<?page no="0"?> narr studienbücher Thomas Herbst Susen Schüller Introduction to Syntactic Analysis A Valency Approach <?page no="1"?> narr studienbücher <?page no="3"?> Thomas Herbst / Susen Schüller Introduction to Syntactic Analysis A Valency Approach Gunter Narr Verlag Tübingen <?page no="4"?> Prof. Dr. Thomas Herbst ist Inhaber des Lehrstuhls Anglistik: Linguistik und Sprecher des Interdisziplinären Zentrums für Lexikografie, Valenz- und Kollokationsforschung an der Friedrich-Alexander-Universität Erlangen-Nürnberg. Susen Schüller, M.A. ist Assistentin am Lehrstuhl Anglistik: Linguistik an der Friedrich- Alexander-Universität Erlangen-Nürnberg. 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. © 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 chlorfrei gebleichtem und säurefreiem Werkdruckpapier. Internet: http: / / www.narr-studienbuecher.de E-Mail: info@narr.de Druck: Gulde, Tübingen Bindung: Nädele, Nehren Printed in Germany ISSN 0941-8105 ISBN 978-3-8233-6390-3 <?page no="5"?> Table of contents Preface ................................................................................................................................................ xi 1 Preliminary remarks about syntactic analysis ............................................... 1 1.1 Syntax and lexis ................................................................................................................ 1 1.2 Sentence and clause ......................................................................................................... 3 1.3 Clause constituents .......................................................................................................... 4 1.3.1 Clause constituents at different levels.............................................................................. 4 1.3.2 Identifying clause constituents ......................................................................................... 6 1.3.3 Relations between constituents ......................................................................................... 9 1.4 The importance of terminology ................................................................................. 12 1.5 The aims of this book .................................................................................................... 14 2 The syntactic framework ........................................................................................ 16 2.1 Survey of the framework sugge sted.......................................................................... 16 2.2 Elements of clause structure ....................................................................................... 18 2.2.1 Subject................................................................................................................................. 18 2.2.2 Adjuncts ............................................................................................................................. 19 2.2.3 Predicate ............................................................................................................................. 20 2.3 Valency .............................................................................................................................. 21 2.3.1 Complements and adjuncts ............................................................................................. 21 2.3.2 Valency complements as constituents in clause structure .......................................... 23 2.4 Phrases and clauses ....................................................................................................... 24 2.4.1 Phrases ................................................................................................................................ 24 2.4.2 The structure of the phrase .............................................................................................. 25 2.4.3 Clauses as verb phrases.................................................................................................... 27 2.5 Coordination and subordination ............................................................................... 27 2.5.1 Coordination...................................................................................................................... 27 2.5.2 Coordinated and subordinate clauses............................................................................ 28 2.5.3 Clause and sentence.......................................................................................................... 29 3 Word classes .................................................................................................................. 31 3.1 Words - word classes .................................................................................................... 31 3.1.1 Problems of classification................................................................................................. 31 3.1.1.1 The arbitrariness of word classes.................................................................................... 31 3.1.1.2 What is a word and what is one word? ......................................................................... 31 3.1.2 Criteria for word classes .................................................................................................. 34 3.1.3 English word classes......................................................................................................... 36 3.2 Word classes occurring in the verbal head-complex ........................................... 37 3.2.1 Lexical verbs ...................................................................................................................... 37 3.2.1.1 Criteria for lexical verbs ................................................................................................... 37 3.2.1.2 Verb forms.......................................................................................................................... 39 <?page no="6"?> Table of contents vi 3.2.2 Modal verbs and primary verbs ..................................................................................... 39 3.2.2.1 Lexical verbs, modal verbs and primary verbs............................................................. 39 3.2.2.2 Criteria for primary verbs................................................................................................ 40 3.2.2.3 Modal verbs ....................................................................................................................... 42 3.2.2.4 Dare, need and ought .......................................................................................................... 43 3.2.3 The verbal head-complex................................................................................................. 43 3.3 Word classes occurring in the head-complex of the noun phrase ................... 45 3.3.1 Lexical nouns ..................................................................................................................... 45 3.3.1.1 Criteria for lexical nouns.................................................................................................. 45 3.3.1.2 Classes of lexical nouns .................................................................................................... 46 3.3.2 Pronouns ............................................................................................................................ 49 3.3.2.1 Word classes in the head-complex of the noun phrase: an overview ....................... 49 3.3.2.2 Subclasses of pure pronouns ........................................................................................... 50 3.3.2.3 Personal pronouns ............................................................................................................ 51 3.3.2.4 Reflexive pronouns ........................................................................................................... 52 3.3.2.5 Reciprocal pronouns......................................................................................................... 52 3.3.2.6 Indefinite pronouns .......................................................................................................... 53 3.3.2.7 Temporal pronouns .......................................................................................................... 53 3.3.2.8 who and which .................................................................................................................... 54 3.3.3 Determiner-pronouns....................................................................................................... 54 3.3.4 Pure determiners............................................................................................................... 56 3.3.5 Lexical nouns, pronouns, determiner-pronouns and pure determiners in the noun phrase ............................................................................................................ 56 3.4 Adjectives ......................................................................................................................... 58 3.5 Adverbs and particles ................................................................................................... 59 3.5.1 Adverbs .............................................................................................................................. 59 3.5.2 The category particles....................................................................................................... 61 3.5.3 Adverbs and particles versus adverbs, prepositions and subordinating conjunctions... 63 3.5.3.1 Traditional adverbs, prepositions and subordinating conjunctions.......................... 63 3.5.3.2 Reasons for the particle category .................................................................................... 64 3.6 Conjunctions .................................................................................................................... 68 3.7 Interjections ...................................................................................................................... 69 3.8 Limits of classification .................................................................................................. 69 3.8.1 Relative and interrogative words ................................................................................... 69 3.8.2 Item-specific properties and special uses ...................................................................... 71 3.8.3 Problems of classification................................................................................................. 72 3.8.4 One word or several words ............................................................................................. 73 3.8.5 Item-specific knowledge and word classes ................................................................... 74 3.9 Survey of differences ..................................................................................................... 74 4 Phrases ............................................................................................................................. 76 4.1 The character of the phrase ......................................................................................... 76 4.1.1 Syntactic units beyond the word .................................................................................... 76 4.1.2 Phrase constituents ........................................................................................................... 77 4.1.3 Heads and head-complexes............................................................................................. 77 4.1.4 Complements and modifiers ........................................................................................... 78 4.2 Noun phrase .................................................................................................................... 78 4.2.1 Reference ............................................................................................................................ 78 <?page no="7"?> vii 4.2.2 The head-complex of the noun phrase: pre-heads and heads .................................... 79 4.2.3 Complements..................................................................................................................... 81 4.2.4 Modifiers ............................................................................................................................ 81 4.2.5 Structure of the noun phrase ........................................................................................... 83 4.2.6 Elliptical noun phrases and special noun phrase construction .................................. 85 4.2.7 Typical elements of noun phrases .................................................................................. 87 4.3 Adjective phrases ........................................................................................................... 87 4.4 Adverb phrases ............................................................................................................... 88 4.5 Particle phrases ............................................................................................................... 89 4.5.1 Structure of the particle phrase ....................................................................................... 89 4.5.2 Headed or non-headed particle phrases........................................................................ 91 5 Clauses ............................................................................................................................. 93 5.1 Constituents of clause structure ................................................................................. 93 5.2 Functions of clauses ....................................................................................................... 95 5.3 Different types of clauses ............................................................................................. 97 5.3.1 Subject - predicate - adjunct ........................................................................................... 97 5.3.2 Finite and non-finite clauses............................................................................................ 98 5.3.3 Active and passive clauses .............................................................................................. 99 5.3.4 Clauses with subject and without subject...................................................................... 99 5.3.5 Word order....................................................................................................................... 100 5.3.6 Wh-clauses........................................................................................................................ 100 5.3.7 Relative clauses ............................................................................................................... 101 5.4 Clause types ................................................................................................................... 102 5.4.1 The notion of clause type ............................................................................................... 102 5.4.2 Finite clauses.................................................................................................................... 103 5.4.3 Non-finite clauses without subject ............................................................................... 