adjectives taking the middle position. 2 Variation is, of course, possible, dependent on the very lexemes used and the general co-text. 3 Physical features may not be purely accidental, but they certainly do not normally count as essential features. This becomes evident in the fact that they are not usually among those expressed by nouns. So we say she is blue-eyed or he is tall, but we do not use nouns for the same physical properties (incidentally, for most of these features we lack nouns anyway and converted adjectives, i.e. adjectives used as nouns, 1 Halliday (1994: 120-124) further distinguishes between attribution (he is a professor), where a person is assigned to a category, and identification (he is the professor that I mean), where two elements are equated. 2 The terms predicative/ attributive adjectives or nouns refer - to be syntactically precise - to the head words of predicative or attributive adjective/ noun phrases. 3 We could, for instance, argue that He is an idiot does not denote a permanent attribute. But to formulate the sentence as if it were essential possibly only provides the statement with the evaluative and emotional power. <?page no="183"?> Describing people 183 e.g. all the blue-eyed, are restricted in usage). Physicalization and visualization, however, entail that visual properties are shifted towards the essential pole in people’s perception. This, for instance, means that a discourse can create physicalization and visualization if it contains a disproportionally high number of noun descriptors in this category or if there is a predominance of the visual category in attributive adjectives. The forms of descriptors do not only have an effect on ideational meaning, but also on the local information structure, in particular the distinction between predicative and attributive phrases. The assigning of features is explicitly asserted in predicative structures. In She is redhaired, for instance, the fact that the woman has red hair is the main piece of information that is communicated. In an attributive descriptor, by contrast, the same fact is presupposed or taken as given. He watched the red-haired woman, for instance, thus is not a statement about the colour of the woman’s hair, but the latter is treated as already known. The difference becomes obvious if the two sentences are negated because only the truth-value of the predicative AP is thereby changed. The attributive AP remains unaffected, which means the woman is still red-haired, whether somebody watched her or not (for the distinction between assertion and presupposition, cf. Lambrecht 1994: ch. 2). 4 A second aspect of the local information structure relevant to predicative APs is the thematic structure and its distinction between the theme of my assertion, i.e. the thing or person I am talking about, and the rheme, i.e. the comment I am making about the theme (cf. Halliday 1994: ch. 3, Lambrecht 1994 - he uses the terms topic/ comment for the same distinction). Metaphorically speaking, the thematic structure is like a filing system, with the themes being the drawers and the rhemes the files I put into the latter. Halliday (1994: ch. 3, cf. also Thompson 1996: 118f.) assumes that in English the theme is defined by its position in the sentence - it always comes first (a claim contested by Lambrecht 1994: 199f.). In the sentences from above - she is red-haired and he watched the red-haired woman - for instance, she and he are the themes, while is red-haired and watched the red-haired woman are the rhemes. Even though consistent themes in a discourse will eventually stand out in our interpretations, within the individual sentence the rheme is high- 4 The terms given and new information are equivalent to the distinction between presuppositions and assertions. They are, however, sometimes used for parts of information, e.g. if she in she is red-haired is claimed to be given information. This usage is problematic as a subject NP does not constitute a whole piece of information, i.e. it does not denote a whole proposition (cf. Lambrecht 1994: 45- 47). <?page no="184"?> Analyses 184 lighted and attracts more attention since it always contains the focus of the assertion. The focus is the element of the information which introduces a new, unpredictable and/ or contrastive aspect (always in relation to background knowledge and/ or previous discourse) (for focus, cf. Lyons 1977: 500f., Lambrecht 1994: ch. 5) and which is usually placed at the end of a sentence. Predicative adjectives always occur in the rheme of a clause and normally are also the focus of the assertion (unless there are other phrases following). So in she is red-haired, red-haired is definitely the part of the rheme that carries most informative weight. Although predicative adjectives thus highlight a certain property, we may wonder whether attributive adjectives are not more effective because they do not attract conscious attention, thus not inviting opposition by readers, and may, as a consequence, be taken for granted. She is red-haired thus emphasizes the red hair but at the same time readers are made consciously aware of this fact. The red-haired lady, on the other hand, mentions the hair colour in passing, as if it was given, and thus does not attract any attention. If this argumentation is valid, then visualization is enhanced if the predominance of visual properties is stronger in attributive APs than in predicative APs. In addition to attributive and predicative adjective phrases, there is also a mixed form where the NP occurs in the predicative position but contains an attributive AP: He is a handsome guy She is the perfect woman Handsome and perfect premodify guy and woman, respectively, but are part of predicative noun phrases. This entails a mixture of the aspects discussed above. The attributive position means that the features will not be interpreted as purely accidental. Unlike other attributive APs, however, I think that these ones do not only stand in the rheme, they usually also constitute the focus of the information, thus being even more accentuated than the nominal head itself. Incidentally, the head positions of such constructions are often taken by relatively general nouns such as person, man or woman or indefinite pronouns such as one, which serve as carriers of the adjectival meaning rather than being rich in semantic content themselves. I will therefore count adjectives in this position as predicative structures. In sum, structural positions can influence our interpretations of descriptions on two planes. Firstly, they may affect the identity status of an attribute, i.e. whether we see it as essential to a person’s identity or more peripheral and accidental. It can be argued that, all things being <?page no="185"?> Describing people 185 equal, nouns will be understood as denoting more essential features, while adjectives are used for more peripheral features. It can further be assumed that attributive structures will also be interpreted as more essential than predicative ones. Secondly, the opposition between attributive and predicative may also influence how much conscious attention we pay to the description processes. Predicative structures, making the description explicit and accentuating it, may thus attract more conscious attention than attributive ones, which presuppose rather than assert the description. This may make the latter more influential in interpretations. Despite this detailed description of syntactic differences (also supposed to create a framework for studies going beyond the current one), my analyses will exclusively focus on attributive adjectives. Although the latter do not represent the most essential features (not the most peripheral either), their importance lies in the fact that they constitute the most implicit and thus probably also the most influential form of description, which justifies this biased choice. Limitations on space also make an exhaustive analysis of all structures impossible (but see my homepage, which also includes lists of predicative descriptors of different types). Translating my general assumptions concerning physicalization and visualization into concrete expectations concerning attributive adjective descriptors for whole persons, I can make the following predictions: • The type and token frequencies of visual physical descriptors will exceed those of other categories or will at least feature prominently in the pornography corpus. • The type and token frequencies of women’s visual physical descriptors will be higher than those of men’s. • In the erotica corpus, the type and token frequencies of visual physical descriptors will be lower relative to those of other categories than in the pornography corpus. • There will be no significant gender bias with regard to the type and token frequencies of visual physical descriptors in the erotica corpus, or at least the gap from women to men will not be as wide as in the pornography corpus. Additionally, I predict other categories of descriptors claimed to contribute to objectification to support the expected trends. <?page no="186"?> Analyses 186 1.1 Technical and practical aspects of the analysis Attributive APs per definition have to be part of NPs, modifying the head noun. The search string can therefore be defined as ‘adjective + noun’ with the noun position being filled by a gender-specific lexeme. As an exhaustive analysis of all gender-specific nouns premodified by adjectives is impossible because they do not share any formal features, I use the select set (or almost the same set) already drawn upon above, viz. woman, girl, lady, wife, girlfriend, mother, daughter, aunt, sister, grandmother, niece for females and man, boy, guy, husband, boyfriend, father, son, uncle, brother, grandfather, nephew for males. These are supposed to encompass the vast majority of gender-specific nouns. In the search string defined above, I do not need to allow for space between adjective and noun because structures with intervening words are impossible: APs are the only premodifiers of nouns, which means that no other phrase (save another adjective phrase) can be closer to the head, and postmodified APs, i.e. adjective phrases where the last word is not the adjective itself, cannot premodify a noun - e.g. *a busy doing other things woman (this even applies to compound adjectives such as the already wet pussy; see below). My search string will produce concordances such as the one below (for female nouns). > like<II> a<AT1> complete<JJ> woman<NN1> as<CSA> newly- > met<VVD> a<AT1> wonderful<JJ> woman<NN1> who<PNQS> mad mpsed<VVN> a<AT1> live<JJ> woman<NN1> ‘s<GE> vagina<NN1> 1> and<CC> a<AT1> strange<JJ> woman<NN1> -<-> naked<JJ> RR> to<II> a<AT1> naked<JJ> woman<NN1> before<RT> .<.> F D> and<CC> a<AT1> young<JJ> girl<NN1> ,<,> perhaps<RR> t > like<II> a<AT1> wild<JJ> woman<NN1> .<.> She<PPHS1> hu > when<CS> a<AT1> new<JJ> girl<NN1> joined<VVD> us<PPIO2 N1> of<IO> a<AT1> teenage<JJ> daughter<NN1> ,<,> but<CCB 1> am<VBM> a<AT1> horny<JJ> lady<NN1> ‘s<GE> wet<JJ> dre The<AT> small<JJ> oriental<JJ> woman<NN1> who<PNQS> was< <NN1> .<.> A<AT1> sensual<JJ> woman<NN1> .<.> A<AT1> wom Concordance 1: Search string ‘adjective immediately followed by woman, lady, girl, etc.’ in the pornography corpus (extract). From these concordances I manually extract all adjectives. Proceeding manually rather than using the Collocation function allows me to ignore those attributive APs that occur inside predicative NPs because these constitute a different category with different functions, as indicated in the theoretical discussions of structures above. The manual approach also allows me to identify cases where there is more than one attributive <?page no="187"?> Describing people 187 adjective, which I can then include in my list, e.g. the small oriental woman. I interpret participles (e.g. blushing or fucked) premodifying nouns (e.g. blushing boy) as adjectives and therefore include them here. This means that I also carry out searches replacing adjectives by present or past participles in the search string because the tagging is not consistent in its treatment of such forms (sometimes they are categorized as adjectives, sometimes as verbal forms). The additional words are then added to the respective lists. Let me turn to two cases that I have included even though they might seem problematic. Firstly, adjectives in the comparative or superlative remain in my lists but are subsumed under their base forms since the focus is on the semantic category of the adjective base and not on its grammatical appearance (for exceptions, see the comments in the data presentation on my homepage). Secondly, the modification of attributive APs is theoretically restricted to intensifiers, but it should not include other phrases or words. We can thus say a very large girl or an incredibly sexy Latin stud, but not *a today angry man or *an in this situation nervous woman. A few adverbs, however, seem to deverify this assumption, e.g. stylishly dressed (in a slim, stylishly dressed man), already hot (in my already hot pussy) or still wet (in my still wet shaft). The question now is whether these are syntactic modifications, i.e. whether attributive adjectives can take certain forms of adverb phrases as premodifiers, or whether these are just instances of compounding so that still wet is actually a compound adjective. The fact that expressions such as those just mentioned often occur with a hyphen (e.g. still wet actually appears as still-wet) indicates that the writers, i.e. native speakers of English, intuitively opt for the compound solution. I therefore count such combinations as single words and consequently as separate types. What are now the semantic classes relevant for descriptions? The following set will be used in the analysis, even though they will not all feature equally prominently. The taxonomy is based not only on a priori intuitive judgement but also on explorations of the corpora. It is finely structured because it is intended to be applicable to descriptors of whole persons (this subchapter) and body parts (next subchapter) and because, ambitiously speaking, it is supposed to be useful in descriptor analyses going beyond the current study. I also have to mention that some of the subcategories will be instrumental in grouping semantically-related expressions more closely together, thus allowing a better synopsis in the data presentation (see my homepage). They will, however, not necessarily feature separately in the analyses. <?page no="188"?> Analyses 188 General descriptors: describe humans only in very general terms (this might mean that they actually do not describe them at all). This group includes nouns such as man, woman, person and pronouns such as one, those, or somebody. While man, woman, girl, or boy belong to the categories of gender and/ or age, too, they are widespread and so general in nature that this multi-category membership (see below) seems justified. Visual descriptors: describe humans in terms of physically visible aspects. The subcategories are: • Shape: describe humans in terms of the qualitative forms of their bodies or their body parts. This covers subaspects such as contour shape, e.g. round, stocky, curvaceous, positional shape, e.g. open, spread, and surface shape, e.g. velvety, furry. • Size: describe humans in terms of the quantitative dimensions of their bodies or their body parts, e.g. tall, slim, large. • Colour: describe humans in terms of the colours of their bodies or their body parts, e.g. dark-skinned, pale, blonde. • Aesthetic: describe humans in terms of the aesthetic qualities of their bodies, e.g. beautiful, beauty, pretty, ugly. • Adhesive: describe humans in terms of physical objects or substances attached to, and thus not really part of, them. This category is split into: - Adhesive/ dress: dressing is essential, e.g. stockinged, welldressed. - Adhesive/ substance: substances on the surface of the body are essential, e.g. dirty, wet. As substances on the body can usually also be felt by touch, the adhesive/ substance category transcends the boundaries between vision and touch and its members are consequently also part of the tactile category below (the only exception are descriptors of cleanliness, which are exclusively visual). • Presence or absence: describe humans in terms of the mere presence or absence of something visual. These features usually involve head, body or facial hair, e.g. bearded, bald, shaved, but also include body parts such as the penis’ foreskin, e.g. uncut, circumcised. <?page no="189"?> Describing people 189 • Others: describe humans in terms of more general visual features or in terms of visual features that cannot be assigned to any of the other categories, e.g. delicate-looking. Tactile descriptors: describe humans in terms of what they feel like. We can distinguish between: • Surface: describe humans in terms of the haptic qualities of their surface (usually the skin), e.g. hard, smooth. Such descriptors sometimes count as visual (surface shape), too. • Temperature: describe humans in terms of their temperatures, e.g. cold, warm. We have to bear in mind, however, that many temperature descriptors are metonymically and/ or metaphorically used for other aspects, e.g. warm and cool for emotions, hot for sexual arousal or positive evaluations. Other physical descriptors: describe humans in terms of nonvisual and non-haptic sensory aspects. The subcategories are: • Auditory descriptors: describe humans in terms of what they sound like, e.g. low-voiced, sweet-sounding. • Olfactory/ gustatory: describe humans in terms of what they smell or taste like, e.g. smelly, perfumed, salty. Physiological descriptors: features relating to the body but not describing - on a primary level - its visuality. Physiological features can be permanent, e.g. strong, weak, or temporary, e.g. hungry, sick. Many of the members of this category have a strong (proprio)perceptive element, which means they denote a physiological state which we perceive via our own senses, e.g. hungry. They thus bear some resemblance to psychological features and are not always easy to distinguish from them. It makes sense to divide the descriptors into the following subcategories. • Body heat: describe humans in terms of the temperature of their bodies. As descriptors in this category denote states that are only perceptible via our temperature sense, I will assign them to the tactile category, too. Many of the descriptors of high heat also denote sexual arousal and are therefore also included in this category (see below). • Health: describe humans in terms of states of health or illness or other medically-relevant aspects, e.g. healthy, well, ill, pregnant. <?page no="190"?> Analyses 190 • Sexuality: describe humans in terms of their sexuality. This mainly includes features of sexual arousal, e.g. hot, randy, aroused, sexual potential, (im)potent, and sexual events, e.g. orgasming. • Others: describe humans in terms of physiological aspects not included in the three subcategories above, e.g. strong. Social descriptors: describe humans in terms of social aspects. The subcategories are: • Social stratum or profession: describe humans in terms of their membership in a social class, e.g. rich, salary elite, middle-class, or their profession, e.g. manager, policewoman. • Social relationships: describe humans in terms of their marital status, family roles, (sexual) preferences, etc., e.g. single, mother, lesbian. • Origin (nationality, ethnicity): describe humans in terms of their ethnic, national or regional origin, e.g. African American, Vietnamese, Londoner. • Age: describe humans in terms of their age, e.g. old, toddler. I classify age as social because I assume that we tend to see it as a bundle of social properties such as practices, attitudes, preferences, etc. rather than as primarily a biological condition of the body. • Gender: describe humans in terms of gender, e.g. woman, man, boy, female. Like age, gender is interpreted as a social feature (as gender, in the narrow sense of the word) and not as a physiological one (not as sex; for the difference between sex and gender, cf. Cameron 1992c: 40, 241). • Social prominence: describe humans in terms of how well known they are within a society, e.g. famous, anonymous, familiar. • Others: describe humans in terms of social features that cannot be assigned to any of the other categories, e.g. guilty, responsible - both features make only sense in the context of a society and its norms. Psychological descriptors: describe humans in terms of psychological aspects. A distinction between personality traits (more permanent psychological features, e.g. ambitious, courageous, <?page no="191"?> Describing people 191 stupid, also including attitudinal properties, e.g. feminist 5 ) and temporary cognitive or affective states (e.g. frightened, certain) appears plausible, but cannot be used successfully because many descriptors could belong to either category. He was friendly might, for instance, be interpreted as a permanent feature of somebody’s character or a momentary feature of his behaviour. I will therefore not use this distinction. Activities: describe humans in terms of situations which they find themselves in or activities which they are engaged in. We can distinguish between: • Active participation: e.g. visiting (visiting aunts) or busy. I interpret active very broadly here, so that I can include many participle adjectives such as trembling or fading, which are not deliberate activities, to be linguistically precise. • Passive participation: e.g. fucked, saved, imagined. This is a residual category because many activities belong to other semantic classes and are consequently included there, e.g. sleeping and orgasming are classified as physiological activities. Minor categories: are classes that do not fit into any of the other large categories and seem to contain few (but sometimes highly frequent) members themselves. • Spatial: describe humans in terms of their spatial relationships to other entities, e.g. close, distant. • Temporal: describe humans and their activities and relations in terms of their positions on a time scale, e.g. last, current. • Epistemological: describe humans’ judgement with respect to truth or falsity, e.g. she is right, he is wrong. This epistemological relation to truth is their main element and they are therefore not counted as psychological. Evaluative descriptors: describe humans in terms of evaluative judgements, e.g. good, perfect, slut. I will try to include only those descriptors whose primary meaning is evaluative, but not those words (or uses of words) that have evaluative/ affective connotations, e.g. nasty, which describes somebody’s character. The two aspects cannot be clearly distinguished in all cases. In some 5 The borderline to social categories is problematic here. <?page no="192"?> Analyses 192 instances a neat distinction from the aesthetic subcategory - which implies an evaluation on a physical beauty scale - is not possible either (e.g. is lovely an ordinary evaluation or an aesthetic judgement). Dependent descriptors: depend on other meaningful elements because they do not carry any significant ideational meaning themselves. Their primary function often is to restrict the reference of the element it goes together with, e.g. only, particular, specific, typical, various (cf. Biber et al. 1999: 509 therefore speak of relational/ classificational/ restrictive descriptors). 6 They may occur as nouns, e.g. part, but they are more commonly expressed by (attributive) adjectives. Unclassifiable: some of the descriptors cannot be assigned to any of the categories given. The classification with the help of this taxonomy will be much more problematic than with body parts. Since the categories used for the latter were based on discrete anatomical areas of the body and thus on at least partly objective external criteria, decisions were mostly unambiguous. With descriptors, on the other hand, the problematic cases abound, not only with boundary cases (e.g. is nasty really evaluative or psychological? ), but also with lexemes that seem to belong to more than one category at the same time (this does not include homonyms or polysemes, e.g. sweet for taste or as evaluation). Black, for instance, is used to describe somebody’s skin colour but also their ethnicity. Or hot denotes body heat as well as sexual arousal. This is why the taxonomy requires, as indicated in the definitions, that I allow for multiple-class membership, which means that in the analysis I will assign certain descriptors to more than one class. There is another problem inherent in my set of classes. As mentioned above, some categories may not be essentially physical or visual, but point to this dimension on a secondary level. This particularly holds for the age, gender and the origin subcategories, where visuality appears to be a signifier. This has to be born in mind when evaluating the quantitative relations of the individual categories. 6 Some of these adjectives have a different meaning in a predicative position, e.g. certain in a certain guy is restrictive, but in they were certain it describes a cognitive state. <?page no="193"?> Describing people 193 1.2 Presenting and discussing results The full lists of attributive adjectives (together with the lists of predicative descriptors) from the pornography corpus can be found on my homepage. Without any further ado, I will turn to the sizes of the different categories. The table below, with the help of the diagram further down, provides all relevant information on lexical variation, listing the type frequencies in the semantic classes. Types Overall % % % General 0 0% 0 0% 0 0% Visual physical 68 25.0% 49 23.9% 117 24.5% Tactile physical 3 1.1% 2 1.0% 5 1.0% Other physical 1 0.4% 2 1.0% 3 0.6% Physiological 18 6.6% 20 9.8% 38 8.0% Social 44 16.2% 32 15.6% 76 15.9% Psychological 64 23.5% 43 21.0% 107 22.4% Activities 23 8.5% 12 5.9% 35 7.3% Minor 10 3.7% 9 4.4% 19 4.0% Evaluative 41 15.1% 37 18.0% 78 16.4% Dependent 7 2.6% 4 2.0% 11 2.3% Unclassifiable 0 0% 0 0% 0 0% Totals 7 272 205 477 Table 1: Absolute and relative numbers of different attributive adjectives (in nonpredicative NPs) functioning as descriptors of women and men in the different semantic categories in the pornography corpus. 7 Percentages do not add up to 100% because of multi-category membership. <?page no="194"?> Analyses 194 0% 5% 10% 15% 20% 25% 30% General Visual physical Tactile physical Other physical Physiological Social Psychological Activities Minor Evaluative Dependent Unidentifiable Female Male Figure 1: Relative sizes of the semantic categories with respect to lexical variation of attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men in the pornography corpus. In full accordance with what I predicted, visual descriptors exceed all other categories in lexical variation. This suggests that visuality will indeed be in a salient position in readers’ conceptualization of human beings, which might be even more effective for creating physicalization and visualization because, following my argumentation from above, attributive adjectives in non-predicative NPs firstly are closer to the centre of a person than predicative ones and secondly are presupposed and downplayed, thus not attracting conscious attention. Even though as adjectives they will not be perceived as essential properties (as nouns would), the data adds weight to the postulated assumption that visualization will play a major role in pornography consumption. As far as the other relevant categories are concerned, there is also great variation in psychological descriptors, which would reduce the effect of visualization a bit, but also in the evaluative category, which, as argued above, might increase the overall objectifying potential of descriptors. The type frequency of the social category is also high, with disproportionally large variation in the subcategory of origin, which bears a strong secondary visual moment. Comparing variation for the two genders does not reveal the expected advantage of female descriptors. They are slightly in the lead, but the gap is not wide enough to provide convincing evidence in favour of the hypothesis that women’s visualization will play a more seminal role in pornography consumption than men’s. Gender differences in other classes prove inconclusive, too, with the objectifying category of evaluative descriptors showing higher percentages for men and the anti-objectifying category of psychological descriptors higher ones for females. <?page no="195"?> Describing people 195 Moving from types to tokens, the table below contains the relevant figures and percentages on how often descriptors occur in the pornography corpus. The relative sizes are graphically represented in the diagram immediately below. Tokens Overall % % % General 0 0% 0 0% 0 0% Visual physical 213 29.4% 129 24.3% 342 27.3% Tactile physical 3 0.4% 2 0.4% 5 0.4% Other physical 1 0.1% 2 0.4% 3 0.2% Physiological 32 4.4% 48 9.0% 80 6.4% Social 185 25.6% 177 33.3% 362 28.8% Psychological 86 11.9% 52 9.8% 138 11.0% Activities 23 3.2% 13 2.4% 36 2.9% Minor 18 2.5% 13 2.4% 31 2.5% Evaluative 128 17.7% 64 12.1% 192 15.3% Dependent 64 8.8% 44 8.3% 108 8.6% Unclassifiable 0 0% 0 0% 0 0% Totals 724 531 1,255 Table 2: Absolute and relative frequencies of attributive adjectives (in nonpredicative NPs) functioning as descriptors of women and men in the different semantic categories in the pornography corpus. 0% 5% 10% 15% 20% 25% 30% 35% General Visual physical Tactile physical Other physical Physiological Social Psychological Activities Minor Evaluative Dependent Unidentifiable Female Male Figure 2: Relative sizes of the semantic categories with respect to token frequencies of attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men in the pornography corpus. The token frequencies reconfirm the impressions gained from the type frequencies and are in line with my expectations. Even though not <?page no="196"?> Analyses 196 ranked first, visual features also occur very often in the pornography corpus, adding further credence to the hypothesis of a strong moment of visualization in pornography, which is, as mentioned, particularly valuable given the effectiveness of the structure analysed. This conclusion is corroborated by the strong presence of evaluative descriptors and of the social subcategories of origin and age with their secondary visual aspects (the comparatively high frequencies of the words young and old are responsible for the dramatic imbalance between types and tokens here) and by the fact that psychological descriptors occur rarely, with their percentage being much lower than in the type column. The data on female and male descriptors is more in line with what was predicted than that on lexical variation, with a greater gap between women and men. Considering that we also find a female advantage in the evaluative category (but also, surprisingly, in the psychological one), we can interpret this at least as qualified support for the assumption that women’s visualization will be stronger in pornography than men’s. The table below, with the ten most frequent attributive adjectives used for males and females, rounds off this section. Women Men young 76 young 79 other 54 old 40 beautiful 37 other 40 old 26 big 24 sexy 20 handsome 21 lovely 18 horny 17 naked 18 black 9 pretty 17 good-looking/ good looking 9 little 11 Japanese 6 gorgeous 10 dark-haired/ dark haired 5 good 5 little 5 real 5 sexy 5 Table 3: The ten most frequent attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men in the pornography corpus (adjectives that occur as often as the 10 th one included) (visual descriptors in bold). The top ten descriptive lexemes reflect the general trends. Attributes of age, namely young and old, are by far the most frequent ones. The most widely represented group among the top ten are visual descriptors with <?page no="197"?> Describing people 197 almost 50% for women and men. Another important class are evaluative descriptors, which can contribute to the overall amount of objectification. These facts underscore the importance of these categories. Gender results are difficult to compare due to the fact that the tenth place in the male column is shared by five words, but female visual descriptors seem slightly higher ranked than male ones. 1.3 Comparing and discussing results On my homepage I provide a detailed presentation of the data. But I will immediately turn to overall lexical variation here. The table below, visually supported by the diagram following, presents all relevant type frequencies in the erotica corpus. The second table then brings together pornography and erotica figures. Types Overall % % % General 0 0% 0 0% 0 0% Visual physical 26 23.8% 34 29.1% 60 26.5% Tactile physical 2 1.8% 1 0.9% 3 1.3% Other physical 1 0.9% 2 1.7% 3 1.3% Physiological 6 5.5% 5 4.3% 11 4.9% Social 28 25.7% 29 24.8% 57 25.2% Psychological 20 18.3% 19 16.2% 39 17.3% Activities 8 7.3% 2 1.7% 10 4.4% Minor 4 3.7% 3 2.6% 7 3.1% Evaluative 10 9.2% 21 17.9% 31 13.7% Dependent 4 3.7% 3 2.6% 7 3.1% Unclassifiable 1 0.9% 0 0% 1 0.4% Totals 109 117 226 Table 4: Absolute and relative numbers of different attributive adjectives (in nonpredicative NPs) functioning as descriptors of women and men in the different semantic categories in the erotica corpus. <?page no="198"?> Analyses 198 0% 5% 10% 15% 20% 25% 30% 35% General Visual physical Tactile physical Other physical Physiological Social Psychological Activities Minor Evaluative Dependent Unclassifiable Female Male Figure 3: Relative sizes of the semantic categories with respect to lexical variation of attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men in the erotica corpus. Types Overall Porn Erotica Porn Erotica Porn Erotica General 0% 0% 0% 0% 0% 0% Visual physical 25.0% 23.8% 23.9% 29.1% 24.5% 26.5% Tactile physical 1.1% 1.8% 1.0% 0.9% 1.0% 1.3% Other physical 0.4% 0.9% 1.0% 1.7% 0.6% 1.3% Physiological 6.6% 5.5% 9.8% 4.3% 8.0% 4.9% Social 16.2% 25.7% 15.6% 24.8% 15.9% 25.2% Psychological 23.5% 18.3% 21.0% 16.2% 22.4% 17.3% Activities 8.5% 7.3% 5.9% 1.7% 7.3% 4.4% Minor 3.7% 3.7% 4.4% 2.6% 4.0% 3.1% Evaluative 15.1% 9.2% 18.0% 17.9% 16.4% 13.7% Dependent 2.6% 3.7% 2.0% 2.6% 2.3% 3.1% Unclassifiable 0% 0.9% 0% 0% 0% 0.4% Table 5: Comparison of the different semantic categories with respect to lexical variation of attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men between the pornography corpus and the erotica corpus. The erotica corpus, too, sees visual descriptors taking the top spot with respect to lexical variation, with the relative size of the category being even larger than in the pornography corpus, which clearly contradicts my prediction. This quantitative comparison therefore provides evidence against my hypothesis that visualization and physicalization will be a less important factor in readers’ interpretation of erotica than in that of <?page no="199"?> Describing people 199 pornography. This is particularly significant given that the structure under scrutiny is the most effective one in making people accept assumptions, even though it does not represent the most essential properties. Some of the other data makes this conclusion seem more plausible. For instance, the relatively large proportions of the origin subcategory, with its visual overtones, might add to visualization in erotica, and the smaller size of the psychological category might fail to relativize its effects. Pornography’s lead with respect to the proportions of evaluative descriptors - also interpreted as an objectifying category - cannot completely reverse this trend. The prediction about gender differences, on the other hand, is corroborated by the percentages found in the tables above. In contrast to the pornography corpus with its consistent advantages of female visual descriptors, there is more variation in this category for men in the erotica corpus, suggesting that, as hypothesized, women will not be more consistently subjected to visualization in erotica than men. This is also underlined by the great male lead in the objectifying category of evaluative descriptors. How do these numbers compare to token frequencies? The two tables - the second one comparing erotica and pornography data - and the diagram will provide answers to this question. Tokens Overall % % % General 0 0% 0 0% 0 0% Visual physical 40 19.7% 75 30.9% 115 25.8% Tactile physical 3 1.5% 1 0.4% 4 0.9% Other physical 1 0.5% 2 0.8% 3 0.7% Physiological 6 3.0% 5 2.1% 11 2.5% Social 83 40.9% 100 41.2% 183 41.0% Psychological 23 11.3% 19 7.8% 42 9.4% Activities 8 3.9% 2 0.8% 10 2.2% Minor 5 2.5% 3 1.2% 8 1.8% Evaluative 13 6.4% 30 12.3% 43 9.6% Dependent 21 10.3% 16 6.6% 37 8.3% Unclassifiable 1 0.5% 0 0% 1 0.2% Totals 203 243 446 Table 6: Absolute and relative frequencies of attributive adjectives (in nonpredicative NPs) functioning as descriptors of women and men in the different semantic categories in the erotica corpus. <?page no="200"?> Analyses 200 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% General Visual physical Tactile physical Other physical Physiological Social Psychological Activities Minor Evaluative Dependent Unclassifiable Female Male Figure 4: Relative sizes of the semantic categories with respect to token frequencies of attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men in the erotica corpus. Tokens Overall Porn Erotica Porn Erotica Porn Erotica General 0% 0% 0% 0% 0% 0% Visual physical 29.4% 19.7% 24.3% 30.9% 27.3% 25.8% Tactile physical 0.4% 1.5% 0.4% 0.4% 0.4% 0.9% Other physical 0.1% 0.5% 0.4% 0.8% 0.2% 0.7% Physiological 4.4% 3.0% 9.0% 2.1% 6.4% 2.5% Social 25.6% 40.9% 33.3% 41.2% 28.8% 41.0% Psychological 11.9% 11.3% 9.8% 7.8% 11.0% 9.4% Activities 3.2% 3.9% 2.4% 0.8% 2.9% 2.2% Minor 2.5% 2.5% 2.4% 1.2% 2.5% 1.8% Evaluative 17.7% 6.4% 12.1% 12.3% 15.3% 9.6% Dependent 8.8% 10.3% 8.3% 6.6% 8.6% 8.3% Unclassifiable 0% 0.5% 0% 0% 0% 0.2% Table 7: Comparison of the different semantic categories with respect to token frequencies of attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men between the pornography corpus and the erotica corpus. The token results see some deviations from the type results. Visual descriptors are ranked high again but they do not reach the top spot. Besides, more in line with what was expected, the relative size of this category is smaller in the erotica corpus than in the pornography corpus. The margin, however, is too small to speak of real support for the <?page no="201"?> Describing people 201 hypothesis that visualization will play a less significant role in erotica than in pornography consumption. The only result suggesting weaker objectification in erotica is the lower percentage of evaluative descriptors. This, on the other hand, is compensated by a great preponderance - even greater than in the pornography corpus - of the social subcategory of age with its secondary visual dimension (this is primarily a result of the high frequencies of the words old and young). Let me just add one comment on the categories not as immediately involved in visualization or objectification: social features top the tables. The predominance of this category is more impressive than in the pornography corpus. This could mean that the trend of presenting primarily the social sides of humans is very prominent in erotica. My predictions about gender differences are not only supported by the type frequencies but also by the token frequencies, which see men in front by more than ten percentage points. This lead is in line with the fact that we also find men ahead of women in the evaluative category, but trailing them in the anti-objectifying psychology category. All this data lends some more weight to the view of a weaker visualization of women (relative to that of men) in erotica than in pornography. Let me finally take a look at the ten most frequent attributive adjectives describing women and men in the erotica corpus. Women Men old 31 little 21 young 18 young 20 other 15 gay 15 beautiful 5 straight 13 little 5 other 12 only 4 old 9 good 3 white 8 nice 3 XY-year-old 5 pretty 3 big 4 straight 3 uncut 4 Table 8: The ten most frequent attributive adjectives (in non-predicative NPs) functioning as descriptors of women and men in the erotica corpus (visual descriptors in bold). The table succinctly reproduces the results gained above on a different plane. Age attributes together with the relationship-related features straight and gay underline the predominance of social descriptors. But <?page no="202"?> Analyses 202 visual features also figure prominently. No gender bias in favour of women, however, becomes apparent, since male visual descriptors, as a matter of fact, claim more of the top ten words than female ones. The main conclusion from the data examined in this subchapter is that most of the predictions have been confirmed. The variation and frequency of visual physical descriptors is high in the pornography corpus. We can also see a female advantage in this category with respect to type and token frequencies in the pornography corpus (even though not a decisive one), something we cannot find in the erotica corpus (the relations are actually reversed there). The only aspect not in line with what was expected is the fact that the erotica stories overall show more lexical variation in visual descriptors and are almost on par with their pornographic counterparts with regard to the token frequencies of this class. The other descriptor categories do not provide a homogeneous picture here either, which they by and large do with respect to the other questions. The data thus adds to the plausibility of all my superordinate hypotheses, with the exception of the one concerned with the difference between pornography and erotica with respect to the overall amount of visualization created by the two discourses. 2 Describing body parts: attributive descriptors While a wide variety of different facets are described in human beings as wholes, their bodies, representing their perceptible sides, are supposed to attract primarily physical descriptors. I can therefore expect a different, broader and richer set of the latter. For this reason, I will tackle the problem differently here, exclusively focusing on visual descriptors and the quantitative relations between their subcategories. This entails that I will refrain from doing an analysis of all descriptors of body parts (apart from short glimpses at the ten most frequent attributes towards the end of the two sections), even though I have the data available and there is no shortage of interesting issues. However, a proper handling of the data appears infeasible to me, with the majority of descriptors being part of two or three classes (e.g. wetness indicating sexual arousal in women - a highly frequent descriptor - is classified as visual (adhesive/ substance), tactile, and physiological (sexual)). Large areas of the categories are thus overlapping, which makes it difficult to reach conclusive results. Why should examining visual subcategories be relevant for the topic of visualization and physicalization? I think there are some attri- <?page no="203"?> Describing body parts 203 butes deserving particular attention because they might positively or negatively contribute to the overall objectifying effect of the visual category. It will hence be highly relevant for my research to see, for instance, how the quantitative features of size compare to the qualitative ones of shape or colour in women and men. I argue that size and in particular superlative size is tightly linked to having or exercising power. As being in a powerful and influential position is not associated with being relegated to an object-like passivity, size could undermine the overall objectifying effect of visualization. Making the owner seem powerful, they can be regarded as having an empowering or re-empowering (given that visualization reduces powerfulness) impact. This link to power being absent from shape and colour attributes, the latter will not have the same mitigating effects on objectification. I will treat this aspect here despite its connection to passivization, a conceptualization dealt with in a separate chapter below. The opposition between aesthetic and neutral features is another pivotal element within the visual category. Thinking of people in terms of aesthetic features bears an additional factor of objectification. Like evaluative descriptors, to which they are very similar, aesthetic descriptors imply the distinction between the subject who passes the judgement and the object which/ who is judged. The aestheticized person is consequently reduced to this object status. Aesthetic features are therefore relevant for the analysis. It might be objected that to include an analysis of descriptors of body parts does not really touch upon the issue of visualization and physicalization, not being concerned with whole human beings. I concede that the overall effect of the conceptualization might be less substantial than if produced by descriptors of whole persons, but I still assume that body parts are not completely dissociated from their owners so that the visualization of the former has implication for the conception of the latter. I also have to emphasize that what I will examine in this subchapter actually marks an area of interaction of visualization and physicalization with fragmentation. Based on the ideas presented above, I make the following predictions for this section: • In the pornography corpus, women will be assigned features of shape, colour and aesthetic judgement in greater variation and in higher frequencies than men, who, in turn, will receive more features of (superlative) size. • This difference will not be as pronounced in the erotica corpus. <?page no="204"?> Analyses 204 My argumentation above suggests that size may reduce the extent of visualization and should therefore, following my general hypotheses, be less frequent in pornography than in erotica. I doubt that this will indeed be the case given pornography’s frantic preoccupation with superlative size. The reason for this might lie in the fact that the association with power can only work in a comparison between groups, but not on an overall scale as power implies differences and imbalances. 2.1 Technical and practical aspects of the analysis To find out which of the body parts represented is assigned further information through an adjective phrase means looking for a subset of those examples found in the chapter on “Fragmentation” (see V.1). I am thus looking for all NPs starting with a - gender-specific - possessive and taking a body part lexeme as their heads with an AP in between. As APs are headed by adjectives and these can be identified by the concordance programme with the help of tags, I search for any string ‘possessive + adjective’. I do not have to include a noun in the sequence since the combination given necessarily marks the beginning of an NP and thus implies the presence of a nominal head. The search string produces concordances such as the following one (for his as the keyword). D> n’t<XX> like<VVI> his<APPGE> young<JJ> wife<NN1> .<.> NN1> lubricates<VVZ> his<APPGE> hairy<JJ> balls<NN2> .<. to<TO> match<VVI> his<APPGE> powerful<JJ> thrusts<NN2> a 1> matched<VVD> his<APPGE> increasing<JJ> thrusts<NN2> howed<VVD> me<PPIO1> his<APPGE> tight<JJ> ass<NN1> my<AP o<TO> meet<VVI> his<APPGE> in-sliding<JJ> cock<NN1> .< lowly<RR> moving<VVG> his<APPGE> huge<JJ> cock<NN1> in<R I21> of<II22> his<APPGE> ultra-hairy<JJ> crotch<NN1> t<II32> of<II33> his<APPGE> twitching<JJ> penis<NN1> ,< I32> of<II33> his<APPGE> still-closed<JJ> door<NN1> 1> inch<NNU1> of<IO> his<APPGE> thick<JJ> prong<NN1> .<. Concordance 2: ‘His + adjective/ participle’ in the pornography corpus (extract). The concordance gives me all occurrences of phrases such as his thick prong or his ultra-hairy crotch, but also phrases with multiple APs, e.g. his handsome steel-like shaft or his delighted and delightful organ - these will of course be counted separately. All phrases not including body parts, e.g. his young wife, have to be eliminated. As in V.1, the different possessives, i.e. his, her, my - in the male and the female subcorpora - and genitive NPs, have to be searched <?page no="205"?> Describing body parts 205 separately. This means that the different lists produced by these searches have then to be combined to form one female and one male list. In order not to omit examples with APs not encompassed by the concordances from above, I also search the corpus for the string ‘possessive + adverb (including intensifiers)’, which should produce all NPs with modified adjective phrases such as my very hard nipples. Again, some of the examples will not include body parts and will therefore be ignored. The rest will be added to the unmodified adjective phrases. As in the previous subchapter, the same searches are carried out for participles so that I also include modifiers interpreted as a verb form by the tagging programme, e.g. her opening snatch. The lists of body parts with their modifying APs are finally edited to ensure that the adjectives that are considered to be part of a complex lexeme, e.g. bountiful charms, big toe or upper arm, are not included. The sets of adjectives are finally categorized according to the classes described above, with only those of the visual category and its subcategories being considered. 2.2 Presenting and discussing results Visual descriptors are not a homogenous class and a closer examination of the subcategories will reveal what visualization actually amounts to and whether there are aspects that may run counter or support tendencies of objectification. As mentioned in the introduction to this subchapter, it is particularly the size category that is interesting in connection with the overall objectifying potential of body part descriptors. Before dealing with the results, let me discuss the relevance of the size category in more depth now that I can illustrate its nature with the help of authentic data (all detailed wordlists can be found on my homepage). For this purpose, I will present the descriptors in a different fashion in the table below. I will focus on the prototypical cases, excluding descriptors of empty space such as tight or deep, which, as a matter of fact, measure the absence rather than the presence of physicality, as well as words that make direct reference to a concrete body part, e.g. long-lashed or fulllipped. The remaining set will be divided into positive size (large, big, high, fat) and negative size descriptors (small, short, thin) (positive and negative are not meant in an evaluative sense). <?page no="206"?> Analyses 206 Positive size abundant; ample; big; bloated; bounteous; bountiful; broad; centerfold lush; distended; endless; engorged; enlarged; enormous; extended; fair-sized; fat; full; generous; giant; good-sized; great; growing; high; huge; large; long; lush; lustswollen; mammoth; monstrous; mountainous; over-developed; overstretched; pleasure-swollen; queensized; rich; sizeable; still-swollen; supple; swelling; swollen; thick; titanic; voluptuous; well-developed; wide abundant; already large; big; bloated; broad; considerable; corpulent; distended; egg-sized; elephantive; engorged; enlarged; enormous; extended; fat; full; gargantuan; giant; good-sized; grapefruit-sized; great; growing; huge; humongous; large; lengthening; lengthy; long; longish; luge; magnum; massive; mighty; monster; monstrous; outstretched; overgrown; oversized; swelling; swollen; tall; thick; tremendous; tumescent; tumescing; turgid; wide Negative size fast-degrading; little; narrow; pencilslim; petite; shallow; short; shriveled; slender; slight; slim; small; smallish; thin; tiny; wispy diminished; dwindling; fast-shrinking; little; minuscule; puny; rapidlydwindling; short; shrinking; shrivelled; slim; small; thin; tiny Table 9: Positive and negative size descriptors of female and male body parts in the pornography corpus. The list shows that although negative size, too, is overworded, the variation is still much higher in the area of positive size, so that size is likely to be conceived of as largeness rather than as smallness, or as more rather than as less. In many cases, we could even speak of superlative size, i.e. size verging on the maximum. This becomes manifest in the fact that many of the adjectives are not gradable and therefore cannot occur in the comparative (e.g. *more monstrous, *more endless, *gianter, *more titanic, *more enormous, but: longer, larger, bigger), but take maximizing intensifiers such as absolutely or fully, which are restricted to adjectives with superlative meanings (cf. Lorenz 1999: 155f.) (e.g. absolutely endless, absolutely monstrous, fully extended, but: *absolutely large, *completely huge, *fully huge). This aspect is further underlined by hyperbolic metaphors such as mountainous, giant, elephantive, titanic, monster/ monstrous, or gargantuan, which take as their sources, i.e. the literal meanings of the expressions, concepts disproportionate in their extensions to genitals. Since large size and extreme size are easily associated with superiority and consequently with power, the emphasis on these properties created by the sort of overlexicalization just discussed will have a re-empowering effect, making readers conceptualize the owners of such ‘supersize’ genitals as power- <?page no="207"?> Describing body parts 207 ful. Obviously the Western culture of growthism, which is opposed to the - according to ecologists - sounder motto of “Small is beautiful,” (cf. Halliday 1992, cit. in Fill 1996a: 9) does not only apply to economics but also to sexuality, where largeness is also (over)emphasized. As may have been noticed in the examples above, size descriptors are not always easy to distinguish from shape descriptors. In particular horizontal size, i.e. slim vs. fat, can in some instances hardly be discerned from shape aspect of the abdominal area. I have tried to base my decision on a judgement on whether the extension or the shaping is foregrounded. Despite a conscientious analysis, a completely consistent categorization is, however, not possible in all cases. The percentages of the nine visual subcategories with respect to lexical variation are given in Table 10 and are graphically represented in Figure 5. Percentages represent the size of a category relative to the overall number of visual descriptors (and not to all descriptors). Types Overall % % % Shape 129 29.7% 86 34.1% 215 31.3% Size 78 18.0% 69 27.4% 147 21.4% Colour 92 21.2% 34 13.5% 126 18.4% Aesthetic 12 2.8% 8 3.2% 20 2.9% Adhesive/ dress 33 7.6% 17 6.7% 50 7.3% Adhesive/ substance 85 19.6% 32 12.7% 117 17.1% Presence/ absence 11 2.5% 8 3.2% 19 2.8% Others 0 0% 0 0% 0 0% Totals 434 252 686 Table 10: Absolute and relative numbers of different attributive adjectives functioning as descriptors of male and female body parts in the eight visual subcategories in the pornography corpus. <?page no="208"?> Analyses 208 0% 5% 10% 15% 20% 25% 30% 35% 40% Shape Size Colour Aesthetic Adhesive/ dress Adhesive/ substance Absence/ presence Others Female Male Figure 5: Relative sizes of the physical visual subcategories with respect to lexical variation of attributive adjectives functioning as descriptors of female and male body parts in the pornography corpus. Shape is by far the most variably lexicalized subcategory, followed by size, with colour and adhesive/ substance (primarily wetness) in third and fourth spot. Given that shape covers a more diverse conceptual field than size, we still see the prominent position that the latter is assigned in pornography. However, as I did not make any predictions concerning overall relations, I will immediately turn to gender differences. My main expectation here was that the visual features attributed to men’s bodies would more often, relatively speaking, represent size than those attributed to women’s bodies. This assumption is supported by the data, which sees males clearly in front in the said category (by almost ten percentage points). Even if the data in Table 10 suggests that the absolute numbers of positive size descriptors (and those of negative size descriptors, too) are the same for women and men, the conclusion holds because it is based on the fact that with fewer modifiers of body parts for men in general, the same number means a much higher share in the overall set of descriptors. As the strong associations of size, in particular large size, with superiority and power might work against the objectifying tendencies, the results suggest that the salience of male visualization should be reduced, which, in turn, provides evidence in favour of the hypothesis of a stronger moment of female visualization being at work in pornography. What about the qualitative categories of shape, believed to sustain the objectifying moment of visualization? It does not correspond to my expectation, because although shape descriptors rank very high in the type columns of both genders, men are in front of women, even though not conspicuously. A closer look at this class shows that there is one <?page no="209"?> Describing body parts 209 particular conceptual area which is densely lexicalized, namely - to an even larger extent as with size - the state of erection (or non-erection) of the penis and thus the state of arousal (or non-arousal), with words such as erect, limp, bulging, dangling, wilting, upright, constantly-erect, erected, icicle, protruding, rapidly-reviving, sagging, semi-erect, stillerect, surging, towering, trouser-tenting, upstanding. The gap between males and females can thus be explained with reference to the fact that erotic excitement features as shape/ position of the penis. As the equivalent to the erect penis is the wet vagina, female sexual arousal features not as shape but as wetness, which is part of the adhesive/ substance category (almost exclusively composed of adjectives denoting wetness). The following list containing these wetness attributes demonstrates the overwording in this area. already-percolating juiced-up slopping squashed already-slick juice-slicked sloppy squirming clammy leaking sloshing squishy come-filled moist sloshy still slippery come-slick moistening smeared superwet cumsoaked moistly-aroused smudgy swampy damp nectar-soaked soaked sweat-covered dank oily soaking sweating dew-soaked oozing soaking-wet sweat-matted dewy overflowing soapy sweat-soaked drenched perspiring sodden sweaty dribbling pouring soggy swimming dripping pussy juice-soaked sopped syrupy dripping wet quickly dampening sopping tangy drooling rapidly-moistening sperm-covered ultra-slippery dropping salivating spit-filled undulating flowing seeping spit-polished water-slicked greasy slick spunk-drenched well-lubed humid slickened spunk-filled well-lubricated juice-coated slimy spunky wet juice-drenched slippery Table 11: Adjectives denoting wetness of female body parts in the pornography corpus. Shape features can thus not be examined in isolation but I have to take results in the adhesive/ substance category into account, which should cause no further problem as both are of a qualitative nature and can <?page no="210"?> Analyses 210 therefore not be regarded as re-empowering. I can hence conclude that while shape attributes on their own do not confirm my prediction, the adhesive/ substance category compensates for this, so that overall the non-empowering descriptors of female bodies exceed those of male bodies in lexical variation, lending more weight to the hypothesis of visual physicality playing a more central role in pornography readers’ conception of women than in their conception of men. In this context, it has to be mentioned that some of the size descriptors of the penis can also be interpreted as reference to the erection and thus as a sign of sexual arousal and would therefore also have to be compared to female wetness descriptors, which could relativize the male lead in this category. However, firstly, this link is not as strong as between shape and arousal, and secondly, the quantitative dimension is not diminished in its significance by this. The assumption that quality plays a more important role for women than for men is underscored by the results of colour descriptors, supposed to lack the re-empowering effect of the size class, too. They show more variation for women than for men, as expected, with 91 different adjective lexemes for the former and 33 for the latter. Mentioning the expressions used to modify the word hair should suffice to demonstrate the overwording operating in this field: auburn, black, blond, blue-black, brown, coal-black, dark, golden, jet-black, light, raven, red, sunbleached, natural blonde, naturally fair, sandy brown (as opposed to blond, black, chestnut, dark brown, dark, sunstreaked for males). The gap strengthens the assumption that qualitative categories will be more salient in the pornographic conceptualization of women than in that of men. Besides, the results perhaps also reconfirm the existence of the stereotype associating colour with women’s bodies (in particular with hair and eyes) more than with men’s. How do aesthetic features fare, which I have claimed to add to the objectifying dimension of visualization, too? They show a marginal advantage of males over females, surprisingly enough. Features of looks, however, rank unexpectedly low and will thus not be relevant for the conceptualization of humans and their bodies. This means that their influence on physicalization and visualization can largely be ignored. (Interestingly, the evaluative category, akin to aesthetic descriptors, does not figure as prominently in body part descriptions either, as my informal comparison of the non-visual data has shown.) I will now turn my attention to the frequencies at which these descriptors occur. The relevant figures can be found in the table below. <?page no="211"?> Describing body parts 211 Tokens Overall % % % Shape 453 21.3% 209 23.5% 662 21.9% Size 593 27.8% 405 45.6% 998 33.1% Colour 346 16.2% 96 10.8% 442 14.6% Aesthetic 114 5.4% 32 3.6% 146 4.8% Adhesive/ dress 186 8.7% 46 5.2% 232 7.7% Adhesive/ substance 405 19.0% 63 7.1% 468 15.5% Presence/ absence 44 2.1% 40 4.5% 84 2.8% Others 0 0% 0 0% 0 0% Totals 2,130 889 3,019 Table 12: Absolute and relative frequencies of attributive adjectives functioning as descriptors of male and female body parts in the eight visual subcategories in the pornography corpus. 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% Shape Size Colour Aesthetic Adhesive/ dress Adhesive/ substance Absence/ presence Others Female Male Figure 6: Relative sizes of the physical visual subcategories with respect to token frequencies of attributive adjectives functioning as descriptors of female and male body parts in the pornography corpus Generally speaking, size is the largest category with respect to tokens, with shape, colour and adhesive/ substance relatively far behind. This underlines the impression of the prominent role that the former generally plays in pornography gained by the type figures. We might take this as a sign of a quantitative orientation being conveyed through pornography. As concerns gender relations - the more important aspect in this particular analysis - the fact that the difference between female and male size descriptors is even larger with respect to token frequencies, with women trailing men by over 17 percentage points, adds significance to <?page no="212"?> Analyses 212 the results from above and fully agrees with what was expected. This strengthens the plausibility of the hypothesis of gender differences with respect to the objectification-mitigating effects of size descriptions in pornography, even though again male sexual arousal descriptors included in the class might relativize the gap a bit. Similarly as with types, we see a male advantage in the shape category. The enormous female lead in the adhesive/ substance category more than compensates for this. Women’s body parts are thus more likely to be described in qualitative terms. This conclusion is underlined by descriptions of colour, which are also more commonly applied to women’s bodies. The qualitative subcategories thus really show the predicted female preponderance, which will add to the generally stronger visualization of women in the discourse of pornography, as hypothesized. As with type frequencies, aesthetic features do not play a role with respect to token frequencies either and can therefore be largely ignored in a discussion of gender differences in visualization. Following the structure of the preceding subchapter, I will conclude by looking at the ten most frequent adjective modifiers. But to bring the results from above into an overall perspective, I will list the top ten body part descriptors, whether visual or not (as mentioned, I have this data available, but have chosen not to examine it here). Women Men wet 135 hard 114 hot 86 stiff 71 naked 74 big 68 tight 71 throbbing 64 soft 64 thick 50 bare 56 long 40 long 53 swollen 37 little 52 other 32 right 51 huge 31 big 48 hairy 30 Table 13: The ten most frequent attributive adjectives functioning as descriptors of female and male body parts in the pornography corpus (visual descriptors in bold). The list illustrates the predominance of physicalization in general and of visualization in particular in the description of body parts. Visual descriptors represent about two thirds of the top ten for both genders. A greater emphasis on female visualization also becomes evident in the <?page no="213"?> Describing body parts 213 fact that there are seven of ten and two in the top three that belong to this category, whereas for males it is only six of ten and the most frequent adjective of this category is placed only third. Tactile descriptors, which I expected to appear in large numbers considering the importance of touch in different conceptions of sexuality, do occur (hot, soft, hard, stiff), but compared to visual ones they are clearly a minority. Significantly, five of the six male visual descriptors denote size, while only three of the seven female ones do. 2.3 Comparing and discussing results I will focus on the question of how the visual subcategories compare to each other and to their counterparts in the pornography corpus (for lists of descriptors, see my homepage). The following two tables and the diagram present all the relevant results on lexical variation. Let me just insert a brief comment on the numbers below: overall the corpus of pornographic stories contains significantly more attributive descriptors for body parts, not only because reference to the body occurs much more frequent, but also because this type of modification is almost twice as common (as calculated with data partly not considered in this book). Types Overall % % % Shape 31 33.3% 38 35.2% 69 34.3% Size 18 19.4% 23 21.3% 41 20.4% Colour 24 25.8% 29 26.9% 53 26.4% Aesthetic 2 2.2% 3 2.8% 5 2.5% Adhesive/ dress 9 9.7% 4 3.7% 13 6.5% Adhesive/ substance 9 9.7% 5 4.6% 14 7.0% Presence/ absence 1 1.1% 3 2.8% 4 2.0% Others 0 0% 3 2.8% 3 1.5% Totals 93 108 201 Table 14: Absolute and relative numbers of different attributive adjectives functioning as descriptors of male and female body parts in the eight visual subcategories in the erotica corpus. <?page no="214"?> Analyses 214 0% 5% 10% 15% 20% 25% 30% 35% 40% Shape Size Colour Aesthetic Adhesive/ dress Adhesive/ substance Absence/ presence Others Female Male Figure 7: Relative sizes of the physical visual subcategories with respect to lexical variation of attributive adjectives functioning as descriptors of female and male body parts in the erotica corpus. Types Overall Porn Erotica Porn Erotica Porn Erotica Shape 29.7% 33.3% 34.1% 35.2% 31.3% 34.3% Size 18.0% 19.4% 27.4% 21.3% 21.4% 20.4% Colour 21.2% 25.8% 13.5% 26.9% 18.4% 26.4% Aesthetic 2.8% 2.2% 3.2% 2.8% 2.9% 2.5% Adhesive/ dress 7.6% 9.7% 6.7% 3.7% 7.3% 6.5% Adhesive/ substance 19.6% 9.7% 12.7% 4.6% 17.1% 7.0% Presence/ absence 2.5% 1.1% 3.2% 2.8% 2.8% 2.0% Others 0% 0% 0% 2.8% 0% 1.5% Table 15: Comparison of the visual subcategories with respect to lexical variation of attributive adjectives functioning as descriptors of female and male body parts between the pornography corpus and the erotica corpus. Although I have made no predictions concerning general relations, it is interesting to see that the only major differences with respect to the overall variation in visual descriptors between the two corpora can be found in the colour category, with a great lead for the erotica corpus, and in the adhesive/ substance category, which sees pornography in front, probably as a result of features of wetness (and thus of sexual arousal) being less prominent in erotica. Otherwise the two corpora are relatively similar, with erotica putting as much emphasis on size as pornography. Proceeding to gender differences, we notice that the erotica corpus resembles the pornography corpus in that it also sees a male advantage in the lexical variation of descriptors of size. As postulated, the gap to women, however, is much smaller, probably because the descriptions of <?page no="215"?> Describing body parts 215 the oversized penis are not as common in the erotica. And the margin is further relativized by a brief look at the list of positive and negative size descriptors in the erotica corpus. Positive size abundant; big; broad; engorged; full; generous; large; long; outstretched; stretched ; voluptuous; swollen; wide big; broad; extensive; fat; full; generous; high; huge; immense; lavish; long; massive; swollen; thick; wide; Negative size small; thin; tiny little; short; slender; small; thin Table 16: Positive and negative size descriptors of female and male body parts in the erotica corpus. The comparatively small number of positive size descriptors for males as well as for females and the relative scarcity of superlative lexemes and hyperbolic metaphors will reduces the associations of power that come with size. The overall effect of size descriptors in undermining the objectifying impact of visualization and physicalization will thus probably be less significant in erotica reading and the female disadvantage will consequently weigh even less heavily. The more qualitatively-oriented categories of shape, colour and beauty do not reveal any clear gender trends, conforming to my expectations that, unlike in the pornography corpus, these descriptors will not consistently feature higher variation for women than for men. It has to be noted, however, that a comparison between the corpora is complicated by the fact that, as we have seen, the male shape descriptors in the pornography corpus also encompass the erect penis and thus sexual arousal, an aspect much less highlighted in the erotica corpus. I would still claim that the data on qualitative and quantitative categories adds plausibility to the hypothesis that the gap in visualization between women and men will not become more substantial in the discourse of erotica, at least not to the same extent as in pornography. <?page no="216"?> Analyses 216 I will now look at token frequencies, presented in the table, the comparative table and the diagram below. Tokens Overall % % % Shape 42 24.4% 56 28.0% 98 26.3% Size 41 23.8% 61 30.5% 102 27.4% Colour 47 27.3% 51 25.5% 98 26.3% Aesthetic 4 2.3% 4 2.0% 8 2.2% Adhesive/ dress 21 12.2% 11 5.5% 32 8.6% Adhesive/ substance 17 9.9% 10 5.0% 27 7.3% Presence/ absence 1 0.6% 3 1.5% 4 1.1% Others 0 0% 4 2.0% 4 1.1% Totals 172 200 372 Table 17: Absolute and relative frequencies of attributive adjectives functioning as descriptors of male and female body parts in the eight visual subcategories in the erotica corpus. 0% 5% 10% 15% 20% 25% 30% 35% Shape Size Colour Aesthetic Adhesive/ dress Adhesive/ substance Absence/ presence Others Female Male Figure 8: Relative sizes of the physical visual subcategories with respect to token frequencies of attributive adjectives functioning as descriptors of female and male body parts in the erotica corpus <?page no="217"?> Describing body parts 217 Tokens Overall Porn Erotica Porn Erotica Porn Erotica Shape 21.3% 24.4% 23.5% 28.0% 21.9% 26.3% Size 27.8% 23.8% 45.6% 30.5% 33.1% 27.4% Colour 16.2% 27.3% 10.8% 25.5% 14.6% 26.3% Aesthetic 5.4% 2.3% 3.6% 2.0% 4.8% 2.2% Adhesive/ dress 8.7% 12.2% 5.2% 5.5% 7.7% 8.6% Adhesive/ substance 19.0% 9.9% 7.1% 5.0% 15.5% 7.3% Presence/ absence 2.1% 0.6% 4.5% 1.5% 2.8% 1.1% Others 0% 0% 0% 2.0% 0% 1.1% Table 18: Comparison of the visual subcategories with respect to token frequencies of attributive adjectives functioning as descriptors of female and male body parts between the pornography corpus and the erotica corpus. While there were some differences between the shape, the size and the colour categories with respect to lexical variation, the token frequencies see them on level terms, all claiming more than a fourth of all visual descriptors of body parts. Furthermore, the proportion of size descriptors in the erotica corpus is below that in the pornography corpus here, indicating that size may still not play the same role in erotica as in pornography, even though the type frequencies suggested otherwise. But let me turn to gender differences, which in this particular case are more important than overall results. In contrast to lexical variation, we do get a clear margin here in favour of men. As expected, however, the male lead is nowhere near as large as that in the pornography corpus and will not be as effective either, given the missing abundance of positive and particularly superlative size descriptors. This category will thus not be as relevant in mitigating the overall objectifying effect of visual descriptors. The data thus provides some support for the assumption that in interpreting erotica, women’s and men’s levels of visualization will not grow as widely apart as in interpreting pornography and that women’s visualization and physicalization will consequently not exceed men’s by far. The shape category shows a male advantage, which is also a result of the erect penis as an indicator of sexual arousal being part of this category, while the corresponding factor, the wet vagina, is part of the adhesive/ substance category, which sees a female lead. But the differences are by no means as marked as in the pornography corpus, which could mean that the influence of these categories on overall visualization is less strong, suggesting that the conclusion from the last paragraph will hold. <?page no="218"?> Analyses 218 As above, I will finally have a brief look at the top ten descriptors overall, to see whether the proportions of visual attributes also exceed other categories in the erotica corpus. Women Men other 14 long 13 left 10 brown 8 white 10 right 7 whole 8 strong 7 bare 7 dark 6 long 7 free 6 small 7 wet 6 dark 6 bare 5 naked 5 left 5 right 5 little 5 smooth 5 swollen 5 soft 5 wet 5 Table 19: The ten most frequent attributive adjectives functioning as descriptors of female and male body parts in the erotica corpus (lexemes occurring as often as the 10 th one included) (visual descriptors in bold). Visual descriptors predominate in the description of body parts in the top ten lists above, suggesting that pornography and erotica might not differ with respect to the amount of visualization and physicalization they create. The facts that the number of visual descriptors for female bodies among the top ten is not larger and that the first two lexemes in the women’s list belong to other categories, unlike in the men’s list, indicate that physicalization and visualization will at least not be a stronger force in the conception of women than in that of men in erotica. Among the visual descriptors, size plays the most important role (with long and small in the women’s list and long, little and swollen in the men’s list), but it is not as predominant as in the pornography corpus and it does not see as wide a gap between men and women, which is in line with what has been found in the full analysis above (even though none of the qualitative categories is extensively represented). Interestingly, tactile qualities are even less relevant in erotica than in pornography, with only one descriptor featuring in the lists. All in all, the current subchapter has seen qualified support for my predictions. Looking at the quantitative relations in the subcategories of <?page no="219"?> Casting a side glance 219 visual physical descriptors, it has found the expectations confirmed that size, which due to its association with power might work against the objectifying potential of visualization, will more often feature in the description of male bodies than in that of female ones in the pornography corpus. The data has also shown that this male dominance in the area of size does not occur in the erotica corpus. Results for the more qualitative categories of colour and wetness - lacking the re-empowering effect of size - have, on the other hand, shown that women gain the upper hand here in the pornography corpus, but not in the erotica corpus. Only the largest qualitative class, viz. shape, deviates from my expectations, with higher type and token frequencies for men, which is probably a result of male sexual arousal featuring as descriptions of the shape or position of the penis. This is, however, over-compensated by the other two qualitative categories mentioned. Overall, the data thus support the hypotheses that women’s visualization, not being as consistently subjected to the re-empowering effects of size, will be stronger than men’s in pornography consumption and that this effect is unlikely to occur in erotica consumption. 3 Casting a side glance: interesting phenomena not examined I would like to round off this chapter by pointing to a few linguistic aspects that would invite thorough investigation in an analysis of visualization and physicalization, but which for reasons of space and resources could not be considered. The most obvious omission in this chapter is the analysis of predicative descriptors, i.e. of predicative nouns, predicative adjectives and attributive adjectives in predicative NPs. Examining these should reveal whether physical and visual features fare better in structures that, according to my theoretical framework, denote more essential features and how the results compare to those gained in the analysis of attributive adjectives, supposed to attract less attention. As a matter of fact, I have carried out these analysis for my dissertation and lists of descriptors are presented on my homepage (for a select set of interesting aspects, cf. also Marko 2006). I leave it to readers to draw their own conclusions from the data. Another field worth examining lies at the intersection between physicalization (but not necessarily visualization), on the one hand, and point of view, on the other. This field is the external expression of internal psychological processes. If, for instance, emotions - in <?page no="220"?> Analyses 220 particular sexual pleasure - are primarily described in terms of vocal expressions, then the person whose inner life is referred to is physicalized rather than subjectified, which, in turn, means that we refrain from taking her or his perspective. Two linguistic elements which can focus on the external signalling of psychological processes are verba dicendi highlighting the vocal quality of an utterance rather than pragmatic or semantic aspects, e.g. shout, cry, grunt, and interjections, i.e. words with purely evaluative but no ideational meanings whose lexical status is unclear (they verge on vocalizations, thus not really being established, lexicalised words - this, incidentally, becomes apparent in the kind of variation found in the expressions, see below). The following table lists all the verbs of saying emphasizing the vocal/ acoustic dimensions. The set is restricted to those lexemes occurring in connection with direct speeches (this is also how I have retraced them in the corpus). babble chorus grizzle murmur slur throat bark chuckle groan pant smooch thunder bellow chunter grumble pronounce sneer wail blast coo grunt quaver snicker wheeze blurt croak guffaw rasp sniff whimper bluster croon gulp roar snivel whine breathe crow gurgle roll snort whisper bristle cry gush scream sob yank bubble drawl hiss screech splutter yell burble erupt hum shout sputter yelp cackle exclaim intone shriek squeak call explode mewl sigh squeal chant gasp moan simper stammer chortle giggle mumble sing stutter Table 20: Verba dicendi highlighting vocal/ acoustic quality in the pornography corpus. The impressively large set suggests that vocal quality is obviously an extremely overworded aspect in conceptualizing people in pornography. As vocal quality will probably be interpreted as physical (auditory rather than visual in this case) manifestation of emotions and other mental processes, thus translating psychological into material phenomena, the high lexical variation (which is also reflected, incidentally, by high token frequencies) could be claimed to contribute to the overall physicalization produced in the discourse of pornography. A thorough quantitative comparison between type and token frequencies of women and men <?page no="221"?> Casting a side glance 221 could allow further insights into the question whether there is a gender bias with respect to (this dimension of) physicalization. Such insights can be complemented by an analysis of interjections, which serve, to a certain extent, as immediate, onomatopoetic representations of what the verbs above lexicalize. The set of interjections in the table below is supposed to prove this point. a aw mmmm ouf uuuummmmmmmm aaaaaaaaaaaaah aye mmmmm ow uuuuuuuuuuuuh aaaaaarrrggghh beeeep mmmmmm oy vaah aaaaah bingo mummmm saah waaah aaaah bzzzt naw scrummm whew aaaarrgghh daaaa nay sh whoa aaaarrrrgh eh noooo sheesh whooo aaagh er o shhh whoop aaah erm oh shoop whoops aaahh gee ohgodohgodohgod ssh woo aaarrrggghh gees ohh ssshhh wow aaarrrggghhh gosh ohhh tut yeeah aagghh ha ohhhh ugh yeeeoowwhhh aah hah ohhhhh uh yep aarrgghh hmm okaaay uhh yesss ah hmmm oof uhhh yessss ahh ho ooh uhhhh yessssss ahhh huh oohh um yo ahhhh huuuuh oooh umm yow ahhhhh hum ooohh ummm yum arggghh jeez ooooh ummmm yup arggghhh jeeze oooohhh ummmmm zummmmm argh mm oops unh arrghhhhhh mmm ouch Table 21: Interjections in the pornography corpus. This list of interjections (or onomatopoetic expressions of vocalizations, if we want to distinguish these), though certainly not extremely original or varied in the sounds they are supposed to imitate, but only in the orthographic diversity, underlines pornography’s emphasis on the expression and thus externalization of feelings, which also could enhance the role that physicalization will play in readers’ interpretations. Again, such cursory remarks would have to be backed up by a detailed analysis of whose voice is represented and in which co-texts. Finally, my analyses of visual physical descriptors can be complemented by an examination of the reverse side of the visual, namely on <?page no="222"?> Analyses 222 the processes of seeing and looking and on the question of how salient they are. Even a casual scan of verbs in the pornography corpus provides me with an extensive list including expressions such as catch a glimpse, eye, flare at, focus, gape at, gawk at, gaze at, get a glimpse of, get a look at, get a view of, get an eyeful of, give somebody a glance, give somebody a look, give somebody a stare, glance at, glare at, glimpse at, leer at, look at, peer at, scowl at, see, squint at, stare at, watch, or, with eyes as subject, be glued to, catch, devour, drink, feast on, take in (with an interesting set of eating metaphors). Such a list, however unsystematically composed, indicates that visual perception is certainly a main area of overwording. This, of course, reflects the general tendency of human cultures to attend to the visual rather than, for instance, the tangible. Such superficial observations need to be elaborated, examining type and token frequencies and also looking at the question whether females or males are subjects of verbs of seeing and looking more frequently. This, however, verges on the field of point of view, which will be treated in the next chapter. 4 Conclusion Thinking of people primarily as physical and visual beings means conceiving of them as objects. Physicalization and visualization can therefore be regarded as objectification. This chapter has examined the question to what extent and whether at all pornography can create these conceptualizations in its consumers, and whether there are any differences between women and men and between pornography and erotica with respect to this issue. For this purpose, I have, linguistically speaking, focused on lexemes employed to describe the visual and physical aspects of people and their bodies. These aspects are considered to be assigned a particular salience in the interpretation process if their descriptors occur relatively often and in great variation in comparison to other categories. The syntactic position of a descriptor is also a factor that can enhance or reduce its visualizing and physicalizing effects: predicative descriptors may be less effective than attributive ones as they attract conscious attention while the latter are presupposed and, if not part of a predicative NP themselves, de-emphasized. I therefore exclusively concentrated on attributive adjectives, describing whole persons or body parts. As regards the question of the general prominence of visualization and physicalization, which was my first hypothesis, this assumption is <?page no="223"?> Conclusion 223 supported by the data because attributive adjectives are ranked at or close to the top. Are women more consistently subjected to physicalization and visualization in pornography? This is the question behind my second hypothesis. Since the findings suggest that women are assigned more varied and more frequent visual descriptors, I conclude that there is indeed evidence in favour of an at least slight gender bias with respect to physicalization and visualization created by pornography. This is corroborated by the fact that in the description of body parts the size category, which as a result of its association with power can relativize visualization, is more strongly represented for men than for women. As for the difference between pornography and erotica, this hypothesis has been refuted by the results. Visual descriptors are not more common in the pornographic discourse, neither with respect to types, nor with respect to tokens. Visualization and physicalization thus appear to be aspects that will feature equally saliently in interpretations of pornography and erotica. The last hypothesis concerned gender differences in erotica, suggesting that there would not be a trend towards more physicalization and visualization of women or the trend would at least be weaker. This assumption has been fully confirmed. Going beyond visual features, we note that by and large the other categories that could enhance physicalization (e.g. other physical descriptors or secondary visual descriptors such as the age, gender or origin subclasses) or objectification in general (e.g. evaluative descriptors) or that mitigate it (e.g. psychological descriptors) support the conclusions drawn above, with the occasional problematic case. Summarizing also means ironing out inconsistencies in the data sets. The partly heterogeneous results encountered may be a consequence of some methodological and metatheoretical faults. Leaving aside problems that will run through all the research presented in this book, e.g. the overpowering majority of female first person narrators in the erotica corpus, I also have to mention the fact that there are countless words and expressions that defy any precise categorization. This is a serious shortcoming of the research, considering that the semantic classes stipulated are supposed to represent the mental categories of language users (even if they usually allow more fuzziness) and not just my conscious attempt to fit words into a procrustean model. An exhaustive analyses of descriptors (including predicative ones) would also suffer from the fact that there seem to be more semantic aspects at play in the different structures, aspects which could not be taken into account and which might be far more complex than suggested. Such <?page no="224"?> Analyses 224 flaws show how far Critical Discourse Analysis - or linguistics in general, for that matter - is still away from providing fully convincing accounts of language patterns in discourses. <?page no="225"?> VII Desubjectification I can imagine the warm smell between your legs. It’s all in the mind, you see. As subjects, we access the world through our minds. Although we can only experience our own cognitive lives immediately, i.e. without any mediating instance, we still assume that other humans are also first and foremost beings constituted and defined by what they perceive, think and feel. Without any great effort, we can empathize with others, interpreting material aspects as signs of what is going on in their minds and thus being able to take their perspectives. Denying this ability and not seeing the material physical existence of other human beings as an index of their subjectivity, i.e. desubjectification, however, means treating them like mindless objects. If pornography manages to induce this kind of thinking in its consumers, then it can rightly be regarded as objectifying. The definition already shows the conceptual kinship between physicalization and desubjectification. While I was concerned with an emphasis on the physical and the visual in the preceding chapter, the current one will deal with the complementary process, i.e. the deemphasizing of cognitive aspects. Desubjectification will be defined negatively in this chapter, which means that I will take it to be the lack of subjectification. I will thus examine desubjectification indirectly by analysing subjectification. The latter can be defined as conceptualizing people primarily in terms of their minds and as interpreting ‘objective’ reality from their perspective. Although there is some correlation between these two aspects, as representing other people’s mental life usually implies taking their perspective, ‘sneaking’ into their subjectivity, this connection does not work in both directions since point of view encompasses aspects going beyond the representation of the mind. The distinction will also have some practical advantages for the analyses later on. Any aspect of content can be argued to be subjectifying if interpreted as a sign of (alleged) preferences of particular people (e.g. if the constant inclusion of fellatio in pornography is said to reveal a male perspective). Such less immediate dimensions of subjectification will, however, not be considered in my study. The psychological processes on which subjectification rests can be divided into perceptions, emotions and thoughts. This differentiation is <?page no="226"?> Analyses 226 relevant considering that there are gender stereotypes about different psychological dimensions (feelings and perception associated with women, thoughts with men). Subjectification is a widely researched phenomenon in linguistics, literary studies and media studies, variously called point of view, perspective, narrative situations (in literary studies) or gaze (in film studies). It is a seminal aspect of discourses because it is the prime factor in guiding readers’ or hearers’ empathy and thus their readiness to identify with particular persons. This is why it has also been treated under the heading of empathy in linguistics (cf. Kuno 1987, Pusch 1991: 109-128, Marko 1995). In its effects, desubjectification is an objectifying process that can have the same discriminating effects as physicalization and visualization. Ignoring the subjectivity of sexual partners will disable us from putting ourselves into their shoes and to accept, let alone understand, their preferences, desires and limits. This lack of empathy may prevent us from experiencing satisfying sexuality in an equal relationship and may, under the right (or wrong) circumstances, also reduce threshold levels for abuse or violence. In a patriarchal society, the problematic consequences of desubjectification will be more acute if affecting women. Subjectification - desubjectification only by implication - is a frequently and broadly discussed aspect of pornography, even though the distinction to other categories related to the objectification of women (and opposing processes for men) is not always made clear. Analysts of pornography have claimed that male subjectivity and the concomitant androcentric perspective are essential elements of pornographic representations (cf. Dworkin 1981: ch. “Power,” Caputi 1994: 16), perhaps reflecting a culture of male-dominated sexual fantasies likely to incorporate a male point of view (cf. Barbach 1985a: x). Kappeler (1986: 52) says this subjectivity is even a twofold one with two male subjects, namely the male character, whose perspective is taken, and the male viewer/ reader, who identifies with the former. Implicitly taking up the reverse side of the coin, i.e. the question of women’s desubjectification, Kaplan (1984, cit. in Hardy 1998: 124) says that the androcentric point of view eventually means that women are denied subjecthood (she focuses on pornographic films and not on all of pornography, though). The importance of the androcentric perspective is also acknowledged - though not explicitly - by writers suggesting that women should explore the genre of pornography for either producing gynocentric versions themselves or for offering subversive readings of existing material, which implies that they assume that today’s porn is, as a matter of fact, <?page no="227"?