104 5.4.4 Non-finite clauses with subject ..................................................................................... 105 5.5 Quasi-clauses ................................................................................................................. 106 6 Valency .......................................................................................................................... 108 6.1 Basic concepts of the valency model ...................................................................... 108 6.1.1 Valency slots .................................................................................................................... 108 6.1.2 Complements and adjuncts ........................................................................................... 109 6.2 Different degrees of optionality ............................................................................... 110 6.2.1 Obligatory, optional and contextually optional complement slots ......................... 110 6.2.2 Three levels of necessity................................................................................................. 112 6.3 Distinguishing between complements and adjuncts ......................................... 113 6.4 Formal realisations of complements ....................................................................... 117 6.4.1 Formal and functional properties in the characterization of complements............ 117 6.4.2 Types of complement ..................................................................................................... 118 6.4.2.1 Phrases and clauses ........................................................................................................ 118 6.4.2.2 Particle complements ..................................................................................................... 119 6.4.2.3 Particle complements or prepositional verbs? ............................................................ 120 6.4.2.4 Unspecified particle phrase complements .................................................................. 121 6.4.2.5 Further complements ..................................................................................................... 122 6.4.2.6 Pro-forms.......................................................................................................................... 122 6.4.2.7 List of complements........................................................................................................ 123 Table of contents <?page no="8"?> Table of contents viii 6.4.2.8 Discontinuous realisations of complements ............................................................... 124 6.4.3 One complement or two? ............................................................................................... 125 6.5 Semantic valency: participants ................................................................................. 126 6.5.1 Participant roles............................................................................................................... 126 6.5.2 Selected semantic roles................................................................................................... 131 6.5.3 Participants and complements ...................................................................................... 134 6.5.3.1 Participant and complement sharing ........................................................................... 134 6.5.3.2 Participant mergers and impersonal uses ................................................................... 135 6.6 Quantitative valency ................................................................................................... 136 6.7 Complement inventories or valency patterns ...................................................... 137 6.7.1 Complement inventory .................................................................................................. 137 6.7.2 Valency patterns and valency constructions............................................................... 138 6.7.2.1 Valency patterns.............................................................................................................. 138 6.7.2.2 Valency constructions..................................................................................................... 139 6.7.2.3 Valency constructions or complement inventories .................................................... 140 6.8 Possible valency carriers ............................................................................................ 141 6.8.1 Verbs, adjectives and nouns - lexical units ................................................................. 141 6.8.2 Particles as valency carriers ........................................................................................... 143 6.8.3 Idioms and other item-based constructions ................................................................ 146 6.8.3.1 Idiomatic combinations.................................................................................................. 146 6.8.3.2 Phrasal verbs.................................................................................................................... 146 7 The meaning of sentences .................................................................................... 148 7.1 Sentence types ............................................................................................................... 148 7.1.1 Sentence types and meaning ......................................................................................... 148 7.1.2 Statements and questions .............................................................................................. 149 7.1.2.1 Major types of statements and questions .................................................................... 149 7.1.2.2 Context-bound types of statements and questions .................................................... 153 7.1.3 Directives.......................................................................................................................... 154 7.1.4 Minor sentence types...................................................................................................... 155 7.2 Sentence meaning ......................................................................................................... 157 7.2.1 Conceptual or propositional meaning ......................................................................... 157 7.2.2 Perspective ....................................................................................................................... 157 7.2.2.1 Thematic meaning........................................................................................................... 157 7.2.2.2 Perspectivization and lexical choice ............................................................................. 158 7.2.2.3 Participant roles and clausal roles ................................................................................ 158 8 Analysis of sentences ............................................................................................. 164 8.1 Clause units .................................................................................................................... 164 8.1.1 Clause units as a combination of structural and valency criteria ............................ 164 8.1.2 Description of SCUs and PCUs ..................................................................................... 166 8.2 Alternative accounts of clause structure ............................................................... 167 8.2.1 PCUs versus objects, predicatives and adverbials ..................................................... 167 8.2.2 Advantages and disadvantages of traditional terminology ..................................... 169 9 Analytic framework ................................................................................................ 173 9.1 Steps in the analysis of sentences ............................................................................ 173 9.2 Examples of syntactic analysis ................................................................................. 179 <?page no="9"?> ix Bibliography ................................................................................................................................ 194 Index ................................................................................................................................................. 201 Glossary .......................................................................................................................................... 207 Abbreviations used in the analysis of sentences ....................................................... 210 Table of contents <?page no="11"?> Preface This book outlines a framework for the analysis of English sentences in which the concept of valency plays a central part. The model of valency theory on which this approach is based is that used in the Valency Dictionary of English (2004) (referred to as VDE) and in the valency pattern bank (www.patternbank.uni-erlangen.de), which is related in many respects to other frameworks, notably the valency models developed in German linguistics. Frequent reference will be made to the standard grammars of English, in particular the Comprehensive Grammar of the English Language (1985) (CGEL) by Randolph Quirk, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik, which has influenced our thinking in many ways, and also to the more recent Cambridge Grammar of the English Language (2002) (CamG), edited by Rodney Huddleston and Geoffrey Pullum. It is obvious that these reference books supply much more detailed accounts of many aspects of the English language than is possible in the present book. The description of many of the categories we employ is based on the approaches taken in these grammars; however, we differ from them in a number of points such as certain aspects of the system of word classes outlined and the emphasis placed on the aspect of valency in the description of sentences. The latter can also be seen as an attempt to take into account the insight promoted by the work of, for instance, John Sinclair and also by construction grammar, that there is no clear dividing line between syntax and lexis. We would also like to mention that most of the example sentences used in this book are taken from two novels by David Lodge, Nice Work and Small World. Although readers may find that the way academic life is described in these novels is not totally dissimilar to the situation of universities elsewhere, there are methodological reasons for this approach: it seemed to us that being forced to find examples for important points of syntactic analysis in two novels of this kind is a useful intellectual exercise, also because studying these texts so closely has drawn our attention to aspects of the English language which we might otherwise have overlooked. In some cases, other sources were also used, namely the British National Corpus and the VDE, which is based on the Cobuild Corpus. For every authentic example, a reference is given (SW + page reference, NW + page reference, BNC, VDE); the very few examples without such a reference were invented to illustrate particular points. Not because it is an integral part of the text type preface but because in our case it is even more appropriate than usual we would like to thank a number of friends and colleagues for their valuable help: in particular we are very grateful to David Heath for reading and commenting on the whole manuscript, to Kevin Pike for looking through large sections of it, and to Dr. Michael Klotz and Peter Uhrig for the many and very valuable discussions we had with them about individual points of analysis and for the great encouragement they have given us concerning the analytical <?page no="12"?> Preface xii apparatus we are trying to develop. Furthermore, we would like to thank Prof. Dr. Ernst Burgschmidt and Dr. Brigitta Mittmann for their comments on parts of the manuscript and Dr. Ian Roe for last-minute assistance. This also applies to all students with whom we discussed the ideas outlined here, especially Christian Hauf, Elisabeth Zinggl and Helena Otto, whom we also wish to thank for their help in the final stages of this book. Finally we would like to express our thanks to Barbara Gabel-Cunningham for her invaluable work on the manuscript. Erlangen, February 2008 Thomas Herbst and Susen Schüller <?page no="13"?> 1 Preliminary remarks about syntactic analysis 1.1 Syntax and lexis One of the most fundamental units of linguistic description is the sentence. This is shown by the fact that from the beginnings of the study of language it was the sentence - and the word - that were considered to be the basic units of analysis. Many of the categories of syntactic analysis that are used today go back to the ancient Greeks, in particular to ideas of the Stoics, of Aristotle, Dionysius Thrax, or Apollonius Dyscolus. 1 The distinction between lexis (the vocabulary) and syntax (the construction of sentences) finds a parallel in the two most important types of book in which languages are generally described: dictionaries provide information about the individual properties of words, whereas grammars deal with the more general rules of syntax and morphology. Nevertheless, there is no sharp dividing line between syntax and lexis: firstly, there is a very close interrelation between particular words and the grammatical constructions in which they appear: for instance, a noun such as spaghetti can only occur as the subject of a sentence with a singular verb; a verb such as manage can be followed by a to-infinitive complement but not by an ing-clause; a verb such as decide is much more likely to be followed by a [that_CL]or [to_ INF]-complement than to be followed by a noun phrase; it is much more likely for a verb such as argue to occur in a passive construction than for a verb such as love. secondly, when we produce language we do not seem to create structures on the basis of syntactic rules and then fill certain positions in the syntactic structure by words, but rather we very often seem to make use of larger chunks of language: this applies to constructions of the type there is, prefabricated chunks such as what I wanted to say or to collocations of the type entirely agree or white coffee. thirdly, it is quite possible that syntactic constructions are acquired and stored in the brain in a way very similar to that in which individual words are acquired and stored. It is interesting to see that these interrelationships between lexis and syntax are in the centre of much of recent work in corpus linguistics (Sinclair 2004ab) and in linguistic theories such as construction grammar (Fillmore 1988, Goldberg 2006, Croft/ Cruse 2004). Although we do not know exactly what happens in the brains of speakers when they formulate utterances, there is very strong evidence to suggest that the decision to use a particular word often entails the use of certain other words or constructions, as the 1 See Robins (1967) or Thümmel (1993). <?page no="14"?> Preliminary remarks about syntactic analysis 2 above examples show. It is certainly not adequate to imagine that the levels of syntax and lexis are completely independent from one another: it is unlikely that when formulating a sentence the human brain should create syntactic structures, which are then filled more or less at random by lexical items that are not subject to any restrictions other than being semantically compatible with each other. Recognising the strong interrelationship between what is traditionally called syntax and lexis does not mean, however, that there is no such thing as syntactic structure. Rather, the categories of syntax represent abstractions of a higher level than lexical items or words. In some cases, there may even be discrepancies between lexical storage and syntactic analysability: for instance, it is very likely that chunks such as make up my/ your/ etc. mind are stored in the brain as one lexical item, but nevertheless they can be analysed into syntactic categories such as word classes or functional categories such as object. 2 To what extent speakers perceive or use such chunks as prefabricated wholes or to what extent they analyse them into parts may depend on the individual case or/ and on the individual person. In this book, we will mainly be concerned with syntactic categories, which we see as abstractions from raw data. Many researchers in the field of language acquisition, particularly those working within the framework of construction grammar such as Tomasello (2003) or Goldberg (2006: 62), believe that the acquisition of syntax involves storage of utterances and abstraction of categories. To a certain extent, at least, the task of the grammarian trying to establish the syntactic categories of a language on the basis of linguistic data is similar to that of a child acquiring the language. This does not mean, however, that the categories arrived at as a result of these abstraction processes will be identical - if they were, there should be fewer discrepancies between the linguistic terminologies employed by different schools than there actually are. This can be partly explained by the fact that linguists analysing a language can hardly do this without preconceptions that arise from established terminology of one sort or another. It is thus one of the purposes of this book to provide an analytical framework for English which is developed on the basis of the analysis of the English language. In some cases this means that deviations from traditional terminology are necessary, as is indicated by Tomasello (2003: 173): Even in a language that is generally considered to have a number of clear categories, like English, in many cases the real situation is that traditional categories from Western linguistics are applied in something less than a thoughtful manner. Nevertheless, we would not like to make any claims as to the cognitive plausibility of the categories developed here, because as yet there does not seem to be empirical 2 Compare Tomasello's (2003: 106) analysis of I dunno: "... in usage-based approaches a given linguistic structure may exist psychologically for the speaker both as a concrete expression on its own - at the bottom of the structural hierarchy, as it were, and, at the same time, as an exemplar of some more abstract construction or constructions." <?page no="15"?> Sentence and clause 3 evidence which is sufficiently detailed to allow such claims. We do hope, however, that highlighting some of the weaknesses of traditional terminology and developing alternatives will provide an impetus for cognitive research along those lines. 1.2 Sentence and clause Problems of definition already start with a central term such as sentence itself. There is considerable variation between the units we generally refer to as sentences as far as shape or structure and even length are concerned. This is immediately apparent when one looks at the beginning of a very successful twentieth century novel: 3 (1) SW0 When April with its sweet showers has pierced the drought of March to the root, and bathed every vein of earth with that liquid by whose power the flowers are engendered; when the zephyr, too, with its dulcet breath, has breathed life into the tender new shoots in every copse and on every heath, and the young sun has run half his course in the sign of the Ram, and the little birds that sleep all night with their eyes open give song (so Nature prompts them in their hearts), then, as the poet Geoffrey Chaucer observed many years ago, folk long to go on pilgrimages. (2) SW0 Only, these days, professional people call them conferences. Both (1) and (2) represent sentences on the basis of two criteria: 1. Punctuation and spelling: they end with a full stop and otherwise only contain elements which are separated by commas or semicolons and the next sentence begins with a capital letter. 2. Structure: they could not be split up into separate units which in themselves could be considered acceptable sentences. These criteria are not entirely satisfactory, however. Punctuation and spelling obviously apply only to the written language and even then punctuation is not always clear. So one could argue that (3) SW5 "No, that's Bob Busby, he's just as bad. Worse, if anything. Been beside himself with excitement for weeks, organizing outings and so forth." could also be represented as (3a) "No, that's Bob Busby. He's just as bad - worse, if anything. …". As far as the spoken language is concerned, defining sentence boundaries is even more difficult. Certainly, intonational cues serve to mark the ends of sentences, but only up to a point. Basing a definition of the sentence on prosodic and intonational criteria does not seem to be possible. Similarly, attempts to define the sentence by saying that it expresses a particular idea or thought are not very promising since it is very difficult to say precisely how many ideas are expressed by a sentence such as 3 Examples marked SW are taken from David Lodge's novel Small World, examples marked NW are taken from Nice Work. <?page no="16"?> Preliminary remarks about syntactic analysis 4 (4) SW0 Rummidge is not Birmingham, though it owes something to popular prejudices about that city. Apparently, it is not possible to define sentence in a satisfactory way. 4 Quirk, Greenbaum, Leech and Svartvik point out in the Comprehensive Grammar of the English Language (CGEL 1985: 2.11) that a sentence is often considered to be "the highestranking unit of grammar" but that this results in a situation where "'grammar' and 'sentence' are mutually defining" because the grammar accounts for the description of the sentences of a language. In fact, for a number of reasons it may be more convenient to focus on the clause rather than the sentence. A clause can be defined as in the Cambridge Grammar of the English Language (CamG 2002: 44) by Huddleston and Pullum as "a syntactic construction consisting (in the central cases) of a subject and a predicate". It should be noted, however, that such a definition is not entirely unproblematic either. Firstly, it is a definition of a prototype and thus allows for non-central cases such as two of the units separated by full stops in (3) or (5), which obviously do not have subjects: (3) SW5 "No, that's Bob Busby, he's just as bad. Worse, if anything. Been beside himself with excitement for weeks, organizing outings and so forth." (5) SW33 "Don't get the wrong idea, Percy" … Secondly, it is difficult to escape the circularity of then defining subject and predicate as clause constituents. Nevertheless it seems to make sense to accept this terminology and regard a sentence as a unit that consists of one or more clauses. The sentence is then distinguished from a clause in that it is not a constituent of any other syntactic unit, i.e. "the largest unit to which we can assign grammatical structure" (Palmer 1971: 73). 5 1.3 Clause constituents 1.3.1 Clause constituents at different levels Assigning a clause grammatical structure means accounting for the structural relations that hold between its parts or, to use a more technical term, its constituents. In order to do that, one has to establish what one considers to be the constituents of a clause, i.e. which types of clause constituent one wishes to identify. 4 For the difficulties of defining sentences see Palmer (1971: 70-82) or Matthews (1981: 26-45). Compare in this context Paul (1919/ 2000: 498): "Der Satz ist der sprachliche Ausdruck, das Symbol dafür, daß die Verbindung mehrerer Vorstellungen oder Vorstellungsmassen in der Seele des Sprechenden sich vollzogen hat ...". 5 Cf. Huddleston (CamG 2002: 44): "We take sentences, like words, to be units that occur sequentially in texts, but are not in general contained one within another." For a very similar definition see Kortmann (2005: 124): "What is understood by sentence is the largest independent (! ) syntactic unit of a language which is not embedded in any larger construction." Again, there are exceptions such as 'I wish you'd put that question to the Vice-Chancellor, Rupert,' says Philip NW87 , which are classified as QUOTE in the Valency Dictionary of English (2004) (VDE). <?page no="17"?> Clause constituents 5 One such level of clause constituent is the word since clauses obviously are made up of words. Although words can be analysed further into morphemes, we will take the word as representing the lowest unit in the analysis of clause structure. It makes sense, however, to establish further syntactic constituents between the level of the clause or sentence and the level of the word, although linguists would not necessarily always agree as to what these constituents are. All speakers of a language seem to be able to identify constituents - at least up to a point. 6 Thus there would be a large measure of agreement amongst speakers of English that a sentence such as (6) SW18 I bought this hat at Heathrow this morning. can sensibly be divided into units such as [6.1] I bought this hat at Heathrow this morning 6A 6B 6C 6D 6E rather than [6.2] I bought this hat at Heathrow this morning 6 α 6 β 6 γ 6 δ Some of the constituents in [6.1] consist of single words whereas others are made up of a number of words. From the point of view of the clause one could thus argue that clauses seem to provide certain structural "slots" which can be filled by particular words or groups of words. From the point of view of the word, one would have to say that different words can combine to form groups of words that can function as constituents at the next higher level of structure in the clause. Such relationships can be very complex. Thus in the case of a clause such as (7) SW0 The modern conference resembles the pilgrimage of medieval Christendom in that it allows the participants to indulge themselves in all the pleasures and diversions of travel while appearing to be austerely bent on self-improvement. 7 it is possible to identify the following four constituents: 6 For evidence as to the psychological reality of phrases see Clark/ Clark (1977: 53-55). Compare also Sinclair/ Mauranen (2006: 6) on chunking. 7 The clause following in that contains further subordinate clauses itself. <?page no="18"?> Preliminary remarks about syntactic analysis 6 [7.1] The modern conference resembles the pilgrimage of medieval Christendom in that it allows the participants to indulge in all the pleasures and diversions of travel while appearing to be austerely bent on self-improvement 7A 7B 7C 7D However, groups of words such as of medieval Christendom or all the pleasures and diversions of travel can also be identified as constituents in this clause. They can, however, be seen as parts of the larger constituents 7C and 7D. 1.3.2 Identifying clause constituents The fact that intuitively (6) can be broken down into units such as 6A - 6E as in [6.1] is reflected by the fact that these units can be replaced by (a) single words or (b) groups of words in the same sentence. [6.3] I bought this hat at Heathrow this morning Morris Zapp had bought something there then The famous professor of English literature who had just arrived from the States had left it somewhere at 9 o'clock 6A 6B 6C 6D 6E All elements occurring in slots 6A, 6B, 6C, 6D and 6E can replace each other syntactically in that they can fill the same slot in the clause. This test of commutation is a very important criterion in this context: in terms of structuralist linguistics, Morris Zapp and The famous professor of English literature who had just arrived from the States can substitute I in (6), had bought can substitute bought etc., which shows that they realise the same constituent in the sentence. 8 A further test to illustrate the possibility of commutation is that the elements in question can be elicited in a wh-interrogative-'question'-construction: 8 For commutation compare Emons (1974). <?page no="19"?> Clause constituents 7 (6a) Who bought this hat at Heathrow this morning? (6b) What did you buy at Heathrow this morning? (6c) Where did you buy this hat this morning? (6d) When did you buy this hat at Heathrow? On the basis of such criteria most of the units identified in [6.3] can be elicited as a relevant type of constituent. This type of constituent is often referred to as a phrase (> Chapter 4). What is remarkable, however, is that constituents such as 6B [bought] or 7B [resembles] cannot be identified as constituents in this way since they cannot be elicited by a question. Questions of the type (6e) What did you do with this hat at Heathrow this morning? (6f) What did you do at Heathrow this morning? do not permit any response of the type (6g) *Buy. But rather (6h) Buy it. (6i) I bought it. This means that [bought this hat] must be seen as a possible constituent of sentence (6), which results in an analysis as in [6.4] I bought this hat at Heathrow this morning 6A 6BC 6D 6E On the other hand, different kinds of syntactic tests also provide arguments for an analysis of the following kind: [6.5] I bought this hat at Heathrow this morning 6A 6BCDE For instance, [bought this hat at Heathrow this morning] can be replaced by a proform of the type (6j) I did. as a reply to a question such as (6k) Did you buy this hat at Heathrow this morning? and similarly, the scope of negation in (6l) I did not buy this hat at Heathrow this morning. <?page no="20"?> Preliminary remarks about syntactic analysis 8 extends to the whole unit 6BCDE [bought this hat at Heathrow this morning], which could thus also be regarded as a constituent of clause (6). Obviously, a phrase such as 6C [this hat] can be seen as a constituent of larger constituents such as 6BC [bought this hat] or 6BCDE [bought this hat at Heathrow this morning]. It is important to see that on the basis of such criteria as commutation and question tests a case can be made out for the identification of quite different kinds of constituents in a clause. One can thus identify three different ways of establishing constituents in a clause such as (6) I bought this hat at Heathrow this morning [6.1] 6A 6B 6C 6D 6E [6.4] 6A 6BC 6D 6E [6.5] 6A 6BCDE Depending on which of these constituents one sees as most important and how one describes the relationships between the various types of constituent, one can thus arrive at different analyses of a sentence. This unfortunately is also reflected in the terminologies that are employed by different linguists. The different possible analyses of example (6) demonstrated above can serve to illustrate why such terminological discrepancies exist. The most familiar categories in the analysis of sentences are probably subject and predicate, which have both been used for centuries. 