> Desubjectification 227 male-centred (cf. Carol/ Pollard 1993, cit. in McNair 1996: 99, Clover 1993, cit. in McNair 1996: 95). Interestingly enough, there are also voices contesting the assumption that pornography is necessarily male-centred, without, however, exonerating it. Beneke (1990, cit. in Hardy 1998: 66f.), Berger, Searles and Cottle (1992: 43), and Hardy (1998: 134) assume that women’s subjective lives have to be represented for the male pornography consumer to see how his sexual desires are appreciated and reciprocated. So there might be a certain tension between the male consumer’s desire to watch the woman character’s body and his wish to read her mind at the same time. This paradoxical situation has to be kept in mind when it comes to interpreting data. The preoccupation with the topic of male subjectivity and the absence (or return) of female subjectivity in pornography may not have resulted in concrete studies of any kind but it shows that research focusing on perspective is overdue. In this chapter, I will look at linguistic items that can be argued to help create subjective positions in order to find out about potential gender differences regarding subjectification and desubjectification. Since it is not clear to what extent it is possible not to adopt anybody’s perspective and whether this would be of any socio-political significance, I will not examine the overall amount of (de)subjectification. Linguistically, the first subchapter will be concerned with verbs and other lexemes denoting or being related to mental life, examining whose minds are thereby represented. The second subchapter will deal with grammatical subjects in the passive voice, which are assumed to have an immediate influence on the creation of perspective in discourse. I expect the linguistic elements examined to support the following assumptions, which are partly formulated on the model of the schematic hypotheses introduced in chapter IV (with the two hypotheses not focusing on gender missing): • The tendency to reduce the amount of subjectivity will be stronger for women than for men in pornography. • There will be no such tendency in erotica or at least the gap from women to men will not be as wide as in pornography. Assuming that desubjectification will be mixed with traditional gender stereotyping, I additionally assume that the gender gap in pornography will become particularly obvious in the ‘masculine’ cognitive processes, but less so in the ‘feminine’ perceptive and emotional processes. As in previous chapters, I will again bring the results of the pornography corpus into perspective by comparing it to data from the erotica <?page no="228"?> Analyses 228 corpus. But a word of caution concerning the legitimacy of such a comparison is in place here. There certainly is some degree of correlation between objectifying features of a discourse and the gender of the first person narrator (in case there is one). This correlation is even a stronger moment in this chapter since a first person narrator is a powerful - possibly even the prime - factor in creating a subjective position because it is difficult to have a first person narrator and still represent events from somebody else’s point of view. This is less of a problem in pornography where I have at least a rough balance between stories told by a female or male first person. But in erotica, there is a strong predominance of women narrators, suggesting that other aspects contributing to subjectification (and desubjectification) will probably not create a completely reverse perspective. This, however, entails that a comparison will have to be considered with due care and caution in this chapter and might be less revealing and thus less relevant than elsewhere. 1 Representing people’s inner lives: mental process verbs Representing the inner lives of humans is the most straightforward mode of creating subjectivity in discourse because it implies a preoccupying focus on the mind and thus on subjectivity. But it is more than that. If we write or read She loves to suck cock, then we are not just concerned with her mind, we are actually entering (or pretend we are) it, looking at the world through her eyes. Representing mental processes thus also means transcending the epistemological boundaries that reality has set for us by allowing access to what is normally hidden inside people’s heads. This section will focus on mental processes and on the question on whose minds we concentrate and into whose minds we ‘sneak’ by finding them represented in the semantic role of the senser (on the importance of mental process verbs for subjectivity and point of view, cf. Short 1996: 268f., Talbot 1997). Mental processes and the senser role are part of the system of transitivity in Functional Grammar (cf. Halliday 1994: ch. 5, Thompson 1996: ch. 5.2), positing that propositions are representations of whole events consisting of processes (what happens or is done; normally represented by the verb), participants (the people, animals or things directly and immediately involved in the process, usually represented by the obligatory complement phrases, esp. subjects and objects), and circumstances (the settings in which the process takes place, represented by the optional adjunct phrases) (cf. Jackson 1990). <?page no="229"?> Representing people’s inner lives 229 There are different types of processes, e.g. processes of acting in the material world (material processes), of using language (verbal processes), of being or having (relational processes), and of perceiving, feeling and thinking (mental processes). Depending on the type of process, participants play different semantic roles. In a material process such as She kissed him, for instance, she actively carries out the process (agent) and he is acted upon (patient). In a mental process such as we are afraid of spiders, we are affected by the fear (senser) and spiders are the object of the emotion (phenomenon). Circumstances also have semantic roles, e.g. temporals, locatives, or instruments. These are not determined by the type of process though. Yesterday thus specifies a point in time, no matter whether it is used in she kissed him yesterday or in we were afraid of spiders yesterday. The connection between participant roles and syntactic functions is determined by the verb as the latter assigns particular roles to subject and object NPs, e.g. to kiss (in the active voice) the agent role to the subject and the patient role to the object, to be afraid of the senser role to the subject and the phenomenon role to the oblique object in the PP. Circumstantial roles, on the other hand, are dependent on the meanings of their phrasal heads, e.g. that of the preposition in prepositional phrases or that of adverbs in adverb phrases. Prepositional phrases with into thus are (usually) locatives, while adverbial phrases with beautifully expresses manner, irrespective of the meaning of the verb. Mental process verbs denote the psychological processes of sensual perception, emotion/ affection and thought/ cognition (cf. Halliday 1994: 118). Below you find two examples of each of the subcategories (the mental process verbs are shaded): Perception: As soon as Gloria saw that I was alright, […] She heard the door banging back and forward. Emotion: Francine loved to suck cocks. […] she definitely appreciated the special flavor of my wife’s cunt cream. Cognition: Sheena considered the prospect. He also understands how I can’t be contained by his cock alone. The central semantic roles in mental processes are, as mentioned, the senser and the phenomenon. The former (bold in the examples above) <?page no="230"?> Analyses 230 is more relevant for the present discussion because it is the sensers’ minds that we are describing in mental processes and that we are thus claiming to have immediate access to. Finding out who appears in this semantic role thus indicates who is subjectified and who is not. To put it in a nutshell, the main objective of this section is to examine who appears in the senser role of mental verbs. Supposing that higher frequencies will lead to a salient status of either a female or a male perspective in pornography and erotica consumers’ interpretations of the texts, I will primarily look at the frequencies of male or female sensers. This is supposed to allow insights into the subjectification and desubjectification of the two genders. My predictions for this subchapter are: • Women will be assigned the senser role of mental verbs less frequently than men in the pornography corpus. • In the erotica corpus, female sensers will not occur less often than male ones or at least the gap from men to women will not be as wide as in the pornography corpus. In addition, pornography’s desubjectification of women will be less compelling in perceptive and emotional processes than in cognitive processes, with gender stereotypes having an impact here. 1.1 Technical and practical aspects of the analysis As a full analysis of all mental verbs is impossible, I have selected twenty-five lexemes representative of the three categories perception, emotion and cognition (see the table below). Perception Emotion Cognition see (1,325 vs. 472) like (462 vs. 195) know (1,455 vs. 679) hear (420 vs. 197) love (370 vs. 171) think (1,113 vs. 556) feel (1,688 vs. 587) enjoy (304 vs. 50) believe (214 vs. 36) notice (186 vs. 75) want (1,294 vs. 663) realize (259 vs. 90) experience (97 vs. 24) wish (86 vs. 56) wonder (207 vs. 92) need (252 vs. 86) imagine (142 vs. 73) hope (152 vs. 48) expect (122 vs. 39) mind (77 vs. 19) be sure (167 vs. 76) be afraid (56 vs. 44) remember (161 vs. 101) be excited (111 vs. 14) forget (137 vs. 55) Table 1: Mental process verbs analysed (with frequencies in the pornography corpus and the erotica corpus). <?page no="231"?> Representing people’s inner lives 231 The selection is based on frequencies and on semantic variety: I want to include the mental process verbs appearing most often in the corpora in addition to examining as wide a range of different processes as possible. The lower number of verbs in the perceptive class is due to the apparent smaller variation found for this type. The label mental process verbs will be understood very broadly, covering adjectives in predicative positions that clearly denote something that is going on in the mind (as opposed to general psychological traits; the two are, as mentioned in the chapter on physicalization and visualization, however, not always easily distinguished from each other), as in: I wasn’t so sure about Morwenna’s herbal magic. He is afraid that he might soil his leather interior. Cindy was excited and held her legs wide apart. Halliday (1994: 121f.) relegates such examples to intensive processes, i.e. processes in which persons, animals or things are assigned to a category (he acknowledges their slightly hybrid status, though, cf. Halliday 1994: 121f., assuming that some cases are mental processes expressed as grammatical metaphors in intensive processes). But assuming that the differences to central members of the mental verb class are not large enough to justify a complete omission, I have, as can be seen from the table, included three mental adjectives (the three used in the examples above, viz. sure, afraid, excited) in the analysis. It has to be noted that many predicative adjectives have an ambiguous status, sometimes occurring as adjectives proper and sometimes as past participles in the passive voice, e.g.: I was excited I was excited by all the lustful looks and comments I’d received As the active voice form of excited would not qualify as mental process in Halliday’s framework but as a material one, the passive form cannot be argued to belong to a different category. On the other hand, it might seem odd to count the adjective excited as a mental process and the past participle excited in the passive voice as a material process, even though they are so close in meaning. But this reveals an obvious weakness in Halliday’s taxonomy. Causative verbs such as excite, which could be paraphrased as ‘to cause someone to be excited’, clearly have a mental side, making claims about what happens in the mind of the patient of the event. As both constructions thus represent a person’s mind, I will not <?page no="232"?> Analyses 232 distinguish between them. As a matter of fact, excited, which is rare in the second form, is the only candidate from these problem cases. With its association with gender stereotypes, the distinction between perception, emotion and thought will be important, although we have to bear in mind that the differences are not always clear-cut. There are many borderline cases, e.g. notice, which might denote a perceptive process but has a strong cognitive slant, or feel, which is perceptive but with a clear emotional dimension. There are also whole semantic subcategories not easily subsumed under the existing three classes, e.g. volition (which I count as emotional even though it also covers cognitive aspects) as represented by the verb want. But Brad wanted more! The subclass of evaluation, represented by verbs such as find, as, for instance, in: She found me physically attractive, as I did her. is also problematic (I tend to count it as cognitive even though it also covers emotional aspects). But as it does not feature any high-frequency members, none of its lexemes has entered the selected set of verbs/ adjectives. The goal of the analysis, as has been said, is to find out how often women and men are assigned the senser role by the verbs from the list. The corpus searches helping in this task are easy to carry out. The search word is simply the verb in its various forms - non-verb forms of the same word are excluded by means of the tags (e.g. the nouns feeling, mind, experience, the adjective loving, or the preposition like). The required searches yield concordances such as the following one for the verb see. JJ> couple<NN1> and<CC> saw<VVD> that<CST> it<PPH1> rried<VVN> ,<,> and<CC> seeing<VVG> me<PPIO1> with<IW> PGE> faces<NN2> and<CC> seeing<VVG> the<AT> pure<JJ> to<TO> try<VVI> and<CC> see<VVI> how<RGQ> far<RR> we<PPIS PIS1> 'd<VHD> been<VBN> seeing<VVG> ,<,> Tracy<NP1> ,<,> > .<.> <^> Bruce<NP1> saw<VVD> his<APPGE> wife<NN1> 's NP1> again<RT> but<CCB> saw<VVD> a<AT1> lot<NN1> of<IO> Z> becoming<VVG> by<II> seeing<VVG> her<PPHO1> like<II> r<NN1> I<PPIS1> can<VM> see<VVI> my<APPGE> shaved<JJ> pu e<AT> guys<NN2> can<VM> see<VVI> daylight<NN1> .<.> They <.> They<PPHS2> can<VM> see<VVI> with<IW> their<APPGE> o Concordance 1: The verb see in the pornography corpus (extract). <?page no="233"?> Representing people’s inner lives 233 From the concordance I first have to exclude examples which are nonmental in meaning (in the case of see, for instance, the meaning of ‘to meet’; with see, I additionally have to distinguish the perceptive uses (vision) from the - metaphorical - cognitive uses ‘to understand’). I then go through the resulting list to manually count the occurrences of women and men as sensers. The same procedure is applied in the search for the adjectives afraid, excited, and sure, which means the basic search string is the word identified as adjective (or as past participle in the case of excited) by the tag. The only difference to the verb searches is that there are more examples that are excluded from consideration, mainly because the adjective sometimes occur in attributive positions, e.g. a sure sign, an excited scream. I am very inclusive in my counting of male or female sensers, not easily discarding examples as unidentifiable, because I assume that whether a senser appears as a first person or second person pronoun, a third person pronoun or any other NP does not affect the subjectifying effect. I do not only include explicit sensers, which usually appear as subjects of mental verbs on the surface, but also examples where the semantic role is implicitly assigned but can - more or less - easily be established from the co-text, as in the examples below: Seeing her tits would have been nice, he admitted […]. Feeling the jizz starting to boil, he moaned urgently […]. It was comforting to know that I [fem.] at least turned him on enough to think about going to bed with me. This requires me to be very diligent in my attempt to identify the gender of the person/ s fulfilling this subjectifying semantic role. This inclusive approach marks a difference to chapters V and VI, where I used a pre-selected set of gender-specific lexemes. But while my focus is on the senser’s gender here, with the search word being fixed, gender was a part of the string supposed to retrieve other structures (body part lexemes and descriptors) from the corpora in the inquiries in the said chapters, which necessitated a limitation. The two gender categories in this chapter do not include cases where the senser role is either taken by a woman and a man, a mixedgender group of any size, or by a gender-neutral word such as everybody, no one, or who (as an interrogative pronoun). I have decided to keep this category separate from the gender-specific classes instead of integrating them into the latter, e.g. by adding a senser role each to the female and the male categories. This marks a different counting strategy <?page no="234"?> Analyses 234 from the one to be applied to semantic roles assigned by verbs of intercourse, physical and social contact in the succeeding chapter on passivization, where I will incorporate the mixed-gender/ gender-neutral class into the gender-specific groups. The reason for choosing two alternatives is that while the mixed-gender category will almost exclusively be composed of references to couples in the intercourse words, there are considerably more occurrences of larger groups or gender-neutral words here so that it would introduce more serious distortions if I counted these for both women and men. I also use a ‘non-identified’ category covering examples where the gender of the senser could not be identified with the help of the co-text or where the senser does not have gender, as is the case for things or body parts, e.g.: My tight, virgin pussy loved every inch of his big prick […] […] my dick knew that her mouth had sucked these loads from another man. The mixed-gender and the non-identified categories, which could be argued to form one class together, will not feature in the analyses below but they are included in the data presentation on my homepage. Distinguishing between women and men might be the primary focus in this subchapter. But it will be necessary to introduce a second, ancillary distinction. There is a difference in the probability at which different NPs might occur in the targeted semantic role. Although we often say that she saw, that she was afraid or that he thinks, we are more likely to say the same things about ourselves as we know more about our own perceptions, feelings and thoughts than about those of others. Subjectification thus, more than any of the processes dealt with in previous chapters, correlates with use of the first person and, in narrative texts, with first person narrators. Since female first person narrators predominate the erotica corpus, as mentioned in the introduction to this chapter, and since this would raise the probability of females occurring more often as sensers there, I will distinguish between narrator-related sensers and narrator-independent ones in my analyses below (even though first persons in direct speech share features with narrator-related sensers, I have nevertheless decided not to categorize them with the latter because they still introduce a perspectivizing element not necessarily agreeing with the narrator’s subjectivity). There are also further problems. How, for instance, shall I deal with examples such as the following sentences, which actually deny that a mental process took place, through a negation in the first example and a counterfactual if-clause in the second? <?page no="235"?> Representing people’s inner lives 235 […] his petite wife didn’t enjoy sex with him. If I had known, I don’t think I would have gone. Negations and counterfactuals are included in my analysis since whether I am saying that somebody enjoys or knows something or does not enjoy or know it does not affect the fact that I am making a statement about their mind and that this statement also presupposes that I somehow have access to it. But what about the following examples? Does he want this as much as I do? “Think about what I’m doing to you - pushing my rod deep into your tunnel, splitting you open […]”. […] but she might think that was uncivilised. Elliot has told me he often feels fear before our lovemaking […] In all of these examples, the mental verb appears in a structure that does not express a straightforward statement. Asking about somebody’s feelings or requesting somebody to perform a mental act does not presuppose that I have knowledge about and thus access to her or his mind. And if I modify a process with a modal verb or relegate it to an object clause of a verb of saying, then I do not claim privileged information about what is going on in other people’s heads either. Can I then include examples such as the ones above? My answer is yes, because no matter how strongly mental verbs are modified, using them with particular persons in the senser role always means that I am talking about their minds and that I am representing their inner lives, thus subjectifying them. It is, however, true that in cases such as those mentioned I am not directly adopting the other persons’ perspectives, which might mean that the extent of subjectification is not as strong. Although this takes away from any conclusions I will arrive at, I still think that I am justified to include such examples in my analysis. 1.2 Presenting and discussing results Detailed data for each of the twenty-five lexemes can be found on my website. Let me start the analysis by looking at the overall results, i.e. the rates at which women and men appear in the senser role with the 25 lexemes selected. In order to arrive at a comprehensive value suitable for comparison, I have calculated two types of percentages: the total percentage is the absolute frequency of the respective category, e.g. <?page no="236"?> Analyses 236 Women total, divided by the overall number of roles considered. It represents the general quantitative relations of the categories, taking into account differences in frequencies between the verbs. The mean percentage, on the other hand, is the arithmetic mean of the percentages (in the respective category) of all 25 verbs and adjectives, i.e. the sum of the 25 percentages divided by 25. It treats all lexemes as equal despite the quantitative imbalances. 50% for women, for example, in both cases means that half of all the sensers - on average or overall - are female. The relevant figures can be found in the table below (the mixed-gender and the non-identified classes are not included). Overall: 10,857 Total Ind. Total Ind. Absolute numbers 5,570 2,656 4,432 1,838 Total % 51.3% 24.5% 40.8% 16.9% Mean % 51.4% 25.8% 40.2% 17.2% Table 2: Absolute and relative frequencies of women and men occurring in the senser role of the 25 mental verbs/ adjectives in the pornography corpus (Ind. = Narrator-independent). The figures reveal that, not conforming to my expectations, approximately half of all senser roles are assigned to women in the pornography corpus, while only about four out of ten to men. If we ignore first persons related to the narrators of the stories, then the gap between the two genders further increases, with women occurring almost one and a half times as often in this semantic role. This data thus does not support my hypothesis that women’s subjectivity will enjoy a less salient status in pornography consumers’ interpretation. This conclusion, however, has to be treated with some caution, considering that, as calculated in the chapter on fragmentation (cf. 1.2), women generally tend to be mentioned about 27% more often than men in pornography (47% more often if we ignore first person pronouns) so that it does not come as a surprise that they should also occur in certain semantic roles more often, even though it is far from obvious how this should affect the female/ male senser ratio quantitatively (it probably is not a linear relationship). While these reservations and limitations are certainly valid, I still claim that the figures suggest that women are at least not less likely to appear in the senser role, as the original prediction said, which, in turn, means, as already mentioned, no support for the hypothesis of women’s desubjectification in pornography. I will now turn to the three categories of perception, emotion and cognition to see how the frequencies from above distribute across them. <?page no="237"?> Representing people’s inner lives 237 The figures can be found in the table and the diagram below. I have again computed total and mean percentages, proceeding along the same lines as above. Mind that the percentages have been calculated with respect to the individual categories, which means that a value of 50% for women, for instance, means that half of the senser roles of this particular category (e.g. cognition or perception) are taken by females. Total Ind. Total Ind. Perception Absolute numbers 1,950 713 1,535 469 3,716 Total % 52.5% 19.2% 41.3% 12.6% Mean % 53.0% 21.5% 38.6% 12.2% Emotion Absolute numbers 1,730 1,068 1,192 690 3,164 Total % 54.7% 33.8% 37.7% 21.8% Mean % 53.8% 32.5% 38.5% 21.7% Cognition Absolute numbers 1,890 875 1,705 679 3,977 Total % 47.5% 22.0% 42.9% 17.1% Mean % 48.2% 21.3% 42.8% 15.1% Table 3: Absolute and relative frequencies of women and men occurring in the senser role of perceptive, emotional, and cognitive verbs/ adjectives in the pornography corpus. 0% 10% 20% 30% 40% 50% 60% Women: Total Women: Narratorindependent Men: Total Men Narratorindependent Perception Emotion Cognition Figure 1: Relative frequencies (total percentages only) of women and men occurring in the senser role of perceptive, emotional, and cognitive verbs/ adjectives in the pornography corpus. The most noticeable fact is that the gaps between women and men, in comparison to the overall frequencies, grow for perceptions and in particular for emotions, while it shrinks for cognition. This is in accordance with what I predicted and could be taken to support the subordinate hypothesis that in pornography interpretation, perceptions <?page no="238"?> Analyses 238 and emotions will be more relevant to the conceptualizations of women than cognition. But it is remarkable that even in the cognitive category women retain the upper hand (a result relativized but not reversed by the women/ men ratio mentioned above), so that there is no dimension of the mind where males can be claimed to be subjectified more persistently. Just as an aside: interestingly enough, the gap between the overall and the narrator-independent percentages are comparatively small for emotional sensers, suggesting that, relatively speaking, feelings are more likely to be attributed to others than thoughts or perceptions, at least in pornography. I will finally turn to the results for the individual words - contained in Table 4 - in order to examine whether there are some which conspicuously defy the trends found above. The percentages represent the size of a category relative to the overall occurrences of the verb. A value of 50% for men in the total line, e.g., thus means that the respective verb takes a male senser half of the times it is used in the corpus. see hear notice experience feel Ind. 17.7 17.4 16.0 13.1 27.4 12.4 26.8 9.2 19.8 9.0 Total 40.0 50.9 47.6 43.3 51.6 35.5 62.9 28.9 63.0 34.6 like love enjoy want wish Ind. 40.3 27.3 32.4 19.2 28.3 19.1 33.9 21.7 34.9 16.3 Total 53.9 39.8 58.9 35.4 51.6 35.2 55.4 38.3 58.1 37.2 need hope mind afraid excited Ind. 38.5 19.4 24.3 23.0 36.4 20.8 32.1 28.6 24.3 21.6 Total 55.6 28.6 50.0 44.7 54.5 39.0 55.4 42.9 45.0 44.1 know think believe realize imagine Ind. 23.2 17.0 22.4 20.5 18.2 9.8 19.7 11.6 22.5 8.5 Total 46.0 39.9 48.2 46.6 42.5 40.2 51.0 44.4 54.9 35.2 wonder expect sure remember forget Ind. 16.9 27.1 23.0 6.6 23.4 15.0 22.4 12.4 21.2 22.6 Total 44.4 52.2 43.4 40.2 55.7 41.3 54.0 37.9 41.6 49.6 Table 4: Relative frequencies of senser roles assigned to women and men by each of the 25 lexemes analysed in the pornography corpus (Ind. = Narrator-independent; all figures are percentages). <?page no="239"?> Representing people’s inner lives 239 Looking at the frequencies for the five perceptive process verbs reveals that the results are very heterogeneous. Contrary to the overall trends, see takes even more males than females as sensers (at least on the total level), while the female advantage is particularly large with notice, experience, and feel. These aspects suggest that even within the area of perception, the subjectification of women and men, though overall not dramatically far apart (at least not in favour of men), differs in details: men are likely to be interpreted more in terms of visual subjects (which - conversely - corresponds to the assumption that women will be positioned as the objects of vision, which has received some support in the chapter on physicalization and visualization), while women’s subjectivity will become more salient with perceptive processes verging on emotions - experience and feel (notice remains an exception here). Emotional processes are a principal domain of female subjectification in pornography and the results for the ten emotional process words do not suggest that this assumption needs to be differentiated. The only two significant deviations are need, with a greater than average female advantage, and excited, which sees women and men almost on par. Due to these words’ lower overall frequencies, drawing definite conclusions remains difficult, but the former result could mean that women’s emotional subjectivity will become even more salient, the more intense or inevitable the process (as is the case for need as opposed to want, wish, etc.). The latter result might at least indicate that that there is no difference between genders with respect to more physiologically-oriented emotional processes - excited might be the lone representative of this hybrid group. As for cognitive processes, the high frequency members of this category, viz. know, think, believe, basically agree with the overall relations. A few interesting, even if not easily explicable, deviations can, however, be found for some low-frequency lexemes. The greater gap between women and men for imagine might point to a stereotyping of women as more concerned with the imagination. The fact that wonder sees males more often in the senser role and the odd converse relations of remember and forget - women are described as remembering, men as forgetting more often - defies any attempt at a sensible interpretation. It does, however, not undermine the overall impression that women are not desubjectified, not even in cognitive processes. <?page no="240"?> Analyses 240 1.3 Comparing and discussing results For detailed data on all verbs/ adjectives, I refer readers to my homepage. I will immediately turn to the overall quantitative relations. The first table below contains all the relevant figures for the erotica corpus, the second one compares the two corpora with respect to how the senser roles are divided between women and men. Overall: 4,498 Total Ind. Total Ind. Absolute numbers 3,078 1,353 1,026 774 Total % 68.4% 30.1% 22.8% 17.2% Mean % 69.1% 30.1% 21.6% 15.4% Table 5: Absolute and relative frequencies of women and men occurring in the senser role of the 25 mental verbs/ adjectives in the erotica corpus. Total Ind. Total Ind. P E P E P E P E Total % 51.3% 68.4% 24.5% 30.1% 40.8% 22.8% 16.9% 17.2% Mean % 51.4% 69.1% 25.8% 30.1% 40.2% 21.6% 17.2% 15.4% Table 6: Comparison of relative frequencies of women and men occurring as sensers of the 25 mental verbs/ adjectives between the pornography corpus and the erotica corpus (Ind. = Narrator-independent). As female first person narrators predominate in the erotica corpus, the total values are skewed and thus misleading, as pointed out, and I will therefore primarily concentrate on the percentages of the narratorindependent sensers. The data in the tables above shows a wide gap between women and men even in the narrator-independent columns, which is particularly significant in view of the fact that a calculation in the “Fragmentation”chapter (cf. section 1.3) has shown that if we exclude first person singular pronouns, men occur slightly more often than women (by 18%). Furthermore, the difference between women and men is apparently larger than in the pornography corpus (where the gap is further relativized by a numerical advantage of references to women). There is thus a stronger emphasis on female sensers, a fact that corresponds to my predictions, and despite the unexpected results from the pornography corpus, which also saw a female lead, the data supports the hypothesis <?page no="241"?> Representing people’s inner lives 241 that the desubjectification of women will be an even less relevant factor in the discourse of erotica than in that of pornography. How does subjectivity distribute across the three categories of perception, emotion, and cognition in erotica in comparison to pornography? The two tables plus the diagram below include the figures relevant for answering this question. Total Ind. Total Ind. Perception Absolute numbers 986 356 249 180 1,355 Total % 72.8% 26.3% 18.4% 13.3% Mean % 72.9% 25.3% 17.4% 12.4% Emotion Absolute numbers 877 476 378 304 1,346 Total % 65.2% 35.4% 28.1% 22.6% Mean % 70.0% 34.9% 23.0% 18.4% Cognition Absolute numbers 1,215 521 399 290 1,797 Total % 67.6% 29.0% 22.2% 16.1% Mean % 66.2% 27.8% 22.2% 13.8% Table 7: Absolute and relative frequencies of women and men occurring in the senser role of perceptive, emotional, and cognitive verbs/ adjectives in the erotica corpus (Ind = Narrator-independent). 0 % 1 0 % 2 0 % 3 0 % 4 0 % 5 0 % 6 0 % 7 0 % 8 0 % W o m e n : T o ta l W o m e n : N a rra to rin d e p e n d e n t M e n : T o ta l M e n : N a rra to rin d e p e n d e n t P e rc e p tio n E m o tio n C o g n itio n Figure 2: Frequencies of women and men occurring in the senser role of perceptive, emotional, and cognitive verbs/ adjectives in the erotica corpus (total percentages only). <?page no="242"?> Analyses 242 Total Ind. Total Ind. P E P E P E P E Perception A 52.5 72.8 19.2 26.3 41.3 18.4 12.6 13.3 B 53.0 72.9 21.5 25.3 38.6 17.4 12.2 12.4 Emotion A 54.7 65.2 33.8 35.4 37.7 28.1 21.8 22.6 B 53.8 70.0 32.5 34.9 38.5 23.0 21.7 18.4 Cognition A 47.5 67.6 22.0 29.0 42.9 22.2 17.1 16.1 B 48.2 66.2 21.3 27.8 42.8 22.2 15.1 13.8 Table 8: Comparison of relative frequencies of women and men occurring in the senser role of perceptive, emotional, and cognitive verbs/ adjectives between the pornography corpus and the erotica corpus (A = Total %; B = Mean %; Ind. = Narrator-independent; all numbers are percentages). The results for the erotica corpus reveals that the percentage gap between women and men in the narrator-independent column remains stable across the three categories (between 13-16 percentage points). If we, however, focus on the ratios of the absolute frequencies (revealing whether female sensers are, e.g., twice as common or only half as common as male ones), the female advantage is lower for emotions than for cognition or perception. Contrary to stereotypes, feelings thus do not seem to be the category that features the highest amount of female subjectification. This is underscored by the fact that the overall genderrelated differences in the emotional category between erotica and pornography are smaller than in the other categories. In addition, we find that the main aspect distinguishing the two discourses is the significantly higher percentage of women in the senser role of cognitive processes in the erotica corpus. These results confirm my predictions, increasing the plausibility of the hypothesis that subjectification of women in erotica readers, in contrast to that in pornography readers, will not follow the stereotypical pattern of ‘women = emotion, but women ≠ thought’. Finally, I will take a closer look at the results for the individual verbs to see whether there are any deviating details worth further attention. The table below contains all the relevant percentages for the erotica corpus, immediately opposed to the corresponding figures from the pornography corpus. 1 1 I will not present the data on the erotica corpus separately here, as I have done above, because there are no special quantitative aspects that I could include in an extra table. <?page no="243"?> Representing people’s inner lives 243 see hear notice P E P E P E P E P E P E Ind. 17.7 27.3 17.4 18.2 16.0 22.8 13.1 15.7 27.4 20.0 12.4 20.0 Total 40.0 60.2 50.9 25.8 47.6 68.0 43.3 20.3 51.6 64.0 35.5 25.3 experience feel P E P E P E P E Ind. 26.8 29.3 9.2 0 19.8 27.3 9.0 8.2 Total 62.9 87.5 28.9 4.2 63.0 85.0 34.6 11.4 like love enjoy P E P E P E P E P E P E Ind. 40.3 47.7 27.3 26.7 32.4 38.0 19.2 29.2 28.3 24.0 19.1 22.0 Total 53.9 63.1 39.8 29.7 58.9 60.8 35.4 29.2 51.6 62.0 35.2 32.0 want wish P E P E P E P E Ind. 33.9 33.3 21.7 23.1 34.9 32.1 16.3 12.5 Total 55.4 64.6 38.3 30.8 58.1 76.8 37.2 16.1 need hope mind P E P E P E P E P E P E Ind. 38.5 25.6 19.4 16.3 24.3 37.5 23.0 14.6 36.4 52.6 20.8 21.1 Total 55.6 60.5 28.6 19.8 50.0 68.8 44.7 22.9 54.5 73.7 39.0 26.3 be afraid be excited P E P E P E P E Ind. 32.1 29.5 28.6 11.4 24.3 28.6 21.6 7.1 Total 55.4 84.0 42.9 15.9 45.0 85.7 44.1 7.1 know think believe P E P E P E P E P E P E Ind. 23.2 30.3 17.0 17.4 22.4 27.5 20.5 18.3 18.2 11.1 9.8 11.1 Total 46.0 66.0 39.9 22.2 48.2 68.7 46.6 23.0 42.5 58.3 40.2 25.0 <?page no="244"?> Analyses 244 realize imagine P E P E P E P E Ind. 19.7 27.8 11.6 15.6 22.5 28.8 8.5 5.5 Total 51.0 71.1 44.4 22.2 54.9 75.3 35.2 16.4 wonder expect sure P E P E P E P E P E P E Ind. 16.9 20.7 27.1 10.9 23.0 28.2 6.6 15.4 23.4 23.7 15.0 15.8 Total 44.4 75.0 52.2 20.7 43.4 51.3 40.2 23.1 55.7 55.3 41.3 32.9 remember forget P E P E P E P E Ind. 22.4 43.6 12.4 10.0 21.2 36.4 22.6 18.2 Total 54.0 79.2 37.9 12.9 41.6 61.8 49.6 23.6 Table 9: Comparison of relative frequencies of women and men appearing in the senser role of the 25 mental verbs/ adjectives analysed between the pornography corpus and the erotica corpus (P = Pornography, E = Erotica; Ind. = Narratorindependent; all figures are percentages). With the proportions of the narrator-independent category varying dramatically between the individual verbs due to very low absolute frequencies, comparisons of differences are problematic. I will, as a consequence, be economical with my comments on patterns and deviations. With respect to perceptive processes, the data in the table above - I again focus on the narrator-independent lines - points to some similarities between the two corpora. The more genuinely perceptive verbs see and hear show a smaller gap between females and males (though not dramatically smaller), while those bordering on emotions, viz. experience and feel, feature higher margins. But even though the differences of specific perceptive processes do not vary tremendously between the two corpora, the conclusion that the subjectification created in erotica will be stronger than that created in pornography - females are clearly in first place for all verbs apart from notice, which has equally high proportions of female and male sensers in the erotica (possibly a result of its preferred use in direct speeches in connection with questions, e.g. “Have you noticed…”, addressed as often to women as to men) - remains intact. Turning to verbs and adjectives for emotions, there is such a great amount of variation among the ten verbs in the erotica corpus that I am not able - and would not dare - to offer a convincing interpretation. It is, however, interesting to see that the deviations found in the pornography <?page no="245"?> Representing people’s inner lives 245 corpus are reversed here, need showing a smaller gap between the numbers of female and male sensers than the average of the emotional category, and excited being much more common with women as sensers, in contrast to the pornography corpus, where I have found the percentages to be almost equal. The heterogeneous set of data only indicates that the emotionalization of women and men does not follow any clear patterns in the erotica. It is also noteworthy that there is a significant difference between the women/ men ratios of the verbs like and love in erotica, with the gap being smaller for the latter (why do women like more than they love? ). A similar lack of apparent patterns is displayed by the cognitive verbs. This heterogeneity becomes most obvious in the fact that think and believe, which are fairly close in meaning (in particular with respect to the epistemic sense of think), differ so widely in their distributions of female and male sensers - think takes more female sensers than average, in believe the percentages are exactly the same. Imagine shows a similar preference for women in the erotica as in pornography, which might indicate that erotica do not escape the stereotypes associated with this process. But there are otherwise too few clear trends to offer a satisfying conclusion. To sum up, there is one central area where we have seen an unexpected result, namely that in the pornography corpus women are assigned the senser role more frequently by mental verbs and adjectives than men. Otherwise the data has produced the expected outcome: the proportion of female sensers is higher in the erotica corpus, and with respect to the mental subcategories, we find women to occur less often as sensers of cognitive processes relative to men than in perceptive and emotional processes in the pornography corpus, a trend reversed in the erotica corpus. The data thus overall lends some weight to the hypotheses that women’s subjectification will be a more salient factor - and their desubjectification consequently a less salient one - in erotica consumption than in pornography consumption and that it conforms to, and thus sustains, gender stereotypes in pornography but not in erotica. The data, however, has failed to produce evidence supporting the hypothesis that generally women’s desubjectification will be more substantial in pornography readers’ interpretations than men’s. Considering that I assume desubjectification to be the most heterogeneous/ contradictory of the global conceptualizations examined in this study, this conclusion does not come as a complete surprise. But the consistency of preference for women goes farther than expected. <?page no="246"?> Analyses 246 2 Marked perspective: passive voice I would like to start this chapter with two examples: John married Mary. Mary married John. We probably agree that the first sentence represents an event from John’s point of view, while the second one takes Mary’s in describing the same event. While the thematic structure and the fact that John is the theme of the first and Mary of the second example might play a role in this change in perspective, the sentences allow for another conclusion, namely that the determining factor is the grammatical subject, which happens to converge with the theme in the structures given. Assuming that subjects play a role in the establishing of point of view, even though the theoretical relation to the thematic structure is far from clear, I will explore the contribution of this linguistic element to the subjectification and desubjectification of women and men in pornography and erotica. Instead of examining all relevant items, however, I will focus on passive voice as a construction which, as a result of its influence on syntactic functions, constitute the marked attempt to establish a certain perspective. Grammatical subjects may directly affect the construction of a perspective because it could be argued that it is their primary function in the information structure to denote the vantage point relative to which an event is represented. This is, for instance, Simon Dik’s position, for whom a subject is “that constituent which refers to the entity which is taken as a point of departure for the presentation of the state of affairs in which it participates” (Dik 1978: 87, cit. in Van Valin/ Lapolla 1997: 246, for problems associated with this view, cf. also Siewierska 1991: 74f.). What is now the significance of passive for grammatical subjects and their role in subjectification? The opposition between active voice and passive voice allows speakers to choose the subject from (at least) two alternatives. According to Tallerman (1998: 22 and 178f.), passive involves the promotion of a phrase to the subject position and the demotion of the original subject, which is either deleted, as in the first example below, or assigned to a more peripheral position, usually a by- PP, as in the second example. This Amazon wanted to be fucked; so did the prim librarian. The severe-looking schoolmam had been replaced by a sensuous temptress. <?page no="247"?> Marked perspective 247 The terms promotion and demotion, despite their process-like character, are not supposed to suggest that the passive voice is a syntactic derivation (or transformation) from the active voice, as was formerly assumed in transformational syntax. Following functional approaches to grammar, active and passive are just considered to constitute two choices. But the choices are not on an equal level because active voice can be argued to be the unmarked structure, which means it is more common and probably more automatized, with its subjects being the default options. The passive voice is consequently the marked construction, occurring less often and possibly requiring more processing efforts, which, in turn, means that its subjects are not the common ones. If grammatical subjects now can be equated with the vantage point of an event, as postulated above, then using passive voice means adopting a perspective that is less common and that requires more effort to be established, probably as a consequence of the fact that it is more difficult to take the view of a patient or a receiver, which are the normal semantic roles of subjects of passive clauses. Passive thus marks an attempt to control perspective rather than attribute it to a default option. It is therefore interesting to see whether the discourses of pornography and erotica employ this structure to establish femaleor male-centred views of the world and, consequently, to desubjectify women or men. Since I assume that a perspective will gain salience the more often it occurs, I will mainly examine the frequencies at which women and men appear as subjects of verbs in the passive voice. All these considerations lead to the following predictions: • Women will appear less often as subjects of verbs in the passive voice than men in the pornography corpus. • In the erotica corpus, this will not be the case or at least the gap from men to women will not be as wide as in the pornography corpus. 2.1 Technical and practical aspects of the analysis I am embarking on a full analysis of passive verbs, which means I will look at all instances of passive voice (with some limitations, as will be discussed below). This requires producing a concordance of all passive verb forms. As the passive voice is formed by ‘be + past participle’, the search string for the concordance is exactly this sequence (both forms can be selected by means of the tags). As intensifiers, adverbs and adverb phrases can intervene between the two elements, e.g. I was so swamped by his sheer sensuality […] or […] he was as deeply embedded <?page no="248"?> Analyses 248 inside me […], and as in questions, a lot of other words can fill the space between the two elements, e.g. Was he really as deeply embedded inside me? , I allow for be and the participle being separated by up to four words. This search yields concordances such as the following one. vy<JJ> redhead<NN1> was<VBDZ> bent<VVN> over<II> a<AT1> and<CC> we<PPIS2> were<VBDR> buried<VVN> alive<JJ> ,<,> mbarrassed<JJ> to<TO> be<VBI> caught<VVN> drooling<VVG> is<APPGE> face<NN1> was<VBDZ> coated<VVN> with<IW> my<AP <NN1> only<RR> to<TO> be<VBI> coated<VVN> with<IW> yet<R IS1> was<VBDZ> completely<RR> spent<VVN> .<.> <^> I<PPIS the<AT> loving<JJ> was<VBDZ> done<VDN> and<CC> we<PPIS2 until<CS> I<PPIS1> was<VBDZ> drenched<VVN> in<II> my<AP <APPGE> pussy<NN1> being<VBG> eaten<VVN> by<II> s<NN2> ca<VM> n't<XX> be<VBI> enjoyed<VVN> by<II> myself had<VHD> never<RR> been<VBN> enjoyed<VVN> before<RT> .< an<AT1> answer<NN1> was<VBDZ> expected<VVN> .<.> I<PPIS1> Concordance 2: Verbs in the passive voice in the pornography corpus (extract). From the resulting concordance, I first exclude all lines with non-human subjects (e.g. an answer was to be expected), being interested in human beings only. I then examine the remaining instances, counting the occurrences of female and male subjects, with WordSmith’s ordering function and the ‘gendered’ subcorpora once again being of help. Some of the passive voice constructions resemble ordinary ‘be + adjective’ strings and even the automatic tagging programme used for my corpora does not consistently distinguish between these two structures. To ensure that all instances of passive voice in my broad definition are included, I therefore also produce control concordances where the position of the participle is taken by an adjective (as the problematic cases are primarily to be found among those forms ending in -ed, e.g. amazed, surprised, frightened, etc., I limit myself to adjectives of this kind), of course, weeding out examples that go beyond the scope of my present interest. English allows for auxiliary be, which normally combines with the past participle form of the lexical verb in the passive, to be replaced by get or - more rarely - by become, to emphasize the dynamic aspect of events (sometimes, arguably, hardly distinguishable from verb plus adjective combinations, though). My sister and I both became totally controlled by the fire burning deep inside us both. As I said, Stretch got traded to another team. <?page no="249"?> Marked perspective 249 In order to include getand become-passives, I also carry out searches where be is replaced by any form of get and become, otherwise proceeding exactly as above. Two structures deserve special attention as they pose some problems for counting. The first of these are non-finite passive constructions such as the following ones, which might be argued not to feature a subject at all: […] with the other guys waiting to be assigned to a golfer […] Now it was my turn to be pleased. […] my worry about being mugged was dissolving in direct proportion to the swelling of my cock. To be syntactically precise, these non-finite clauses (shaded) lack an overt subject. Can such clauses contribute to the creation of a perspective then? I think they can because we automatically establish an invisible subject that corresponds to some element in the immediate co-text of the clause (invisible means ‘not occurring in the pre-verb position reserved for subjects’ - of course, it appears overtly somewhere else), e.g. in the first example, the guys functions as the subject, in the second and third ones I does (derived from the possessive pronoun my) (I will skip the syntactic arguments for this solution, they are usually dealt with in connection with the concepts of PRO and control, cf. Haegeman 1991: ch. 5). In the counting, I therefore have to identify which element this covert subject is linked to. If it is linked to none - there are cases of ‘free’ subjects, e.g. To err is human - I will regard the subject as unclassifiable. The second problem are instances of participles being conjoined by and, or, or but, e.g.: He looked at her with amazement, […] enjoying it, yet ashamed at being taken and used. As the effect on perspective just happens once (I do not believe that the effect will be intensified by the conjoining), I count such structures as just one subject. Further problematic cases are less concerned with establishing the subject but with the more general question whether they constitute passive constructions at all or whether they just resemble them. The first one of these questionable structures involves lexicalizations of passive forms, where past participles have become or are becoming adjectives. I have already mentioned that in particular the past participles of verbs denoting emotions, e.g. surprised, confused, upset, shocked, etc., have <?page no="250"?> Analyses 250 almost acquired the status of adjectives (note that I treated some of them as adjectival descriptors in chapter VI). As they, however, can still be argued to correspond to active clauses - e.g. I am confused because something or someone has confused me - I will take a broader definition of passives here and include them. This is also justified by the fact that many of these verbs do in fact occur with prepositional phrases with by, an indication that they are still treated as the passive voice of the verb. […] and this female seemed more shocked by Gina standing there. […] She was most surprised by the strength in his grip Pseudo-passives (my term), i.e. constructions that look like passives formally, but do not really - or only very remotely and vaguely - correspond to an active version, are also problematic. Typical examples include the past participle forms of the verbs go, head, finish, done: Her roommate was gone for a week […] […] a good-looking blonde waitress that was headed towards our table […] after Julia and Peter were finished. "We are not done yet." Stephanie was used to working around the ponds by herself. There are, of course, cases where the same verbs can be used for a genuine passive clause, as, for instance, in: […] leaving Jean shaking and begging to be finished off. Generally speaking, however, these structures do not involve the promotion of an object into subject position and thus cannot be used to alter the perspective on a situation. Pseudo-passives will therefore not feature in the analysis. A final deviant group of passives are raising verbs such as say or suppose. […] when he’s supposed to be playing golf and then […] the rich members were said to lay big tips on the caddies. […] a happy hour I was destined to spend pressed between them. The passive voice of such verbs does not involve the promotion of an object into subject position but raising the subject of a subordinate (nonfinite) clause (he and the rich members in the examples above) to the subject status in the superordinate clause (cf. Haegeman 1991: 306f.). <?page no="251"?> Marked perspective 251 Although the new subject is not a ‘passive’ element (in the semantic sense of passive) as is the case with prototypical passive voice, raising does involve promoting an element to a higher-order position and thus imposing its perspective on the event. Raising verbs will thus be included in the examination. A last group of passive constructions will be excluded from considerations not for theoretical but for practical reasons. These are past participle clauses without any form of be, become or get, e.g.: Taken by surprise, I screamed out and pulled Steve’s hair with both hands. Overwhelmed with the notion of earning $100 for something she enjoyed, Francine agreed. It is impossible to consider them, because the only defining element is the occurrence of the past participle. Searching for past participles without any restrictions, however, would underspecify the target structures completely, giving me thousands of present perfect tenses, for instance. I will again differentiate between female and male subjects, relegating all mixed-gender or gender-neutral ones to a separate category (called Other below) or, if the gender could not be established, classifying them as unidentifiable. I will also retain the distinction between narrator-related and narrator-independent subjects in order to avoid - once again - neglecting the influence of the narrator’s gender, particularly in the erotica corpus. 2.2 Presenting and discussing results The table below contains all important figures concerning passive constructions with human beings as subjects. The relative frequencies represent the size of the respective category relative to the overall numbers of passive clauses (absolute frequency of the category divided by the overall number of passive clauses). A value of 50% for men thus, e.g., means that half of all passive clauses take a male subject. <?page no="252"?> Analyses 252 Overall: Other Unid. 1,794 Total Ind. Total Ind. Absolute numbers 977 607 607 296 202 8 % 54.5% 33.8% 33.8% 16.5% 11.3% 0.4% Table 10: Absolute and relative frequencies of women and men occurring as subjects of passive verbs in the pornography corpus (unid. = unidentified). Contrary to my expectations, women appear more regularly as the subjects of passive clauses than men in the pornography corpus, even if we take into account that they are mentioned 27% more frequently (47% if we just consider narrator-independent expressions). This further undermines the hypothesis of desubjectification being a more salient aspect in the conceptualization of women than in that of men in pornography consumption. The data may indicate that female subjectivity is more important for pornography than hypothesized. But there might be a second reason for the female advantage in the results above, namely a tension between the objectifying conceptualizations of passivization (to be discussed in the next chapter) and desubjectification. Passive voice, as described, involves the promotion of a participant usually associated with semantic roles considered to be more passive (meaning-wise), e.g. the patient or the receiver. The next chapter will reveal that women tend to be assigned such roles more consistently than men, which entails that pornography consequently has fewer chances to make men the subjects of passive clauses. This also provides a logical explanation of why males thus show lower percentages than females in the table above. However, I do not think that it can explain away the fact that the data does not even remotely agree with the prediction, viz. that men would gain the upper hand, so that the conclusion that there is no support for the desubjectification hypothesis in pornography can be maintained. 2.3 Comparing and discussing results I will now try to find out whether the second hypothesis, concerned with the differences between pornography and erotica, receives more support than the first one. In the first table below, you find the relative frequencies of female and male subjects of passive clauses in the erotica corpus. The second table compares the percentages of the two corpora. <?page no="253"?> Marked perspective 253 Overall: Other Unid. 557 Total Ind. Total Ind. Absolute numbers 327 145 161 138 62 7 % 58.7% 26.0% 28.9% 24.8% 11.1% 1.3% Table 11: Absolute and relative frequencies of women and men occurring as subjects of passive verbs in the erotica corpus. Total Ind. Total Ind. P E P E P E P E % 54.5% 58.7% 33.8% 26.0% 33.8% 28.9% 16.5% 24.8% Table 12: Comparison of relative frequencies of women and men occurring as subjects of passive verbs between the pornography corpus and the erotica corpus (P = Pornography, E = Erotica, Ind. = Narrator-independent). The gap between women and men appearing as subjects of passive clauses in the erotica corpus, which is very wide for the total percentages, almost disappears if we neglect all instances of the first person narrator taking this position. This could, however, be interpreted as a consequence of men generally appearing more often as narrator-independent persons in the erotica corpus than women. What comes as a surprise in the results above, particularly regarding my expectations, is that in comparison to the narrator-independent frequencies of the pornography corpus - and it is these I will be focusing on - the quantitative difference between female and male passives is much lower in the erotica corpus. We, however, have to consider that the women/ men ratios (with respect to narrator-independent expressions) differ widely (47% in favour of women in pornography, 18% in favour of men in erotica), so that this gap may eventually disappear, although again this is difficult to prove mathematically as it remains unclear how the different ratios can be correlated. The data can thus not really be interpreted as providing extra evidence in support of the hypothesis that women’s perspective will play a more pervasive role for erotica readers than for pornography readers and that women’s desubjectification will consequently be a less relevant aspect in erotica. However, we have to consider in this context that, as will be shown in the next chapter, women do not appear more often in passive roles than men in the erotica corpus, which means that, unlike the pornography data, the results from above are not affected by passivization. The fact that I have ignored narratordependent cases above, which weigh heavily in the erotica corpus, will <?page no="254"?> Analyses 254 be another factor contributing to the above conclusion, which therefore has to be treated with some caution. All in all, the results have failed to provide support for my predictions concerning subjects of passive clauses. The percentage of women appearing in the subject position is higher than that of men in the pornography corpus. The data thus does not add to the plausibility of the hypothesis that women’s perspective will be less important than men’s in the discourse of pornography. It might even be taken to lend further weight to the assumption that desubjectification is a relatively heterogeneous dimension in pornography because the female perspective is needed for the full appreciation of male sexual desires. Regarding differences between the two discourses under examination, the data also remains inconclusive. While some of the results may be interpreted as counterevidence to my hypotheses, they can, however, also be taken to indicate that the assumption about the contribution of passive voice to the creation of subjectification was too optimistic. In this context, I have to mention the alternative view on passive by Henley, Miller and Beazley (1995). The authors conducted experiments showing that the passive voice if used in the description of sexual violence reduces the responsibility of the perpetrator and the harm to the victim in the interpretations of readers. For them, the said grammatical construction thus reduces the agency of the subordinated even further and add to their powerlessness. Subjectification therefore comes at the cost of passivization, which means that there is some tension between two global conceptualizations - the one objectifying (passivization), the other not (subjectification). This tension could, in turn, result in heterogeneous linguistic data. 3 Casting a side glance: interesting phenomena not examined The problem with many aspects involved in the creation of subjectification is that they cannot adequately be dealt with in a transtextual analysis. There neither is a pre-defined set of expressions that I could look for (apart from a few deictic expressions), nor would they, had I found a way to trace them, make it easy to identify the perspective they imply because intensive co-textual investigation would be required. This, in turn, would prevent me from quantifying the effects that they will have on consumers and thus from bringing them in relation with other aspects analysed. But one paragraph from the pornography corpus should suffice <?page no="255"?> Casting a side glance 255 to show how prominent a female perspective can be, even if it is analysed co-textually. Daphne revelled in the feeling of John’s strong arms around her, the sensations of his taut, muscular body pressed against the feminine softness of her own body causing a pleasant tingling in her loins. The tingling developed deliciously into a divine dampness as her desire deepened, and she longed for more than a mere embrace. She longed to be satisfied as a woman should be satisfied. She loved to feel his penile barge navigating its way up her oozing love canal. “I’ve got some super news, Daph,” John told her excitedly. “Mummy’s going away next weekend and I’ve got the house to myself.” “I’ll be there,” she replied, swooning at the thought of spending an entire weekend alone with her dashing captain. Daphne does not only appear repeatedly as the theme (most importantly, as the paragraph theme) and as the subject of clauses, it is also her mind that is constantly referred to, e.g. in the mental verbs revel, long, be satisfied, love, swoon (in the not fully literal sense used in the passage), the mental nouns feeling, sensation, tingling, thought, and the evaluative adjectives and adverbs pleasant, divine, dashing, and deliciously. So this is a very clear example of a female perspective being created in the readers. Interestingly enough, the anti-objectifying effect of the female point of view is undermined by a metaphor constellation strongly passivizing females. In the participle clause his penile barge navigating its way up her oozing love canal, the penis is represented as a vehicle (I will classify this as a marginal case of a tool metaphor in the next chapter, cf. section 2.2) and the vagina as its direction (classified as a passage metaphor below). As the penile barge is moved by its owner, i.e. by John, and the vagina only provides the path for the movement, these metaphors suggest a conceptualization of the sexual act being initiated and performed by John, with Daphne only as the receiving end. Let me finally just add that an analysis of desubjectification could, of course, also examine the most obvious tool for creating perspective, namely the use of first person pronouns, which usually means first person narrators. Since the pornography corpus features similar numbers of female and male first person narrators, reflected by almost equal numbers of first person pronouns - 13,965 for women and 13,304 for men - there is not much difference to be found here. But in the erotic corpus with its clear preponderance of female first person narrators and the concomitant advantage for female first person pronouns - 7,184 vs. <?page no="256"?> Analyses 256 1,013 - this assigns a clear prominence to a female perspective. But the analyses have shown that there are also independent factors contributing to subjectification. 4 Conclusion To avoid thinking of people in terms of their inner mental lives is part of seeing them as objects. Desubjectification is thus a form of objectification. As an overall amount of desubjectification is not really conceivable, I have concentrated on gender differences only, dealing with the question of whether pornography and erotica can create a distinction between women and men with respect to the degree of desubjectification in their readers’ minds. To research this linguistically, I have examined the frequencies of women and men being assigned the senser role of a select set of mental process verbs and adjectives and the number of times they take the grammatical subject function in passive clauses. I have taken a negative approach in this chapter, analysing linguistic phenomena creating subjectivity to see whether they are less common for women than for men and then drawing conclusions with respect to desubjectification. My first hypothesis was that the analysis of the linguistic structures mentioned would suggest that women’s perspective and their subjectivity should play a less significant role compared to men’s. The data overall does not support this assumption. Women consistently appear more frequently than men in structures associated with subjectivity. It is thus women’s subjectification rather than their desubjectification that will gain a salient status in pornography consumption. To a certain extent, this conflict of perspectives was to be expected - and the hypothesis was thus tentative right from the beginning - since a female point of view could not be completely absent, as pornography users also need a reconfirmation that what they are doing to women is appreciated. My second hypothesis said that women’s desubjectification would be lower in erotica than in pornography. The data does not allow a definite conclusion concerning the main point of the hypothesis, but if we take into account how often women and men are mentioned in the two corpora, then we at least see larger margins between female and male results in the erotica corpus than in the pornography corpus. This applies to the proportions of female senser roles, but less so to female grammatical subjects. <?page no="257"?> Conclusion 257 I also included analysis of an extra hypothesis, namely that the subjectification of women would follow stereotypical patterns, which means more emphasis on emotions and perceptions than on thought, in pornography but not in erotica. Although this question could only be examined in the part on mental process verbs, the data indicates that this assumption is plausible. In sum, of all the objectifying conceptualizations examined in this study, women’s desubjectification seems to be the one that is least likely to be a salient element in pornography consumption. There may be aspects revealed in finer details, which cannot be examined with corpus methods, but the data analysed in this chapter does not suggest that pornography is characterized by a strictly male perspective. But given the mentioned necessity to show consumers that women love what men are doing to them this does not necessarily mean that pornography is not discriminatory. The other aspects examined suffice for such an effect. It has to be conceded that many of the results might have been blurred by theoretical problems. Firstly, there is the preponderance of female first person narrators in the erotica corpus, which, of course, produces an a priori advantage for a female perspective and therefore has to be considered in connection with all other phenomena. In some cases, it might be argued that perspectivization through narratorindependent cases is just a minor aspect, particularly in erotica. Secondly, particularly the foundation for the assumption that grammatical subjects contribute to perspectivization is not solid, which also reduces the validity of the results gained in the analyses concerned with this element. <?page no="259"?> VIII Passivization The animal in me wanted to fuck my sister’s ears off They found themselves making verbal love to one another Megin knelt at Jill’s joybox and began using her able tongue The ability to act intentionally, being the initiating source of events and imposing one’s will on the world, i.e. agentivity, is a genuinely human quality. Though some animals may be claimed to possess it to a limited extent, it is completely absent from objects. Inanimate things are rather characterized by passivity, which always finds them at the receiving end of events, being at the mercy of other entities’ decisions and whims. Human beings can, of course, be conceived of as active or as passive, depending on their roles in particular situations. But we should at least expect the two poles to equal each other over time. But thinking of humans primarily and consistently as inactive recipients, i.e. passivization, means putting them on par with passive objects. If pornography successfully creates passivization in its consumers, then it is justly regarded as objectifying. Passivization can mainly manifest itself in our thinking of (partners in) interpersonal relationships (including sexual relationships, which are the main subject matter of erotic discourses 1 ). Passivizing one partner in such a relationship often necessitates agentizing the other. Passivity and agentivity thus entertain an inverse correlation with each other: if one partner is more active, the other one is more passive. 2 This makes it difficult to speak of an overall amount of passivization, as we did with the global conceptualizations in some of the previous chapters. The link between sexuality and objectification, which feminism often claims to be the special feature of pornography (cf. Campaign Against Pornography and Censorship “Policy Statement,” repr. in Chester/ Dickey 1988: 270-273, cit. in Smith 1993: 73 [no further details given there], Jensen 1995: 299), seems a particularly pertinent point in passivization. I will therefore put more emphasis on sexual relationships and contact in the current chapter. 1 The words partner and relationship should not be understood as necessarily indicating fixed, stable or non-temporary bonds between people. 2 This claim has to be taken with a grain of salt, since situations where both partners are seen as active are, of course, easily conceivable. <?page no="260"?> Analyses 260 Passivization will have similar discriminatory effects on (sexual) relationships as other objectifying tendencies. Since active roles allow more freedom to realize our own potential and our own ideas and also to choose whether and how to do so, they are without doubt more important. If we perceive our sexual partners as passive and ourselves as active, we might feel tempted to impose our will more strongly - or even more forcefully - on them. This may undermine attempts to enjoy a reciprocal relationship based on egalitarian principles and it will definitely impede searches for equal and balanced sexuality. This particularly applies to the passivization of women as the ideology of passive females and active males is deeply ingrained in our - patriarchal - culture. Although only Andrea Dworkin establishes an explicit link between objectification and passivization - most notably in the chapter “Objects” of her book Pornography (1981) - the prominence of the distinction between passivity and agentivity is widely acknowledged in the literature. Many writers on pornography point out that there is a culturally omnipresent script of passive women and active men in sexuality (cf. Berger/ Searles/ Cottle 1992: 35, Caputi 1994: 2, Smith 1993: 84, Wareing 1994, cit. in Mills/ White 1997: 238). There are, however, diverging views on the source of this unequal division of roles. While some (cf. Dworkin 1981, Johann/ Osanka 1989: 25) consider pornography to play a pivotal role in the construction of the active/ passivescript, others blame anti-pornography feminists if not for creating it in the first place, then for maintaining it (cf. Heise 1995: 124ff., Bristow 1997: 157). In all accounts of the active/ passive-dichotomy, its link to inequality is implied, but never mentioned separately except by Kappeler (1986: 50): “Social relationships are relations between subjects: if there is exchange or communication, each partner is and remains a subject or agent of action, or a subject of speech and communication. The roles are reciprocal, the situation is one of intersubjectivity.” As the quote shows, she even finds a connection to the issue of subjectification. Although no research, to my best knowledge, has been undertaken to investigate passivization, the prominence it has gained in the theorizing of pornography and sexuality proves it a topic that is well worth a more profound analysis. In the current chapter, I will look at linguistic items supposed to contribute to the construction of certain groups as more passive than others. The first subchapter will be concerned with syntactic patterns and the concomitant semantic role allocation of various verbs denoting sexual intercourse, (other forms of) physical contact and social encounters. The second subchapter will deal with conceptual metaphors for the genitals, trying to establish whether particular categories of figurative <?page no="261"?> Positioning sexual partners 261 meanings can be associated with passivity or agentivity. I will examine to what extent these linguistic structures can render particular conceptualizations more salient by looking at the frequencies and lexical variation of particular elements in the corpora. I expect the data analysed in this chapter to support the following hypotheses: • Women will be subjected to passivization more strongly than men in pornography. • The imbalance between the two genders with respect to passivization will not be as clear in erotica as in pornography (even though I believe that inequality is so ingrained in general conceptions of sexuality that erotica will not reverse the trend that women are seen as passive in sexuality). Having mentioned that inequality seems particularly delicate in the area of sexuality, I would like to add a subhypothesis, namely that all tendencies to passivize will be stronger, the more sexual the aspect described (this will primarily be of relevance to the first subchapter). 1 Positioning sexual partners: semantic role patterns Language provides us with a large repertoire of tools for making subtle distinctions in positioning (sexual) partners with respect to agentivity and passivity, particularly in the representations of events in which both partners are involved. The probably most important of these devices are the semantic roles that NPs referring to human beings are normally assigned syntactically by verbs, as has already been acknowledged - even though not by a linguistically profound analysis - by Catharine MacKinnon, “man fucks woman. Subject, verb, object” (1982: 541, cit. in Cameron/ Kulick 2003: 29). This subchapter will analyse semantic role patterns, concentrating on a number of verbs denoting intercourse, other physical contact and social encounters, assuming that these are the most significant processes that people in sexual discourses are involved in, no matter whether they are partners in permanent or temporary relationships (my ideas for this chapter have been partly influenced by Manning 1997, cf. also Cameron/ Kulick 2003: 29f.). Before discussing semantic role patterns in more depth, I have to mention one prerequisite assumption underlying the ensuing theoretical preliminaries, namely that passivity versus agentivity does not form a complementary opposition, but that it rather represents the poles of a <?page no="262"?> Analyses 262 scale. So it is possible to speak of one construction being more passivizing and the other one more agentizing. The following five clauses are supposed to provide a preview of the linguistic differences that will be analysed in this subchapter. we fucked right there he fucked her hard and fast he fucked her tight cunt he fucked into me All of the four clauses contain the verb fuck in the sense of ‘having intercourse’. But it appears in different verb constellations, 3 which means the syntactic and semantic patterns of the environments vary, with the two sexual partners appearing in different syntactic functions (subject, object, prepositional phrases) and consequently in different semantic roles (agents, patients, locatives). Additionally, the partners may feature as wholes or merely as parts of their bodies (her tight cunt). I will argue below that changes in syntactic functions and semantic roles in combination with preferences for holistic or fragmented references determine the degree of passivization. My first linguistic focus will be the constellations that particular verbs enter into and which semantic roles are assigned to different persons or classes of persons by these constellations. To summarize what I have said about semantic roles in the chapter on desubjectification, they refer to the parts that participants (those directly involved in an event) and circumstances (the settings of the event) play in the event that the whole clause describes. The verb determines which participant roles are required (e.g. fuck normally requires an agent and a patient) and these are formally realized by the subject and the different kinds of objects (e.g. in fuck in the active voice, the agent is realized as the subject NP and the patient as the object NP). Circumstantial roles are independent from the verb and are optional. They are determined by the phrasal heads of the element taking the role (e.g. a PP with the head after realizes a temporal role, irrespective of the co-text). Which semantic roles will now be relevant for my analysis in this subchapter? As mentioned above, my main focus will be on intercourse, physical contact and social encounters, suggesting that I will almost ex- 3 I adopt this term from Carlota Smith (1991, cit. in Lenz 1997: 119), even though she focuses on verbs and their complements only. As the boundaries between complements and adjuncts are not always easy to draw, I define verb constellations in broader terms to include the latter. <?page no="263"?> Positioning sexual partners 263 clusively be concerned with material processes (the only exceptions are verbal processes, which I will comment on below). The only participant roles of relevance for this category - for the present analysis - are agents and patients (beneficiaries/ receivers do not figure prominently and will therefore be ignored; but see the paragraph on verbal processes below). Anyhow, I fucked her In this example, for instance, the event is represented as initiated by the first person (I is thus the agent) and affecting the third person (her is thus the patient). Sexual partners do, however, not always take participant roles, as the examples below prove. My husband was fucking with Trang every chance he got. Then I fucked slowly into her. he fucked in and out of her he kissed along my shoulders While I pinched and fondled at her ass cheeks […] The agent in the four clauses above is always the subject, but the other partner is not represented as a participant at all but rather as part of the circumstances. In the first clause, Trang occurs in the circumstantial role of comitative, i.e. the person together with whom or in the company of whom I do something. They normally appear as with-PPs (cf. Halliday 1994: 156). In the other four clauses, the second person involved occupies the circumstantial role of locative, with some internal differences, though: into her is a directive, i.e. the direction of an event, in and out of her is both a directive and a source, i.e. where the event has originated, and along my shoulders is a path, i.e. the way along which an event progresses. Directives, sources and paths imply motion. At her ass cheek, finally, is a location, i.e. where something happens. This role does not imply motion. Directives are realized by PPs with into, onto, to, etc., sources by PPs with from, out of, off, etc., paths by PPs with along, around, etc., 4 and locations by PPs with on, at, in, etc. (cf. Jackson 1990: 49f.). The prepositional phrases can be replaced by corresponding adverbial phrases such as there, here (which we might also regard as pro-forms for prepositional phrases). 4 Paths can also be expressed by PPs with a preceding NP, e.g. Louise began kissing a trail down and over the younger woman’s smooth, flat midriff, or The one hand stroked a path downward to her slightly swollen mound. <?page no="264"?> Analyses 264 Besides, (parts of) sexual partners may also appear in the circumstantial role of instrument, i.e. the thing with the help of which something is done, as can be seen in the example below, where with your tongue is the instrument. […] she said, whimpering, “fuck me with your tongue.” Instruments normally occur in prepositional phrases headed by with (cf. Jackson 1990: 54f.). But this role is of less relevance to me here since its occurrence almost always implies that the whole person is the agent in the same event. One of the two social activity verbs to be examined in this subchapter is talk. In Halliday’s (1994) taxonomy of process types, this verb does not qualify as a material process but belongs to verbal processes. Strictly speaking, the latter assign a different set of semantic roles, namely sayer to the person (or other entity) producing a semiotic message, receiver to the person who is the recipient of the message, and verbiage to the message itself, whether it is defined by content or wording (cf. Halliday 1994: 140f.). Despite differences between sayers and agents, there are also apparent parallels concerning agentivity, as sayers are also initiators of action. With regard to passivity, we can see similarities between patients and receivers (even though we can also find parallels between verbiages and patients). This entails that I am justified to equate sayers with agents and treat receivers in verbal processes as similar to patients in material ones. Roughly speaking, the semantic roles mentioned form a cline with respect to agentivity and passivity. The agent is, of course, associated with the active pole and the patient with the passive one. But the other roles can also be assessed with respect to this opposition. The role of comitative, though not fully agentive, still is inclined towards this end of the scale - after all, doing something together with somebody else entails that the second person also does it, e.g. take a walk with Jane means that Jane also takes a walk. While the patient clearly is assigned a passive role, I would argue that the roles of directive, source, path, and location can be considered to be even more passive. If partners are represented not as objects of an act but rather as the direction in which the act progresses or the place in which it takes place, then they are even more marginalized, not having any direct impact on the event whatsoever. Further and finer distinctions could be made: we might, for instance, argue that locations are even farther away from participant roles than paths, sources, and particularly directives, which usually constitute more integral components of events. This is reflected in the very fact that the latter often cannot be omitted, e.g. put requires a directive and come a <?page no="265"?> Positioning sexual partners 265 directive and/ or a source. And it also shows in constructions where directives, sources or paths occur together with locations, but the former are positioned closer to the verb: He fucked into her on the sofa *He fucked on the sofa into her All these things considered, I am left with the following agentivity-scale (> = ‘more agentive than’): agent > comitative > patient > directive/ path/ source > location An additional complication is introduced to this scale by the fact that all roles may be taken by phrases that denote the body or a body part of one of the sexual partners. Such a fragmented representation reduces the level of involvement of the whole person: so if it is a vagina that is fucked, then the woman’s participation is further minimized and she will thus be interpreted as even more passive. And if it is a penis that is doing the fucking, then the man’s active role is also mitigated. Taking fragmentation into account, the modified agentivity-scale looks as follows: Agent (holistic > fragmented) Comitative (holistic > fragmented) Directive/ path/ source (holistic > fragmented) Location (holistic > fragmented) Table 1: Agentivity scale, with decreasing agentivity from top down. The status of the categories on the lower end is certainly open to debate as it is difficult to assess whether a fragmented directive will be perceived as less passive than a whole location. Since some of these roles are very rare, such questions do not need to cause us too much concern. One problem of the scale is that its design suggests that there are equally large steps in decreasing agentivity or, conversely, in increasing passivity from top downwards. This would imply that the difference in passivity between location and directive is as wide as that between directive and comitative. But the former is probably much smaller than the latter, which means that the scale might convey a distorted view of the actual relations. The triple line between comitative and directive/ path/ source is supposed to indicate that this is the primary dividing line. Simplifying matters - and this will be necessary in the large-scale comparisons in the actual analysis - I can posit the top two semantic roles to be agentizing, and the lower two to be passivizing. Decreasing agentivity <?page no="266"?> Analyses 266 Postulating that the scale represents to what extent semantic roles (in combination with holistic or fragmented reference) may be responsible for passivization presupposes that all verb constellations can theoretically contain all of the roles mentioned. This, however, is not the case. Some verbs for sexual, physical or social contact do not denote any kind of movement, e.g. meet. They can therefore, for instance, not take directives, paths, or sources. What is more, however, some verbs do not take patients, e.g. sleep or make love. Sleep (in the sense of intercourse) necessarily takes a with-PP and thus a comitative role, make love (usually) a to-PP and thus a directive. The question now is whether the significance of these roles for the scale is as large as with other verbs. In other words, does make love to XY passivize XY more than fuck XY or sleep with XY? I assume that the fact that the two prepositions to and with are strongly conditioned by the verb so that they are almost part of it reduces the circumstantial role and almost makes the NP inside the PP an object and (perhaps) consequently a patient. This also becomes evident in the fact that the prepositions (in particular the to in make love) have almost lost their original meanings in such constructions. This is one of the reasons why making too fine a distinction between roles is problematic in the analysis. The relevance of the scale for the study of passivization of women and men seems obvious. If females occur repeatedly in positions assigned the more passive roles, i.e. patients, directives, sources, paths, and locations, but rarely in any associated with active roles, i.e. with agents and, to a lesser extent, with comitatives, then this division of role allocation will become salient and discourse recipients will interpret women as passive (sexual) participants. If women, however, are more often assigned active than passive roles, they will be conceptualized as capable of initiating sexual and other activity and of progressing according to their own desires and goals. The same, of course, also holds for men. The task is thus to find out about the frequencies of semantic roles for females and males in verb constellations. Translating my hypotheses into linguistic structures, I make the following predictions with respect to the structures to be examined: • Women will occur more often in roles ranked at the lower end of the agentivity scale than men in the pornography corpus. • There will be no significant gender bias in the erotica corpus or the gap from women to men will not be as wide as in the pornography corpus. Assuming that sexualization will make passivization more acute, I also expect all these tendencies to be more pronounced the ‘more sexual’ the <?page no="267"?> Positioning sexual partners 267 meaning of the verb, e.g. they will be more clearly visible in verbs denoting intercourse than in verbs denoting social activities, and also in the verbs with a sex-exclusive focus than in the verbs which take a more comprehensive view of sexual activity. 