9 All three analyses identify the identical first constituent 6A [I], to which the term subject can be applied. Thus it is tempting to refer to one of the remaining parts of the sentence with the term predicate, which as a result is used for different types in different frameworks. In what one could call a "broad use" of the term predicate, which is employed in the Comprehensive Grammar of the English Language by Quirk, Greenbaum, Leech and Svartvik, for instance, the predicate corresponds to a constituent such as bought this hat at Heathrow this morning in [6.5], i.e. it is that part of the sentence which is not the subject. 10 The "narrow view" corresponds to a definition of the predicate as a constituent such as in [6.4] based on a threefold distinction between subject, predicate and adjuncts, which will be outlined further in 2.2. 11 There is a certain overlap between the use of the term predicate and the term verb phrase, as can be seen from the following table: 12 9 For the different use of the term predicate in Stoic philosophy see Hülser (1996). 10 See, for instance, Robins ( 2 1971: 240). 11 Note that Kortmann (2005: 131) uses the term predicate to refer to a constituent such as 6B in [6.1]. Compare also Haegeman (1991). 12 Up to a point these terminological differences can be taken to reflect different views of sentence structure. However, it must also be understood that if the terms predicate and verb phrase are used to refer to the same unit, this does not necessarily mean that they are used synonymously since the term verb phrase is often used to describe formal units (verb phrase, noun phrase etc.), whereas predicate is often contrasted with a functional unit such as subject. <?page no="21"?> Clause constituents 9 predicate verb phrase CGEL 6BCDE 6B Haegeman 6B 6BCDE Kortmann 6B 6B 1.3.3 Relations between constituents The fact that a unit that can be identified as a constituent can both be part of a larger constituent and consist of smaller constituents means that the relationships between the different types of constituent must be accounted for in any syntactic analysis. It is perhaps not surprising that different linguistic theories should have opted for different kinds of solutions, which is also reflected in the use of different terminologies, as outlined above. The status of a particular constituent in a particular syntactic description varies considerably depending on the overall principles of the approach. If we look at a clause such as (6m) I bought this hat [6.3] 6A 6B 6C the relations between the three constituents 6A, 6B and 6C can be represented in such a way that (a) 6A is at a higher level than 6B and 6C (immediate constituent analysis) (b) 6B is at a higher level than 6A and 6C (dependency analysis) (c) all three are at the same level of analysis (linear analysis). Immediate constituent analysis: Analysis (a) is typical of the phrase structure analysis employed in American structuralist linguistics and early versions of generative grammar. Applying principles outlined by Bloomfield (1933), it was common to divide sentences into their immediate constituents 13 , generally following the principle of binary divisions. 14 In the case of a sentence such as (6m) I bought this hat. a first division would lead to the immediate constituents I and bought this hat as in [6.6] I ⎮ bought this hat. 13 Bloomfield (1933), who was mainly concerned with morphology, illustrated the principle of immediate constituent analysis using the sentence Poor John ran away. See also Nida (1964). For an outline of the problems of immediate constituent analysis see Palmer (1971) or Matthews (1993). 14 For a discussion of the idea of binary divisions see Matthews (1993: 54-55) or Palmer (1971). <?page no="22"?> Preliminary remarks about syntactic analysis 10 which can then be split up further into bought and this hat, which in the next step is divided up into this + hat. This kind of constituent analysis can be represented in the form of a phrase structure tree of the following kind: [6.7] Sentence I bought this hat If one takes the word as the final level of syntactic analysis, I, bought, this and hat can then be called the ultimate constituents of the sentence. Such a model represents part-whole-relationships and establishes a hierarchy in which the whole sentence is at the top and the smallest constituents identified are at the bottom. According to these principles, constituents such as 6A [I] and 6BC [bought this hat] could be identified as immediate constituents of (6m). Since 6C [this hat] is a constituent that is the result of a further division, it can be seen as taking a lower place in the hierarchy of the clause structure. (6) is a good example to illustrate the fact that the principle of binary divisions cannot be easily applied to all clauses since there does not seem any sensible way of dividing 6BCDE [bought this hat at Heathrow this morning] into two constituents, as illustrated in [6.8] and [6.9]: [6.8] bought this hat at Heathrow this morning 6BC 6DE [6.9] bought this hat at Heathrow this morning 6B 6CDE Binary division results in constituents of the kind 6DE [at Heathrow this morning] or 6CDE [this hat at Heathrow this morning], both of which are counterintuitive and do not meet the criteria for constituents established above. 15 It is for this reason that the principle of binary divisions has not been accepted by other schools of thought, in particular by dependency grammar. 15 Similar problems arise with sentences containing trivalent verbs such as (2). <?page no="23"?> Clause constituents 11 Dependency analysis: Dependency models - following the ideas outlined by the French structuralist Tesnière (1959) - do not establish a hierarchy of elements in terms of part-whole relationships, but attribute a higher place in a hierarchy to such elements on which the occurrence of certain other elements depends. Thus, in (2) SW0 ... professional people call them conferences. professional can be seen as being dependent upon people in the sense that the word professional can only occur in a sentence such as (2) together with a word such as people, whereas the occurrence of people is not dependent upon the occurrence of professional. Similarly, one could argue that in (2) the constituent call governs the constituents professional people, them and conferences in that these constituents require a constituent such as call in order to be able to occur in a sentence of this kind. Since bought in (6m) differs from call in (2) in that it requires only two dependent elements, the verb is often regarded as the highest element of clause structure. Without discussing details of the nature of dependency relations here, it is important to realise that a dependency approach represents an analysis of type (b) (page 9) in which 6A [I] and 6C [this hat] are represented at the same hierarchical level and are governed by 6B [bought]. 16 [6.10] bought I hat this Linear analysis: Finally, analysis (c) is represented, for example, by the approach taken in the Comprehensive Grammar of the English Language by Quirk, Greenbaum, Leech and Svartvik (1985), where a clause such as (6m) is described as representing the basic clause type SVO, where the constituents identified as 6A, 6B and 6C above would all be at the same level: 16 For the principles of dependency grammar see Tesnière (1959/ 2000), Herbst/ Heath/ Dederding (1980: 32-46) and Heringer (1993). For an evaluation of the principles of dependency and constituency see Matthews (1981: 71- 95). <?page no="24"?> Preliminary remarks about syntactic analysis 12 [6.11] I bought this hat S V O 6A 6B 6C [6.10] and [6.11] can also serve to illustrate the fact that up to a point such different accounts are simply a matter of perspective. In [6.10], the verb buy is seen as the governing element of the clause in that it determines its structure on the basis of its valency requirements, whereas in [6.11] a particular type of clause pattern is established which can only occur with a verb with valency features such as those possessed by the verb buy, for instance. Up to a point, one could argue that these two models represent the same facts but the overall structure looks rather different. We cannot go into the details of the various approaches and possible combinations of these approaches here but we hope that what this discussion has revealed is that there is no immediately obvious way to analyse clause structure. Firstly, there are very good reasons for identifying different kinds of constituents. Secondly, there are very good reasons for establishing different kinds of relations between the constituents identified, which again may result in different positionings of one particular element in the hierarchy of clause structure suggested. It would be intriguing to know to what extent these different views of the hierarchy of clause elements are reflected in cognitive or psychological terms, but for the time being there is not enough evidence to provide a conclusive solution. Furthermore, of course, cognitive or psychological plausibility is not the only criterion to go by. 1.4 The importance of terminology What this short discussion of possible syntactic constituents has shown is that there are considerable differences between different linguistic schools as to which kind of segmentation process is appropriate and which kind of terminology ought to be used. This is mostly true of a kind of intermediate level since there seems to be widespread agreement amongst linguists that the highest element of syntactic analysis should be the sentence and that syntactic analysis should not go beyond the level of words. Even if there may be a certain amount of disagreement as to how to define categories such as sentence or word, this is not where the main differences between different approaches to syntax lie. The purpose of any model of syntactic analysis is to identify the constituents of a sentence or of a clause and to describe them with respect to their formal properties (i.e. with respect to their morphology or their position in the clause) their function <?page no="25"?> The importance of terminology 13 and the relations between them. Different schools of linguistics have approached syntactic analysis from different angles and reached different conclusions as to what is the most appropriate way of accounting for clause structure. Partly, such differences have to do with the general frameworks in which analyses are being carried out and the purpose the descriptions are to serve. Categories that are useful for the teaching of a language as a foreign language to foreigners do not necessarily coincide with the categories and abstractions that native speakers form in their minds (if that is what they do). Similarly, a structuralist approach looking at language as a system of signs and relations may yield different categories from a cognitive approach that tries to account for what goes on in the brain. Also, linguists who try to account for properties of all languages may differ in their use of terminology from those who develop the categories for the analysis of one specific language on the basis of the analysis of this language. The great variety of approaches to syntactic analysis has also resulted in a great variety of different terminologies. As a consequence, the same or very similar concepts are often given different labels - semantic roles, deep cases, or theta roles (> Chapter 7) are a case in point. What is worse, however, is that very often the same and often established terms are used to mean different things, as the discussion of the term predicate has shown, but there are many other cases of this kind: for instance, Quirk, Greenbaum, Leech and Svartvik (1985: 4.3), in CGEL, consider tense "strictly as a category realised by verb inflection" and consequently identify only two tenses for English - present tense and past tense. Many teaching grammars, however, also identify a future tense for English - probably because in Latin there was an inflected future tense, which has given rise to a teaching tradition which is based on Latin rather than on English and presents the facts of English in a rather odd way. A similar example is that of the gerund, a term which is used in Latin grammar to identify a particular morphological form (such as laudandi), which is distinct from a present participle (such as laudans). In English, no such morphological distinction exists, but nonetheless some school grammars make use of the term gerund, although the standard linguistic reference grammars of English, the Comprehensive Grammar of the English Language (1985) and the Cambridge Grammar of the English Language (2002), reject it for very good reasons. 17 It is interesting to see that de Saussure (1894/ 1964: 95 and 96) already spoke of the "utter inadequacy of current terminology" and "the need to reform it". 18 One of the most fundamental principles to follow is that the categories used in the description of one language should result from the analysis of that language and not be used because they are appropriate to the analysis of another language. 19 17 Compare also Nida (1964: 22): "The historical and comparative points of view are most often responsible for the confusion in the descriptive analysis of the language." See also 3.2. 18 "Cela finira malgré moi par un livre où, sans enthousiasme ni passion, j'expliquerai pourquoi il n'y a pas un seul terme employé en linguistique auquel j'accorde un sens quelconque." See Culler (1976: 15). 19 Cf. Burgschmidt and Götz (1974: 26-31) and Herbst (2005). See also Sinclair (2004b). <?page no="26"?> Preliminary remarks about syntactic analysis 14 What is particularly important in our context are terms that are used for functional categories and for word classes because here striking discrepancies of usage can be found between the different frameworks. For instance, some grammarians see it as a defining property of the categories direct object and indirect object that they can function as the subject of a corresponding passive clause (Aarts/ Aarts 1982/ 1988), whereas others such as Quirk, Greenbaum, Leech and Svartvik (1985: 727) say that the "object of an active clause may generally become the subject of a passive clause", where the word generally allows for a number of important exceptions (see Chapter 8.2.2). As far as word classes are concerned, some approaches account for the different uses of words such as that in (3a) "No, that's Bob Busby, he's just as bad - worse, if anything. …". and (4) SW0 Rummidge is not Birmingham, though it owes something to popular prejudices about that city. by saying that in (3a) that is a member of a word class pronoun, whereas in (4) it is a member of a word class determiner. This is the policy taken in CGEL, for example. Others simply treat (3a) and (4) as different uses of one word, which is then classified as a pronoun. We will come back to this kind of problem in more detail in our discussions of word classes, but what we want to show with these examples is that many terms, including many important and established terms, are used in different ways by different people. This is important for two reasons: firstly and rather obviously, one must be sure to know in which sense a particular term is being used within a particular framework: if one consults a dictionary and a word is classified as an adverb or a determiner, one has to know what is meant by these terms if one wants to use the word correctly. Secondly, these examples also show that it makes sense to think about terminology and define categories in a precise and appropriate way. This is one of our objectives in this book. While we will largely be making use of established terminology - mainly terminology based on CGEL and valency theory -, we will try to define all terms unambiguously. (In some cases this means a slight modification of terminology readers may be acquainted with - or think they are acquainted with -, but we hope that whenever a new term is introduced, the reasons for doing so will be clear). 