1.1 Technical and practical aspects of the analysis An exhaustive analysis of all verbs for intercourse, physical contact and social encounters - and it is these I will be concentrating on - is impossible, firstly because they do not share any formal aspects and can therefore not be tracked by corpus methods, and secondly because taking apart all the relevant verb constellations is beyond an individual linguist’s capacity. I will therefore confine myself to the following fourteen verbs. Sexual Physical Social make love have sex sleep with fuck shag screw (120 vs. 54) (53 vs. 10) (13 vs. 9) (951 vs. 82) (16 vs. 0) (76 vs. 0) kiss touch stroke caress fondle hug (590 vs. 280) (241 vs. 169) (204 vs. 80) (181 vs. 30) (93 vs. 10) (42 vs. 27) talk meet (277 vs. 153) (239 vs. 121) Table 2: Verbs whose constellations are analysed with respect to passivization (with frequencies in the pornography corpus and the erotica corpus). The choices have been motivated by semantic criteria and/ or frequencies (even though the select set itself features partly very wide gaps). I have opted for expressions that seem typical representatives of their categories (move or press, for instance, do not typically occur as sexual or physical verbs even though there are constructions such as to move further into somebody or to press one’s body against somebody, where they do have sexual meanings), without, however, being too specific in meaning either (as, e.g., tickle, lick, suck would be). In the sexual category, there is another factor relevant for my choice. There appears to be a division between verbs which primarily highlight the physical dimension of sex, presenting intercourse as a highly energetic material process (e.g. fuck or shag), and verbs which conceive of sexuality as encompassing physical, emotional and social dimensions of human beings (e.g. make love or sleep with). There probably is more to the meaning differences between these two groups, e.g. that the former group has strong intertextual associations with pornography and similar discourses - therefore they <?page no="268"?> Analyses 268 are often considered to be ‘slang’, notwithstanding the vagueness with which this notion is normally used - while the latter are the unmarked choices in everyday language about sexuality. But the main difference, in my view, remains that the ‘harder’ expressions (to borrow the first element of the term hard-core) are more exclusively sex-focused in meaning, whereas the ‘softer’ ones add semantic nuances that could mitigate the purely sexual. I therefore expect the latter to occur in different - presumably more reciprocal and less passivizing - constructions than the more ‘slangy’ lexemes. To take this distinction into consideration, I have included three expressions from the former (fuck, screw, shag) and three from the latter category (make love, sleep with, have sex). Interestingly, the former are much more frequent in pornography than in erotica, as opposed to the latter. Applying the above criteria (frequencies, typicality and nonspecificity) to the category of social encounters, we soon realize that the number of verbs fulfilling these requirements is limited. I therefore have decided to include just two, a choice also justified by the fact that the primary focus of pornography is sexuality and physical contact. Let me just add that I have dropped an obvious choice, viz. the verb marry, because with its three relatively distinct structures (to marry, to get married, to be married (as a state)) it poses some insurmountable problems for the kind of analysis proposed above. The goal of the research, as has been pointed out, is to find out how often the fourteen verbs from above occur in which constellations and how often they assign the different semantic roles to women and men. For this purpose, a detailed description of the verb constellations is required. The necessary corpus searches are relatively easy to carry out. The search word is simply the verb in its various forms. With the help of the tags I can also exclude instances where the same form occurs in a non-verb class, e.g. he was a real fucking machine, we missed the fucking plane, applying fervent kisses to its velvet tip, after a long talk with Laura, or the first sweet touch was quite magical. Searches based on these principles produce concordances such as the following one for fuck. <?page no="269"?> Positioning sexual partners 269 <VVD> .<.> I<PPIS1> fucked<VVD> up<RP> with<IW> one<MC1> was<VBDZ> just<RR> fucking<VVG> the<AT> hottest<JJT> br > had<VHD> just<RR> fucked<VVN> her<PPHO1> ,<,> and<CC> would<VM> keep<VVI> fucking<VVG> her<PPHO1> .<.> You<PPY I<PPIS1> kept<VVD> fucking<VVG> her<PPHO1> ,<,> and<CC> was<VBDZ> like<II> fucking<VVG> a<AT1> tub<NN1> of<IO> let<VVI> me<PPIO1> fuck<VVI> her<PPHO1> ,<,> instead<RR let<VVI> me<PPIO1> fuck<VVI> him<PPHO1> ,<,> too<RR> ,< ve<VHI> minded<VVN> fucking<VVG> her<PPHO1> myself<PPX1> > could<VM> n’t<XX> fuck<VVI> again<RT> because<CS> I<PP > does<VDZ> n’t<XX> fuck<VVI> me<PPIO1> are<VBR> the<AT> Concordance 1: The verb fuck in the pornography corpus (extract). I go through the individual concordances line by line and mark each occurrence of a particular structure in a matrix such as the one below. If the verb is used in another sense (in a non-sexual sense in the case of fuck, for instance), the example is not considered. to fuck they fuck her body fucks his body they [ ] fuck her body fucks him they [ ] fuck his body fucks her they fuck each other her body fucks her they fuck with each other his body fucks him he fucks her she fucks she fucks him he fucks he fucks him she fucks into: his body she fucks her he fucks into: her body he fucks her body she fucks along: his body she fucks his body he fucks along: her body she fucks her body she fucks on: his body he fucks his body he fucks on: her body she fucks with him she fucks her own body into: his body he fucks with her he fucks his own body into: her body she fucks with her she fucks her own body he fucks with him he fucks his own body his body fucks her body he fucks himself her body fucks her body she fucks herself his body fucks his body Unidentifiable Table 3: Matrix for counting the different verb constellations and semantic roles (for the verb fuck). The matrix contains the possible constellations in schematic form (e.g. she fucks his body represents an example such as Mary fucked his prick). The relevant figures concerning semantic roles, which are of primary interest in the current chapter, can without too much effort be derived <?page no="270"?> Analyses 270 from the matrix, which is so detailed because it was originally intended to account for a broader range of aspects. As far as the identification of genders is concerned, I will proceed along similar lines as in the chapter on desubjectification, which means I will be almost as inclusive as there (but see below for ellipsis). Most of the constellations found in the corpora fall into the clear patterns included in the agentivity scale. There are, however, a few structures that pose problems to an analysis operating on the basis of the scale. In the following, I will discuss how I handle those that gain some prominence in the corpus due to their relatively high frequencies. The first complication for the agentivity scale, particularly regarding the quantification of semantic roles, are constructions in English with two or more people (or any other entities) represented as agents and patients (or whatever the second role) at the same time. The first of these constructions uses a reciprocal pronoun (each other or one another) together with a plural subject, as in the example below: I saw her and Nan embracing each other. The second construction, which, though not always possible, works with the majority of verbs under analysis in this chapter, just uses a plural subject, e.g.: I stifled her groan as we kissed passionately. We fucked like this for a long time. […] but it was only when we met and started talking, that I knew I liked you well enough. […] and we made love all through the night. I call this construction reciprocal ergative. Ergative verbs are verbs that allow transitive and intransitive constructions with the object of the former and the subject of the latter fulfilling the same semantic role, e.g. I break the glass vs. the glass breaks (cf. Halliday 1994: 163f.; he speaks of ergative relationships between the transitive and the intransitive constructions rather than of ergative verbs). The verbs in the examples above are ergative because they allow transitive and intransitive constructions with a change in the semantic role of the subject. The intransitive version is, however, exceptional because the aforementioned change is only partial as the people (or things) in the subject NP actually fulfil both the agent and the patient role in a reciprocal combination at the same time. We fucked thus melts I fucked you and You fucked me into one construction. It is the perfect representation of reciprocity as the two roles are quasi merged into one, unlike the construction with reciprocal <?page no="271"?> Positioning sexual partners 271 pronouns, where the two roles are at least formally represented separately. Both reciprocal ergatives and constructions with reciprocal pronouns pose some serious problems to counting. Active and passive roles converge to a certain extent in the two structures. So analysing, for instance, they kiss (in a dual heterosexual relation) into one female agent and one male agent would ignore the fact that both are patients, too, i.e. that they also ‘undergo’ kissing. The solution should therefore be to also have one female patient and one male patient. This, however, would unduly overrepresent the two constructions mentioned since one occurrence would assign four different semantic roles (two agents and two patients). I have therefore decided to divide the numbers of semantic roles in reciprocal ergatives and verb constellations involving reciprocal pronouns by two. (In case all participants belong to the same gender, e.g. the girls kissed, one agent and one patient are counted for the respective class.) A second case difficult to quantify are structures with mixedgender or gender-neutral (the latter are very rare) participants, e.g. Jenny and Peter/ people were talking to him. Despite the risk of falling prey to inconsistency, I still count the roles as full for both genders, since I think that there is more emphasis on the individual roles assigned to the different genders than in reciprocal constructions. This strategy also marks a contrast to the one applied in the analysis of desubjectification in the preceding chapter, where I omitted these cases from consideration because they diffused the construction of a gendered perspective. But I suppose that in my research on passivization, the mere appearance of a semantic role adds to the overall amount of salience of this role, even if it applies equally to both genders. Let me just add that difficulties such as this one indicate more general problems with quantification, pointing to the necessity to study such issues and the effects of different solutions on the results more thoroughly. In the meantime, I have to make a decision, however unsatisfying it may appear. A third problematic constellation are reflexive processes, where people manipulate themselves (expressed by a reflexive pronoun) or their own bodies. She touched herself, hands roaming around that hard body. I stroked my own tits […] In my view, autoerotic processes are sufficiently distinct from interpersonal sexual, physical or social processes. They cannot meaningfully be claimed to involve a distinction of active and passive, considering that the same person is assigned both the agent and the patient role. I have <?page no="272"?> Analyses 272 therefore decided to leave them out of consideration as far as agentivity and passivity are concerned. Fourthly, there are some verbs which allow a constellation where one partner and her or his body occurs twice, in two different roles. Some of these represent a mixture of reflexive and ordinary constructions, as in the examples below. When he touched his fingers to my clit, I came […] […] he began to fuck his prick steadily into my hot, itching pussy. The agent’s body takes the patient role (his fingers, his prick) and the partner is represented in a directive role (to my clit, into my hot, itching pussy) (theoretically, the patient role could also be assigned to a reflexive pronoun, e.g. He ground himself against me. But none of the verbs included occurs in this construction.). Although I assume that the intervening patient moves the second partner even farther away from the agent, thus making her or him seem even more passive, I will still count them as directive participants. Furthermore, I will not consider the patient roles as it is taken by the same person or a part of her or him that also occurs in the agent position. The following pattern, where one partner appears as patient and as locative (usually a location or directive - the distinction is not easy to make), presents a similar problem. […] as I kissed him full on the mouth. As in the construction discussed above, the second semantic role is not taken into account for the passivity scale, as it would not add to the passive position created by the patient role. Fifthly, many of the verbs under consideration feature in elliptical constellations where just the agent is mentioned, with other concrete or indefinite persons - whether they occur as patients or other passive roles - being implied but not explicitly represented. Ellipsis is mainly used when the agent’s sexual performance or general sexual skills are emphasized or if the frequencies of sexual encounters are foregrounded, with the second sexual partner being irrelevant. He growled and fucked even harder […] he fucked with such a fury. […] he screwed quickly. She had made love a couple of times in the year since Colin had left. <?page no="273"?> Positioning sexual partners 273 The verb talk also regularly occurs in constellations with sayer and verbiage, but without a receiver and thus a second human being. […] the parking garage she was talking about […] I have decided to ignore the patient role, but include agents and sayers nevertheless, which is justified considering that the constructions completely background the former while particularly emphasizing the latter. As a final source of problems I would like to mention agentless constructions in the passive voice. When she’s getting fucked the way she likes. My pussy had never been touched. Although agentless passive constitute the opposite of the kind of ellipsis described above - in the latter the patient is missing but implied, in the former the agent is missing but implied - I have decided to treat them similarly. I will not count the agent role for the agentivity scale because if it is not mentioned explicitly, it is presented as irrelevant (the only exception are imperatives, which are not really elliptical to me, which means that I will establish the gender of the second person addressed). Even if a gender could be inferred from the context - if a heterosexual man talks of being fucked, we can assume that the agent is supposed to be female - the agent role is de-emphasized and I therefore feel justified not to consider it. All fourteen verbs occur in negations or counterfactual constructions. As in the previous chapter, I include such instances because the factuality of the activity is of less relevance than the positioning of women and men relative to the event. 1.2 Presenting and discussing results To gain a general view, I will first look at the overall frequencies of the different roles, comparing both total percentages (assigning more weight to the more frequent verbs) and mean percentages (assigning equal weight to all verbs), calculated in the same way as in the chapter “Desubjectification” (cf. VII.1.2) for women and men. A value of 50% for women, for instance, means half of the semantic roles that women are assigned in the constellations examined belong to the respective category (overall or on average). Although the agentivity scale allows for finer distinctions between different roles, the results from above suggest that the focus should be on the agent/ patient opposition. As the <?page no="274"?> Analyses 274 different locative roles are relatively rare I subsume them under one heading. Roles with fragmentation are included in the respective categories, but I have added a separate column showing how often roles are fragmented in general. Table 4 contains all the relevant figures. Agent Com. Patient Loc. Frag. Absolute frequencies 1,224 35 1,770.5 115.5 708 Total % 38.9% 1.1% 56.3% 3.7% 22.5% Mean % 45.9% 5.4% 44.1% 4.7% 20.1% Absolute frequencies 1,645 42 741.5 51.5 393 Total % 66.3% 1.7% 29.9% 2.1% 15.8% Mean % 58.4% 8.6% 29.3% 3.6% 18.9% Table 4: Absolute and relative frequencies of different semantic roles assigned to women and men by the fourteen verbs in the pornography corpus (Com. = Comitative, Loc. = Locative, Frag. = Fragmented). Although the two types of percentages partly differ considerably, the overall relations remain the same. Men are twice as likely to take the agent role than the patient role, while women are equally likely to occur in either role (if we look at the mean percentages) or may even be more likely (if we consider the total percentages) to appear as patients. These figures are in full agreement with my predictions, suggesting that the hypothesis that women will be conceptualized as more passive than men in the discourse of pornography seems well-founded. This is further corroborated by the fact that women also appear more frequently in the even more passivizing locative roles and by the higher amount of fragmentation of females, a feature also contributing to conceptions of passivity. I will now look at the results for the individual verbs in order to find out how these aspects of passivizing and agentizing distribute across the different words and whether my prediction that the sexual nature of the events denoted by the verbs should have an impact on the extent of passivization (for information on the exact distribution of patterns for each verb, go to my homepage). Table 5 contains the absolute and relative frequencies of agentizing (agents and comitatives) and passivizing (all others) semantic roles assigned to women and men by the fourteen verbs under analysis in the pornography corpus. Figure 1, immediately below the table, represents the relations visually. <?page no="275"?> Positioning sexual partners 275 make love have sex sleep with % % % % % % Agentizing 64 53.8 55 63.2 50 100 42 100 12 100 12 100 Passivizing 55 46.2 32 36.8 0 0 0 0 0 0 0 0 Totals 119 87 50 42 12 12 shag fuck screw % % % % % % Agentizing 1.5 12.5 10.5 87.5 152 16.7 710 86.4 20.5 29.7 47.5 70.9 Passivizing 10.5 87.5 1.5 12.5 756 83.3 112 13.6 48.5 70.3 19.5 29.1 Totals 12 12 908 822 69 67 kiss touch stroke % % % % % % Agentizing 317 44.6 262 59.3 98 41.4 107 61.8 99.5 55.9 63.5 46.4 Passivizing 393 55.4 180 40.7 139 58.6 66 38.2 78.5 44.1 73.5 53.6 Totals 710 442 237 173 178 137 caress fondle hug % % % % % % Agentizing 79 43.4 80 60.2 52 50.0 36 50.7 26 57.8 17 41.5 Passivizing 103 56.6 53 39.8 52 50.0 35 49.3 19 42.2 24 58.5 Totals 182 133 104 71 45 41 talk meet % % % % Agentizing 151 61.4 138 62.7 136.5 50.0 106.5 48.2 Passivizing 95 38.6 82 37.3 136.5 50.0 114.5 51.8 Totals 246 220 273 221 Table 5: Absolute and relative frequencies of agentizing and passivizing semantic roles assigned to women and men by the fourteen verbs in the pornography corpus (the figures in the columns under % are percentages). <?page no="276"?> Analyses 276 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% W M W M W M W M W M W M m ake lo ve h a ve sex sleep with sha g fuck screw 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% W M W M W M W M W M W M W M W M kiss touch stroke caress fondle hug talk meet Figure 1: Proportions of agentizing and passivizing roles assigned to women and men by the fourteen verbs in the pornography corpus (W = ; / M = ; ; dark colours at bottom → agentizing roles, light colours at top → passivizing roles). We find the most dramatic differences between women and men in the three ‘slangy’ sex verbs, with females taking passive roles much more frequently than males (e.g. 83.3% vs. 13.6% for fuck), who, conversely, feature significantly more often as agents (e.g. 86.4% vs. 16.7% for fuck). This trend is not effective in the remaining three verbs for intercourse, with make love seeing females only slightly more often in passive roles and with have sex and sleep with finding neither gender in any passive role. These results fully correspond to what I predicted, suggesting that the exclusion of social and emotional sides of sex (in shag, fuck and screw) or their incorporation (in make love, sleep with and have sex) has an impact on whether intercourse will be primarily conceptualized as a unidirectional process starting in men and ending in women or as a multi-directional one where all participants can perform <?page no="277"?> Positioning sexual partners 277 any role. Taking into consideration the extreme quantitative advantage of fuck, we might, however, conclude that the conception with women in the passive position will take precedence in pornography users’ thinking of sex. Three of the six verbs of physical contact show higher percentages of male active roles and female passive roles, viz. touch, kiss, and caress. The gaps are not as extreme as with the stronger sexual verbs, but they are still larger than initially expected. The reason for this could be that the processes that these verbs denote are not understood as independent and general activities in their own right, but rather as integral parts of sexual activity, with a special focus on their relation to particular body parts, which also becomes manifest in the high proportion of fragmented participants (the three verbs regularly occur with a body part, e.g. he touched her breasts, I kissed his cockhead). These results may lead us to conclude that physical contact, too, will partly be interpreted as being performed by men upon women. The verbs hug and stroke, however, do not fit into this pattern as women are assigned agentizing roles and men passivizing ones more frequently. Of the six physical contact verbs hug is probably the one that is least firmly integrated into sexuality. With its greater distance to intercourse, its passivizing potential seems also reduced. I find it, however, difficult to explain the figures for stroke. The relatively high proportion of female agents may be the result of the verb describing what the hand does to the male physique, possibly better than fondle, touch or caress. As was predicted, the two social activity verbs meet and talk do not display any significant differences in semantic role attribution between women and men, suggesting that passivization is a tendency primarily affecting the conceptualization of sexuality. 1.3 Comparing and discussing results To examine whether there are consistent patterns of passivization, I will look at semantic roles that the fourteen verbs assign to women and men in the erotica corpus, always in comparison to the results gained in the preceding section for the pornography corpus. The table below contains the absolute and relative frequencies (total and mean percentages) of the different roles. <?page no="278"?> Analyses 278 Agent Com. Patient Loc. Frag. Absolute frequencies 545.5 33.5 507 7 211 Total % 49.9% 3.1% 46.4% 0.6% 19.3% Mean % 53.5% 4.1% 36.4% 6.0% 16.0% Absolute frequencies 422.5 22.5 325 10 113 Total % 54.2% 2.9% 41.7% 1.3% 14.5% Mean % 52.0% 4.3% 34.9% 8.9% 17.3% Table 6: Absolute and relative frequencies of different semantic roles assigned to women and men by the fourteen verbs in the erotica corpus (Com. = Comitative, Loc. = Locative, Frag. = Fragmented). Ag. Com. Pat. Loc. Frag. P E P E P E P E P E Total % 38.9 49.9 1.1 3.1 56.3 46.4 3.7 0.6 22.5 19.3 Mean % 45.9 53.5 5.4 4.1 44.1 36.4 4.7 6.0 20.1 16.0 Total % 66.3 54.2 1.7 2.9 29.9 41.7 2.1 1.3 15.8 14.5 Mean % 58.4 52.0 8.6 4.3 29.3 34.9 3.6 8.9 18.9 17.3 Table 7: Comparison of relative frequencies of different semantic roles assigned to women and men by the fourteen verbs between the pornography corpus and the erotica corpus (Ag. = Agent, Com. = Comitative, Pat. = Patient, Loc. = Locative, Frag. = Fragmented, P = Pornography, E = Erotica; all figures are percentages). The total and mean percentages differ considerably, probably as a consequence of the partly dramatic gaps in frequencies between the verbs analysed. But as predicted, women are not as consistently relegated to the patient role in the erotica corpus as in the pornography corpus, with a total percentage of 46.4% (as opposed to 56.3% in pornography) and a mean percentage of 36.4% (as opposed to 44.1%), and they are significantly more likely to take the agent role, with a total percentage of 49.9% (as opposed to 38.9% in pornography) and a mean percentage of 53.5% (as opposed to 45.9%). The percentages of women and men are also fairly similar (at least if we consider both total and mean percentages in combination). We only see differences in the frequencies of female and male patients, but they are not at all as wide as in the pornography corpus. All this lends support to my hypothesis that women’s passivization will be a much weaker force in erotica than in pornography. It is now interesting to see how strongly this tendency is reflected in the results for the individual verbs and also whether my subhypothesis <?page no="279"?> Positioning sexual partners 279 concerning the connections between sexual meaning and passivization, can be upheld if tested against the data. To answer these questions, I will look at the role patterns for women and men for each of the fourteen verbs (percentages of specific semantic roles are presented on my homepage). Table 8 presents all the relevant figures, visually supported by a diagram. Table 9 further down compares these results to those from the pornography corpus. make love have sex sleep with % % % % % % Agentizing 30.5 56.5 21.5 50.0 6 100 7 100 10 100 8 100 Passivizing 23.5 43.5 21.5 50.0 0 - 0 - 0 - 0 - Totals 54 43 6 7 10 8 shag fuck screw % % % % % % Agentizing 0 - 0 - 29 42.6 47 62.7 0 - 0 - Passivizing 0 - 0 - 39 57.4 28 37.3 0 - 0 - Totals 0 0 68 75 0 0 kiss touch stroke % % % % % % Agentizing 149.5 43.3 127.5 64.7 86.5 46.3 61.5 59.1 41.5 47.2 34.5 60.5 Passivizing 195.5 56.7 69.5 35.3 100.5 53.7 42.5 40.9 46.5 52.8 22.5 39.5 Totals 345 197 187 104 88 57 caress fondle hug % % % % % % Agentizing 7 31.8 16 76.2 5 71.4 2 28.6 15 50.0 9 47.4 Passivizing 15 68.2 5 23.8 2 28.6 5 71.4 15 50.0 10 52.6 Totals 22 21 7 7 30 19 talk meet % % % % Agentizing 94.5 66.5 58.5 43.0 78 58.2 40 37.7 Passivizing 47.5 33.5 77.5 57.0 56 41.8 66 62.3 Totals 142 136 134 106 Table 8: Absolute and relative frequencies of agentizing and passivizing semantic roles assigned to women and men by the fourteen verbs in the erotica corpus (all figures are percentages). <?page no="280"?> Analyses 280 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% W M W M W M W M W M W M make love have sex sleep with shag fuck screw 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% W M W M W M W M W M W M W M W M kiss touch stroke caress fondle hug talk meet Figure 2: Proportions of agentizing and passivizing roles assigned to women and men by the fourteen verbs in the erotica corpus (W = ; / M = ; ; darker colours at bottom → agentizing roles, lighter colours at top → passivizing roles). <?page no="281"?> Positioning sexual partners 281 Ag. Pass. Ag. Pass. Ag. Pass. Ag. Pass. make P 53.8 46.2 63.2 36.8 kiss P 44.6 55.4 59.3 40.7 love E 56.5 43.5 50.0 50.0 E 43.3 56.7 64.7 35.3 have P 100 0 100 0 touch P 41.4 58.6 61.8 38.2 sex E 100 0 100 0 E 46.3 53.7 59.1 40.9 sleep P 100 0 100 0 stroke P 55.9 44.1 46.4 53.6 with E 100 0 100 0 E 47.2 52.8 60.5 39.5 shag P 12.5 87.5 87.5 12.5 caress P 43.4 56.6 60.2 39.8 E 0 0 0 0 E 31.8 68.2 76.2 23.8 fuck P 16.7 83.3 86.4 13.6 fondle P 50.0 50.0 50.7 49.3 E 42.6 57.4 62.7 37.3 E 71.4 28.6 28.6 71.4 screw P 29.7 70.3 70.9 29.1 hug P 57.8 42.2 41.5 58.5 E 0 0 0 0 E 50.0 50.0 47.4 52.6 talk P 61.4 38.6 62.7 37.3 E 66.5 33.5 43.0 57.0 meet P 50.0 50.0 48.2 51.8 E 58.2 41.8 37.7 62.3 Table 9: Comparison of relative frequencies of agentizing and passivizing semantic roles assigned to women and men between the pornography corpus and the erotica corpus (P = Pornography, E = Erotica, Ag. = Agentizing, Pass. = Passivizing; all figures are percentages). Starting with the verbs for intercourse, we immediately see that, given the rare occurrence of sleep with and have sex and the complete absence of shag and screw in the erotica corpus, a full comparison between the percentages is impossible. But a look at the two most frequent ones, make love and fuck, reveals interesting details. Like in the pornography corpus, the exclusive sexual focus of fuck correlates with a preponderance of women occurring in passive roles. But the effect of the sexual content is, as expected, much less severe, as the gap between women and men with respect to passive and active roles is significantly smaller (women occur 42.6% of the times in an active role with fuck in the erotica corpus, as opposed to only 16.7% of the times in the pornography corpus), and the contrast to the patterns found for make love less pronounced (the erotica corpus shows a higher percentage of women in active roles here). This lends weight to the hypothesis that the link between sexualization and passivization of women will not be as close in erotica as in pornography. The difference becomes even more marked if we take into account the quantitative predominance of the verb fuck in the pornography corpus, where it occurs eight times as often as make <?page no="282"?> Analyses 282 love (951 vs. 120), because this means that the unidirectional conception of sexuality with women in passive roles, which the use of fuck propagates, will also become more salient. By contrast, fuck is used just one and a half times as often as make love in the erotica corpus (82 vs. 54), which means that the different conceptualizations of female and male roles in sexuality inherent in them receive almost equal weight. As far as the verbs of physical contact are concerned, the results from the erotica corpus yield a surprising - even if inconsistent - pattern: except for fondle and hug, the verbs in this group show a significant preference for women to appear in passive rather than in active roles, a preference that is partly clearer than in the pornography corpus. I might have expected a smaller gap between pornography and erotica in women’s passivization than for intercourse words. What I did certainly not expect, however, was that there would indeed be higher frequencies of passivizing roles being assigned to women in the erotica. Although this can be interpreted to indicate that in the area of physical contact the hypothesis of a stronger element of female passivization in pornography is not valid, I assume that there is a second reason, namely the interference by perspective and subjectification, which creates some kind of tension and produces heterogeneous data. Focusing on subjectivity entails an emphasis on sensory experience and thus on receiving rather than on acting. Assuming that in its attempt to subjectify and thus de-objectify women, erotica may consequently use women in passive roles with verbs of physical contact. This would then be a case of two anti-objectifying conceptualizations - agentizing and subjectifying - working against each other to produce problematic results for the analysis of passivization. If we progress to the two social verbs, we find a marked advantage of female over male agentivity in the erotica corpus, an advantage which we did not see as distinctly in the pornography corpus (the difference is striking for men with the verb talk - seemingly men are more often talked to and do less of the talking themselves in erotica in contrast to pornography). The data suggests that passivization of women will be a less relevant factor in erotica than in pornography for the area of social encounters. Whether this contrast is less pronounced than with the verbs with a stronger physical and sexual focus, as predicted, is difficult to evaluate because of the inconsistencies and heterogeneity in the data. If we just compare the percentages to those of fuck, as the most common sexual verb, then the prediction is confirmed. But this can only be taken as very limited evidence for the hypothesis that differences in women’s passivization between the two discourses will be smaller for social than for sexual activity. <?page no="283"?> Representing women as empty space and men as tools 283 How can all the results discussed in this subchapter in retrospect be evaluated with respect to my initial predictions? Some but - it has to be admitted - not all of the expectations have been confirmed. With respect to semantic role allocations, the findings show that women appear significantly more frequently in passive roles than men in pornography, while the distribution of roles is almost equal in the erotica corpus. If we look at the link between sexualization and passivization, we do see that the gaps between females and males are very wide for intercourse words (and in particular for the ‘hard’ sex verbs) and shrink or are reversed for social words in the pornography corpus. This trend has also been noted in the erotica corpus, but it is not as pronounced. The only exception are physical contact verbs, which do not show the expected differences to sex words and also behave similarly in both corpora as regards semantic roles for women and men. These results mean that all my hypotheses concerning passivization - women being more passivized than men in pornography but not in erotica, a stronger element of female passivization in more sexually-oriented events in pornography, which does not apply as clearly to erotica - do receive support from the data with the exception of the assumption that women would be conceptualized as less passive in physical contact than in sexuality by both pornography and erotica readers. The integration of physical contact into intercourse and the interference of subjectification may be responsible for the latter deviation. 2 Representing women as empty space and men as tools: conceptual metaphors for genitals The means for passivization are not restricted to semantic roles. Creating associations between persons and ‘active’ or ‘passive’ concepts can also contribute to passivization or work against it. Metaphors are among the most effective instruments in evoking such associations. I will therefore be concerned with the qualitative and quantitative analysis of metaphors for male and female genitalia in this subchapter, assuming that this is one of the most sensitive areas for role differentiation. Generally speaking, a metaphor is the reference to a concept with the help of another concept. The concept to which I refer, i.e. the target, does not belong to the same conceptual domain 5 as the meaning of the 5 Kövecses (2002: 247f.) defines conceptual domain as the knowledge “of any coherent segment of experience.” Roughly speaking, this means it is a specific area of thought, cognitively speaking, or meaning, linguistically speaking. <?page no="284"?> Analyses 284 expression I am using, i.e. the source. The target is associated - by convention or by a spontaneously triggered process - with the source by virtue of semantic correspondences resting on relations of similarity. If we use the source stick, for instance, to mean ‘penis’, the association is formed as a result of the similarities in shape between sticks and erect penises and may extend to the functions of the stick, e.g. ‘sticking’ it somewhere or beating somebody with it. The approach to the study of metaphors most fertile for a Critical Discourse Analysis is the theory of conceptual metaphors (= CM) as initially formulated by George Lakoff and Mark Johnson in their groundbreaking book Metaphors We Live By (1980) (cf. also Lakoff 1987, Goatly 1997, Kövecses 2002). It has progressed from focusing on the originality and aesthetic value of metaphors as ornamental devices, as has always been common in literary studies, to regarding them as primarily cognitive - and only indirectly linguistic and literary - phenomena. The theory of conceptual metaphor rests on the basic assumption that metaphors fundamentally mean thinking of, and thus structuring, one concept in terms of another. Metaphorical linguistic expressions such as stick thus only realize underlying conceptual metaphors, e.g. THE MALE GENITALS ARE LONG HARD OBJECTS (I adopt the now common practice of using small caps for conceptual metaphors) (cf. Kövecses 2002: 4). It is widely acknowledged in the field that conceptual metaphors do not just constitute single and isolated connections between target and source concepts but involve whole conceptual domains being associated with each other (e.g. the penis with particular types of objects). These usually surface in an ever-growing number of individual metaphorical linguistic expressions (not only stick, but also staff, pole, rod, etc.). The theory of conceptual metaphors is interested in these expressions as far as they point to underlying connections of domains rather than in their originality, thus focusing on commonness and patterns rather than on uniqueness and creativity. Another traditional trope, viz. metonymy, is often subsumed under the discussion of conceptual metaphors (cf. Ungerer/ Schmid 1996: ch. 3, Kövecses 2002: ch. 11). The difference to metaphors is that in metonymies, the source and the target do not belong to different conceptual domains but are associated with each other on the basis of relationships of neighbourhood and/ or causality, e.g. Shakespeare can be used metonymically for his works, love for intercourse, etc. Metonymy is also primarily a mode of thinking and we can thus also postulate the existence of conceptual metonymies, i.e. consistent patterns of associations between different aspects within conceptual domains, e.g. THE PRO - <?page no="285"?> Representing women as empty space and men as tools 285 DUCT IS ITS PRODUCER (as in the Shakespeare example) or THE EVENT IS AN ACCOMPANYING EVENT / EMOTION (as in the love example). Though conceptual metonymies will not be the main object of investigation in this section, I can nevertheless not afford to ignore them, firstly because they are not always easy to distinguish from conceptual metaphors, and secondly because they sometimes interact with the latter (see below). Even though my discussion, like most accounts of conceptual metaphors, may have given the impression that we are dealing with entities that can be clearly delineated, this is far from the truth. Firstly, we can postulate conceptual metaphors for different levels of specificity. Stick for ‘penis’ thus realizes THE MALE GENITALS ARE LONG HARD OB - JECTS , but it can also be argued to point to the more general THE MALE GENITALS ARE OBJECTS . Secondly, there are compound metaphorical expressions, where different parts of a complex lexeme realize different CMs or conceptual metonymies (which may or may not be easily matched). In love tool, for instance, THE MALE GENITALS ARE TOOLS (= tool) combines with the conceptual metonymy SEX IS AN ACCOM - PANYING EMOTION (= love). Thirdly, and perhaps most interestingly, there are also linguistic expression that seem to manifest two different conceptual metaphors. Stick, for instance, realizes both THE MALE GENITALS ARE LONG HARD OBJECTS and THE MALE GENITALS ARE WEAPONS , two CMs that are not embedded within each other (there are long hard objects that do not qualify as weapons, e.g. tower, and there are weapons that are not long hard objects, e.g. sex bomb). I call such cases portmanteau metaphorical expressions, as one form on the surface realizes two or more underlying elements. Portmanteau metaphorical expressions may also combine conceptual metaphors with conceptual metonymies. To make love thus draws upon the metaphor SEX IS WORK / PRODUCTION and the metonymy SEX IS AN ACCOMPANYING EMOTION . For research in CDA, the aspect of structuring one conceptual domain in terms of another is the most important idea of the theory described above. Structuring in this context means the different weight that is attributed to internal elements of a domain (similar in meaning to conceptualization). With an infinite number of potential areas of semantic correspondence between concepts, the very fact that I use a particular source concept entails the highlighting of specific elements in the target at the cost of others, e.g. the solid, lengthy shape of the erect penis in stick as opposed to its soft, lengthy shape when flaccid, or to its sensitivity. Conceptual metaphors thus allow for the foregrounding of certain aspects and the backgrounding of others in entire conceptual domains (e.g. shape over sensitivity in sexual organs), greatly affecting the <?page no="286"?> Analyses 286 structure of the schema of the domain and schemata contained within it. It seems obvious that structuring in this sense leaves ample space for ideological influences because it changes our perception of what is essential. In addition to the conceptual reorganization, CMs can also affect evaluations and attitudes inherent in schemata, mostly by masking or exaggerating the affective impact of certain concepts. In SEX IS ATTACK - ING , intercourse is not only conceptualized differently, it also inherits some of the negative connotations of the source concept, which might make it seem like a tough act (dysphemism). In SEX IS PLAY , on the other hand, intercourse adopts some of the positive connotations, which might conceal the emotionally problematic status of sexuality in our culture (euphemism) (cf. Goatly 2000: 132, who rightly mentions the hiding or expressing of emotions as one of the main motives for using metaphors). Despite its immediate appeal to CDA, the theory of conceptual metaphors has perhaps been too static, treating the underlying associations as a prioris and linguistic expressions only as their reflexes. In its constructionist orientation, CDA - as propagated in this book - assumes that correspondences between targets and domains are not given but are constantly created in discourse. By using certain metaphorical linguistic expressions I might be able to create, modify or sustain conceptual metaphors of discourse participants, thus affecting their schemata. This places more emphasis on metaphorical language as it actually occurs in discourse (as opposed to introspective collections) and also invites a quantitative approach, presuming that the more different expressions are linked to a particular conceptual metaphor and the more often these occur, the more salient its position will be and the more salient the associations between the two conceptual areas. How does the theory of conceptual metaphor translate into concrete research procedures? Basically, I have to search for all metaphorical linguistic expressions denoting a particular target area, e.g. male and female genitals, in order to find out whether they can be grouped into semantically defined classes, e.g. persons, animals, or valuable objects. If such classes can be established, they will be interpreted as underlying conceptual metaphors. As mentioned above, a quantitative approach is called for as I can assume that if numerous different expressions can be found for one class (types), this shows that the conceptual metaphor is still very productive and that in their striving for creative and innovative expressions language users still fall back onto the predominant but possibly unconscious underlying patterns. Conceptual metaphors featuring great lexical variation probably also have a stronger <?page no="287"?> Representing women as empty space and men as tools 287 effect on recipients’ interpretations, enhancing the salience of the restructuring of one conceptual domain in terms of the other. Although I believe that type frequencies are highly important, token frequencies will also play a part because the more often a certain kind of metaphorical expression is encountered, the more it will increase the effects on understanding and thus, in turn, on thinking. What do conceptual metaphors have to do with passivization? I will examine metaphorical expressions for one target area, namely for, as mentioned, male and female genitals. There are a number of conceptual metaphors that seem to be used frequently, e.g. THE MALE GENITALS ARE PERSONS (John Thomas, Willy, bishop), THE FEMALE GENITALS ARE ANIMALS (pussy, beaver), THE MALE GENITALS ARE ANOTHER PART OF THE BODY (muscle, finger), THE FEMALE GENITALS ARE PLANTS (bush, flower), etc. (cf. Richter 1987, 1993, Cameron 1992b). Some of the source domains can be assessed with respect to agentivity. Persons and animals are thus probably better able to initiate action than plants. The clearest cases, however - and these are the ones I will be focusing on - are THE FEMALE GENITALS ARE EMPTY (or NEGATIVE ) SPACE and THE MALE GENITALS ARE TOOLS . Negative space metaphors, realized in expressions such as hole, tunnel, box, or passage, imply a strong and obvious element of female passivization. Empty space is defined by the absence rather than the presence of matter. Such concepts, as a result of their ‘emptiness’, automatically invoke the association of filling them with something, an idea that bears a strong element of passivity. The use of the negative space metaphor for the vagina will thus contribute to a passivizing conceptualization of the female genitals and thus of females in general. THE MALE GENITALS ARE TOOLS , as realized in expressions such as tool, hammer, or sword, on the other hand, represents the most apparent example of a CM agentizing men. Tools are objects that are used by persons to perform an act. They may thus not be active by themselves, but they assign agentivity to their users, in the case of the penis this means to men (for a more differentiated view of the concept of tools, see below). If the two conceptual metaphors just mentioned are indeed the predominant forms of representing female and male genitals respectively in pornography, then this will no doubt contribute to the passivization of women and to the agentizing of men in readers’ minds. My task in this section is to find out how quantitatively prominent - in both types and tokens - the tool metaphor is for male genitals and the empty space metaphor for female genitals. I will, however, not concentrate on these two CMs exclusively but will also examine other <?page no="288"?> Analyses 288 kinds of metaphors and metonymies that can be argued to mitigate or enhance passivizing tendencies. Person and animal metaphors, for instance, as indicated, make the ‘owners’ of the genitals seem active, while food or clothes as source areas imply that they are rather the passive objects of the acts the partner performs upon them (‘eat’ them or ‘wear’ them, respectively). Toys and musical instruments, too, are passivizing rather than agentizing, since they usually imply that the genitals are not ‘played’ by the owner but by the partner, thus rendering the former passive rather than active. Metaphors and metonymies that define the genitals as location (geography/ geology, buildings, and area metonymies) may also have a passivizing effect. As already discussed in connection with locative semantic roles in the first subchapter, locations only say where an event takes place but cannot be seen as active participants in it. So if the sexual organs are conceptualized as places, then this further passivizes the owner. Metonymies which use accompanying emotions (as in words such as love button) may also increase the impression of passivity by adding a gentle, affective and thus potentially harmless element. There are also some distinctions within the tool and empty space metaphors relevant to the evaluation of their passivizing or agentizing impact. The most important tool subcategory is THE MALE GENITALS ARE WEAPONS , e.g. sword, lance for ‘penis’ (alternatively, man trap for THE FEMALE GENITALS ARE WEAPONS ) because it additionally realizes another conceptual metaphor, namely SEX IS VIOLENCE . 6 This means that sexuality is not only conceived of as work performed by the active man with the help of his ‘tool’, but also as an act of violence, hurting and harming the woman. The aspect of violence thus enhances the effects of agentization (and passivization by implication) even further by adding power and subordination to the distinction between active and passive roles. G ENITALS ARE PARTS OF MACHINES , on the other hand, has the opposing effect, suggesting that the ‘owner’ of the genitals as a whole is a tool or a machine to be used by her or his partner, e.g. the button (= ‘clitoris’) may ‘turn on’ the woman. Among the empty space metaphors, the subcategory of openings, e.g. hole, entrance or slit, is probably strongest in its passivizing effect as a result of the implication that openings invite something to enter through it (this to a certain extent also applies to passages). 6 It may also be relevant to consider further conceptual metaphors for sexuality in general that are brought into play by certain expressions, e.g. SEXUALITY IS WORK or SEXUALITY IS PLAY . <?page no="289"?> Representing women as empty space and men as tools 289 Even though other conceptual metaphors may prove more difficult to assess with regard to the dimension of passivity, I will nevertheless present an exhaustive categorization of the elements found in the corpora in my analyses, drawing upon a full taxonomy of categories and subcategories of conceptual metaphors and metonymies. This taxonomy will be presented in the “Technical and practical aspects of the analysis”section. Resting on the hypotheses concerned with gender differences formulated in the introduction to this chapter, my expectations for this subchapter are: • The type and token frequencies of THE FEMALE GENITALS ARE EMPTY SPACE and THE MALE GENITALS ARE TOOLS metaphors will be the quantitatively predominant categories in the representations of female and male genitals in the pornography corpus. • The same tendencies will not be found to be as prominent in the erotica corpus. I will also look at the quantitative dimensions of further categories and subcategories that can be considered to interact with agentization and passivization. I assume that such categories of metaphors and metonymies will be in line with the above-mentioned tendencies. 2.1 Technical and practical aspects of the analysis In order to examine the salience of the conceptual metaphors described above, I have to find all expressions for the respective target categories in the corpus, i.e. all lexemes for the penis and the vagina. I have already gathered data on male and female genitals in the chapter on fragmentation and I will draw upon these lists. They include just expressions used after possessives and so I miss out on a number of lexemes that happen not to occur in this combination, but I do not see any other systematic way of tracking genital lexemes. I will not limit myself to penises and vaginas, but I have decided to include the full lists of expressions for female and male genitals even though this means including lexemes for parts of the genitals. This implies that I will also examine expressions for the male testicles or the female clitoris, because they are anatomically distinct from the penis and the vagina and the sets of linguistic expressions used for them might relativize the agentive potential of the penis and the passivizing potential of the vagina. <?page no="290"?> Analyses 290 What are now the conceptual metaphors and metonymies that can be expected to figure prominently in the two corpora? The taxonomy to be presented below is based on previous accounts of metaphors of sexuality (most notably Richter 1987, 1993 and Cameron 1992b) and informal explorations of my corpora. Literal: words without figurative meanings (synchronically, this is - some of the words, of course, initially were metaphorical expressions, but with the original meanings being lost, the figurative meaning has developed into the literal one). This includes: • General terms for the whole genitals: e.g. penis, vagina. • Terms denoting particular parts of the genitals: e.g. pubic hair, vaginal muscle. Personification: the source is a person, e.g. John Thomas or Willy for ‘penis’. Animals: the source is an animal, part of an animal, or something clearly associated with animals, e.g. pussy (= cat) for ‘vagina’, monster or serpent’s tail for ‘penis’, nest for ‘vagina’. Plants: the source is a plant or part of a plant (including fruit), e.g. bush for ‘pubic hair’, nuts for ‘testicles’. Body: the source is a part of the body different from the genitals. Body metaphors are not always easy to distinguish from metonymies related to the body. The main difference is that in the latter, the source is part of the target. So if I used swelling blood vessels for the penis, then this would be a case of pars pro toto and thus a typical case of metonymy. Bones and muscles, however, are not part of the penis, so boner or love muscle are categorized as body metaphors. Tools: the source is a tool or a machine or parts of them. Subcategories included in this conceptual metaphor are: • General tools: the source is a tool or a machine - in the general sense of the word - which does not belong to the specific categories below, e.g. tool or love machine for ‘penis’. • Weapons: the source is a tool used to hurt or kill people or is otherwise associated with violence and warfare, e.g. lance, spear for ‘penis’, love trap for ‘vagina’. • Parts of tools: the source is a part of a tool or a machine, e.g. button, love handle. Mind, however, that I include only con- <?page no="291"?> Representing women as empty space and men as tools 291 cepts here that imply that the ‘owner’ of the genitals is the tool or the machine, e.g. button for ‘clitoris’ suggests that the woman as a whole is the machine that is switched on if the button is pressed. Toys: the source is something you can play with. These include toys proper (in the narrow sense of the word), e.g. pogo stick for ‘penis’ (pogo stick is a device with a handle bar and foot pieces attached to a stick. It has an inbuilt spring which allows the user to move around in a jumping fashion), sporting equipment, e.g. balls for testicles, and musical instruments, e.g. flute for ‘penis’. In contrast to the tool metaphors, to which they show some kind of kinship, the ‘player’ is not necessarily the ‘owner’ 7 (e.g. flute rather carries implications of fellatio and thus of a woman ‘playing the flute’), and they also point to SEX IS FUN (more specifically PLAY / SPORTS / MUSIC ) rather than to SEX IS WORK . Valuables: the source is something valuable and precious, e.g. treasure box for ‘vagina’. Secrets and mysteries: the source is something secret and hidden from view (at least for most people), often with implications of mysteriousness, e.g. privates, feminine mysteries. Geography/ geology: the source is a geographical or geological phenomenon, e.g. mound, valley, for various parts of the genitals. Shapes: the source is a concept that is primarily defined by its spatial appearance, though I admit that this is but a fuzzy description of the phenomena covered by this category, e.g. node, triangle, folds. The borderlines to neighbouring categories such as empty space or geography/ geology are not always easy to define. Folds, for instance, could very well also be conceived of as a geographical phenomenon or as negative space (is the fold the groove rather than the protuberant piece? ). Empty space: the source is something defined primarily by the absence of matter. There are four subcategories: • Openings: the empty space is seen as an entrance to some other area, e.g. entrance, slit for ‘vagina’ (as opposed to three dimensional concepts such as cave). 7 This is also often true of the ‘user’ in parts of tools metaphors, i.e. the love handles are used by the partner not by the owner. <?page no="292"?> Analyses 292 • Passages: the empty space implies movement through it, e.g. tunnel, passage for ‘vagina’. • Containers: the empty space is seen as being enclosed by a definable entity, i.e. by a container, e.g. box, conch for ‘vagina’. • Others: sources that do not fall into any of the above, e.g. cave for ‘vagina’. Buildings: the source is a building or a part of a building, e.g. pussy walls. Food: the source is something edible, e.g. lollipop for ‘penis’, honey pot for ‘vagina’. Clothes: the source is something that can be worn, e.g. cock crown. Metonymy: the source is causally or locally connected to the target. The most important subcategories are: • Superordinates: the source concept is more general in meaning, thus including the genitals, e.g. thing, organ for ‘penis’. Superordinates are, strictly speaking, only a marginal case of metonymy. I still assume that they are more similar to the other metonymies than to metaphors. • Area/ position: the source is the place where the genitals are, e.g. down there for the female genitals, sometimes relative to some other areas, e.g. cockend, tip. 8 The expressions falling into this category are often modified by further elements, e.g. sensitive spot. • Pars pro toto: the source is a part of the genitals but represents the whole, e.g. pink flesh for ‘vagina’. • Neighbourhood: the source concept is in the vicinity of the genitals, e.g. groin for both female and male genitals. 8 It might be objected - and rightly so, I admit - that words such as tip or cockend do not represent metonymies and thus figurative meanings at all. However this issue is resolved - and the question of a distinction between literal and figurative meanings is a tough one to make - all expressions in this group highlight the conceptual domain of place and this probably justifies assigning them to the same category. <?page no="293"?> Representing women as empty space and men as tools 293 • Qualities/ states: the source concept is a quality that the genitals have, e.g. 15 inches for ‘penis’, sweetness for ‘vagina’, or a state that they are in, e.g. erection for ‘penis’. • Activities: the source is a thing or person carrying out an activity, e.g. throbber for ‘penis’. • Accompanying emotions: the source is an emotion that accompanies the act. This category is not used to denote genitals, but rather to modify other elements in compound metaphorical expressions, e.g. love rocket, passion pit. Due to the complex nature of some metaphorical expressions, many of the lexemes for genitals will be listed in more than one category. While this is unproblematic for cases of portmanteau expressions such as cave, which is both empty space and a geographical phenomenon, the situation with compound metaphorical expression is more intricate. The question with words such as pussylips or ballbag, for instance, is whether the metaphorical nature of the modifying elements should also be considered. Strictly speaking, these two lexemes do not realize the conceptual metaphors THE GENITALS ARE ANIMALS and THE GENITALS ARE TOYS , respectively because the expressions as wholes do not refer to animals or toys, only the modifying element pussy and ball do. But I have nevertheless decided to include such words also in the categories of the modifying elements because animals and toys do at least play a role in these lexemes. This, however, has to be born in mind when it comes to the interpretation of results. 2.2 Presenting and discussing results A detailed presentation of data can be found on my homepage, allowing readers to get a more profound view of the metaphorization operating in the discourse of pornography. But I will immediately concentrate on the comparison of the sizes of the conceptual metaphors relevant to passivization, i.e. primarily on empty space and tool metaphors. I will start my analysis by looking at lexical variation in the various classes. Table 10 below contains the absolute and relative type frequencies (relative to the overall numbers of lexemes for the female and the male genitals, i.e. 50% means that the respective category takes half of all different lexemes used for the genitals; with multi-category membership possible, the percentages do not add up to 100%). The diagram visualizes the relations between the different categories. <?page no="294"?> Analyses 294 Types % % General 41 13.2% 21 10.0% Wholes 9 2.9% 5 2.4% Parts 32 10.3% 16 7.7% Persons - - 16 7.7% Body 33 10.6% 15 7.2% Animals 29 9.4% 17 8.1% Plants 18 5.8% 11 5.3% Tools 22 7.1% 67 32.1% General 2 0.6% 31 14.8% Weapons 8 2.6% 34 16.3% Parts of tools/ machines 12 3.9% 4 1.9% Toys 1 0.3% 9 4.3% Valuables 10 3.2% 3 1.4% Secrets and mysteries 34 11.0% 2 1.0% Geography/ geology 21 6.8% 1 0.5% Shape 19 6.1% 3 1.4% Empty space 90 29.0% 22 10.5% Openings 36 11.6% 5 2.4% Passages 23 7.4% 7 3.3% Containers 18 5.8% 10 4.8% Others 13 4.2% - - Buildings 13 4.2% 1 0.5% Food 9 2.9% 24 11.5% Clothes 3 1.0% 7 3.3% Metonymies 105 33.9% 53 25.4% Superordinates 9 2.9% 6 2.9% Area/ position 49 15.8% 8 3.8% Pars pro toto 10 3.2% 6 2.9% Neighbourhood 3 1.0% 3 1.4% Qualities/ states 8 2.6% 21 10.0% Activities 3 1.0% 2 1.0% Accompanying emotions 33 10.6% 5 2.4% Others 2 0.6% 2 1.0% Totals 310 209 Table 10: Absolute and relative numbers of different expressions realizing the different conceptual metaphors and metonymies in the pornography corpus. <?page no="295"?> Representing women as empty space and men as tools 295 0% 5% 10% 15% 20% 25% 30% 35% 40% General Persons Body Animals Plants Tools Toys Valuables Secrets and mysteries Geography/ geology Shape Empty space Buildings Food Clothes Metonymies Female Male Figure 3: Relative sizes of the individual conceptual metaphors and metonymies with respect to lexical variation in the pornography corpus. Looking at the results for women, we see that negative space is, as predicted, the metaphorical category showing the greatest amount of lexical variation. The fact that almost a third of all lexemes fall into this class proves that this is an overworded and thus very productive CM, which will create a strong link between female genitals and thus women and empty space in pornography consumers’ minds. This means that the data increases the plausibility of the hypothesis of a strong moment of female passivization in pornography. This impression is enhanced by the fact that the opening expressions, which may have the strongest passivizing potential because of their close association with a person being entered, are the largest subcategory. There are a number of figures in the table above which support this conclusion. Very striking is the fact that two categories concerned with locations (in a wide sense) are densely lexicalized, in particular in comparison to those used for the male genitals. Almost every fifth expression for the female genitals uses an area metonymy - whose vagueness may reflect our culture’s difficulty in defining female sexual organs - and approximately 7% of all expressions draw on a concept from geography and/ or geology (I could add the building category, where women also have a clear advantage over men). As already discussed in connection with locative semantic roles in the first part of this chapter, locations only define where an event can take place but cannot be seen as active participants in it. If women’s sexual organs are <?page no="296"?> Analyses 296 conceptualized as places, then this means that they and consequently also women themselves will be interpreted as passive. Perhaps we can additionally count the secrets and mysteries CM, which contains many lexemes for female genitals, as enhancing if not passivization then the overall amount of objectification since secrets and mysteries are unknown only to the male would-be explorer, who is thus put into the subjective position in the process. Among the conceptual metaphors that can be argued to weaken passivization, THE FEMALE GENITALS ARE ANIMALS shows some lexical variation. It has to be considered, however, that the category owes its high value primarily to the numerous expressions with pussy as the modifying element (e.g. pussylips), which means there is an association with animality, but the whole expression does not refer to an animal and thus not necessarily to an active being. T HE FEMALE GENITALS ARE TOOLS may not appear to be an underlexicalized CM at first sight either, but a closer look reveals that most of the expressions belong to the part of tool/ machine category, i.e. they all are concerned with buttons or knobs (for the clitoris), which represent, as argued above, the passivizing rather than agentizing side of the tool metaphor. And the fact that person metaphors are not used at all for the female sexual organs in the pornography corpus underlines the impression that passivization remains, as postulated, a strong force and is not really reduced by agentizing CMs. It is interesting to note that among the categories that interact with passivization and agentization, the conceptual metonymy using accompanying emotions to refer to the accompanied act shows a particularly great difference in lexical variation between women and men. The emotion words exclusively refer to positive feelings such as love or passion. The overlexicalization for women’s genitals in this category (even though it is doubtful whether overlexicalization is the appropriate term here, given that there is not that much variation in the metonymies themselves, because the emotion word primarily serves as the modifier of another figurative expression, e.g. in love button) adds an element of gentleness and perhaps even harmlessness to female genitals and thus by extension maybe also to women themselves. This might lead to conceptions of women not only as passive, but also as harmless and ‘nice’. As far as conceptual metaphors for men are concerned, THE MALE GENITALS ARE TOOLS is, as expected, similarly predominant with respect to lexical variation as the empty space metaphors for women, claiming about a third of all types. Together with the fact that the ‘passive’ subcategory, i.e. parts of tools/ machines, hardly plays any role at all, this points to a conception of males as the active users of their genitals and thus as the agents in sexuality, which agrees with my hypotheses. <?page no="297"?> Representing women as empty space and men as tools 297 It is striking that almost half of the tool metaphors fall into the subcategory of the CM THE MALE GENITALS ARE WEAPONS . The metaphorical expressions cover traditional devices for stabbing and hitting (the correspondence being a similarity of the movements involved in stabbing and hitting and in intercourse), e.g. arrowhead, baseball bat of flesh, fuck rod, fucking rod, fucking stick, fuckstick, meatstick, prick, prod, prong, ramrod, rod, shaft, spear, spike, staff, stalk, stick, sword, but also firearms or missiles (the correspondence being a similarity between orgasming and exploding), e.g. cum-gun, love rocket, male missile, missile, pocket rocket, sex bomb. These expressions are distinct signs of an underlying conception of males not only as agents but also as aggressive (sexual) agents and thus of a link between sexuality and violence. This additional - socially problematic - dimension of men’s agentization stands in stark contrast to the above-mentioned gentleness and harmlessness incorporated into women’s passivization through the high variation of expressions using emotion words such as love, passion or pleasure for the female genitals. We might also wonder what the effect of the CM SEXUALITY IS ( MECHANICAL ) WORK , which is implied by the remaining tool expressions, is on the overall objectification created by metaphors. T HE MALE GENITALS ARE PERSONS and THE MALE GENITALS ARE ANIMALS are also represented by a number of different lexemes. But the lexical variation is not so high as to suggest that these CMs can substantiate the trend of males’ agentization. As regards opposing tendencies, expressions in the empty space category show some lexical variation, but their type frequency is still significantly lower than that of their female counterparts. The percentage is further relativized by the fact that the subcategory of openings, which particularly strongly highlights the moment of passivity, plays a minor role for men. Container terms, on the other hand, such as ballbag and nut sack also show that empty space is primarily reserved for the testicles and not for the penis, the arguably more important part of the male physique in sexually explicit material. Surprisingly, men are predominant in the food class, also having a passivizing potential, mainly as a result of compounds with meat and variation of sausages as sources. Other conceptual metaphors and metonymies, especially those drawing upon locative concepts, can be ignored due to their low frequencies. I will now turn to the token figures to see whether the relations in lexical variation are also reflected in the frequencies of occurrence. Table 11 contains the relevant data, graphically complemented by a diagram. <?page no="298"?> Analyses 298 Tokens % % General 1,890 40.3% 207 4.7% Wholes 1,206 25.7% 129 2.9% Parts 684 14.6% 78 1.8% Persons - - 484 10.9% Body 232 4.9% 179 4.0% Animals 1,401 29.8% 1,933 43.5% Plants 68 1.4% 43 1.0% Tools 53 1.1% 891 20.0% General 2 0.0% 364 8.2% Weapons 14 0.3% 715 16.1% Parts of tools/ machines 37 0.8% 37 0.8% Toys 1 0.0% 319 7.2% Valuables 11 0.2% 7 0.2% Secrets and mysteries 51 1.1% 2 0.0% Geography/ geology 76 1.6% 4 0.1% Shape 34 0.7% 17 0.4% Empty space 671 14.3% 39 0.9% Openings 490 10.4% 10 0.2% Passages 62 1.3% 12 0.3% Containers 82 1.7% 17 0.4% Others 37 0.8% - - Buildings 33 0.7% 1 0.0% Food 20 0.4% 98 2.2% Clothes 27 0.6% 12 0.3% Metonymies 523 11.1% 409 9.2% Superordinates 10 0.2% 30 0.7% Area/ position 82 1.7% 11 0.2% Pars pro toto 44 0.9% 17 0.4% Neighbourhood 144 3.1% 130 2.9% Qualities/ states 29 0.6% 210 4.7% Activities 101 2.2% 2 0.0% Accompanying emotions 77 1.6% 7 0.2% Others 49 1.0% 2 0.0% TOTALS 4,695 4,447 Table 11: Absolute and relative frequencies of expressions realizing the different conceptual metaphors and metonymies in the pornography corpus. <?page no="299"?> Representing women as empty space and men as tools 299 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% General Persons Body Animals Plants Tools Toys Valuables Secrets and mysteries Geography/ geology Shape Empty space Buildings Food Clothes Metonymies Female Male Figure 4: Relative sizes of the individual conceptual metaphors and metonymies with respect to token frequencies in the pornography corpus. In the female token column, THE FEMALE GENITALS ARE EMPTY SPACE upholds its high position among CMs, second only to the animal metaphor (and also trailing the non-figurative general category, which features many of the high-frequency lexemes), which is in line with my expectations. This conceptual metaphor thus not only gives rise to nonce formations, but has also led to some expressions gaining an established status - this applies in particular to hole and slit. The very fact that there is not just a host of empty space expressions but that these also occur frequently will certainly add to the salience of the conception of female genitals as negative space, which, following my argumentation from above, will result in women’s passivization in the porn consumers’ thinking, as hypothesized. This conclusion is further underscored by the fact that the most effectively passivizing subcategory - the opening metaphor - predominates. None of the other passivizing metaphors features prominently in the results above so that we do not obtain further support for the hypothesis from them. Among the conceptual metaphors that can be argued to weaken the effect of passivization, there is, however, one that almost tops the tables (only non-figurative expressions occur more often), namely, as already mentioned, THE FEMALE GENITALS ARE ANIMALS . This fact might mitigate the passivization created by the empty space metaphor. But we have to bear in mind that the sole reason for this first position is the high frequency of the word pussy (1,246 occurrences). As the association probably mainly works via the similarity of a cat’s fur with a <?page no="300"?> Analyses 300 woman’s pubic hair rather than via other, more agent-related aspects, and as cats are pets and thus linked to domesticity and control (for different correspondences in the animal-for-woman metaphors, cf. Whaley/ Antonelli 1983/ 84), the relativizing impact may, however, not be as strong as the proportion might suggest. Besides, there are also the abovementioned compounds where pussy only functions as a modifier (e.g. pussy walls, pussy muscles). As regards men’s CMs, the tool metaphor is, as I predicted, also ranked very high, just trailing animals, with about a fifth of all expressions. T HE MALE GENITALS ARE TOOLS is thus not only a very productive conceptual metaphor but also one that is often encountered in pornography, putting the male owners into the position of those actively working. This provides further evidence in favour of my hypothesis of men being likely to be conceptualized as more active than women. The high frequency of weapons (as tools of warfare) underlines the effects - men being conceived of as violent agents - already discussed in connection with lexical variation above. Even though their high token frequencies are the result of the predominance of individual lexemes, namely of cock and dick (whose status as animal and person CMs may be doubted), without, however, resting entirely on these, THE MALE GENITALS ARE PERSONS and THE MALE GENITALS ARE ANIMALS are also quantitatively well represented conceptual metaphors. As persons and animals can be thought of as active participants in events, they enhance the agentizing effects of the tool metaphor just discussed, thus adding to the credence of the hypothesis that men will feature as more active in pornography consumers’ thinking. The only CM from those supposed to potentially undermine agentization is THE MALE GENITALS ARE TOYS because toys are things that usually the sexual partner plays with, which puts the owner into a passive position. Its high rank is a result of the high frequencies of the word balls and compound expressions with it. 2.3 Comparing and discussing results Leaving it to the readers to appreciate details of the data presented on my homepage, I will immediately turn to the examination of the sizes of the relevant conceptual metaphors in the erotica corpus in comparison to those found in the pornography corpus. I will again start with lexical variation. Table 12 and Figure 5 present the type frequencies. The table immediately below compares the results from the erotica corpus with those from the pornography corpus. <?page no="301"?> Representing women as empty space and men as tools 301 Types % % General 20 28.2% 6 17.1% Wholes 4 5.6% 2 5.7% Parts 16 20.8% 4 11.4% Persons - - 2 5.7% Body 8 11.3% - - Animals 4 5.6% 2 5.7% Plants 6 8.5% - - Tools 2 2.8% 4 11.4% General - - 2 5.7% Weapons - - 2 5.7% Parts of tools/ machines 2 2.8% - - Toys - - 2 5.7%% Valuables 3 4.2% - - Secrets and mysteries 7 9.9% - - Geography/ geology 6 8.5% 1 2.9% Shape 2 2.8% 4 11.4% Empty space 15 21.1% - - Openings 10 14.1% - - Passages 1 1.4% - - Containers 3 4.2% - - Others 1 1.4% - - Buildings 3 4.2% - - Food 2 2.8% 1 2.9% Clothes 1 1.4% - - Metonymies 19 26.8% 15 42.9% Superordinates 3 4.2% 1 2.9% Area/ position 9 12.7% - - Pars pro toto 1 1.4% 3 8.6% Neighbourhood 2 2.8% 2 5.7% Qualities/ states 4 5.6% 8 22.9% Activities - - 1 2.9% Accompanying emotions 1 1.4% - - Others 2 2.8% - - Totals 71 35 Table 12: Absolute and relative numbers of different expressions realizing the different conceptual metaphors and metonymies in the erotica corpus. <?page no="302"?> Analyses 302 0 % 5 % 1 0 % 1 5 % 2 0 % 2 5 % 3 0 % 3 5 % 4 0 % 4 5 % General Persons Body Animals Plants Tools Toys Valuables Secrets and mysteries Geography/ geology Shape Empty space Buildings Food Clothes Metonymies F e m a le M a le Figure 5: Relative sizes of the individual conceptual metaphors and metonymies with respect to lexical variation in the erotica corpus. Types Porn Erotica Porn Erotica General 13.2% 28.2% 10.0% 17.1% Wholes 2.9% 5.6% 2.4% 5.7% Parts 10.3% 20.8% 7.7% 11.4% Persons - - 7.7% 5.7% Body 10.6% 11.3% 7.2% - Animals 9.4% 5.6% 8.1% 5.7% Plants 5.8% 8.5% 5.3% - Tools 7.1% 2.8% 32.1% 11.4% General 0.6% - 14.8% 5.7% Weapons 2.6% - 16.3% 5.7% Parts of tools/ machines 3.9% 2.8% 1.9% - Toys 0.3% - 4.3% 5.7% Valuables 3.2% 4.2% 1.4% - Secrets and mysteries 11.0% 9.9% 1.0% - Geography/ geology 6.8% 8.5% 0.5% 2.9% Shape 6.1% 2.8% 1.4% 11.4% Empty space 29.0% 21.1% 10.5% - Openings 11.6% 14.1% 2.4% - Passages 7.4% 1.4% 3.3% - Containers 5.8% 4.2% 4.8% - Others 4.2% 1.4% - - Buildings 4.2% 4.2% 0.5% - <?page no="303"?> Representing women as empty space and men as tools 303 Food 2.9% 2.8% 11.5% 2.9% Clothes 1.0% 1.4% 3.3% - Metonymies 33.9% 26.8% 25.4% 42.9% Superordinates 2.9% 4.2% 2.9% 2.9% Area/ position 15.8% 12.7% 3.8% - Pars pro toto 3.2% 1.4% 2.9% 8.6% Neighbourhood 1.0% 2.8% 1.4% 5.7% Qualities/ states 2.6% 5.6% 10.0% 22.9% Activities 1.0% - 1.0% 2.9% Accompanying emotions 10.6% 1.4% 2.4% - Others 0.6% 2.8% 1.0% - Table 13: Comparison of conceptual metaphors and metonymies with respect to lexical variation between the pornography corpus and the erotica corpus. An erotica-internal comparison between the different CMs for women’s genitals with respect to lexical variation reveals similar trends as have been found in the pornography corpus. Empty space is the predominant metaphorical category - the most clearly passivizing subcategory of openings claiming a first place - with metaphors and metonymies concerned with place and secrets and mysteries also showing some diversity in their expressions. If we, however, compare these results with those from the pornography corpus, then we see that my expectations are fulfilled because all of the passivizing categories just mentioned trail their pornographic counterparts significantly (except the geography/ geology metaphors, where there is an advantage for the erotica corpus). The data thus suggests that although the CMs found in erotica will also lead to a salient link between female genitals and negative space and consequently between women and passivity, the hypothesis that this association will be weaker in erotica consumption than in pornography consumption remains tenable. There are a few results in the agentizing categories potentially relativizing this conclusion, e.g. the proportions of the animal and the tool metaphors are lower in the erotica corpus than in the pornography corpus. But as the figures are not significantly high generally and as the animal metaphor for women in the pornography corpus has been shown to consist mainly of pussy compounds, which cannot be considered to have a substantial agentizing potential, these results will not constitute a real challenge. A first glance at the relations between the different categories for males shows results resembling those from the pornography corpus: tools are clearly in the lead in the conceptual metaphors for male <?page no="304"?> Analyses 304 genitals - weapons taking an important position within the category - with sizeable proportions of person and animal metaphors. The results still correspond to my expectations because the erotica figures of all these categories are far below the respective values in the pornography corpus (e.g. for tools, 11.4% vs. 33.0%). This supports my hypothesis that the extent of conceptualizing men as active (and partly also as violent) will be smaller in those reading erotica than in those consuming pornography. The absence of any empty space metaphors in the erotica corpus will probably not really undermine this impression, considering that it probably primarily points to a limited obsession with the scrotum in the latter. Remarkably enough, one of the most richly lexicalized categories in the erotica corpus (with a percentage that is twice as high as in the pornography corpus) is THE MALE GENITALS ARE THEIR QUALITIES / STATES metonymy. As the source is practically limited to states of penile erection, this metonymy will result in a strong emphasis on sexual arousal, a conclusion contradicting results from the chapter on physicalization and visualization, where this aspect did not play an important part in the description of the body parts in the erotica corpus. In view of the fact that genitals generally are much less often referred to in the latter, we should, however, perhaps be cautious not to read too much into this result. Do the token frequencies allow insights differing from those gained in the examination of lexical variation? The figures contained in the following tables (the second one comparing pornography and erotica data) and the diagram will provide the basis for answering this question. <?page no="305"?> Representing women as empty space and men as tools 305 Tokens % % General 118 40.4% 77 22.7% Wholes 81 27.7% 72 21.2% Parts 37 12.7% 5 1.5% Persons - - 40 11.8% Body 23 7.9% - - Animals 21 7.2% 126 37.2% Plants 9 3.1% - - Tools 2 0.7% 16 4.7% General - - 7 2.1% Weapons - - 9 2.7% Parts of tools/ machines 2 0.7% - - Toys - - 23 6.8% Valuables 3 1.0% - - Secrets and mysteries 11 3.8% - - Geography/ geology 11 3.8% 1 0.3% Shape 4 1.4% 4 1.2% Empty space 21 7.2% - - Openings 16 5.5% - - Passages 1 0.3% - - Containers 3 1.0% - - Others 1 0.3% - - Buildings 3 1.0% - - Food 2 0.7% 2 0.6% Clothes 1 0.3% - - Metonymies 50 17.1% 52 15.3% Superordinates 3 1.0% 1 0.3% Area/ position 14 4.8% - - Pars pro toto 6 2.1% 5 1.5% Neighbourhood 18 6.2% 18 5.3% Qualities/ states 5 1.7% 27 8.0% Activities - - 1 0.3% Accompanying emotions 1 0.3% - - Others 6 2.1% - - Totals 292 339 Table 14: Absolute and relative frequencies of expressions realizing the different conceptual metaphors and metonymies in the erotica corpus. <?page no="306"?> Analyses 306 0 % 5 % 1 0 % 1 5 % 2 0 % 2 5 % 3 0 % 3 5 % 4 0 % 4 5 % General Persons Body Anim als Plants Tools Toys Valuables Secrets and mysteries Geography/ geology Shape Empty space Buildings Food Clothes M etonymies F e m a le M a le Figure 6: Relative sizes of the individual conceptual metaphors and metonymies with respect to token frequencies in the erotica corpus. Porn Erotica Porn Erotica General 40.3% 40.4% 4.7% 22.7% Wholes 25.7% 27.7% 2.9% 21.2% Parts 14.6% 12.7% 1.8% 1.5% Persons - - 10.9% 11.8% Body 4.9% 7.9% 4.0% - Animals 29.8% 7.2% 43.5% 37.2% Plants 1.4% 3.1% 1.0% - Tools 1.1% 0.7% 20.0% 4.7% General 0.0% - 8.2% 2.1% Weapons 0.3% - 16.1% 2.7% Parts of tools/ machines 0.8% 0.7% 0.8% - Toys 0.0% - 7.2% 6.8% Valuables 0.2% 1.0% 0.2% - Secrets and mysteries 1.1% 3.8% 0.0% - Geography/ geology 1.6% 3.8% 0.1% 0.3% Shape 0.7% 1.4% 0.4% 1.2% Empty space 14.3% 7.2% 0.9% - Openings 10.4% 5.5% 0.2% - Passages 1.3% 0.3% 0.3% - Containers 1.7% 1.0% 0.4% - Others 0.8% 0.3% - - <?page no="307"?> Representing women as empty space and men as tools 307 Buildings 0.7% 1.0% 0.0% - Food 0.4% 0.7% 2.2% 0.6% Clothes 0.6% 0.3% 0.3% - Metonymies 11.1% 17.1% 9.2% 15.3% Superordinates 0.2% 1.0% 0.7% 0.3% Area/ position 1.7% 4.8% 0.2% - Pars pro toto 0.9% 2.1% 0.4% 1.5% Neighbourhood 3.1% 6.2% 2.9% 5.3% Qualities/ states 0.6% 1.7% 4.7% 8.0% Activities 2.2% - 0.0% 0.3% Accompanying emotions 1.6% 0.3% 0.2% - Others 1.0% 2.1% 0.0% - Table 15: Comparison of conceptual metaphors and metonymies with respect to token frequencies between the pornography and the erotica corpus. Corresponding to the assumptions formulated at the beginning of this subchapter, the empty space metaphor for female genitals occurs much less often - approximately half as frequently - in the erotica corpus than in the pornography corpus. This means that the tendency to conceptualize female sexual organs as openings, passages or containers and thus their owners as passive recipients will be weaker in erotica than in pornography, which provides supportive evidence for one of my hypotheses. There are, however, results seriously challenging this conclusion. Firstly, other passivizing categories such as metaphors and metonymies concerned with place or the secrets and mysteries metaphors are used more frequently in the erotica corpus. And secondly, the animal CM, which could be regarded as promoting agentization and mitigating passivization, sees the erotica corpus trailing its pornography counterpart by far. The former aspect, however, might not be as problematic, considering the relatively small sizes of the classes. And the latter one will not seriously reduce the validity of the conclusion from above either, because we have to take into account that, as mentioned, the high rank of this CM in pornography is primarily a result of the high frequencies of the word pussy and its compounds. This word highlights the connection between the furry side of cats and pubic hair rather than that between feline liveliness and human activity, which will undermine any potentially agentizing effect. Turning our attention to the male genitals, we see that expressions drawing upon tools as source, in agreement with the predictions, occur significantly less frequently in the erotica corpus than in the porno- <?page no="308"?> Analyses 308 graphy corpus, which also applies to the subcategory of weapons. Since tools, as argued above, can make their user appear more active, this result increases the plausibility of the hypothesis that erotica will create less agentization of men in its readers than pornography. This conclusion is supported by the fact that erotica also fall behind pornography in another agentizing category, viz. THE MALE GENITALS ARE ANIMALS , which, incidentally, is the largest CM in both corpora, a result owing a lot to the high-frequency word cock. The other metaphors that are supposed to contribute positively or negatively to men’s agentization are either insignificant in size or do not show great gaps between the two collections of stories. To conclude, the linguistic data examined in this subchapter has shown all predictions to be plausible: we have seen extensive lexical variation and high frequencies of occurrence of expressions realizing the conceptual metaphor THE FEMALE GENITALS ARE EMPTY SPACE , which I consider to create a passive conception of women, and of those realizing the conceptual metaphor THE MALE GENITALS ARE TOOLS , which I assume to contribute to an active conception of men, in the pornography corpus. Type and token frequencies of the respective categories in the erotica corpus are significantly below the pornography levels. Most other conceptual metaphors and metonymies are in line with these findings. The data thus lends considerable weight to the hypotheses that women will be conceptualized as passive and men as active by pornography readers and that the gap will not be as wide for erotica readers. 3 Casting a side glance: interesting phenomena not examined Although there are further linguistic features relevant to the current topic (e.g. verb constellations and their role in representing relationships as involving an unequal ‘division of labour’ - as a matter of fact, this is a prerequisite to passivization - conceptual metaphors for other body parts or events, etc.), I just want to mention one aspect, viz. conceptual metaphors for sexual intercourse, because it seems a topic that suggests itself in a study of pornography. The passivizing and agentizing potential of intercourse expressions is more difficult to pin down on metaphors than with genitals. We find various prominent conceptual metaphors and conceptual metonymies for this target domain, e.g. SEX IS PLAY / SPORTS (to cock’n’roll, to score, etc.), SEX IS WORK (to work on somebody, to get <?page no="309"?> Casting a side glance 309 it done, etc.) (cf. Terry 1983/ 84, Richter 1987, 1993; examples are mine), or SEX IS ITS ACCOMPANYING ACTS (e.g. to go to bed with somebody, to sleep with somebody). These, however, do not per se distinguish between passive females and active males or vice versa, since such distinctions could only be created by the constellations in which these expressions occur and the distribution of semantic roles. But are there conceptual metaphors that can be argued to contribute to active or passive conceptualizations of women and men irrespective of the syntactic environment? Two that clearly seem to passivize women and agentize men are SEX IS FILLING (e.g. to fill, to stuff) and SEX IS ENTERING (to enter, to penetrate) and the conceptual metonymy SEX IS ( USING ) THE MALE GENITALS (mainly expressed in verbs derived by conversion from nouns for the penis or the testicles, e.g. to dick, to ball - the latter may, of course, themselves be metaphorical expressions) because they seem to (ontologically) bar an interpretation of the activity starting in women. The opposite trend is represented by SEX IS GRABBING (to grab, to cling to, to take in) and SEX IS ENCLOSING (to enclose, to house, to accommodate), where only females can take the active position. 9 If one of these sets can be shown to take precedence over the other, then this will also have an effect on the conceptualization of either women or men as more passive in sexuality. There are, of course, a number of side issues raised by such an analysis. I could, for example, also investigate whether further superordinate CMs are operating in this field, e.g. SEX IS VIOLENCE / WARFARE with many of the intercourse expressions drawing upon (forcefully) attacking as source domain, e.g. to invade for entering or to entrap for enclosing. (For a detailed account of the problematic nature of research into these metaphors, cf. Marko 2003.) Let me finally show how passivization becomes manifest in coherent text passages, which indicates that it could also - even though not as systematically - be examined co-textually. Luckily, with a few well-calculated jabs in the juiciest part of her cunt, she started to shake, her body quaking with the delicious spasms of erotic release. “Oh, baby, baby,” she whined, “fuck that rod into me! ” As she came, she grinded her narrow, milkwhite hips into me while jets of my creamy jism erupted in her lust-satiated hole. 9 Incidentally, the feminist Susan Brownmiller (1976: 372) suggests using the word enclose to represent intercourse. <?page no="310"?> Analyses 310 Firstly, the weapon metaphor rod for the penis - in connection with jabs, describing its action - presents the male not only as active but also as violent (the highly energetic orgasm metaphor erupt, even though not referring to an intentional act, may add to this effect). This is underlined by the constellation fuck that rod into me, which sees the male as the (invisible) agent doing something with his ‘weapon’, while the woman is only represented as the direction of the act. The sequence of involuntary movements, all occurring within one sentence (she started to shake, her body quaking with delicious spasms of erotic release) and the empty space metaphor hole (plus the adjective juicy, which belongs to the food CM because it describes the property of something edible) will certainly support the impression of the woman playing the passive role in the sexual event described by the passage. The clause she grinded her narrow, milk-white hips into me will undermine this a little, but it will enhance the general aggressive conception of sexuality created in this paragraph. For the purpose of contrast, I will add a passage from the erotica corpus, which suggests a more evenly balanced relationship, centring around the reciprocal ergative we had made love, but with many further actions taking both partners as agents. Both of us looked toward the television set at the same time, attracted by a loud announcer. Then the credits rolled by. We had made love for a full hour and missed the entire documentary. “I’m sorry,” I said, smiling. “I’m not,” he said, and he took my hand and we kissed. “So when can we do this again? ” he asked. “Not now, that’s for sure? ” I said, in mock exasperation. And we both laughed. As it turned out, Lawrence and I made love more times together than I ever have with another man. 4 Conclusion Conceptualizing people as passive and not able to decide upon the future course of events means seeing them as similar to objects. Passivization can therefore be regarded as a form of objectification. This chapter has examined whether passivization is a stronger factor in pornography consumers’ thinking of women or men and whether this might distinguish pornography from erotica. To study these questions, I have linguistically focused on verb constellations (of fourteen select verbs) and the semantic roles they assign to women and men. I have further examined the lexical variation and the frequencies of occurrence of metaphorical expressions for genitals supposed to promote female <?page no="311"?