1.5 The aims of this book This book is an attempt to present an approach to the analysis of English sentences that provides an exhaustive description of their structure, which is based on established descriptive categories. The more specific parts of the framework outlined are based on the principles of valency theory, in particular in the form in which it has been applied to English in the Valency Dictionary of English by Herbst, Heath, Roe and Götz (VDE). As far as the terminology is concerned, this book aims to develop a terminological framework that can be applied consistently to the analysis of English <?page no="27"?> The aims of this book 15 sentences. An attempt will be made to explain the terminology used in some detail. This is done for two main reasons: Even the brief discussion of some central terms of syntactic analysis in this chapter has shown that different approaches to syntax vary considerably in both the terms they employ and the meanings they attribute to particular terms. Many students who begin to study foreign languages at German universities have not received sufficient instruction in formal grammatical analysis to make them feel confident about categories such adverb, adverbial, participle, preposition or conjunction. In this introductory outline we will thus provide short definitions of all terms that we consider central to the framework developed and give references to sources where certain aspects are dealt with in more detail than seems appropriate here. We will concentrate on the aspects of the framework developed here, but at the same time try to make clear where our approach differs from others. <?page no="28"?> 2 The syntactic framework 2.1 Survey of the framework suggested One aim of syntactic analysis is to identify the constituents of the clause and to establish the relations between them. Following the principles outlined in Chapter 1, sentences such as (1) NW384 Robyn thinks. (2) SW18 I bought this hat at Heathrow this morning. (3) SW0 Only, these days, professional people call them conferences. (4) NW47 She puts her soiled breakfast things into the sink ... can be segmented into the following constituents: 1 (1') [Robyn] | [thinks]. (2') [I] | [bought] | [this hat] | [at Heathrow] | [this morning]. (3') [Only] | [these days] | [professional people] | [call] | [them] | [conferences]. (4') [She] | [puts] | [her soiled breakfast things] | [into the sink] These constituents will have to be described with respect to their function (or syntactic role) in the clause, their formal properties and their semantic contribution to the meaning of the clause as a whole (Chapters 6 and 7). All models of syntactic analysis are abstractions from sentences as actually used in communication. Different models of syntax differ in the kinds of abstractions they make and which kind of starting point they regard as most appropriate in terms of the generalizations they wish to make. In the model proposed here, the structure of a 1 We can assume, up to a point at least, that such an analysis is carried out by speakers when they hear or read a sentence. Following the approach of Sinclair and Mauranen's Linear Unit Grammar (2006), one can assume that speakers first divide sentences into chunks which do not necessarily coincide with the constituents of a constituent analysis. Thus in the case of There is a widespread feeling in the country that universities are 'ivory tower' institutions, whose staff are ignorant of the realities of the modern commercial world. NW85 a first superficial analysis could result in chunks of the following kind There | is | a widespread feeling | in the country | that | universities | are | 'ivory tower' institutions,| whose staff | are | ignorant | of the realities | of the modern commercial world. A more complex analysis, however, would have to account for such facts as that of the modern commercial world belongs to realities, that whose staff are ignorant of the realities of the modern commercial world is closely related to ivory tower institutions and that the whole clause that universities are ivory tower institutions whose ... is related to feeling (and not to the immediately preceding country). <?page no="29"?> Survey of the framework suggested 17 clause is seen as being determined by properties of syntactic constructions and by syntactic properties of the lexical items occurring in the clause, in particular the valency of verbs. This is in line with the view, expressed in Chapter 1, that language can be described as an inventory of different types of units, namely lexical items including their collocational and valency properties as well as syntactic items such as particular types of clauses, phrases or other syntactic constructions. 2 For instance, sentences (1-4) can be said to represent the same syntactic item in terms of a particular sentence because we can analyse them as consisting of a subject (> 2.2.1), which is followed by a predicate (> 2.2.3). Subject Predicate (1'') Robyn thinks (2'') I bought this hat at Heathrow this morning (3'') Only these days professional people call them conferences (4'') She puts her soiled breakfast things into the sink At the same time, the structure of these sentences is also related to properties of the verbs occurring in the predicate so that it makes sense to describe the constituents in the grey boxes as complements of the verb, which means that the valency properties of the verbs will be taken into account. A sentence such as (3) can then be seen as a realisation of a particular syntactic item - the sentence type declarative-'statement'construction - and of a valency pattern of the lexical item call. 3 As far as the constituents at Heathrow, this morning, only and these days are concerned, these are neither typical of particular sentence types nor dependent on the valency of the verb. Under certain conditions, such adjuncts can occur in any kind of structure, which is why the greatest generalization in their description is achieved by describing the use of different classes of adjuncts (such as place adjuncts, time adjuncts etc.) irrespective of, say, the sentence type. The overall analysis then takes the following form: 2 Since the term construction is used in construction grammar in terms of a fixed combination of form and meaning (Fischer/ Stefanowitsch 2006: 4), we use the term syntactic item here in order to refer to the formal side of constructions. Note that the term clause type is not used in the sense of the seven clause types of CGEL (1985: 2.16) here. 3 This is very much in line with Goldberg's (2006: 21) view that "an actual expression typically involves the combination of many different constructions". Compare also Fischer/ Stefanowitsch (2006: 6-7). <?page no="30"?> The syntactic framework 18 Subject Predicate (1'') Robyn thinks (2'') I bought this hat at Heathrow this morning (3'') Only these days professional people call them conferences (4'') She puts her soiled breakfast things into the sink It is the purpose of this chapter to introduce basic notions such as subject, predicate, valency, and to give a general outline of the structure of the clause and the phrase. Most of these notions will be discussed in more detail in the chapters that follow: thus Chapters 4 and 5 deal with phrases and clauses and Chapter 6 gives a much more precise account of the phenomenon of valency. Although terms such as verb or noun are used in this chapter, they will be defined in Chapter 3. 4 2.2 Elements of clause structure 2.2.1 Subject The distinction between subject and predicate is made in almost all approaches to language and it can be seen as central to the nature of clause structure. Firstly, with the exception of cases such as (5) NW87 Don't be absurd. the overwhelming majority of English sentences contains subjects. Secondly, most sentences contain predicates which are closely related to the subject in that they contain a verb that shows concord with the subject. Thus a singular subject co-occurs with a singular verb (Robyn thinks) and a plural subject co-occurs with a plural verb (… professional people call ...). Thirdly, the position of the subject with respect to the predicate plays a central role with respect to the identification of the sentence type, which will be discussed in Chapter 7. In declarative-'statement'-constructions such as (1) NW384 Robyn thinks. (3) SW0 Only, these days, professional people call them conferences. the subject normally precedes the predicate in English, whereas in particular interrogative-'question'-constructions the subject follows the predicate or an operator verb which is part of the predicate: (6) NW333 Is it? (7) NW326 How shall I get the manuscript to you? 4 Admittedly, this entails the danger of circularity, but we attempt to keep this to a minimum. <?page no="31"?> Elements of clause structure 19 The subject can thus be defined as a clause constituent that shows concord with the verb of the predicate, precedes the predicate in declarative-'statement'-constructions, is realised by the nominative form of pronouns (I, he, she, they). 5 2.2.2 Adjuncts In the light of the strong interrelationship between subject and predicate, adjuncts are less central to the description of clause structure since they are not structurally related to either the subject or the predicate of a clause in the same way. This is relevant with respect to two criteria: Firstly, from a structural point of view, adjuncts are purely optional: it is not unusual for a clause not to contain an adjunct, as in (1) NW384 Robyn thinks. (8) NW214 She had plenty of work: her teaching, her research, her administrative duties in the Department. Furthermore adjuncts such as at Heathrow, this morning or only and these days can be omitted from sentences such as (2) and (3) without making them ungrammatical or ill-formed: (2a) I bought this hat. (3a) Professional people call them conferences. Secondly, many types of adjunct show more positional mobility than other constituents of a clause: (2b) This morning, I bought this hat at Heathrow. (2c) At Heathrow, I bought this hat this morning. (3b) Only, professional people call them conferences these days. A clause can contain several adjuncts, which can occur in different positions in the structure of the clause: (9) NW50 Anyway, you'll never get a job here, Robyn. Semantically, adjuncts often provide information about circumstances such as place (at Heathrow), time (this morning), causal relationships etc., which is why they have been described as circumstantial elements 6 in dependency grammar, but they can 5 Although it is quite possible that certain semantic roles typically realized by subjects play an important role in the identification of the role of subject in language acquisition (cf. Tomasello 2003: 167-169), we use the term subject for a clause constituent that is defined in terms of such formal criteria. 6 Compare also Tesnière's (1959) notion of circonstants. Compare also Halliday's (1970: 146-147) description of "participant functions" and "circumstantial functions" or Fillmore's (2007: 133) distinction between "core frame elements" and "peripheral frame elements". <?page no="32"?> The syntactic framework 20 also express causal relationships as in (10) or have a textual function like instead in (11), only in (3) or anyway in (12): (10) NW383 Since the English Department is bulging with students, and some of the smaller departments in the Faculty are on the brink of disappearing altogether, there's a chance that we may be able to replace Rupert after all, in spite of the new round of cuts. (11) NW42 Though urged by the school to apply for a place at Oxbridge, she chose instead to go to Sussex University, as bright young people often did in the 1970s, because the new universities were considered exciting and innovatory places to study at. (3) SW0 Only, these days, professional people call them conferences. (12) NW218 They don't pay us all that much, anyway. It is important to note that the reason why adjuncts are generally treated as not being central to clause structure is not that their semantic contribution to the sentence or their role in creating textual cohesion is marginal. However, since the occurrence of adjuncts does not depend on the valency of the governing verb (> 2.3), it seems more appropriate to describe the use of adjuncts in terms of the different categories of adjunct irrespective of different types of clause structure. 2.2.3 Predicate Of the three clause constituents - subject, predicate, adjuncts -, the predicate is perhaps the most complex. What the predicates in the following examples have in common is that they all contain at least one verb: (1) NW384 Robyn thinks. (2) SW18 I bought this hat at Heathrow this morning. (4) NW47 She puts her soiled breakfast things into the sink ... (13) NW326 I'm catching the 9.45 shuttle to Heathrow. (14) NW360 He must have been really impressed by your book. We call these verbs or groups of verbs the verbal head-complex of the predicate. The first verb of the head-complex shows concord with the subject of the clause (see 2.2 above). The structure of the verbal head-complex, which will be described in more detail in 3.2.3, is independent of the verbs occurring in it, as can be shown quite easily: (2d) He must have bought this hat at Heathrow this morning. (2e) I'm buying a hat this afternoon. The elements that follow the verbal head-complex, however, could not be so easily exchanged because their occurrence depends on the presence of particular verbs: (13a) *She is catching her breakfast things into the sink. (4a) *She puts this hat. (4b) *She puts really impressed by your book. <?page no="33"?> Valency 21 In other words: this is the point at which a description of sentence structure can no longer be achieved without referring to the properties of lexical items. We refer to the constituents of the predicates of the sentences above which are outside the verbal head-complex as complements. Their occurrence is dependent on the last verb of the verbal head-complex, which will be called the head or predicator. 