> Conclusion 311 passivization ( THE FEMALE GENITALS ARE EMPTY SPACE ) and male agentization ( THE MALE GENITALS ARE TOOLS ). My hypothesis predicted that the textual aspects analysed would lead to a view of women as passive and men as active participants in relationships in pornography, with differences growing in sex-focused events. Women appear much more frequently in passive roles and less frequently in active ones than men and women’s genitals are preferably described by reference to empty space, which invites filling, thus being conceived of as receiving and passive, while male genitals often take tool metaphors, putting men into the user position, which is considered active. These significant differences indicate that the hypothesis is wellfounded and plausible. The said differences are also more obvious in pure intercourse verbs and get blurred the more social the event. The only exception is physical contact, which is partly taken as belonging to intercourse, consequently resembling the former in the results. The second hypothesis, saying that female passivization, particularly in exclusively sexual events, would be weaker in erotica than in pornography, is supported by data on both semantic role allocation and conceptual metaphors. The only qualification of this support concerns physical contact, where women surprisingly often take passive roles in the erotica, which is probably a result of the intervention of subjectification and its emphasis on women’s sensual perception of tenderness. Despite the partly straightforward results, a few methodological and metatheoretical problems distort the picture. In particular the gross differences in sample sizes (the frequencies of the verbs analysed differ extremely, both on intraand inter-corpus levels), which make comparisons difficult, and the failure to anticipate the differentiations between physical contact verbs are main shortcomings in the research done in this chapter. <?page no="313"?> Conclusion: Questions, Answers and Evaluations I set out to answer questions about pornography’s potentially harmful nature by dissecting its language, demonstrating in the process how my version of Critical Discourse Analysis in combining an interpretative approach with a scientific orientation was the appropriate tool for researching such a phenomenon. Now eight chapters, 300 pages, and hundreds of data items later, it is time to draw conclusions, reviewing how the findings yielded by the analyses can contribute to answering the initial questions and critically reflecting upon the project’s achievements as well as its failures and shortcomings. I will divide my conclusions into three parts, the first one dealing with the questions raised by the study of the pornography debate and the second one with the answers provided by the analyses, with the third one evaluating the metatheoretical and methodological conditions of the study. Questions What is the background to the questions? Researching pornography - in this book defined as the excessive explicit representation of sexuality in different semiotic modes for commercial purposes - meant that I was entering an ongoing debate about its uses and effects and, working within a critical and thus politically-committed and socially-sensitive paradigm, I could not ignore the fact that theories, arguments and studies all centre on the pivotal question of pornography’s harmfulness. I therefore chose this as the starting point for my project. In my conception of what harm amounts to I adopted ideas proposed by the feminist camp in the debate, suggesting that pornography, like any other discourse, plays a pertinent role in creating beliefs and attitudes in its consumers, and that those created by this form of sexually explicit material discriminate against women (or, for that matter, any other group or humans in general), relegating them to inferior and subordinate roles (= subordination), and reducing them to mindless objects, body parts and mere passive recipients in sexuality <?page no="314"?> Conclusion: Questions, Answers and Evaluations 314 (= objectification). Men holding such beliefs and attitudes may resort to harmful action more readily, with rape being the most extreme form of such behaviour. What are the questions? Although subordination may be more pivotal with respect to harm, it partly overlaps with objectification and appears to be fuzzier, so that I decided to concentrate on the latter only. Being an umbrella concept rather than a homogeneous phenomenon itself, objectification was divided into four subareas, since viewing humans as objects involves conceptualizing them as parts rather than as wholes (= fragmentation), as material and visual rather than as mental or spiritual (= physicalization and visualization), as mindless (in the literal sense of ‘without a mind’) rather than as endowed with subjectivity (= desubjectification), and as passive rather than as active (= passivization). These four conceptualizations can be assumed to provide ample space for ideological operations. I therefore took them as my general areas of questions, asking whether they generally appear very often in pornography and whether they apply more substantially to women than to men, particularly in comparison to another sexually explicit discourse, namely erotica. How were these questions approached? These general questions were translated into concrete predictions about linguistic elements and structures, which were then tested against data produced from a 680,000-word corpus of pornographic short stories and a 240,000-word corpus of erotica stories. The data was analysed by a combination of qualitative and quantitative methods. Answers Is there fragmentation? In studying fragmentation, I focused on nouns denoting body parts, examining primarily their lexical variation and their frequencies, but also the amount of genitive NP modification (claimed to indicate immediate attention to bodies rather than to whole persons). Even though some of the data was not as homogeneous as expected, the results have added to the plausibility of the main hypotheses, suggesting that fragmentation is <?page no="315"?> Conclusion: Questions, Answers and Evaluations 315 indeed an important factor in pornography, particularly in comparison to erotica, and that women are more likely to be fragmented than men in pornography, even if not by far, while such a gender bias cannot be found in erotica. The analyses have also shown the trends to be partly more pronounced in connection with sexual body parts in pornography, pointing to a strong interdependence of fragmentation and sexualization, which, however, applies to both women and men equally as no difference has been found between the genders with respect to this question (probably as a result of the omnipresent power of the penis). Is there physicalization and visualization? The chapter on physicalization and visualization was mainly concerned with personal descriptors, i.e. elements specifically used to assign properties to people, concentrating on attributive adjectives, which were assumed to be less attention-attracting than predicative descriptors and thus possibly more effective. The analysis looked at descriptors of whole persons and of body parts, drawing upon an elaborate taxonomy of categories in order to compare visual physical descriptors to other types with respect to lexical variation and frequency. The data suggests that physicality and visuality are indeed important features in the conceptualizations of people in pornography, even if not the only ones. This trend, however, is not restricted to pornography but applies equally strongly to erotica. A hypothesis that has been confirmed by the results is that women are more likely to be physicalized and visualized than men by pornography readers, even though the gap is not very wide. Erotica do not feature such gender differences. Is there desubjectification? In my chapter on desubjectification, I examined mental verbs to see who appears in the semantic role of the senser and whose inner lives are thereby represented more consistently. I also looked at who occurs as grammatical subject, with an emphasis on passive voice subjects, assuming a link between this syntactic function and the point of view created. As an overall amount of desubjectification is inconceivable, the chapter was exclusively concerned with gender differences. The analyses have failed to produce evidence in favour of the hypothesis that women are desubjectified more strongly than men in pornography. As a matter of fact, most of the data suggests that female subjectivity and a female perspective are salient factors in this discourse. Female subjectification <?page no="316"?> Conclusion: Questions, Answers and Evaluations 316 has been shown to follow gender stereotypes in putting more emphasis on emotions and perceptions than on cognition. As for differences between the two discourses under scrutiny, the results suggest - albeit not clearly - that, in accordance with my hypotheses, female subjectification will be a more central force in erotica than in pornography and, by highlighting the cognitive dimension over the emotional and the perceptive ones, will not follow stereotypical patterns. Is there passivization? The linguistic focus of my study of passivization was on passive and active semantic roles and on conceptual metaphors for female and male genitals implying passivity or agentivity of the ‘owners’. The data produced by the analyses lends some weight to the assumptions considered. Passivizing women is a prominent feature of the pornographic discourse, which particularly applies to events with a strong sexual and physical emphasis, but less to social encounters. This trend to relegate women to passive positions is not as obvious in erotica. All in all, is there objectification? All these results and conclusions reviewed, what is now the ultimate answer to the question of whether pornography is objectifying, particularly to women? Pressed to return a verdict of guilty or not guilty, I would probably resign from my responsibility as a member of the jury because the analyses have revealed a situation that requires distinctions that would not be captured by a simplifying either-or decision. My differentiating and discerning answer would be that pornography can be considered to objectify women in the sense of fragmenting, passivizing, and physicalizing and visualizing them. Apart from passivization, the differences between women and men, however, are not extreme, suggesting that objectification extends to the latter, even though the discriminatory social effects - ranging from ignoring wishes and preferences to sexual aggression - in a patriarchal culture will apply more severely to women than to men. As for desubjectification, my conclusion is that denying women their subjectivity does not seem to be an essential component of pornography’s objectification. But this does not come as a complete surprise. It might even lie in the very nature of pornography that it cannot exclude a female point of view because this would prevent readers from seeing how male sexual action is actually appreciated and welcomed by women. <?page no="317"?> Conclusion: Questions, Answers and Evaluations 317 Although there is a moment of doubt in my answers, because the data examined has hardly ever revealed overly impressive differences, I think the results are relevant. Considering that I have analysed a very mild version of pornography and one that might even be targeted at a slightly more sophisticated consumership than its primarily imagerybased counterparts, the trends observed here can be expected to be more pronounced in sexually explicit material of a more hard-core and more visual nature. Evaluations My project also pursued a second goal, namely to demonstrate, in analysing the primary question, that my own approach, in combining hermeneutic principles with scientific rigour, could be a valuable elaboration of the common research practice of Critical Discourse Analysis and that it could be a promising alternative to other methods applied in the study of pornography. What are the conclusions concerning this dimension of my research? I think that my study will contribute to the metatheoretical and methodological basis of Critical Discourse Analysis, not because it has been a success story throughout - quite the contrary as will be explained immediately - but because as one of the few large-scale, corpus-based and scientifically-oriented studies in CDA, it has highlighted problems that can arise in such an approach. As these problems can be associated with one of the five principles of scientific research applied (empirical data, quantitative data, comparative data, model-based, hypothesesbased), I will discuss them in connection with the latter in the following. What kind of problems have arisen in connection with the empirical (transtextual) orientation? The corpus approach with its concomitant emphasis on empirical details of language has shown that the attempt to be all-encompassing rather than selective, e.g. analysing all expressions for genitals or all attributive descriptors, created serious problems because a lot of the data was unwieldy and ‘defiant’ and could not easily be fitted into a descriptive model. This has produced a partly rather heterogeneous picture. I might not go so far as to claim that comprehensiveness is a serious mistake on my part (apart from being a sign of overambition), but it points to the need for tools more adjusted to the analysis of large bodies of textual <?page no="318"?> Conclusion: Questions, Answers and Evaluations 318 data as most analytical instruments have only been tried in small-scale studies. The transtextual focus on formal elements, though definitely a valuable alternative to existing methods used in CDA, in certain areas proved to be too limited. Particularly in the area of perspective and subjectification, which are also created in intricate interactions between elements whose values cannot with any ease be determined from their formal appearances, a stronger integration of co-textual forms of analysis is called for. The consistency of the transtextual orientation of the analyses might also have given the - not completely wrong - impression that texts have been atomized, fragmented into unrecognizable snippets. Although I have described the problems inherent in a co-textual approach (in particular the fact that the researcher’s own interpretative processes play a more prominent role), I admit that a few complementary analyses of larger passages, demonstrating the relevance of the atomistic data, would have benefited my study, at least as far as its readability is concerned. What kind of problems have arisen in connection with quantification? Judging from the awkwardness involved in evaluating some of the data, attempting to steer a middle course between an exhaustively quantitative and statistical analysis and a thoroughly qualitative examination of the material, I might have run aground. I still believe that opting for statistics proper would have created insurmountable obstacles for me as an individual researcher (call it ignorance! ). But relying mainly on the validity of figures and percentages often turned out to bar a definite judgement on the hypotheses. This is why I had to resort to scientifically half-hearted descriptors such as suggest, indicate, support, and add plausibility to instead of the preferred set of prove or (de)verify. It may lie in the nature of research in the humanities and the social sciences that the question of plausibility is more relevant than the question of truth, but it still made the process of drawing conclusions concerning the general assumptions difficult. There were also problems involved in representing data appropriately, particularly if the correlations between different factors were not linear and thus not really describable in terms of percentages only. The influence of the proportions of first-person-narrators on different frequencies could, for instance, not be quantified. <?page no="319"?> Conclusion: Questions, Answers and Evaluations 319 Another serious problem encountered in the research was concerned with principles of counting. Especially the chapters on desubjectification and passivization featured structures that could not simply - in a one-to-one fashion - be counted for one or the other category, which required taking decisions that may partly appear arbitrary and not easy to explain. Despite the merits of a quantitative approach, it also - to a certain degree - sacrificed clarity for its sound scientific orientation. In many parts, the ‘thicket of figures’ prevented a more immediately revealing - and thus potentially entertaining - look at what the discourse of pornography could look like (a case of not seeing the wood for the trees). What kind of problems have arisen in connection with the comparison to erotica? Generally, I assume that choosing erotica as a comparative discourse was the right decision because it represented the ideal balance between being similar and still distinct. But the composition of the erotica corpus turned out to be problematic. The smaller size reduced the overall comparability because percentages based on low absolute frequencies are more prone to distortions. The predominance of first person narrators, required to represent the essence of erotica, also hampered comparisons as a consequence of the unpredictable interactions of various linguistic elements with the narrative voice. What kind of problems have arisen in connection with the model of the idealized discourse participant? The study’s tendency to get lost in explaining the linguistic intricacies of the phenomena to be examined might fuel the general criticism that could be levelled against my approach. It could be claimed, following Widdowson’s argumentation against CDA (1995), that the model of the idealized discourse participant is only a sophisticated (if at all) legitimation for doing more ‘hard-core’ (in both senses) linguistic text analysis, without thereby, however, getting closer to the actual meanings created in the discourse. I admit that this objection is partly valid, particularly because in many of its linguistic details, the model presented is not warranted by empirical psycholinguistic research (the connection between grammatical subjects and perspective, for instance), and can therefore not really support the claim that likely interpretations can be reconstructed from textual data alone. I do, however, believe that such <?page no="320"?> Conclusion: Questions, Answers and Evaluations 320 general criticism goes too far in its dichotomization of potential meanings and actual interpretations, sometimes in an almost postmodern spirit insinuating that the latter are not at all related to the former. So even if my approach is text analysis in disguise of discourse analysis, the potential meanings it discovers cannot be discarded as completely irrelevant. On a more concrete level, my analyses have exposed shortcomings on many levels of linguistic theory with hardly any independent studies being available on, for instance, the connections between grammatical subjects and perspective, or the different roles of predicative and attributive descriptors, to name just two prominent issues. I may have been ill-advised to go ahead with my attempt to theorize these aspects myself, because not being my actual focus, they could not be investigated with the same diligent care that would have been applied had these been the actual topics of my study. This might have led to questionable theoretical accounts, which, in turn, probably contributed to the heterogeneity of some of the data. The taxonomies of categories, particularly the one used in the chapter on physicalization and visualization, were the most serious drawback in the theoretical framework. These categories are postulated to ultimately represent the mental classes that readers draw upon in their interpretations. While language users can be claimed to allow for a great deal of fuzziness in categorizations (as research on prototypes has proved), the task of the scientifically-oriented linguist (aiming at generalizations) is to clearly assign each linguistic element to one (or two or three, if I allow for multi-category membership) class. The numerous elements that resisted such classification showed that the rigid grids that the taxonomies provided probably were not appropriate to really simulate or reconstruct the interpretation process. It goes without saying that such doubts in categorization also erode the validity and reliability of the outcome of the analyses to some extent. What kind of problems have arisen in connection with the hypotheses? The main objection that could be voiced with respect to this principle is, to put it simply, that I have chosen the wrong hypotheses and thus the wrong issues for analysis. Although I do not believe that I opted for the ‘wrong’ topics because a selective focus is required in this type of research and any choice has its pros and cons, I would like to mention at least a few alternatives and what they could have amounted to. <?page no="321"?> Conclusion: Questions, Answers and Evaluations 321 My feminist orientation may have prevented me from tackling serious and interesting questions raised by other camps in the pornography debate and thus from drawing a wider picture of the social roles of the pornographic discourse. Morality and pornography’s alleged attempt to corrupt and pervert it could have been a topic in the conservative vein, with evaluative polyvalence - i.e. the simultaneous presence of positive and negative judgements - in words such as slut (Tell that slut that she is a goddess) or dirty being a starting point. Fictionalization, i.e. how pornography signals that it should not be taken as factual, on the other hand, would have been a subject matter with a more liberal slant (I assume, however, that the opposite trend is more pervasive, considering the strong focus on ‘realistic’ discourses such as letters or diaries and the merging of editorial personnel and fictional characters mentioned in chapter IV). A further critical point that could be raised is why I have not looked at the more concrete harmful effects instead of taking the detour via the influence on knowledge. The ‘safety valve’ or catharsis theory could have provided a relevant point of departure since I could have tried to dis/ prove it in my analyses. Though I consider this to be an important question, I doubt that it lends itself easily to a discourse analysis along the lines used in this book. Firstly, such an issue requires a theory of how the consuming of a text can be a form of authentically experiencing sexuality (or something similar), because not knowledge but only such experience can serve as a safety valve. Secondly, even the possibility of catharsis seriously challenges - and would therefore require refining - the social constructionist foundation of CDA because it implies that certain beliefs and attitudes become established in consumers’ minds through their reading or watching, but that for reasons also related to consumption, they then refrain from acting on the basis of this knowledge. And thirdly, and most practically, the catharsis theory, laying emphasis on consequences remote from the text, necessitates - more than my topics - examining the pornography users directly. All in all, how can the methodological and metatheoretical design of the study be evaluated? Despite all the shortcomings mentioned, I still believe that to study pornography adhering to the five metatheoretical and methodological principles mentioned was the right decision. And even though my project might be considered an elaborate and sophisticated form of grand-style failure, it has at least pointed the way for further discourse <?page no="322"?> Conclusion: Questions, Answers and Evaluations 322 analytical research into pornography, concentrating on smaller issues, based on better-developed linguistic theories of the individual items to be examined, finding a better balance between qualitative and quantitative methods, and combining research of discursive practices and textual analysis (preferably involving several researchers) without, however, losing the desire to undig and uncover ideologies from texts. Such research is certainly far from redundant because pornography has not stopped posing serious questions. <?page no="323"?> References Abel, Gene G./ David H. Barlow/ Edward B. Blanchard/ Donald Guild (1977). “The Components of Rapists’ Sexual Arousal.” Archives of General Psychiatry 34. 895-903. Abramson, Paul R./ Haruo Hayashi (1984). “Pornography in Japan: Crosscultural and Theoretical Considerations.” In: Malamuth/ Donnerstein (1984). 173-183. Agar, Michael (1991). “Writing Left in America.” Discourse Processes 14: 2. 261-276. Ageton, Suzanne (1983). Sexual Assault among Adolescents. Lexington, MA: Lexington Books. Alderson, Charles/ Alexander H. Urquhart (1984a). “Introduction: What is Reading? ” In: Alderson/ Urquhart (1984b). xv-xxviii. Alderson, Charles/ Alexander H. Urquhart (eds.) (1984b). Reading in a Foreign Language. London & New York: Longman. Anderson, John (1995, 4 th edition). Cognitive Psychology and its Implications. New York: W.H. Freeman and Company. Assiter, Alison (1989). Pornography. Feminism and the Individual. London: Pluto Press. Assiter, Alison/ Avedon Carol (eds.) (1993). Bad Girls and Dirty Pictures: The Challenge to Reclaim Feminism. London: Pluto Press. Attorney General’s Commission on Pornography and Obscenity Final Report (1986). Volumes 1 and 2. Committee chaired by Henry E. Hudson, Washington, D.C.: United States Department of Justice. Augoustinos, Martha/ Iain Walker (1995). Social Cognition. An Integrated Introduction. Thousand Oaks, CA, et al.: Sage. Baddeley, Alan (1990). Human Memory. Theory and Practice. Hove et al.: Lawrence Erlbaum. Baird, Robert M./ Stuart E. Rosenbaum (1991a). “Introduction.” In: Baird/ Rosenbaum (1991b). 7-15. Baird, Robert M./ Stuart E. Rosenbaum (eds.) (1991b). Pornography. Private Right or Public Menace. Buffalo, NY: Prometheus. Bakhtin, Mikhail (1994). The Bakhtin Reader. Selected Writings of Bakhtin, Medvedev, and Voloshinov. Ed. Pam Morris. London et al.: Edward Arnold. Barbach, Lonnie (1985a). “Introduction.” In: Barbach (1985b). ix-xxii. Barbach, Lonnie (ed.) (1985b [originally 1984]). Pleasures. Women Write Erotica. New York: Harper Perennial. Barbaree, Howard/ William Marshall (1991). “The Role of Male Sexual Arousal in Rape: Six Models.” Journal of Consulting and Clinical Psychology 59: 5. 621-630. <?page no="324"?> References 324 Baron, Larry (1990). “Pornography and Gender Equality: An Empirical Analysis.” The Journal of Sex Research 27: 3. 363-380. Baron, Larry/ Murray Straus (1984). “Sexual Stratification, Pornography, and Rape in the United States.” In: Malamuth/ Donnerstein (1984). 185-209. Baron, Larry/ Murray Straus (1987). “Four Theories of Rape: A Macrosociological Analysis.” Social Problems 34: 5. 467-489. Baron, Larry/ Murray Straus (1989). Four Theories of Rape in American Society. A State-level Analysis. New Haven, CT: Yale UP. Barry, Kathleen (1995). The Prostitution of Sexuality. New York & London: New York UP. Bartlett, Frederic C. (1932). Remembering: A Study in Experimental and Social Psychology. Cambridge: CUP. Bataille, Georges (1962). Eroticism. Transl. Mary Dalwood. London: Calder and Boyars. Baudrillard, Jean (1990). Seduction. Transl. Brian Singer. Basingstoke: Macmillan. Beaugrande, Robert de/ Wolfgang Dressler (1981). Introduction to Textlinguistics. London & New York: Longman. Becker, Judith/ Robert Stein (1991). “Is Sexual Erotica Associated with Sexual Deviance in Adolescent Males? ” International Journal of Law and Psychiatry 14. 85-95. Becker, Michael A./ Donn Byrne (1985). “Self-regulated Exposure to Erotica, Recall Errors, and Subjective Reactions as a Function of Erotophobia and Type A Coronary-prone Behavior.” Journal of Personality and Social Psychology 48. 760-767. Beneke, Timothy (1990). “Intrusive Images and Subjectified Bodies: Notes on Visual Heterosexual Porn.” In: Kimmel (1990). 168-187. Bensinger, Terralee (1992). “Lesbian Pornography: The Re/ Making of (a) Community.” Discourse 15: 1. 69-93. Berger, Peter/ Thomas Luckmann (1971 [originally 1966]). The Social Construction of Reality. Harmondsworth: Penguin. Berger, Ronald/ J. Patricia Searles/ Charles Cottle (1992). Feminism and Pornography. New York et al.: Praeger. Biber, Douglas/ Stig Johansson/ Geoffrey Leech/ Susan Conrad/ Edward Finegan (1999). Longman Grammar of Spoken and Written English. Harlow: Longman. Boston Women’s Health Book Collective (1971). Our Bodies, Ourselves. New York: Simon and Schuster. Brannigan, Augustine/ Sheldon Goldenberg (1987). “The Study of Aggressive Pornography: The Vicissitudes of Relevance.” Critical Studies in Mass Communication 4. 262-283. Brigman, William E. (1983). “In Depth: Pornography: Pornography as Political Expression.” Journal of Popular Culture 17: 2. 129-134. Bristow, Joseph (1997). Sexuality. London & New York: Routledge. Britton, Bruce K./ Arthur C. Graesser (eds.) (1996). Models of Understanding Text. Mahwah, NJ: Lawrence Erlbaum. <?page no="325"?> References 325 Brown, Dan/ Jennings Bryant (1989). “The Manifest Content of Pornography.” In: Zillman/ Bryant (1989). 3-24. Brown, Gillian/ George Yule (1983). Discourse Analysis. Cambridge: CUP. Brownmiller, Susan (1976 [originally 1975]). Against Our Will. Men, Women and Rape. Harmondsworth: Penguin. Burr, Vivien (1995). An Introduction to Social Constructionism. London & New York: Routledge. Burstyn, Varda (ed.) (1985). Women against Censorship. Vancouver: Douglas and McIntyre. Bussmann, Hadumod (1996). Routledge Dictionary of Language and Linguistics. Ed. and transl. by Gregory Trauth and Kerstin Kazzazi. London & New York: Routledge. Byer, Curtis O./ Louis W. Shainberg/ Grace Galliano (1999). Dimensions of Human Sexuality. New York: McGraw-Hill. Caldas-Coulthard, Carmen/ Malcolm Coulthard (eds.) (1996). Texts and Practices. London & New York: Routledge. Cameron, Deborah (1990a). “Discourse of Desire: Liberals, Feminists, and the Politics of Pornography in the 1980s.” American Literary History 2: 4. 784- 798. Cameron, Deborah (1990b). “Demythologizing Sociolinguistics: Why Language does not Reflect Society.” In: Joseph/ Taylor (1990). 79-93. Cameron, Deborah (1992a). “Pornography - What is the Problem? ” Critical Quarterly 34: 2. 3-11. Cameron, Deborah (1992b). “Naming of Parts: Gender, Culture, and Terms for the Penis among American College Students.” American Speech 67: 4. 367- 382. Cameron, Deborah (1992c, 2 nd edition). Feminism and Linguistic Theory. Basingstoke & London: Macmillan. Cameron, Deborah/ Don Kulick (2003). Language and Sexuality. Cambridge: CUP. Cameron, Deborah/ Elizabeth Frazer (1992). “On the Question of Pornography and Sexual Violence: Moving beyond Cause and Effect.” In: Itzin (1992c). 359-383 [repr. in Cornell (2000). 240-253]. Caplan, Pat (1989a). “Introduction.” In: Caplan (1989b). 1-30. Caplan, Pat (ed.) (1989b [originally 1987]). The Cultural Construction of Sexuality. London & New York: Routledge. Caputi, Jane (1992). “Advertising Femicide. Lethal Violence against Women in Pornography and Gorenography.” In: Radford/ Russel (1992). 203-221. Caputi, Mary (1994). Voluptuous Yearnings. A Feminist Theory of the Obscene. Lanham, MD & London: Rowman and Littlefield Publishers. <?page no="326"?> References 326 Carmen, Arlene/ Colleen Dewhurst/ Lisa Duggan/ Betty Friedan/ Richard Green/ Leanne Katz/ Max Lillienstein/ Barry Lynn/ Ann Welbourne-Moglia/ Donald Mosher/ Eve Paul/ Harriet Piopel/ Anthony Schulte/ Kurt Vonnegut (1986). The Meese Commission Exposed. Proceedings of a National Coalition against Censorship. Public Information Briefing on the Attorney General’s Commission on Pornography January 16. 1986. New York: National Coalition Against Censorship. Carol, Avedon (1994). Nudes, Prudes and Attitudes. Pornography and Censorship. Cheltenham: New Clarion Press. Carol, Avedon/ Nettie Pollard (1993). “Changing Perceptions in the Feminist Debate.” In: Assiter/ Carol (1993). 45-56. Carroll, David W. (1999, 3 rd edition). Psychology of Language. Pacific Grove, CA, et al.: Brooks/ Cole Publishing Company. Check, James/ Ted Guloien (1989). “Reported Proclivity for Coercive Sex Following Repeated Exposure to Sexually Violent Pornography, Nonviolent Dehumanizing Pornography, and Erotica.” In: Zillmann/ Bryant (1989). 159-184. Chester, Gail/ Julienne Dickey (1988). Feminism and Censorship. London: Prism. Chouliaraki, Lilie/ Norman Fairclough (1999). Discourse in Late Modernity. Rethinking Critical Discourse Analysis. Edinburgh: Edinburgh UP. Christensen, Ferrel M. (1990). Pornography: The Other Side. New York et al.: Praeger. Clover, Carol (1993). “Introduction.” In: Gibson/ Gibson (1993). 1-4. Cole, Susan (1995). Power Surge. Sex, Violence and Pornography. Toronto: Second Story Press. Comolli, Jean-Louis (1980). “Machines of the Visible.” In: de Lauretis/ Heath (1980). 121-142. Cornell, Drucilla (1995). The Imaginary Domain. Abortion, Pornography, and Sexual Harassment. London & New York: Routledge. Cornell, Drucilla (ed.) (2000). Feminism and Pornography. Oxford et al.: OUP. Coulthard, Malcolm (1985, 2 nd edition). An Introduction to Discourse Analysis. London & New York: Longman. Court, John H. (1984). “Sex and Violence: A Ripple Effect.” In: Malamuth/ Donnerstein (1984). 143-172. Court, John H. (1985). “What Then Shall We Say and Do? ” Bible Advocate December 1985. 22-26. Cowan, Gloria (1990). “What is Degrading in Pornography: Through Women’s Eyes.” Paper presented at the annual meeting of the Western Psychological Association, Los Angeles, CA. Cowan, Gloria / Robin R. Campbell (1994). “Rape Causal Attitudes among Adolescents.” The Journal of Sex Research 32: 2. 145-153. Cowan, Gloria/ Carole Lee/ Daniella Levy/ Debra Snyder (1988). “Dominance and Inequality in X-rated Videocassettes.” Psychology of Women Quarterly 12. 299-311. <?page no="327"?> References 327 Cowan, Gloria/ Kerri F. Dunn (1994). “What Themes in Pornography Lead to Perceptions of the Degradation of Women? ” The Journal of Sex Research 31: 1. 11-21. Crawford, Patricia (1994). “Sexual Knowledge in England, 1500-1750.” In: Porter/ Teich (1994b). 82-106. Curtis, James E. (1972). The Sociology of Knowledge. A Reader. New York et al.: Praeger. Dant, Tim (1991). Knowledge, Ideology and Discourse. A Sociological Perspective. London & New York: Routledge. Day, Gary/ Clive Bloom (eds.) (1988). Perspectives on Pornography. Sexuality in Film and Literature. London: Macmillan. de Lauretis, Teresa/ Stephen Heath (eds.) (1980). The Cinematic Apparatus. New York: St. Martin’s Press. Denhière, Guy/ Serge Baudet (1989). “Cognitive Psychology and Text Processing: From Text Representation to Text-world.” Semiotica 77-1/ 3 (special edition). 271-293. Dietz, Park Elliot/ Alan Sears (1987/ 88). “Pornography and Obscenity Sold in ‘Adult Bookstores’: A Survey of 5132 Books, Magazines, and Films in Four American Cities.” Journal of Law Reform 21: 1 & 2. 7-46. Dik, Simon (1978). Functional Grammar. Amsterdam: North-Holland. Dillon, George L. (1982). “Whorfian Stylistics.” Journal of Literary Semantics 11: 2. 73-77. Dines, Gail (1995). “‘I Buy It for the Articles’: Playboy Magazine and the Sexualization of Consumerism.” In: Dines/ Humez (1995). 254-262. Dines, Gail (1998a). “Dirty Business. Playboy Magazine and the Mainstreaming of Pornography.” In: Dines/ Jensen/ Russo (1998). 37-63. Dines, Gail (1998b). “Living in Two Worlds. An Activist in the Academy.” In: Dines/ Jensen/ Russo (1998). 163-166. Dines, Gail/ Jean M. Humez (eds.) (1995). Gender, Race and Class in Media. A Text Reader. Thousand Oaks, CA, et al.: Sage. Dines, Gail/ Robert Jensen/ Ann Russo (1998). Pornography. The Production and Consumption of Inequality. London & New York: Routledge. Dodson, Betty (1974). Liberating Masturbation: A Meditation on Self-Love. Self-published. Dodson, Betty (1996). Sex for One: The Joy of Selfloving. New York: Crown. Donnerstein, Edward (1984). “Pornography: Its Effect on Violence against Women.” In: Malamuth/ Donnerstein (1984). 53-81. Donnerstein, Edward (1989). Testimony in Her Majesty the Queen against Fringe Product, Inc. and 497906 Ontario Limited (1989, November 21). Court File No. 4079, Ottawa, Ontario, Canada. Donnerstein, Edward/ Daniel Linz (1986). “The Question of Pornography.” Psychology Today (December 1986). 56-59. Donnerstein, Edward/ Daniel Linz (1988). “The Methods and Merits of Pornography Research (Colloquy).” Journal of Communication 38: 2. 180-184. <?page no="328"?> References 328 Donnerstein, Edward/ Daniel Linz/ Stephen Penrod (1987). The Question of Pornography. Research Findings and Policy Implications. New York: Free Press. Donnerstein, Edward/ Leonard Berkowitz (1981). “Victim Reactions in Aggressive Erotic Films as a Factor in Violence against Women.” Journal of Personality and Social Psychology 41. 710-724. Dority, Barbara (1991). “Feminist Moralism, ‘Pornography’, and Censorship.” In: Baird/ Rosenbaum (1991b). 111-116 [excerpts from The Humanist 49 (November-December 1989). 8-9, 46]. Drakeford, John W./ Jack Hamm (1973). Pornography: The Sexual Mirage. Nashville, TN: Thomas Nelson. Duncan, David F. (1991). “Violence and Degradation as Themes in ‘Adult’ Videos.” Psychological Reports 69: 1. 239-240. Dunn, Charles (1987). Constitutional Democracy in America. Glenview, IL: Scott, Foresman and Comp. Dunn, Sara (1990). “Voyages of the Valkyries: Recent Lesbian Pornographic Writing.” Feminist Review 34. 161-170. Dworkin, Andrea (1981). Pornography. Men Possessing Women. New York: Perigee. Dworkin, Andrea (1987). Intercourse. London: Secker and Warburg. Dworkin, Andrea/ Catharine MacKinnon (1988). Pornography and Civil Rights: A New Day for Women’s Equality. Minneapolis: Organizing Against Pornography. Dworkin, Ronald (1991). “Pornography, Feminism, and Liberty.” In: Baird/ Rosenbaum (1991b). 164-173 [originally (1991). The New York Review of Books (August 15). 12-15]. Eagleton, Terry (1991). Ideology. An Introduction. London: Verso. Easton, Susan M. (1994). The Problem of Pornography. Regulation and the Right to Free Speech. London & New York: Routledge. Edwards, Derek (1996). Discourse and Cognition. Thousand Oaks, CA, et al.: Sage. Eggins, Suzanne (1994). Introduction to Systemic Functional Linguistics. London: Pinter. Einsiedel, Edna F. (1988). “The British, Canadian, and U.S. Pornography Commissions and their Use of Social Science Research (Social Science and Pornography).” Journal of Communication 38: 2. 108-121. Ellis, Henry Havelock (1936). Studies in the Psychology of Sex. A reissue of the 8 vol. set [1905-1928], with the parts arranged in a different order, and with new preliminary matter and cumulative indices. With a new foreword by Havelock Ellis. New York: Random House. Ellis, Richard (1988). “Disseminating Desire: Grove Press and ‘The End(s) of Obscenity’.” In: Day/ Bloom (1988). 26-43. Ertel, Henner/ Rainer Hagen/ Gisela Ripp/ Hand Schulenburt/ Alf Wagner (1990). Erotika und Pornographie. Repräsentative Befragung und psychophysiologische Langzeitstudie zu Konsum und Wirkung. München: Psychologie Verlags Union. <?page no="329"?> References 329 Fairclough, Norman (1989). Language and Power. London & New York: Longman. Fairclough, Norman (1992a). Discourse and Social Change. Cambridge et al.: Polity Press. Fairclough, Norman (ed.) (1992b). Critical Language Awareness. London & New York: Longman. Fairclough, Norman (1995). Critical Discourse Analysis: The Critical Study of Language. London & New York: Longman. Fairclough, Norman (1996). “A Reply to Henry Widdowson’s ‘Discourse Analysis: A Critical View’.” Language and Literature 5: 1. 49-56. Fairclough, Norman/ Ruth Wodak (1997). “Critical Discourse Analysis.” In: van Dijk (1997d). 258-284. Farr, Robert M./ Serge Moscovici (eds.) (1984). Social Representations. Cambridge & Paris: CUP/ Maison des Sciences de l’Homme. Fill, Alwin (1996a). “Ökologie der Linguistik - Linguistik der Ökologie.” In: Fill (1996b). 3-16. Fill, Alwin (ed.) (1996b). Sprachökologie und Ökolinguistik. Referate des Symposions “Sprachökologie und Ökolinguistik” an der Universität Klagenfurt, 27.-28. Oktober 1995. Editorial assistance by Hermine Penz. Tübingen: Stauffenburg. Firth, John Rupert (1935). “The Technique of Semantics.” Transactions of the Philological Society. 36-72. Fisher, William A./ Donn Byrne/ Leonard A. White/ Kathryn Kelley (1988). “Erotophobia-Erotophilia as a Dimension of Personality.” The Journal of Sex Research 45. 123-151. Fisher, William A./ Guy Grenier (1994). “Violent Pornography, Antiwoman Thoughts, and Antiwoman Acts. In Search of Reliable Effects.” The Journal of Sex Research 31: 1. 23-38. Fletcher, Charles R./ Paul van den Broek/ Erik J. Arthur (1996). “A Model of Narrative Comprehension and Recall.” In: Britton/ Graesser (1996). 141-163. Foucault, Michel (1984). “What is an Author? ” In: Rabinow (1984). 101-120. Foucault, Michel (1991). Die Ordnung des Diskurses. Transl. Walter Seitter. With an essay by Ralf Konersmann. Frankfurt a.M.: Fischer [originally (1972). L’ordre du discours. Paris: Gallimard]. Foucault, Michel (1992, 6 th edition). Der Wille zum Wissen. Sexualität und Wahrheit 1. Transl. Ulrich Raulff/ Walter Seitter. Frankfurt a.M.: Suhrkamp [originally (1976). Histoire de la sexualité, 1: La volonté de savoir. Paris: Gallimard]. Foucault, Michel (1993a, 3 rd edition). Der Gebrauch der Lüste. Sexualität und Wahrheit 2. Transl. Ulrich Raulff/ Walter Seitter. Frankfurt a.M.: Suhrkamp [originally (1984). Histoire de la sexualité, 2: L’usage des plaisirs. Paris: Gallimard]. Foucault, Michel (1993b, 3 rd edition). Die Sorge um sich. Sexualität und Wahrheit 3. Transl. Ulrich Raulff/ Walter Seitter. Frankfurt a.M.: Suhrkamp [originally (1984). Histoire de la sexualité, 3: Le souci de soi. Paris: Gallimard]. <?page no="330"?> References 330 Foucault, Michel (1994). Überwachen und Strafen. Die Geburt des Gefängnisses. Transl. Walter Seitter. Frankfurt a.M.: Suhrkamp [originally (1975). Surveiller et punir: naissance de la prison. Paris: Gallimard]. Foucault, Michel (1995, 7 th edition). Archäologie des Wissens. Frankfurt a.M.: Suhrkamp [originally (1969). L’Archéologie du savoir. Paris: Gallimard]. Fowler, Roger (1991). Language in the News. Discourse and Ideology in the Press. London & New York: Routledge. Fowler, Roger (1996). “On Critical Linguistics.” In: Caldas-Coulthard/ Coulthard (1996). 3-14. Fowler, Roger/ Bob Hodge/ Gunther Kress/ Tony Trew (1979). Language and Control. London et al.: Routledge & Kegan Paul. Friday, Nancy (1973). My Secret Garden. Sexual Fantasies of Women. New York: Trident Press. Gagnon, John H. (1977). Human Sexualities. Glenview, IL: Scott & Foresman. Gagnon, John H./ Richard G. Parker (1995a). “Conceiving Sexuality.” In: Gagnon/ Parker (1995b). 3-16. Gagnon, John H./ Richard G. Parker (eds.) (1995b). Conceiving Sexuality. Approaches to Sex Research in a Postmodern World. New York & London: Routledge. Gagnon, John H./ William Simon (1973). Sexual Conduct. The Social Sources of Human Sexuality. New York: Aldine de Gruyter. Gardiner, Judith Kegan (1993). “What I didn’t Get to Say on TV about Pornography, Masculinity, and Representation.” New York Law School Law Review 38. Symposium: The Sex Panic. 319. Garnham, Alan/ Jane Oakhill (1996). “The Mental Models Theory of Language Comprehension.” In: Britton/ Graesser (1996). 313-339. Garside, Roger (1987). “The CLAWS Word-tagging System.” In: Garside/ Leech/ Sampson (1987). 30-41. Garside, Roger/ Geoffrey Leech/ Geoffrey Sampson (eds.) (1987). The Computational Analysis of English: A Corpus-based Approach. London: Longman. Gee, James Paul (1992). The Social Mind. Language, Ideology, and Social Practice. New York et al.: Bergin & Garvey. Gee, James Paul (1999). An Introduction to Discourse Analysis. Theory and Method. London & New York: Routledge. Geer, James H./ Matthew McGlone (1990). “Sex Differences in Memory for Erotica.” Cognition and Emotion 4. 71-78. Gentry, Cynthia (1991). “Pornography and Rape: An Empirical Analysis.” Deviant Behavior: An Interdisciplinary Journal 12. 277-288. George, William H./ Peter A. Lopez (1995). “Men’s Enjoyment of Explicit Erotica: Effects of Person-Specific Attitudes and Gender-specific Norms.” The Journal of Sex Research 32: 4. 275-288. Gergen, Kenneth J. (1999). An Invitation to Social Construction. Thousand Oaks, CA, et al.: Sage. Gibson, Pamela C./ Roma Gibson (eds.) (1993). Dirty Looks: Women, Pornography, Power. London: BFI. <?page no="331"?> References 331 Giddens, Anthony (1992). The Transformation of Intimacy. Sexuality, Love and Eroticism in Modern Societies. Cambridge: Polity Press. Giobbe, Evelina (1995). “Surviving Commercial Sexual Exploitation.” In: Dines/ Humez (1995). 314-318. Goatly, Andrew (1997). The Language of Metaphor. London & New York: Routledge. Goatly, Andrew (2000). Critical Reading and Writing. London & New York: Routledge. Goldman, Harry (1983). “In Depth: Pornography: Introduction.” Journal of Popular Culture 17: 2. 123-128. Goldstein, Michael (1973). “Exposure to Erotic Stimuli and Sexual Deviance.” Journal of Social Issues 29. [No page numbers given in source]. Gottdiener, Mark (1985). “Hegemony and Mass Culture: A Semiotic Approach.” American Journal of Sociology 90: 5. 979-1001. Gramsci, Antonio (1971). Selection from the Prison Notebooks. Ed. and transl. Q. Hoare and G. Nowell Smith. London: Lawrence and Wishart. Grey, Antony (1993). Speaking of Sex. London & New York: Cassell. Griffin, Susan (1981). Pornography and Silence. Culture’s Revenge against Nature. New York: Harper and Row. Gutting, Gary (1994a). “Introduction: Michel Foucault: A User’s Manual.” In: Gutting (1994b). 1-27. Gutting, Gary (ed.) (1994b). The Cambridge Companion to Foucault. Cambridge: CUP. Habermas, Jürgen (1995 [originally 1981]). Theorie des kommunikativen Handelns. Band 1: Handlungsrationalität und gesellschaftliche Rationalisierung. Band 2: Zur Kritik der funktionalistischen Vernunft. Frankfurt a.M.: Suhrkamp. Haegeman, Liliane (1991). Government and Binding Theory. Oxford & Cambridge, MA: Blackwell. Halliday, Michael A.K. (1992). “New Ways of Meaning. The Challenge to Applied Linguistics.” In: Pütz (1992). 59-95 [originally (1990). Journal of Applied Linguistics 6. 7-36]. Halliday, Michael A.K. (1994, 2 nd edition). An Introduction to Functional Grammar. London et al.: Edward Arnold. Halliday, Michael A.K./ Christian M.I.M. Matthiessen (1999). Construing Experience through Language. A Language-based Approach to Cognition. London & NewYork: Cassell. Hamilton, Mykol Cecilia (1985). Linguistic Relativity and Sex Bias in Language: Effects of Masculine “Generics” on the Imagery of the Writer and the Perceptual Discrimination of the Reader. Dissertation, University of California: Ann Arbor, MI. Hampshire, Stuart (1978). “The Illusion of Sociobiology.” New York Review of Books 25. 64-69. Hardy, Simon (1998). The Reader, the Author, his Woman and her Lover. Soft- Core Pornography and Heterosexual Men. London & Washington: Cassell. <?page no="332"?> References 332 Harley, Trevor A. (2008, 3 rd edition). The Psychology of Language. From Data to Theory. Hove et al.: Psychology Press. Hartley, Nina (n.d.). Nina’s Official Web Site. [Online] http: / / www.nina.com (accessed January 2008). Harvey, Keith/ Celia Shalom (1997a). “Introduction.” In: Harvey/ Shalom (1997b). 1-17. Harvey, Keith/ Celia Shalom (eds.) (1997b). Language and Desire. Encoding Sex, Romance and Intimacy. London & New York: Routledge. Hebditch, David/ Nick Anning (1988). Porn Gold: Inside the Pornography Business. London: Faber and Faber. Heilbrun, Alfred B./ David. T. Seif (1988). “Erotic Value of Female Distress in Sexually Explicit Photographs.” The Journal of Sex Research 24. 47-57. Heise, Lori L. (1995). “Violence, Sexuality, and Women’s Lives.” In: Gagnon/ Parker (1995b). 109-134. Henderson, Lisa (1992). “Lesbian Pornography: Cultural Transgression and Sexual Demystification.” In: Munt (1992). 173-191. Henley, Nancy M./ Michelle Miller/ Jo Anne Beazley (1995). “Syntax, Semantics, and Sexual Violence. Agency and the Passive Voice.” Journal of Language and Social Psychology 14: 1-2. 60-84. Hill, Judith M. (1991). “Pornography and Degradation.” In: Baird/ Rosenbaum (1991b). 62-75 [originally (1987). Hypatia 2: 2. 39-54]. Hite, Shere (1976). The Hite Report: A Nationwide Study on Female Sexuality. New York: Macmillan. Hodge, Robert/ Gunther Kress (1993, 2 nd edition). Language as Ideology. London & New York: Routledge. Howard, Robert W. (1992). “Classifying Types of Concept and Conceptual Structure: Some Taxonomies.” European Journal of Cognitive Psychology 4: 2. 81-111. Howitt, Dennis/ Guy Cumberbatch (1990). Pornography: Impacts and Influences. A Review of Available Research Evidence on the Effects of Pornography. London: Home Office Research and Planning Unit. Hunt, Lynn (ed.) (1993). The Invention of Pornography: Obscenity and the Origin of Modernity, 1500-1800. New York: Zone Books. Hunter, Ian/ David Saunders/ Dugald Williamson (1993). On Pornography. Literature, Sexuality and Obscenity Law. Basingstoke, London & New York: Macmillan & St. Martin’s Press. Hutchby, Ian/ Robin Wooffitt (1998). Conversation Analysis. Principles, Practices and Applications. Cambridge et al.: Polity Press. Hyde, Janet Shibley/ John DeLamater (1999, 7 th edition). Understanding Human Sexuality. New York: McGraw-Hill. Itzin, Catherine (1992a). “Pornography and Capitalism: The UK Pornography Industry.” In: Itzin (1992c). 76-87. Itzin, Catherine (1992b).“‘Entertainment for Men’: What It Is and What It Means.” In: Itzin (1992c). 27-53. Itzin, Catherine (ed.) (1992c). Pornography. Women, Violence and Civil Liberties. Oxford: OUP. <?page no="333"?> References 333 Itzin, Catherine/ Corinne Sweet (1992). “Women’s Experience of Pornography: UK Magazine Survey Evidence.” In: Itzin (1992c). 222-235. Jackson, Howard (1990). Grammar and Meaning. A Semantic Approach to English Grammar. London & New York: Longman. Jary, David/ Julia Jary (1995, 2 nd edition). Collins Dictionary of Sociology. Glasgow: HarperCollins. Jensen, Robert (1995). “Pornography and the Limits of Experimental Research.” In: Dines/ Humez (1995). 298-305. Jensen, Robert (1998a). “The Pain of Pornography.” In: Dines/ Jensen/ Russo (1998). 155-163. Jensen, Robert (1998b). “Using Pornography.” In: Dines/ Jensen/ Russo (1998). 101-146. Jensen, Robert/ Gail Dines (1998). “The Content of Mass-Marketed Pornography.” In: Dines/ Jensen/ Russo (1998). 65-100. Johann, Sara Lee/ Franklin Mark Osanka (1989). Sourcebook on Pornography. Lexington, MA & Toronto: Lexington Books. Johnson-Laird, Philip N. (1983). Mental Models. Cambridge, MA: Harvard UP. Joseph, John E./ Talbot Taylor (eds.) (1990). Ideologies of Language. London: Routledge. Juffer, Jane (1998). At Home with Pornography. Women, Sex, and Everyday Life. New York & London: New York UP. Kaplan, E. Ann (1984). “Is the Gaze Male? ’ In: Snitow/ Stansell/ Thompson (1984b). 321-338. Kappeler, Susan (1986). The Pornography of Representation. Cambridge: Polity Press. Keating, Charles H. Jr. (1991). “A Dissent to The Report of the Commission on Obscenity and Pornography.” In: Baird/ Rosenbaum (1991b). 26-35 [from The Report of the Commission on Obscenity and Pornography (1970). 513- 516, 536-537, 542-549]. Kelly, Liz (1988). Surviving Sexual Violence. Minneapolis, MN: U of Minnesota P. Kendrick, Walter (1987). The Secret Museum: Pornography in Modern Culture. New York: Viking Penguin. Kettemann, Bernhard/ Georg Marko (eds.) (2003). Expanding Circles, Transcending Disciplines, and Multimodal Texts. Tübingen: Narr. Kettemann, Bernhard/ Georg Marko (eds.) (2006). Planing, Gluing and Painting Corpora. Inside the Applied Corpus Linguist’s Workshop. Frankfurt a.M.: Peter Lang. Kimmel, Michael S. (ed.) (1990). Men Confronting Pornography. New York: Crown Publishers. King, Bruce M. (1998). Human Sexuality Today. Englewood Cliffs, NJ: Prentice Hall. King, Helen (1994). “Sowing the Fields: Greek and Roman Sexology.” In: Porter/ Teich (1994b). 29-46. Kinsey, Alfred C./ Wardell B. Pomeroy/ Clyde E. Morton (1948). Sexual Behavior in the Human Male. Philadelphia: PA: W.B. Saunders. <?page no="334"?> References 334 Kinsey, Alfred C./ Wardell B. Pomeroy/ Clyde E. Morton/ Paul H. Gebhard (1953). Sexual Behavior in the Human Female. Philadelphia, PA: W.B. Saunders. Kipnis, Laura (1996). Bound and Gagged. Pornography and the Politics of Fantasy in America. New York: Grove Press. Kirsch-Rosenkrantz, Janis/ James H. Geer (1991). “Gender Difference in Memory for a Sexual Story.” Archives of Sexual Behavior 20: 3. 295-305. Kövecses, Zoltán (2002). Metaphor. A Practical Introduction. Oxford: OUP. Krafft-Ebing, Richard Freiherr von (1984 [1 st edition originally 1894]). Psychopathia Sexualis (mit besonderer Berücksichtigung der konträren Sexualempfindungen; eine medizinisch-gerichtliche Studie für Ärzte und Juristen). Ed. Alfred Fuchs. Reprint of 14 th edition 1912. With contribution by Georges Bataille. München: Matthes and Seitz. Krafka, Carol (1985). Sexually Explicit, Sexually Violent, and Violent Media: Effects of Multiple Naturalistic Exposures and Debriefing on Female Viewers. Doctoral dissertation, University of Wisconsin, Madison, WI. Kress, Gunther (1993). “Against Arbitrariness: The Social Production of the Sign as a Foundational Issue in Critical Discourse Analysis.” Discourse & Society 4: 2. 169-191. Kress, Gunther (1996). “Representational Resources and the Production of Subjectivity: Questions for the Theoretical Development of Critical Discourse Analysis in a Multicultural Society.” In: Caldas-Coulthard/ Coulthard (1996). 15-31. Kress, Gunther/ Theo van Leeuwen (1996). Reading Images. The Grammar of Visual Design. London & New York: Routledge. Krueger, Richard A./ Mary Anne Casey (2000). Focus Groups. Thousand Oaks, CA, et al.: Sage. Kuhn, Annette (1995). “Lawless Seeing.” In: Dines/ Humez (1995). 271-278. Kuno, Susumo (1987). Functional Syntax: Anaphora, Discourse and Empathy. Chicago: U of Chicago P. Kutchinsky, Berl (1970). “Toward an Explanation of the Decrease in Registered Sex Crimes in Copenhagen.” In: U.S. Commission on Obscenity and Pornography, Technical Report, Vol. 8. Kutchinsky, Berl (1991). “Pornography and Rape: Theory and Practice? Evidence from Crime Data in Four Countries where Pornography is Easily Available.” International Journal of Law and Psychiatry 14: 1-2. 47-64. Lab, Steven P. (1987). “Pornography and Aggression: A Response to the U.S. Attorney General’s Commission.” Criminal Justice Abstracts 19. 301-321. LaHaye, Tim (1991). “The Mental Poison.” In: Baird/ Rosenbaum (1991b). 177- 182 [excerpt from (1982). The Battle for the Family. No place given: Feming H. Revell Company]. Lakoff, George (1987). Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago: U of Chicago P. Lakoff, George/ Mark Johnson (1980). Metaphors We Live By. Chicago: U of Chicago P. <?page no="335"?> References 335 Lambrecht, Knud (1994). Information Structure and Sentence Form. Topic, Focus, and the Mental Representations of Discourse Referents. Cambridge: CUP. Lane, Frederick S. (2000). Obscene Profits. The Entrepreneurs of Porn in the Cyber Age. London & New York: Routledge. Langacker, Ronald (1987). Foundations of Cognitive Grammar. Volume 1: Theoretical Prerequisites. Stanford, CA: Stanford UP. Laumann, Edward O./ John H. Gagnon (1995). “A Sociological Perspective on Sexual Action.” In: Gagnon/ Parker (1995b). 183-213. Lawrence, David Herbert (1936). Phoenix. The Posthumous Papers of D.H. Lawrence. Edited by Harry T. Moore. London: Heinemann. Lawrence, Kelli-an/ Edward S. Herold (1988). “Women’s Attitudes Toward and Experience with Sexually Explicit Materials.” The Journal of Sex Research 24. 161-169. Lederer, Laura (ed.) (1980). Take Back the Night. Women on Pornography. New York: William Morrow and Company. Lee, David (1992). Competing Discourses. Perspective and Ideology in Language. London & New York: Longman. Leech, Geoffrey/ Roger Garside/ Michael Bryant (1994). “CLAWS4: The Tagging of the British National Corpus.” In: Proceedings of the 15 th International Conference on Computational Linguistics (COLING 94) Kyoto, Japan. 622-628 [online: http: / / www.comp.lancs.ac.uk/ computing/ research/ ucrel/ papers/ coling1994paper.pdf (accessed January 2008)]. Lenz, Friedrich (1997). Diskursdeixis im Englischen. Sprachtheoretische Überlegungen und lexiko-grammatische Analysen. Tübingen: Max Niemeyer Verlag. Levinson, Stephen C. (1983). Pragmatics. Cambridge: CUP. Linz, Daniel/ Edward Donnerstein (1990). “Evidence on the Causal Connection between Exposure to Penthouse Magazine and Anti-social Conduction.” Testimony before the Indecent Publications Tribunal, Wellington, New Zealand. Linz, Daniel/ Edward Donnerstein/ Stephen Penrod (1988). “The Effects of Long-term Exposure to Violent and Sexually Degrading Depictions of Women.” Journal of Personality and Social Psychology 55. 758-768. Linz, Daniel/ Neil Malamuth (1993). Pornography. Newbury Park, CA: Sage. Longino, Helen E. (1991). “Pornography, Oppression, and Freedom: A Closer Look.” In: Baird/ Rosenbaum (1991b). 84-95 [originally (1980). In: Lederer (1980). 40-54]. Lorde, Audre (1980 [originally 1978]). “Use of the Erotic: The Erotic as Power.” In: Lederer (1980). 295-300. Lorenz, Gunter R. (1999). Adjective Intensification - Learners versus Native Speakers. A Corpus Study of Argumentative Writing. Amsterdam & Atlanta, GA: Rodopi. Lovelace [Marciano], Linda, with Mike McGrady (1980). Ordeal. New York: Berkley. Lyons, John (1977). Semantics (2 vols.). Cambridge: CUP. <?page no="336"?> References 336 MacDonald, Scott (1995). “Confession of a Feminist Porn Watcher.” In: Dines/ Humez (1995). 307-313. MacKinnon, Catharine (1982). “Feminism, Marxism, Method and the State: An Agenda for Theory.” Signs 7: 3. 515-544. MacKinnon, Catharine (1989). “Sexuality, Pornography, and Method: ‘Pleasure under Patriarchy’.” Ethics 99. 314-346. MacKinnon, Catharine (1992a). “Pornography, Civil Rights and Speech.” In: Itzin (1992b). 456-511. MacKinnon, Catharine (1992b). “Does Sexuality Have a History? ” In: Stanton (1992b). 117-136. MacKinnon, Catharine (1993). Only Words. Cambridge, MA: Harvard UP. Malamuth, Neil M. (1986). “Predictors of Naturalistic Sexual Aggression.” Journal of Personality and Social Psychology 50: 5. 953-962. Malamuth, Neil M. (1989). “The Attraction to Sexual Aggression Scale: Part One.” Journal of Sex Research 26: 1. 26-49. Malamuth, Neil M./ Barry Spinner (1980). “A Longitudinal Content Analysis of Sexual Violence in the Best-selling Erotic Magazines.” Journal of Sex Research 16: 3. 226-237. Malamuth, Neil M./ James V.P. Check (1983). “Sexual Arousal to Rape Depictions: Individual Differences.” Journal of Abnormal Psychology 92: 1. 55-67. Malamuth, Neil M./ Edward I. Donnerstein (eds.) (1984). Pornography and Sexual Aggression. Orlando, FL: Academic Press. Malamuth, Neil M./ Robert L. Sockloskie/ Mary P. Koss/ Jeffrey S. Tanaka (1991). “Characteristics of Aggressors Against Women: Testing a Model using a National Sample of College Students.” Journal of Consulting and Clinical Psychology 59: 5. 670-681. Malamuth, Neil M./ Victoria Billings (1984). “Why Pornography? Models of Functions and Effects (Studies on Sex and Violence).” Journal of Communication 34: 3. 117-129. Manning, Elizabeth (1997). “Kissing and Cuddling. The Reciprocity of Romantic and Sexual Activity.” In: Harvey/ Shalom (1997b). 43-59. Manning, Rita C. (1991). “Redefining Obscenity.” In: Baird/ Rosenbaum (1991b). 153-163 [excerpt from an article that originally appeared in 1988 in Journal of Value Inquiry 22. 193-205]. Marko, Georg (1995). “Empathy in Women’s Magazines.” Arbeiten aus Anglistik und Amerikanistik 20: 2. 335-349. Marko, Georg (1997). Always Greener on the Other Side. A Critical Discourse Analysis of PR Material by the Austrian and the British Green Parties. Lancaster University: M.A. Dissertation. Marko, Georg (2003). “Nasty Data. Problems with Linguistic Data in Critical Discourse Analytical Research.” In: Kettemann/ Marko (2003). 163-176. Marko, Georg (2006). “‘… was an incredibly sexy Latin stud’. Critically Analysing Descriptors in a (Pornography) Corpus.” In: Kettemann/ Marko (2006). 175-203. Masters, William H./ Virginia E. Johnson (1966). Human Sexual Response. Boston: Little, Brown. <?page no="337"?> References 337 Masters, William H./ Virginia E. Johnson/ Robert Kolodny (1992, 4 th edition). Human Sexuality. New York: Harper Collins Publishers. Matacin, Mala L./ Jerry M. Burger (1987). “A Content Analysis of Sexual Themes in Playboy Cartoons.” Sex Roles 17: 3-4. 179-186. McCormack, Thelma (1985). “Making Sense of the Research on Pornography.” In: Burstyn (1985). 183-205. McIntosh, Mary (1981). “The Homosexual Role.” In: Plummer (1981). 30-44 [originally (1968). Social Problems 16: 2. 182-192]. McKay, H./ D. Dolff (1984). “Working Papers on Pornography and Prostitution. Report #13. The Impact of Pornography: An Analysis of Research and Summary Findings.” Policy Programs and Research Branch, Research and Statistics Section, Ottawa, Canada: Department of Justice. McKenzie-Mohr, Doug/ Mark P. Zanna (1990). “Treating Women as Sexual Objects: Look to the (Gender Schematic) Male who has Viewed Pornography.” Personality and Social Psychology Bulletin 16: 2. 296-308. McNair, Brian (1996). Mediated Sex. Pornography and Postmodern Culture. London et al.: Edward Arnold. Miall, David S. (1989). “Beyond the Schema Given: Affective Comprehension of Literary Narratives.” Cognition and Emotion 3: 2. 55-78. Michelson, Peter (1986). “Women and Pornorotica.” Another Chicago Magazine 16. 131-176. Mills, Sara/ Christine A. White (1997). “Discursive Categories and Desire. Feminists Negotiating Relationships.” In: Harvey/ Shalom (1997b). 222-244. Morgan, Robin (1980). “Theory and Practice: Pornography and Rape.” In: Lederer (1980). 134-140. Morokoff, Patricia J. (1985). “Effects of Sex Guilt, Repression, Sexual ‘Arousability’, and Sexual Experience on Female Sexual Arousal During Erotica and Fantasy.” Journal of Personality and Social Psychology 49: 1. 177-187. Morris, Madeline (1985). “Governmental Regulation of Pornography: Rhetoric of Harm.” Paper to the American Sociological Association. Munt, Sally (ed.) (1992). New Lesbian Criticism. New York: Columbia UP. Myers, Kathy (1995). “Towards a Feminist Erotica.” In: Dines/ Humez (1995). 263-270. Nemes, Irene (1992). “The Relationship between Pornography and Sex Crimes.” The Journal of Psychiatry and Law 20: 4. 459-481. O’Halloran, Kieran (2003). Critical Discourse Analysis and Language Cognition. Edinburgh: Edinburgh UP. O’Toole, Laurence (1998). Pornocopia. Pornography, Sex, Technology and Desire. London: Serpent’s Tail. Oakes, Michael P. (1998). Statistics for Corpus Linguistics. Edinburgh: Edinburgh UP. Pally, Marcia (1993). “Out of Sight and out of Harm’s Way.” Index of Censorship, no. 1. 4-6. Pally, Marcia (1994). Sex and Sensibilities. Reflections on Forbidden Mirrors and the Will to Censor. Hopewell, NJ: The Ecco Press. Palmer, Frank R. (1986). Mood and Modality. Cambridge: CUP. <?page no="338"?> References 338 Parker, Thomas (1991). “The Impact of Pornography on Marriage.” In: Baird/ Rosenbaum (1991b). 183-190 [excerpt from the 1989 Christian Life Commission Annual Seminar Proceedings]. Pease, Allison (2000). Modernism, Mass Culture, and the Aesthetics of Obscenity. Cambridge: CUP. Pennycook, Alastair (2001). Critical Applied Linguistics. A Critical Introduction. Mahwah, NJ & London: Lawrence Erlbaum. Plummer, Ken (1995). Telling Sexual Stories. Power, Change and Social Worlds. London & New York: Routledge. Plummer, Kenneth (ed.) (1981). The Making of the Modern Homosexual. London: Hutchinson. Popp, Wolfgang (1989). “Ist das Pornographie? Ein literaturwissenschaftliches Gutachten und sein Kontext.” Forum Homosexualität und Literatur 7. 71- 96. Porter, Roy (1994). “The Literature of Sexual Advice Before 1800.” In: Porter/ Teich (1994b). 134-157. Porter, Roy/ Mikuláš Teich (1994a). “Introduction.” In: Porter/ Teich (1994b). 1- 26. Porter, Roy/ Mikuláš Teich (eds.) (1994b). Sexual Knowledge, Sexual Science. The History of Attitudes to Sexuality. Cambridge: CUP. Posner, Michael I. (ed.) (1989). Foundations of Cognitive Science. Cambridge, MA & London: MIT Press. Potter, Jonathan/ Margaret Wetherell (1987). Discourse and Social Psychology. Beyond Attitudes and Behaviour. London & Newbury Park, CA: Sage. Prechtl, Peter/ Franz-Peter Burkard (eds.) (1996). Metzler Philosophie Lexikon. Begriffe und Definitionen. Stuttgart & Weimar: Metzler. Pryor, John B./ Lynnette M. Stoller (1994). “Sexual Cognition Processes in Men High in the Likelihood to Sexually Harass.” Personality and Social Psychology Bulletin 20: 2. 163-169. Pusch, Luise F. (1991 [originally 1984]). Das Deutsche als Männersprache. Frankfurt a.M.: Suhrkamp. Pütz, Martin (ed.) (1992). Thirty Years of Linguistic Evolution: Studies in Honour of Rene Dirven on the Occasion of his Sixtieth Birthday. Philadelphia & Amsterdam: John Benjamins. Qualls, C. Brandon/ John P. Wincze/ David H. Barlow (eds.) (1978). The Prevention of Sexual Disorders: Issues and Approaches. New York: Plenum. Quirk, Randolph/ Sidney Greenbaum/ Geoffrey Leech/ Jan Svartvik (1985). A Comprehensive Grammar of the English Language. London & New York: Longman. Rabinow, Paul (ed.) (1984). The Foucault Reader. New York: Pantheon. Radford, Jill/ Diana Russell (eds.) (1992). Femicide. The Politics of Woman Killing. Buckingham: Open UP. Rayner, Keith/ Alexander Pollatsek (1989). The Psychology of Reading. Englewood Cliffs, NJ, et al.: Prentice Hall. Renkema, Jan (1993). Discourse Studies. An Introductory Textbook. Amsterdam et al.: John Benjamins. <?page no="339"?> References 339 Report of the Commission on Obscenity and Pornography (1970). Committee chaired by William B. Lockhart, New York: Bantam Books. Richter, Alan (1987). The Language of Sexuality. Jefferson, NC & London: McFarland & Company. Richter, Alan (1993). Dictionary of Sexual Slang. Words, Phrases, and Idioms from AC/ DC to Zig-zig. New York et al.: John Wiley & Sons. Rickheit, Gert/ Hans Strohner (1993). Grundlagen der kognitiven Sprachverarbeitung. Tübingen & Basel: Francke. UTB 1735. Riffe, Daniel/ Stephen Lacy/ Frederick Fico (1998). Analyzing Media Messages: Using Quantitative Content Analysis in Research. Mahwah, NJ: Lawrence Erlbaum. Rose, Diana (2001). Visual Methodologies. An Introduction to the Interpretation of Visual Materials. London et al.: Sage. Rosen, Ruth (1994). “Not Pornography! ” Dissent 41 (Summer). 343-345. Rosenfield, Lawrence W. (1973). “Politics and Pornography.” Quarterly Journal of Speech 59. [No page numbers given in source]. Rouse, Joseph (1994). “Power/ Knowledge.” In: Gutting (1994b). 92-114. Royalle, Candida (2000). “Porn in the USA.” Put together from interviews with Candida Royalle conducted by Anne McClintock in New York, 1991-1993, and an oral presentation by Royalle at Columbia University. In: Cornell (2000). 540-550 [originally (1993). Social Text (Winter). 23-32]. Rubin, Gayle (1995). “Misguided, Dangerous and Wrong. An Analysis of Anti- Pornography Politics.” In: Dines/ Humez (1995). 244-253. Rumelhart, David E. (1980). “Schemata: The Building Blocks of Cognition.” In: Spiro/ Bruce/ Brewer (1980). 33-58. Russell, Diana E. (1998). Dangerous Relationships. Pornography, Misogyny, and Rape. Thousand Oaks, CA, et al.: Sage. Russell, Diana E. (1980). “Pornography and Violence: What does the New Research Say.” In: Lederer (1980). 218-238. Russo, Ann (1998a). “Feminists Confront Pornography’s Subordinating Practices.” In: Dines/ Jensen/ Russo (1998). 9-35. Russo, Ann (1998b). “‘Feeding People in all their Hungers.’ One Woman’s Attempt to Link the Struggle against Violence with the Struggle for Sexual Freedom.” In: Dines/ Jensen/ Russo (1998). 147-153. Ryan, Ann/ Alison Wray (eds.) (1997). Evolving Models of Language. Clevedon: Multilingual Matters Saeed, John I. (1997). Semantics. Oxford et al.: Blackwell. Salamun, Kurt (ed.) (1986, 2 nd edition). Was ist Philosophie? Tübingen: J.C.B. Mohr (Paul Siebeck). UTB 1000. SAMOIS (1987 [originally 1981]). Coming to Power. Writings and Graphics on Lesbian S/ M. Boston: Alyson Publications. Saunders, Richard M./ Peter J. Naus (1993). “The Impact of Social Content and Audience Factors on Responses to Sexually Explicit Videos.” Journal of Sex Education and Therapy 19: 2. 117-130. Savitz, Leonard/ Norman Johnson (1976). “Crime and Society.” Psychology Today 61. 6. <?page no="340"?> References 340 Schacter, Daniel L. (1989). “Memory.” In: Posner (1989). 683-725. Schank, Roger C./ Robert P. Abelson (1977). Scripts, Plans, Goals and Understanding. Hillsdale, NJ: Lawrence Erlbaum. Schill, Thomas R./ James Chapin (1972). “Sex Guilt and Males’ Preference for Reading Erotic Magazines.” Journal of Consulting and Clinical Psychology 39. 516. Schmidt, Gunter (1985). “Pornotopia - Schlaraffenland der Sexualität.” Pro Familia Magazin 6. 20-22. Schwarz, Monika (1992). Einführung in die Kognitive Linguistik. Tübingen: Francke. UTB 1636. Scott, Joseph/ Loretta Schwalm (1988a). “Pornography and Rape: An Examination of Adult Theater Rates and Rape Rates by State.” In: Scott/ Hirschi (1988). 40-53. Scott, Joseph/ Loretta Schwalm (1988b). “Rape Rates and the Circulation Rates of Adult Magazines.” Journal of Sex Research 24. 241-250. Scott, Joseph/ Steven Cuvelier (1987). “Violence in Playboy Magazine: A Longitudinal Content Analysis.” Archives of Sexual Behavior 16. 279-288. Scott, Joseph/ Steven Cuvelier (1993). “Violence and Sexual Violence in Pornography: Is it Really Increasing? ” Archives of Sexual Behavior 22: 4. 357-371. Scott, Joseph/ Travis Hirschi (eds.) (1988). Controversial Issues in Crime and Justice. Beverly Hills, CA: Sage. Segal, Lynne (1990). “Pornography and Violence: What the ‘Experts’ Say.” Feminist Review 36. 29-41. Segal, Lynne/ Mary McIntosh (eds.) (1993). Sex Exposed: Sexuality and the Pornography Debate. New Brunswick, NJ: Rutgers UP. Semino, Elena (1997). Language and World Creation in Poems and Other Texts. London & New York: Longman. Senn, Charlene (1993). “Women’s Multiple Perspectives and Experiences with Pornography.” Psychology of Women Quarterly 17: 3. 319-341. Senn, Charlene/ Lorraine Radtke (1986). “A Comparison of Women’s Reactions to Violent Pornography, Non-violent Pornography, and Erotica.” Paper presented at the Canadian Psychological Association, June 1986, Toronto. Senn, Charlene/ Lorraine Radtke (1990). “Women’s Evaluations of and Affective Reactions to Mainstream Violent Pornography, Nonviolent Pornography, and Erotica.” Violence and Victims 5: 3. 143-155. Short, Mick (1996). Exploring the Language of Poems, Plays and Prose. London: Longman. Siewierska, Anna (1991). Functional Grammar. London & New York: Routledge. Silbert, Mimi H./ Ayala M. Pines (1984). “Pornography and Sexual Abuse of Women.” Sex Roles 10: 11/ 12. 857-869. Singer, Murray (1990). Psychology of Language. An Introduction to Sentence and Discourse Processing. Hillsdale, NJ: Lawrence Erlbaum. Singer, Murray/ Fernanda Ferreira (1983). “Inferring Consequences in Story Comprehension.” Journal of Verbal Learning and Verbal Behavior 22: 4. 437-448. <?page no="341"?> References 341 Smith, Anna Marie (1993). “‘What is Pornography? ’: An Analysis of the Policy Statement of the Campaign against Pornography and Censorship.” Feminist Review 43. 71-87. Smith, Carlota (1991). The Parameter of Aspect. Dordrecht: Kluwer. Smith, Don D. (1976). “The Social Content of Pornography (Explicit Sex - Liberation or Exploitation? ).” Journal of Communication 27: 1. 16-24. Smith, Edward E. (1995). “Concepts and Categorization.” In: Smith/ Osherson (1995). 3-33. Smith, Edward E./ Daniel N. Osherson (eds.) (1995, 2 nd edition). Thinking. An Invitation to Cognitive Science, Volume 3. Cambridge, MA & London: A Bradford Book - MIT Press. Snitow, Ann Barr/ Christine Stansell/ Sharon Thompson (1984a). “Introduction.” In: Snitow/ Stansell/ Thompson (1984b). [No page numbers given in source]. Snitow, Ann/ Christine Stansell/ Sharon Thompson (eds.) (1984b [originally 1983]). Powers of Desire: The Politics of Sexuality. London: Virago. Soble, Alan (1985). Pornography, Marxism, Feminism, and the Future of Sexuality. New Haven, CT: Yale UP. Soble, Alan (1991). “Defamation and the Endorsement of Degradation.” In: Baird/ Rosenbaum (1991b). 96-107 [excerpt from an article that originally appeared in 1985 in Social Theory and Practice 11: 1. 61-87]. Spender, Dale (1985, 2 nd edition). Man Made Language. London et al.: Routledge and Kegan Paul. Spiro, Rand/ Bertram C. Bruce/ William F. Brewer (eds.) (1980). Theoretical Issues in Reading Comprehension. Hillsdale, NJ: Lawrence Erlbaum. Sprinkle, Annie (1998). Post Porn Modernist - My Twenty Five Years as a Multi-media Whore. San Francisco, CA: Cleis Press. Stanton, Domna C. (1992a). “Introduction: the Subject of Sexuality.” In: Stanton (1992b). 1-46. Stanton, Domna C. (ed.) (1992b). Discourses of Sexuality. From Aristotle to AIDS. Ann Arbor, MI: The U of Michigan P. Steffensen, Margaret S./ Chitra Joag-Dev (1984). “Cultural Knowledge and Reading.” In: Alderson/ Urquhart (1984b). 48-61. Steinem, Gloria (1980 [originally 1978]). “Erotica and Pornography: A Clear and Present Difference.” In: Lederer (1980). 35-39. Stillings, Neil A./ Steven E. Weisler/ Christopher H. Chase/ Mark H. Feinstein/ Jay L. Garfield/ Edwina Rissland (1995, 2 nd edition). Cognitive Science. An Introduction. Cambridge, MA & London: A Bradford Book - MIT Press. Stoller, Robert (1976). “Sexual Excitement.” Archives of General Psychiatry 33. 899-909. Stoller, Robert (1991). Porn: Myths for the Twentieth Century. Yale: Yale UP. Strossen, Nadine (1995). Defending Pornography. Free Speech, Sex, and the Fight for Women’s Rights. New York et al.: Scribner. Stubbs, Michael (1983). Discourse Analysis. The Sociolinguistic Analysis of Natural Language. Oxford et al.: Blackwell. Stubbs, Michael (1996). Text and Corpus Analysis. Computer-assisted Studies of Language and Culture. Oxford & Cambridge, MA: Blackwell. <?page no="342"?> References 342 Stubbs, Michael (1997). “Whorf’s Children: Critical Comments on Critical Discourse Analysis.” In: Ryan/ Wray (1997). 100-116. Surgeon General’s Workshop on Pornography and Public Health (1986). Report prepared by Edward Mulvey and Jeffrey Haugaard: U.S. Department of Health and Human Services. Talbot, Mary (1997). “‘An Explosion Deep Inside Her.’ Women’s Desire and Popular Romance Fiction.” In: Harvey/ Shalom (1997b). 106-122. Tallerman, Maggie (1998). Understanding Syntax. London et al.: Edward Arnold. Tang, Isabel (1999). Pornography. The Secret History of Civilisation. London: Channel 4 Books. Terry, Roger L. (1983/ 84). “A Connotative Analysis of Synonyms for Sexual Intercourse.” Maledicta 7. 237-253. Thomas, Jenny (1995). Meaning in Interaction. An Introduction to Pragmatics. London: Longman. Thompson, Bill (1994). Soft Core. London: Cassell. Thompson, Geoff (1996). Introducing Functional Grammar. London et al.: Edward Arnold. Tisdale, Sallie (1995 [originally 1994]). Talk Dirty to Me. An Intimate Philosophy of Sex. London & Basingstoke: Pan Books. Toolan, Michael J. (1988). Narrative. A Critical Linguistic Introduction. London & New York: Routledge. Toolin, Cynthia (1983). “In Depth: Pornography: Attitudes Toward Pornography: What Have the Feminists Missed? ” Journal of Popular Culture 17: 2. 167-174. Torfing, Jacob (1999). New Theories of Discourse. Laclau, Mouffe and Zizek. Oxford: Blackwell. Treibel, Annette (2000, 5 th edition). Einführung in die soziologischen Theorien der Gegenwart. Opladen: Leske & Budrich. Trostle, Lawrence C. (1993). “Pornography as a Source of Sex Information for University Students: Some Consistent Findings.” Psychological Reports 72: 2. 407-412. Trudgill, Peter (1974). Sociolinguistics. An Introduction to Language and Society. Harmondsworth: Penguin (Pelican Book). Turner, Robin (1999). “Debating Pornography: Categories and Metaphors.” [Online] http: / / neptune.spaceports.com/ ~words/ debating.html (accessed April 2007). Ungerer, Friedrich/ Hans-Jörg Schmid (1996). An Introduction to Cognitive Linguistics. London & New York: Longman. van Dijk, Teun (1987). Communicating Racism. Ethnic Prejudice in Thought and Talk. Newbury Park, CA, et al.: Sage. van Dijk, Teun (1988). News as Discourse. Hillsdale, NJ: Lawrence Erlbaum. van Dijk, Teun (1990). “Discourse & Society: A New Journal for a New Research Focus.” Discourse & Society 1: 1. 5-16. van Dijk, Teun (1993). “Principles of Critical Discourse Analysis.” Discourse & Society 4: 2. 249-283. <?page no="343"?> References 343 van Dijk, Teun (1996). “Discourse, Power and Access.” In: Caldas- Coulthard/ Coulthard (1996). 84-104. van Dijk, Teun (1997a). “The Study of Discourse.” In: van Dijk (1997c). 1-34. van Dijk, Teun (1997b). “Discourse as Interaction in Society.” In: van Dijk (1997d). 1-37. van Dijk, Teun (ed.) (1997c). Discourse as Structure and Process. Discourse Studies 1. A Multidisciplinary Introduction. London & Thousand Oaks, CA & New Delhi: Sage. van Dijk, Teun (ed.) (1997d). Discourse as Social Interaction. Discourse Studies 2. A Multidisciplinary Introduction. London & Thousand Oaks, CA & New Delhi: Sage. van Dijk, Teun (1998). Ideology. A Multidisciplinary Approach. Thousand Oaks, CA, et al.: Sage. van Dijk, Teun/ Walter Kintsch (1983). Strategies of Discourse Comprehension. New York: Academic Press. van Leeuwen, Theo (1993). “Genre and Field in Critical Discourse Analysis: A Synopsis.” Discourse & Society 4: 2. 193-223. van Valin, Robert D. Jr./ Randy J. LaPolla (1997). Syntax. Structure, Meaning and Function. Cambridge: CUP. Vance, Carole S. (ed.) (1992 [originally 1984]). Pleasure and Danger: Exploring Female Sexuality. With a new introduction by the editor. London: Pandora. Wales, Katie (ed.) (1994). Feminist Linguistics in Literary Criticism. Woodbridge: English Association/ D.S. Brewer. Wareing, Shan (1994). “‘And then he kissed her’. The Reclamation of Female Characters to Submissive Roles in Contemporary Fiction.” In: Wales (1994). 117-136. Weaver, Mary Jo (1991). “Pornography and the Religious Imagination.” In: Baird/ Rosenbaum (1991b). 191-213 [originally (1989). In: Susan Grubar/ Joan Hoff (eds.) (1989). For Adults Users Only. Bloomington, IN: Indiana UP]. Weeks, Jeffrey (1977). Coming Out: Homosexual Politics in Britain from the Nineteenth Century to the Present. London: Quartet. Weinstein, James (1998). Hate Speech, Pornography, and Radical Attacks on Free Speech Doctrine. Boulder, CO: Perseus Books, Westerview Press. Werlen, Iwar (1989). Sprache, Mensch und Welt. Geschichte und Bedeutung des Prinzips der sprachlichen Relativität. Darmstadt: Wissenschaftliche Buchgesellschaft. Werth, Paul (1999). Text Worlds: Representing Conceptual Space in Discourse. Harlow: Longman. Whaley, C. Robert/ George Antonelli (1983/ 84). “The Birds and the Beasts. Woman as Animals.” Maledicta 7. 219-229. Whorf, Benjamin Lee (1956). Language, Thought and Reality. Selected Writings of Benjamin Lee Whorf. Ed. John B. Carroll. Foreword by Stuart Chase. Cambridge, MA: MIT Press. Widdowson, Henry (1979). Explorations in Applied Linguistics. Oxford: OUP. <?page no="344"?> References 344 Widdowson, Henry (1995). “Discourse Analysis: A Critical View.” Language and Literature 4: 3. 157-172. Widdowson, Henry (1996). “Reply to Fairclough: Discourse and Interpretation: Conjectures and Refutations.” Language and Literature 5: 1. 57-69. Widdowson, Henry (2004). Text, Context, Pretext. Critical Issues in Discourse Analysis. Oxford: Blackwell. Wierzbicka, Anna (1992). Semantics, Culture, and Cognition. Oxford: OUP. Williams, Bernard (1979). Report of the Committee on Obscenity and Film Censorship. Cmnd. 7772, London: Her Majesty’s Stationery Office. Williams, Linda (1999, expanded paperback edition [originally 1989]). Hard Core. Power, Pleasure, and the “Frenzy of the Visible.” Berkeley, CA, et al.: U of California P. Wilson, W. Cody (1978). “Can Pornography Contribute to the Prevention of Sexual Problems? ” In: Qualls/ Wincze/ Barlow (1978). [No page numbers given in source]. Wodak, Ruth (1996). Disorders of Discourse. London & New York: Longman. Wodak, Ruth/ Bernd Matouschek (1993). “‘We are Dealing with People whose Origins One can Clearly Tell Just by Looking’: Critical Discourse Analysis and the Study of Neo-Racism in Contemporary Austria.” Discourse & Society 2: 4. 225-248. Wodak, Ruth/ Florian Menz/ Richard Mitten/ Frank Stern (1990). “Wir sind alle unschuldige Täter! ” Diskurshistorische Studien zum Nachkriegsantisemitismus. Frankfurt a.M.: Suhrkamp. Wodak, Ruth/ Rudolf de Cillia/ Martin Reisigl/ Karin Liebhart/ Klaus Hofstätter/ Maria Kargl (1998). Zur diskursiven Konstruktion nationaler Identität. Frankfurt a.M.: Suhrkamp. Wood, Linda A./ Rolf O. Kroger (2000). Doing Discourse Analysis. Methods for Studying Action in Talk and Text. Thousand Oaks, CA, et al.: Sage. Zillmann, Dolf (1984). Connections between Sex and Aggression. Hillsdale, NJ: Lawrence Erlbaum. Zillmann, Dolf/ Jennings Bryant/ Rodney A. Carveth (1981). “The Effect of Erotica Featuring Sadomasochism and Bestiality on Motivated Intermale Aggression.” Personality and Social Psychology Bulletin 7. 153-159. Zillmann, Dolf/ Jennings Bryant (1982). “Pornography, Sexual Callousness, and the Trivialization of Rape.” Journal of Communication 32: 4. 10-21. Zillmann, Dolf/ Jennings Bryant (1984). “Effects of Massive Exposure to Pornography.” In: Malamuth/ Donnerstein (1984). 115-138. Zillmann, Dolf/ Jennings Bryant (1988a). “Effects of Prolonged Consumption of Pornography on Family Values.” Journal of Family Issues 9: 4. 518-544. Zillmann, Dolf/ Jennings Bryant (1988b). “Pornography’s Impact on Sexual Satisfaction.” Journal of Applied Social Psychology 18: 5. 438-453. Zillmann, Dolf/ Jennings Bryant (1988c). “A Response [to Donnerstein and Linz (1988)] (Colloquy).” Journal of Communication 38: 2. 185-192. Zillmann, Dolf/ Jennings Bryant (eds.) (1989). Pornography: Recent Research, Interpretations, and Policy Considerations. Hillsdale, NJ: Lawrence Erlbaum. <?page no="345"?> References 345 Sources Pornography corpus Dear Diary (1994: 5). Malibu: Gold Coast Press. Down-n-Nasty (1995: September). New York: Vanity Publishing Company. Erotic Stories (1994: May and 1995: February and April). London: Northern Shell. Erotica Pen Pals (1994: 6). Baltimore: Komar Co. Experience Readers’ Erotic Letters (n.d.). London: Bellright Ltd. Experience (1994: 8). London: Bellright Ltd. Family Secrets (1995: August). New York: Vanity Publishing Company. Hustler Fantasies (1995: September). Beverly Hills: HG Publications. Luv Letters (1994: 5). Malibu: Gold Coast Press. Mid-night (1995: January). Teaneck: Second Wind Publications. Night (1995: October). Teaneck: Candy Publications. P.S. I Love You (1994: 10). Malibu: Gold Coast Press. Penthouse Forum (1995: July). New York: Forum International Ltd. Silky Secrets (1994: 3). Baltimore: Komar Co. Turn-Ons (1995: October). Portchester: AJA Publishing Corp. Vibrations (1994: 6). London: Bellright Ltd. Wet (1995: October). Port Chester: Sportomotic Ltd. Wild (1994: 3). Baltimore: Komar Co. Sources erotica corpus Barbach, Lonnie (ed.) (1985 [originally 1984]). Pleasures. Women Write Erotica. New York: Harper Perennial. Bright, Susie (ed.) (1994). The Best American Erotica 1994. New York: Touchstone Book by Simon & Schuster. Slung, Michelle (ed.) (1995 [originally 1994]). Fever. Sensual Stories by Women Writers. New York: Harper Perennial. <?page no="347"?> Index To facilitate orientation, I have marked the passages in bold which contain the central information on a specific concept. Words of omnipresent importance to my study, such as pornography, erotica, corpus, etc., are only included with a reference to the place where they are introduced and/ or defined. For further definitions and explanations of important terms - in particular the non-canonical ones - see the glossary on my homepage http: / / unigraz.at/ georg.marko adjective phrase/ AP 145, 172-173, 178, 181-182, 184, 186, 204, 205, attributive 120, 178-223, 315 predicative 182-184, 219, 231 adverb phrase/ AdvP 119, 145, 187, 229, 247 agentivity scale 265-266, 270, 273 agentization 259, 262, 265, 274- 282, 287-289, 296-297, 300, 303, 307-308, 311 archaeology of knowledge 78 assertion 120, 183, 184 catharsis 32, 40, 321 Christian moralists/ conservatives 21, 23-33, 37, 44, 46, 49-50, 57, 321 civil right ordinance 37 clinical studies (research on pornography) 50, 52-53 collocation 113, 147-148, 186 concept (cognitive psychology) 19, 26-27, 40, 44, 70-72, 75, 86-88, 98, 118, 283-287, 290-293, 295, 297 conceptual structure (level of meaning) 118-119, 124, 129, 130 connotation 120, 191, 286 content analysis (research on pornography) 20, 57, 61-64 contextual world (level of meaning) 118, 121, 124 corpus (analysis) (definition) 93, 112-115 correlational studies (research on pornography) 50-52 co-textual analysis 128 Critical Discourse Analysis/ CDA 12-13, 15-16, 49, 65, 67-101, 103, 111, 115-116, 122-224, 284-286, 313, 317-319, 321 critical theory 84-89 Critical Theory (Frankfurt School) 85 descriptor 129, 172, 178, 179-223, 233, 250, 315, 317-318, 320 desubjectification 131, 133, 225- 257, 262, 270-271, 273, 314-315, 316, 319 discourse (analysis) (definition) 78- 81 Dworkin, Andrea 21, 33, 36-37, 39, 41, 130, 142, 226, 260, 328 dysphemism 286 <?page no="348"?> Index 348 empirical (data) 22, 49-50, 55, 57, 64, 67, 92-95, 101, 103, 111-112, 317, 319 episodic knowledge/ memory 70 erotica (definition) 38-39 euphemism 286 evaluative structure (level of meaning) 118-119, 123, 129, 130 evaluative survey (opinion poll) (research on pornography) 57-59 experimental studies (research on pornography) 50, 53-56, 62 Fairclough, Norman 67-68, 80-82, 85, 87, 91-92, 94-95, 101, 118, 122, 123, 144, 326, 329, 344 feminism/ feminist 21-23, 33-46, 50-51, 56-57, 59, 63, 69, 83, 88, 91, 109, 130, 133, 142, 191, 259- 260, 309, 313, 321 anti-censorship feminism 23, 41- 46 anti-pornography feminism 23, 25, 33-40, 41-46, 52, 57-58, 63, 69, 260 First Amendment (free speech) 24, 27, 30, 37-38 first person narrator 138, 145, 154, 175, 223, 228, 234, 240, 253, 255, 257, 319 Foucault, Michel 34, 57, 78, 80, 85- 86, 111, 329, 330, 331, 338 Fowler, Roger 16, 67-68, 83, 90, 91, 330 fragmentation 36, 89, 99, 131-133, 138, 141-175, 203-204, 236, 240, 262, 265-266, 274, 277,-278, 289, 314-316, 318 full scenario (level of meaning) 117-118, 120-121, 125, 129 generativism 82 global conceptualization 99, 130- 132, 137, 178, 245, 254, 259 Halliday, Michael A.K. 82-83, 125, 182-183, 207, 228-231, 263-264, 270, 331 harm (issue) 22-25, 26-30, 34, 36- 37, 40, 47, 49, 53, 58-59, 61, 64, 68-70, 79, 84-86, 88, 91, 102, 130, 254, 313-314, 321 hermeneutics 14, 49-50, 56-57, 63- 64, 67-69, 84, 91-93, 99-100, 127, 317 Hodge, Bob 67, 330, 332 idealized discourse participant 97- 98, 116-129, 319 identifiability 120 ideology 13, 15, 41, 47, 86-89, 90- 92, 100, 123, 260, 286, 314, 322 inference 73, 117, 120-125, 129, 144 information structure (level of meaning) 117-118, 120, 123, 125, 143-144, 165, 181, 246 global information structure 118, 121 local information structure 118, 120, 129, 130, 165, 183 intensifier 114, 187, 205-206, 247 interactive structure (level of meaning) 119, 123 interjection 220-221 Internet 12, 20, 21, 127 intersubjectivity 64, 67, 92-94, 96, 100, 137, 260 intertextuality 81, 96, 105, 108, 110, 122-123, 144, 157, 267 Kress, Gunther 67, 94-95, 330, 332, 334 Lawrence, D.H. 27, 31, 44, 335 liberals 23-24, 28-32, 33-46, 50, 52, 57, 321 MacKinnon, Catharine 19-21, 33- 38, 130, 142, 261, 328, 336 <?page no="349"?> Index 349 marked/ unmarked 246-247 meaning actual meaning 80, 83, 319 contextual meaning 117-118, 120, 125 evaluative meaning 117, 144 ideational meaning 117-118, 120-121, 123, 125, 144, 165, 183, 192, 220 interactive meaning 117 meaning potential 80, 83 mental process adjective 231-245 mental process verbs 228-245, 255-257, 315 metaphor 144, 233, 255, 283-310 conceptual metaphor 260-261, 283-311, 316 source 284-297, 304, 307, 309 target 283-292, 308 metonymy 284-285, 290, 292, 295, 304 conceptual metonymy 284-285, 296, 308-309 modality 119-120, 129, 235 multimodality 94, 111 noun phrase/ NP 114, 119, 145- 146, 165-171, 181-184, 186, 204- 205, 219, 229, 232-234, 261-263, 266, 270 genitive NP 145-147, 165-171, 174, 204, 314 predicative NP 184, 186, 193- 201, 219, 222 objectification 15, 36, 45, 88, 100, 115, 130-133, 141, 174-175, 177- 178, 180, 185, 194, 197, 199, 201, 203, 205, 208, 210, 212, 215, 217, 219, 222-223, 225-226, 228, 252, 254-257, 259-260, 282, 296-297, 310, 314, 316 obscenity laws 27, 37 orthographic word 103, 108, 112, 114 overwording/ -lexicalization 144, 152, 156-157, 165, 206, 209-210, 220, 222, 295, 296 passive voice 227, 231, 246-254, 273, 315 passivization 89, 131-133, 138, 173, 203, 234, 252-255, 259-311, 314, 316, 319 perspective/ point of view 107, 117, 120-121, 131, 219-220, 222, 225- 228, 230, 235, 246-247, 249-257, 271, 282, 315-316, 318-320 physicalization and visualization 131, 133, 177-224, 225-226, 231, 239, 304, 314-316, 320 pornography (definition) 19-22 power 34, 36-37, 39, 42-43, 47, 62, 67, 86-91, 101, 157, 165, 174, 182, 203-204, 206, 208, 215, 219, 223, 226, 254, 288, 315 pragmatics 84, 220 prepositional phrase/ PP 119, 145, 179, 181, 229, 232, 246, 250, 262- 264, 266, 269 presupposition 120, 129, 183, 185, 194, 222 process types intensive process 231 material process 179, 229, 231, 263-264, 267 mental process 79, 220, 228-245, 256-257 mental process (emotion) 225- 245, 257, 316 mental process (cognition) 116- 117, 124, 125, 133, 225-245, 257, 316 mental process (perception) 222, 225-245, 257, 316 verbal process 229, 263-264 proposition 118-119, 125, 183 rape 36, 51-54, 58, 59, 61, 104, 314 rape myths 36, 52-53 <?page no="350"?> Index 350 reciprocal ergative 270-271, 310 reciprocal pronoun 270-271 salience 121-122, 124, 129-130, 142-144, 149-158, 164-166, 171- 172, 180, 194, 208, 210, 222, 230, 236, 239, 245, 247, 252, 256-257, 261, 266, 271, 282, 286-287, 289, 299, 303, 315 Sapir-Whorf-Hypothesis 35, 83 schema 55, 71-78, 99, 121, 286 script 75, 78, 260 semantic grid (level of meaning) 117-118, 120, 122, 124-125, 129- 130 semantic knowledge/ memory 70- 74, 77, 81, 93-97, 100-101, 116, 121-124, 128, 130 semantic roles 119, 138, 172-173, 228-229, 233-236, 247, 252, 260- 283, 288, 295, 309-311, 315-316 agent 95, 133, 172, 229, 260- 278, 296, 297, 300, 310 circumstance 226, 228-229, 262- 264, 266 comitative 263-266, 274, 278 directive 84, 138, 263-266, 272 instrument 229, 264 location 263-265 locative 229, 262-263, 272, 274, 278, 288, 295 participant 121-122, 124, 129, 132, 179, 228-229, 252, 262-277 path 255, 263-265, 88 patient 95, 100, 172, 229, 231, 247, 252, 262-274, 278 phenomenon 226, 229, 254, 256- 257 receiver 247, 252, 263-264, 273 sayer 264, 273 senser 228-245, 256, 315 source 263-266 temporal 262 verbiage 264, 273 sex crimes 15, 50-52 sexology 49-56, 57, 63-65, 67 sexual aggression/ violence 23, 33, 36-37, 53, 56, 61, 78, 104, 177, 254, 316 social constructionism 33-35, 42, 56-57, 69-80, 96, 116, 321 sociolinguistics 84 structuralism 50, 82 subject (grammatical) 222, 229, 246-257, 261-263, 270, 315 subjectification 225-230, 235, 239, 242, 244-246, 254-257, 260, 282, 283, 311, 315, 318 subordination (social) 36, 100, 130, 288, 313, 314 tagging/ tags 114-115, 145-146, 148, 187, 204, 205, 232, 233, 247, 248, 268 CLAWS-tagger 114 testimonial studies (research on pornography) 57, 59-61 textual world (level of meaning) 118, 121, 124 thematic structure 120, 183, 246 focus 121, 138, 184 rheme 120, 183-184 theme 120, 183, 246, 255 transitivity 228 transtextual analysis 128-130, 254, 317-318 van Dijk, Teun 68, 71-72, 75, 78- 79, 86-88, 90-91, 119, 122, 125, 329, 342, 343 van Leeuwen, Theo 94-95, 334, 343 verb phrase/ VP 145 Wodak, Ruth 68, 90-91, 329, 344 WordSmith 104, 112-113, 147-148, 248