2.3 Valency 2.3.1 Complements and adjuncts The analysis of the different types of complement that can occur with a particular predicator is at the centre of valency theory, which will be outlined in more detail in Chapter 6. The fundamental idea of valency theory, which developed out of Tesnière's model of dependency theory, is that the structure of a sentence is largely determined by the verb. This means that certain properties of the verb have a determining influence on the structure of a clause. Thus there are many English verbs, for example think or apply, which only require one further element, for which valency theory also uses the term complement, to form a grammatical active declarative clause: (1) NW384 Robyn thinks. (15) NW52 Thus a three-year lectureship in English literature was advertised, Robyn applied, was interviewed along with four other equally desperate and highly qualified candidates, and was appointed. Verbs such as think or apply occur in sentences such as (1) or (15) with only one complement, which is the subject of the clause. Other verbs, however, require further complements. For instance, verbs such as like, want or afford do not usually occur in sentences of the type (16a) *Robyn likes. (17a) *The students want. (18a) *I couldn't afford. but rather in sentences such as (16) NW60 What Robyn likes to do is to deconstruct the texts, to probe the gaps and absences in them, to uncover what they are not saying, to expose their ideological bad faith, to cut a cross-section through the twisted strands of their semiotic codes and literary conventions. (17) NW60 What the students want her to do is to give them some basic facts that will enable them to read the novels as simple straightforward reflections of 'reality', and to write simple, straightforward, exam-passing essays about them. (18) NW62 I couldn't afford the time, even if I had the money. Thus, for example, a constituent such as the time is required for the verb afford to be used in a grammatical sentence. As (18) is an active declarative clause, the same <?page no="34"?> The syntactic framework 22 applies to I, which is the subject of the sentence. The description of valency focuses on the complements of a valency carrier (a verb, adjective or noun). Since the occurrence of adjuncts is independent of the predicator, they are not part of the valency specification of a verb. In fact, the distinction between complements and adjuncts is central to valency theory. It is important to note that a simple deletion test will not suffice to distinguish complements from adjuncts. From a grammatical point of view, adjuncts (such as even if I had the money) can always be deleted without making a sentence ungrammatical: (18b) I couldn't afford the time. However, this is also true of some elements that must be regarded as verb-specific and thus are not adjuncts. 7 For instance, in (19) NW383 'You did say, back in January,' says Robyn, 'that if a job came up, I should apply for it.' for it could certainly be omitted but nevertheless the fact that a for-construction can be used has to be seen as a feature of the verb apply. Similarly, the underlined constituents of (20) to (23) are determined in their form by the verb although the verb can be used without them as in (24): (20) NW185 Tell me one person you know, I mean know, not know of, somebody you talked to in the last week, who earns less than six thousand a year. (21) NW15 So tell me about structuralism. (22) NW65 I have to tell them that the only way we can meet our targets is an absolute freeze. (23) NW65 It's no fun being Dean these days, I can tell you. (24) NW325 Robyn told him. On this basis, complements can be distinguished from adjuncts as follows: A complement is a clause constituent that is determined by the governing element in that it must occur if the governing element is to be used in a grammatical sentence or it is determined in its form by the governing element. An adjunct is a clause constituent that is not determined by the valency of a governing element. This distinction reflects what was said from a slightly different perspective about the role of adjuncts in clause structure in comparison with the more central role of subject and predicate. 8 7 For the problems of deletion tests and a distinction between different types of necessity see 6.2.2. 8 Compare, for instance, Lyons's (1968: 334) distinction between nucleus and periphery of a clause, or the elimination of "optional adverbials" in the clause types established in CGEL (1985: 2.13-2.19). <?page no="35"?> Valency 23 2.3.2 Valency complements as constituents in clause structure As will be shown in more detail in Chapter 6, a valency description of a verb, an adjective or a noun can thus be seen as a list of all possible complements these respective lexical units can take. Since such a valency description is an abstraction from actual language use, all possible formal realisations of the complements as they occur in different types of clause must be specified. It is part of the abstractions involved in such a valency description that the complement realised by Robyn in (25) NW61 Robyn has been told that he once had a beard, and he is forever fingering his chin as if he missed it. corresponds to him in (24) or me in (21): (24) NW325 Robyn told him. (21) NW15 So tell me about structuralism. In the model outlined in Chapter 6, which is based on the approach taken by Herbst, Heath, Roe and Götz in the Valency Dictionary of English (VDE) (2004), such a complement will be described in two respects: (a) that it expresses a semantic role that can be described as ' RECEIVER OF INFORMA - TION ', (b) that it can be realised either by the subject of a passive clause or a complement that is part of the predicate in an active clause. Such a valency description is thus not independent of the categories subject and predicate. Rather, we can say that the valency complements of the verb are realised as the subject of the clause and as constituents of the predicate. Mono-, divalent and trivalent uses of verbs can then be described as follows: (1'') clause structure subject predicate Robyn thinks monovalent use complement 1 predicator (2'') clause structure subject predicate adjunct adjunct I bought this hat at Heathrow this morning divalent use complement 1 predicator complement 2 adjunct adjunct <?page no="36"?> The syntactic framework 24 (4'') clause structure subject predicate She puts her soiled breakfast things into the sink trivalent use complement 1 predicator complement 2 complement 3 In these sentences one of the valency complements is always realised by the subject of the clause. How many valency complements occur in the predicate depends on the valency of the verb: in the case of a trivalent verb as in (4), it is two; in the case of a divalent verb as in (2), it is one; and in the case of a monovalent verb no valency complement occurs in the predicate at all. One can thus argue that the form of the predicate is determined by the valency of the verb since the verb can open up valency slots in the predicate. If one combines the structural aspect and the valency aspect terminologically (> 8.1), the following clause units can be identified: subject complement unit (SCU) (I in (2)), predicate complement units (PCU) (this hat in (2) and her soiled breakfast things and into the sink in (4)), predicate head unit (PHU) (bought in (2)), adjunct units (AU) (at Heathrow and this morning in (2)). 9 2.4 Phrases and clauses 2.4.1 Phrases On the basis of the distinctions made above, we can now say that a clause consists of a predicate head unit realised by a verbal head-complex (> 2.2.3), complements, which function either as the subject of the clause (SCU) or as constituents of the predicate (PCUs), adjuncts (AUs). Both complement units and adjunct units can be realised by clauses or phrases: 9 This kind of representation has the advantage of describing the structure of a clause in English as consisting of subject, predicate and any number of adjuncts. Alternatively, one could identify several types of clause and attribute sentences such as (1), (2) and (4) to a monovalent, a divalent and a trivalent clause type. In such a view, the valency properties of the verb would then have to match those of the clause type. An approach along those lines is taken in the clause types identified in CGEL (1985: 2.16) and in a different form in the argument structure constructions identified in construction grammar (Goldberg 2006), for example. Up to a point, however, representing clause structure in the one form or the other is mainly a matter of perspective. What is important is that in either approach specifications of lexical items will have to be combined with syntactic items. <?page no="37"?> Phrases and clauses 25 (2) SW18 I bought this hat at Heathrow this morning SCU phrase PCU phrase AU phrase AU phrase (26) NW84 Sometimes I think the VC takes his economy drive a little too far AU phrase SCU phrase PCU clause (27) NW244 When she got home, Robyn telephoned Charles AU clause SCU phrase PCU phrase In the terminology we have employed it is possible to distinguish between phrase and clause by saying that a phrase is a clause constituent that is not headed by a verb. 2.4.2 The structure of the phrase The structure of the phrase largely parallels that of the clause. It will be analysed here as containing a head-complex containing the head (and possibly pre-heads), valency complements of the word functioning as head, modifiers (which are independent of the valency of the word forming the head). These phrase constituents can be exemplified by looking at the following adjective phrases: (28) NW83 Don't want to make the Arts Faculty look incompetent. (29) NW61 She is always uncertain how to address her Head of Department. (30) NW61 'Philip' seems too familiar, 'Professor Swallow' too formal, 'Sir' impossibly servile. These phrases are classified as adjective phrases because it is quite clearly adjectives that form the essential elements of these constituents: in the case of (28), the adjective phrase consists only of an adjective anyway, and in the case of (29) and (30) it is the adjectives that cannot be deleted whereas the other elements of the phrase can: (29a) She is always uncertain. (30a) 'Philip' seems familiar, 'Professor Swallow' formal, 'Sir' servile. <?page no="38"?> The syntactic framework 26 The elements of a phrase that in a given context can stand for the phrase on its own will be called its head-complex. If the head-complex consists of a single word, this word is called the head of the phrase. The structure of these phrases can then be described as follows: (29') head complement uncertain how to address her Head of Department (30') premodifier head too familiar The clause how to address her Head of Department in (29) is a valency complement of uncertain, whereas too and impossibly are modifiers, which are independent of the valency of the adjectives functioning as heads. This can be shown by the fact that too also combines with uncertain, but the [how_to_CL]-construction does not combine with familiar, for example: (29b) She is always too uncertain how to address her Head of Department. (30b) *'Philip' seems too familiar how to address her Head of Department. The same types of constituent can be found in other types of phrase, too (> Chapter 4). In the case of noun phrases, however, the head-complex of the phrase can consist of more than one word. For instance, noun phrases such as the one underlined in (31) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. cannot be realised by one of the words contained in it in the same way as this was possible in the case of adjective phrases in (29) and (30): (31a) She was a teacher. (31b) *She was a. (31c) *She was teacher. Thus, a teacher will be seen as the head-complex of the noun phrase in (31). If the head-complex contains more than one word, the right-most element of the headcomplex is called its head, all other elements are called pre-heads. (31') pre-head premodifier head postmodifier a popular and conscientious teacher whose optional courses on women's writing were oversubscribed <?page no="39"?> Coordination and subordination 27 2.4.3 Clauses as verb phrases In the descriptive framework we have outlined so far, there are obvious parallels between the structure of the phrase and the structure of the clause. Both contain a head-complex which contains the head (possibly preceded by pre-heads). Both can contain complements, which depend on the valency of the head, and elements which are not dependent on valency properties called adjuncts or modifiers. One could thus argue - and, in fact, it has been argued - that clauses should be regarded as verb phrases. 10 Despite obvious parallels between clauses and phrases, there are also important differences: for instance, as was pointed out in 2.2, different types of clause can be identified, whose characteristics (such as whether a subject is required or not) are independent of the properties of the governing word. It is for this reason that the term verb phrase is avoided in the terminological framework outlined here. 11 It is obvious, however, that different types of phrase and clause can only be properly described after the word classes that function as heads have been defined independently of phrase structure, which will be done in Chapter 3. 2.5 Coordination and subordination 2.5.1 Coordination There is a further parallel between phrases and clauses: the valency complements of phrases and clauses, adjuncts and modifiers can be realised by the same kinds of linguistic unit, namely ⌐ phrases such as I, this hat, at Heathrow, this morning in (2), ⌐ clauses such as how to address her head of Department in (29), ⌐ coordinated phrases or clauses as popular and conscientious in (31). (2) SW18 I bought this hat at Heathrow this morning. (29) NW61 She is always uncertain how to address her Head of Department. (31) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. Coordination refers to a situation in which one constituent is realised by more than one constituent of the same rank, when each of these constituents could realise the constituent on its own: (31a) She was a popular teacher ... (31b) She was a conscientious teacher ... 10 Cf. Fillmore (1988: 43): "I treat a clause or sentence as a maximal verb-headed phrase." Compare also CamG's (2002: 24) analysis of the predicate as "the head of a clause". 11 See also the different uses of the term verb phrase indicated in Chapter 1. <?page no="40"?> The syntactic framework 28 In (31), there is coordination of two adjective phrases to form a single adjective phrase. Coordination can occur between elements at various levels, for instance, between different types of phrase, as is shown by the following examples: 12 (32) SW23 Jakobson cites the gradation of positive, comparative and superlative forms of the adjective as evidence that language is not a totally arbitrary system. [coordination of adjective phrases] (33) NW351 Now we have linguistics, media studies, American Literature, Commonwealth Literature, literary theory, not to mention about a hundred new British writers worth taking seriously. [coordination of noun phrases] (34) SW17 Have you really and truly come all the way from London by taxi? [coordination of adverb phrases] Similarly, there is coordination of clauses or parts of clauses: (35) NW214 The mornings were dark, dusk came early, and the sun seldom broke through the cloud cover in the brief interval of daylight. (36) NW215 At home she read, she took notes, she distilled her notes into continuous prose on her word-processor, she marked essays; at the University she lectured, she gave seminars and tutorials, she counselled students, interviewed applicants, drew up reading-lists, attended committee meetings, and marked essays. (36) shows the complexity of coordination since what is being coordinated is two clauses (beginning with at home and at the University), which themselves contain a sequence of coordinated Subj__Pred units. A similar case is presented by (37) NW214 The winter term at Rummidge was of ten weeks' duration, like the autumn and summer terms, but seemed longer than the other two because of the cheerless season. 2.5.2 Coordinated and subordinate clauses The possibility of coordinating different clauses is one of the crucial differences between sentence and clause. On the basis of the definition of sentence and clause given in Chapter 1, it is possible to distinguish between sentences, which, like (2), consist of a single clause, and sentences, which, like (36) and (38), consist of more than one clause. (2) SW18 I bought this hat at Heathrow this morning. (38) NW325 Send me your manuscript and I'll have a look at it. (39) NW84 Sometimes I think the VC takes his economy drive a little too far. If a sentence contains more than one clause, one can distinguish between coordination and subordination. Coordinated clauses belong to the same sentence without one being a constituent of the other as in 12 Cf. Hausser (2006: 115-136). <?page no="41"?> Coordination and subordination 29 (40) SW0 For not all conferences are happy, hedonistic occasions; not all conference venues are luxurious and picturesque; not all Aprils, for that matter, are marked by sweet showers and dulcet breezes. (38) NW325 Send me your manuscript and I'll have a look at it. A subordinate clause is a constituent of a larger clause (termed superordinate clause) or of a constituent of a larger clause. Thus the VC takes his economy drive a little too far in (39) and ignoring the dismal weather (in 41) (neither in italics) are subordinate clauses of the superordinate clauses (the scope of the superordinate clause is indicated by underlining). 13 (39) NW84 Sometimes I think the VC takes his economy drive a little too far. (41) NW214 She drove to and fro between her cosy little house and her warm, well-lit room at the University, ignoring the dismal weather. 2.5.3 Clause and sentence In traditional accounts, for instance in CGEL (1985: 14.2-4), the distinction between coordination and subordination is sometimes used as the basis for distinguishing between complex sentences, which contain clause subordination, compound sentences, which contain coordination of clauses, and simple sentences, which contain neither coordination nor subordination. In view of the different ways in which clauses and phrases can be structured, such a distinction does not really seem particularly useful, however. 14 It seems more appropriate to make a distinction between independent clauses, which are not constituents of another clause and dependent (or subordinate) clauses, which are constituents of another clause or phrase. The term sentence will be used to refer (a) to any independent clause and (b) to a combination of two or more independent clauses involving coordination. 15 Thus, for example, (38) is a sentence involving coordination of two sentences, whereas (39) is a sentence containing a dependent clause. 13 Similarly, in a sentence such as Dismay had been already plainly written on many faces when they assembled the previous evening for the traditional sherry reception. SW1 we would analyse they assembled the previous evening for the traditional sherry reception as a dependent clause (> 2.5.3), which depends on the particle when (> 4.5). In traditional grammar, the term subordinate clause is often used for the whole constituent (when they ... reception). 14 Compare CamG (2002: 45). 15 Note that this definition even applies to cases such as Life was short, criticism long. NW215 where criticism long is not a constituent of life was short, although it could not stand on its own. Compare CGEL's (1985: 14.2) use of the term main clause in this context. <?page no="42"?> The syntactic framework 30 sentence sentence <I> (independent clause) sentence <II> (independent clause) (38) NW325 Send me your manuscript and I ll' have a look at it PHU PCU1 PCU2 LU SCU PHU PCU sentence (39) NW325 Sometimes I think the VC takes his economy drive a little too far (dependent clause) AU SCU PHU PCU <?page no="43"?> 3 Word classes 3.1 Words - word classes 3.1.1 Problems of classification 3.1.1.1 The arbitrariness of word classes The concept of word classes is firmly established in grammars, dictionaries and in linguistic theory. Nevertheless, the most important fact one has to realise about word classes is that word classes - like most categories used for the description of language - do not exist as such in the language but that they are constructs developed by grammarians in order to describe certain properties of words in more general terms. Which word classes one identifies in the description of a particular language thus not only depends on the language itself, but also on the purpose of the description, i.e. the type of generalization one wishes to make. 3.1.1.2 What is a word and what is one word? Identifying word classes in English involves a number of problems. For instance, if one takes the word as the basic unit of syntactic analysis, then of course one should be perfectly clear about what one means by word. This is not as straightforward as it may seem since spelling is not a particularly reliable criterion for English. For instance, one might argue about how many words a sentence such as (1) SW28 Of course, it all goes back to Peirce. actually consists of - 7 or 8. While Sinclair (1991: 110) gives of course as an example of the "apparently simultaneous choice of two words", some dictionaries such as the Longman Dictionary of Contemporary English ( 4 2003) or the Collins Cobuild English Dictionary ( 2 1995) have a separate entry for of course (adverb) but others such as the Oxford English Dictionary ( 2 1989), the Oxford Advanced Learner's Dictionary ( 7 2005) or Langenscheidt Collins Großwörterbuch Englisch (2004) do not. Similarly, none of these dictionaries contains a separate entry for at all as in: (2) SW28 ... what, with the greatest respect, is the point of our discussing your paper if, according to your own theory, we should not be discussing what you actually said at all, but discussing some imperfect memory or subjective interpretation of what you said? Whether one analyses combinations such as at all or of course as multi-word lexemes, which can be assigned a word class label, or whether one takes them to be free combinations of two different words is a relatively arbitrary decision, which is certainly influenced by factors of semantic analysis. From a syntactician's point of view, this is a relatively marginal problem since what one defines as a single word or a combination of words is more of a lexicological than a syntactic issue. <?page no="44"?> Word classes 32 What is more problematic, however, is whether the all in sentence (1) should be seen as a form of the same word all as the all in sentences such as (3 - 9) (1) SW28 Of course, it all goes back to Peirce. (3) SW0 ... not all Aprils, for that matter, are marked by sweet showers and dulcet breezes. (4) SW43 The assumption was that all the most interesting people were concentrated into a few institutions, like Harvard, Yale, Princeton and suchlike, and in order to get into the action you had to be at one of those places yourself. (5) SW37 That's about all I can say in its favour. (6) SW51 You know we all thought it was an absolutely super book, Philip. (7) SW31 Oh, it's not as bad as all that. (8) SW17 Have you really and truly come all the way from London by taxi? (9) SW22 "Structuralism? " said Dempsey, coming up with a sherry for Angelica just in time to hear Persse's plea, and all too eager to show off his expertise. It is a central question for the architecture of a syntactic framework for English whether one should or could describe all uses of a word such as all by attributing them to one word class or whether one should assign them to different word classes on the basis of the different kinds of contexts in which they appear and of properties which they share with other words occurring in these positions. If one decides on the latter approach, this creates the problem of whether one is talking about several different words with the form all or one word all, for which different uses can be observed. Up to a point, such decisions may be influenced by the degree of difference between different uses and considerations of meaning. Thus the case for analysing the following uses of that in terms of different word classes may be stronger in the case of (12) than in the case of (10) and (11), for example: (10) NW348 Faculty Board would never accept that. (11) NW353 Of course, in that case, we shouldn't get any new staff. (12) NW347 Robyn became aware that Philip Swallow was talking to her. One could also take the view that psychological validity should be considered in this context. If lexical access procedures of the mind could be shown to be related to word class in some way or another, then a word class system that is psychologically appropriate would be a highly relevant factor for any theory of the mental lexicon. 1 Translated into somewhat simpler language this leads to questions such as the fol- 1 For a comparison of the role of word classes in generative grammar and in cognitive linguistics compare Behrens (2005). Behrens (2005: 182) writes: "Eine Lösung ist denkbar, wenn man nicht davon ausgeht, dass Wortarten ontologische Primitiva sind, sondern emergente, sich herausbildende Kategorien, die das Resultat von kontextbasierter Sprachverarbeitung sind (emergentism)." <?page no="45"?> Words - word classes 33 lowing: Is there any evidence to suggest (or, in the absence of evidence, is it plausible to assume) that a word such as all is stored in different places in the mental lexicon (in which case word class would be prior to the word) or is it stored in one place with information about the different ways it can be used (in which case the word would be prior to the word class)? Furthermore, one might ask whether what applies to a word such as all must apply in the same way to a word such as that or to representatives of the so-called major word classes, in particular those that are characterised by inflectional properties. In other words, do the different uses of all have to be accounted for in the same way as the uses of a word such as mind in (13) SW20 Never mind that now. (14) SW23 And one look at a tree-diagram makes my mind go blank. or in the case of lecture in (15) SW24 In the event, not many people did like Morris Zapp's lecture, and several members of the audience walked out before he had finished. (16) SW78 I'm lecturing on Hazlitt. nouns verbs lecture event realize verb use noun use architecture introduce lecture lecture tree-diagram decide While such questions are certainly central to the design of a word class system within cognitive linguistics, they need not necessarily be asked in a structuralist approach. In both cases, however, the implications of certain decisions for the remaining architecture of the grammar must be considered: if one opts for different words all rather than one word all in English, then one must also provide a mechanism for relating them, for example, in terms of a particular kind of word formation process; this is not necessary if one accounts for such different occurrences as different uses which, on the whole, is the view that we tend to take in this book. 2 2 For a discussion of the problem of a word occurring in different word classes see Leisi/ Mair (1999: 86): "Bei vielen englischen Wörtern ist die Wortart also neutral oder latent. Sie ergibt sich erst in einem gegebenen Text: Die Kategorie der Wortart ist im Englischen also weithin nur noch eine funktionelle und keine formale mehr." Compare also Kastovsky (1982: 76-80) in this context. <?page no="46"?> Word classes 34 3.1.2 Criteria for word classes The system of word classes introduced in this book is based on the word class labels used in the main twentieth-century reference grammars and dictionaries. Where we do not follow established terminology, this is explicitly pointed out. 3 Although the word classes identified in modern linguistics resemble those of traditional grammar, the basis of classification is different. Traditional grammar has often been accused of intermingling the levels of form and content and drawing upon morphological and especially semantic criteria in its definitions of word classes. It is worth noting, however, that such traditional semantic descriptions still form the basis of the terminology introduced at German primary schools, for instance, where terms such as Namenwort, Tunwort or Wiewort are used. Although it cannot be denied that many nouns denote persons or things, that many verbs describe actions or states and most adjectives express qualities, this is by no means true of all the elements of these word classes. 4 Counter-examples are easy to find. Thus in (17) SW3 Dismay had been already plainly written on many faces when they assembled the previous evening for the traditional sherry reception. one could argue that a noun such as dismay expresses a state rather than a thing but one could probably make out a case for seeing sherry reception as a kind of activity rather than an entity. In fact, Robins (1971: 219) argues that such semantic criteria were only introduced because the word classes established on formal grounds for Ancient Greek and Latin were then transferred to languages in which they were no longer formally justified. 5 It is for this reason that modern structuralist linguistics emphasizes the necessity of keeping the levels of form and meaning apart. Since semantic criteria are obviously inappropriate, a word class system for English will have to be based on formal criteria, in particular morphological properties and distributional criteria such as the positions in which members of the word class can occur in a clause. The problem with such criteria, however, is that as a rule it does not seem possible to establish one criterion or a set of such properties that apply to all members of a word class and only to those members. If one takes a word such as ultimate, for instance, one would tend to classify it as an adjective on the grounds that it shows one property which 3 For the origins of the present system of word classes in Greek philosophy see Robins (1966/ 2000). 4 See Allerton (1990: 89). 5 Compare Chapter 1.4. Cf. Robins (1971: 218-219) about the development of word classes: "... though for the most part they could be justified formally in Latin and Ancient Greek, in which languages they were first worked out, several of them were rather hazily defined in terms of types of meaning or philosophical abstractions (verbs being said to stand for actions or for being acted on, and nouns for things or persons). Because such extra-linguistic criteria were taken as the basis of this word classing (though it could originally have been quite well grounded on formal criteria), these word classes were applied to the material of other languages without regard for their suitability ...". Cf. also Behrens (2005: 183): "Die Wortarten sind also ein Epiphänomen syntaktischer Konstruktionen (...) und damit emergente und sprachspezifische Konstrukte. Eine solche Sichtweise vermeidet Probleme einer starren Zuordnung oder des Überstülpens unpassender Kategorien, wie es historisch gesehen geschah, als die Wortarten des Lateinischen auf das Deutsche übertragen wurden." <?page no="47"?> Words - word classes 35 can be seen as a central feature of adjectives: it can be used attributively, i.e. before a noun in a noun phrase (> 3.4). (18) SW26 The dancer teases the audience, as the text teases its readers, with the promise of an ultimate revelation that is infinitely postponed. The problem with a word such as ultimate is, however, that it does not show any of the other criteria generally used to define the class of adjectives (> 3.4): (18a) *The revelation was ultimate. (18b) *The revelation of the text was more ultimate than that of the dancer. (18c) *The revelation seemed very ultimate. On the other hand, classifying ultimate as an adjective on the basis of the single criterion that it can occur in attributive position is not entirely satisfactory since this is a feature that adjectives share with nouns such as Jane Austen and tennis in: (19) SW24 I think I can say in all modesty I was the Jane Austen man. (20) SW26 The tennis analogy will not do for the activity of reading - it is not a to-and-fro process, but an endless, tantalising leading on, a flirtation without consummation ... Statistically, cases such as ultimate make up a relatively small number of words which are generally subsumed under the class of adjectives so that the value of the word class concept should not be questioned altogether. Nevertheless, it is important to bear in mind two facts about word classes: Word classes have to be regarded as constructs designed to meet particular descriptive or theoretical needs. The assignment of individual words to particular word classes may also involve a certain amount of arbitrariness, depending on the importance attributed to factors of semantic and phraseological analysis of particular items. In the approach outlined in this book, we will make a number of further assumptions which one must be aware of: word classes are identified on the basis of formal criteria such as morphological and distributional properties. In order to avoid a situation in which a word class label is no more than a statement about the function of a particular word in a phrase or clause in a different kind of terminology, it is the general aim of the classification to subsume under one word class words that share more than one feature. Nevertheless, although we will try to restrict the number of such cases to a minimum, there are a number of words where, within the framework of this book, it may seem appropriate and economical to say that they can occur in different word classes, in other words, where different uses of the same word will be described in terms of different word classes. Furthermore, it will not be regarded as the aim of the description to assign a word class label to every single word use in the language. If one believes in the principle that word class assignment ought to enable one to make generalizations then there is little point in classifying, for example, the use of very in <?page no="48"?> Word classes 36 (21) NW82 At the very moment when they were writing about these problems, Marx and Engels were writing the seminal texts in which the political solutions were expounded. as an adjective since it has very little in common with other adjectives. It may thus be more appropriate to describe such uses of a word as special uses outside its normal word class membership (> 3.8.2). 6 Finally, as far as the criteria used for the establishment of word classes are concerned, it is obvious that in many cases there is no one single characteristic of a word class that would apply to all its members and only to its members. 3.1.3 English word classes The following word classes (or parts of speech, a term often used in English on the model of Latin partes orationes) will be defined for English: lexical verbs write, say, … primary verbs be, do, have. modal verbs can, could, may, might, shall, should, will, would, must. lexical nouns April, month, windowpanes, snow, lawn, campus, ... pure pronouns I, you, each other, someone, … determiner-pronouns each, some, many, one, two, third, ... pure determiners a, every, no, the. adjectives cruel, grimy, unseasonable, ... adverbs however, not 7 , now, recently, silently, ... particles at, here, of, outside, to, through, whether, with, ... conjunctions and, but, or, nor. interjections oh, … 6 Another case in point could be uses of the in constructions of the type The more phonemes, the more emphasis SW23 . In a similar context, Fillmore, Kay, and O'Connor (1988: 508) raise the question: "Do we, indeed, have the right to describe the the here as the definite article? " However, one could argue that this use of the finds a parallel in uses of the type ... looking much the worse for wear SW3 and classify the in such uses as an adverb. 7 Note that CGEL (2.34) describes not (and also to when used as an infinitive marker) as "words of unique function". <?page no="49"?> Word classes occurring in the verbal head-complex 37 Most of these word class labels are commonly used in linguistics and large reference grammars such as CGEL. The one major deviation from the traditional system is the use of the term particle for words which are often classified as prepositions and subordinating conjunctions (> 3.5). It is important to bear in mind that these classes differ greatly in character. For instance, the classes of lexical verbs, nouns, adjectives and adverbs are much larger than any of the other classes (with the exception of numerals, which form a fixed system). It is also much more common for, say, a new noun to be created than a new pronoun. Thus these four classes are often referred to as open classes, whereas the term closed classes is used for the others. 8 3.2 Word classes occurring in the verbal head-complex 3.2.1 Lexical verbs 3.2.1.1 Criteria for lexical verbs In English, lexical verbs (referred to as full verbs in CGEL) can be defined on the basis of the following criteria, which are mostly morphological in character: 9 [A] Lexical verbs are morphologically marked with respect to the grammatical categories person and number in that third person singular present tense forms are distinguished from other present tense forms by the suffix {S}. (14) SW23 And one look at a tree-diagram makes my mind go blank. (22) SW22 It all goes back to Saussure's linguistics. [B] Lexical verbs have -ing-participles (also referred to as present participles) marked by {-ing} 10 . (23) SW18 Unless you're giving one yourself, of course. (24) SW18 I wouldn't want to discourage you from hearing my paper this afternoon. [C] Lexical verbs have past tense forms, which so-called regular verbs form by means of an {ed}-suffix. (25) SW20 Does everybody know this young man, who kindly looked after me when I arrived? So-called irregular verbs can make use of different morphological means to distinguish present tense and past tense forms: (26) SW29 I mean, the idea of romance as a narrative striptease, the endless leading on of the reader ... 8 Compare CGEL (1985: 2.39-2.41), where numerals and interjections are seen as "additional classes". 9 Cf. CGEL (1985: 3.2). 10 Note that CamG (2002: esp. 80-83) refers to these forms as gerund-participles but rejects a distinction between gerund and participle. For the inappropriateness of the term gerund in English see also CGEL (1985: 17.54), Esser (1992) and Herbst (2005). <?page no="50"?> Word classes 38 (27) SW20 "When I said 'Take a cab','" said Bob Busby reproachfully to Morris Zapp, "I meant from Heathrow to Euston, not from London to Rummidge." (28) SW10 Did Mr Pabst find you? (29) SW10 I was found in the toilet of a KLM Stratocruiser flying from New York to Amsterdam. In some cases, irregular verbs do not have distinct morphological forms for present tense and past tense: (30) NW33 When you put a stop to all that, when the place is silent and empty, all that is left is a large, ramshackle shed -cold, filthy and depressing. (31) NW55 She put this calmly and rationally to Charles one day, and calmly and rationally he accepted it. [D] Lexical verbs have past participles (also referred to as -ed-participles or -enparticiples) 11 , which so-called regular verbs form by means of an {-ed}-suffix. (15) SW20 ... several members of the audience walked out before he had finished. In the case of irregular verbs, the -ed-participle can take different morphological forms: (32) SW21 I've forgotten. (33) SW24 So that when each commentary was written, there would be nothing further to say about the novel in question. In some cases, the -ed-participle can be morphologically identical with other forms of the verb: (8) SW17 Have you really and truly come all the way from London by taxi? (34) SW25 By that I don't just mean that if successful it would have eventually put us all out of business ... [E] Lexical verbs have a base form, which is used as an infinitive form or a present tense form (with the exception of the third person singular) or also in imperative-'directive'-constructions: (35) SW24 Then you're not going to like my lecture, Philip. (36) SW15 I think I'll write a poem instead. (37) SW15 So tell me about structuralism. While criteria A - E referred to the inflectional morphology of lexical verbs, certain derivational morphemes can also be taken as indications of word class status: [F] If a word ends in -efy , -ify (pronounced / / ) or in the morpheme {<ize> / / }, it is a verb. 12 11 The term -en-participle refers to the fact that some irregular verbs make use of an -en-suffix and thus distinguish the participle suffix from the past tense -ed-suffix. <?page no="51"?> Word classes occurring in the verbal head-complex 39 (38) SW11 Her name badge identified her as Miss Sybil Maiden, of Girton College, Cambridge. 3.2.1.2 Verb forms The different morphological forms of lexical verbs can occur in the following uses: base form like, think, write, tell, mean, find, put finite 13 present tense (except 3rd person singular) imperative present subjunctive non-finite bare infinitive to-infinitive {S}-form makes, goes finite 3rd person singular present tense past tense form looked, arrived, meant, found, put finite past tense -ing-participle/ present participle giving, hearing non-finite -ed-participle/ past participle finished, forgotten, written, come, put non-finite 3.2.2 Modal verbs and primary verbs 3.2.2.1 Lexical verbs, modal verbs and primary verbs Apart from lexical verbs two further classes of verbs will be identified - modal verbs and primary verbs. 14 They can be distinguished from lexical verbs by two important characteristics: firstly, primary verbs and modal verbs can occur immediately before not followed by another verb: (39) NW87 Don’t be absurd. (20) SW26 The tennis analogy will not do for the activity of reading - it is not a to-and-fro process, but an endless, tantalising leading on, a flirtation without consummation ... 12 Compare Muthmann (2002: 317 and 99-100) and Stein (2007: 82 and 94-95). In the case of -ize this is to be taken as a purely morphological and not as an orthographic criterion because of words such as prize, size, life-size, half-size etc. 13 For the distinction between finite and non-finite see CamG (2002: 1173) and CGEL (1985: 3.52), where a scale of finiteness is given. 14 For primary verbs and modal verbs see CGEL (1985: 3.31-51). For a discussion of auxiliaries see CamG (2002: 92-115). For the so-called NICE constructions (Negation, Inversion, Code, Emphasis) as characteristics of auxiliaries see CamG (2002: 92). Compare Palmer (1974: 18-28). <?page no="52"?> Word classes 40 (24) SW18 I wouldn't want to discourage you from hearing my paper this afternoon. Secondly, they can precede the subject of a clause to form interrogative-'yes-noquestion'-constructions or wh-interrogative-'question'-constructions (> 7.1.2.1): (40) SW5 Is that right? Lexical verbs cannot do this but require do-support if no other primary or modal verb is present: (41) NW87 I'm sure I don't know what you're talking about. (42) SW27 Don't worry, if they don't know already, I'll tell them. (43) NW383 What does that mean? (44) NW358 Don't you want to know what the salary is? Verbs that are distinguished in their use from lexical verbs in this way are often referred to as auxiliaries. The term auxiliary is sometimes used as a term for a formal word class, as in CamG (2002: 92), and sometimes, as in CGEL (1985: 2.29), as a term to describe a particular function of words. We will follow CGEL and say that this function (for which we will introduce the term pre-head in 3.2.3) 15 can be performed by two subclasses of verbs, namely ⌐ modal verbs (> 3.2.2.3) and ⌐ primary verbs, which also have the characteristics of lexical verbs. 3.2.2.2 Criteria for primary verbs The class of primary verbs consists of the verbs be, have and do, which belong to the 20 most frequent words of the English language. 16 They share five criteria with lexical verbs: Primary verbs [A] are morphologically marked with respect to the grammatical categories person and number in that they distinguish a third person singular present tense form (is, has, does) - and in the case of the verb be also a first person singular present tense form (am) - from the other present tense forms (are, have, do). [B] have -ing-participle forms marked by {-ing}. [C] distinguish between present tense and past tense forms (was/ were, had, did). [D] have distinct past participle forms (been, had, done). 15 However, it would hardly be acceptable to take the unit of the verbal head-complex as the basis of identifying the three classes of verb until the verbal head-complex itself has been defined independently of the notion of verb. It is for this reason that we will establish these word classes on the basis of criteria that are independent of the notion of the verbal head-complex. 16 In the BNC, forms of be are the second most frequent lemma to be found, have ranges 8th and do twentieth. <?page no="53"?> Word classes occurring in the verbal head-complex 41 [E] have a base form, which is used as an infinitive form or a present tense form (with the exceptions given under A) or also in imperatives. Furthermore, they can [F] occur immediately before not followed by another verb (39) NW87 Don’t be absurd. [G] precede the subject of a clause to form interrogative-'yes-no-question'constructions or wh-interrogative-'question'-constructions (> 7.1.2.1): (40) SW5 Is that right? (8) SW17 Have you really and truly come all the way from London by taxi? (43) NW383 What does that mean? [H] occur immediately before non-finite forms (bare infinitives or participles) of lexical verbs in sentences. (45) SW27 I have listened to your paper with great interest, Morris. (46) SW27 Your mind has lost none of its sharpness since we first met. It could be argued that be, do and have should be seen as members of two word classes since in some cases they are clearly like lexical verbs and in other cases they function as pre-heads. Thus, certainly, the lexical element decreases from (47) to (50): (47) SW10 I did my MA thesis on Shakespeare's influence on T.S. Eliot. (48) SW28 And as it looks as if we have done our duty for today, shall we all adjourn for a drink? (49) SW29 "Professor Zapp, I did so enjoy your lecture," she said, with a greater degree of enthusiasm than Persse had expected or could, indeed, bring himself to approve. (50) SW28 Does your mother know you're away out listening to that sort of language? Similarly, have in (48) or had in (49) have primarily a grammatical function, which is not the case in (51) SW28 "I have just one question," said Philip Swallow. (52) SW21 Rumour has it that Zapp had an affair with Hilary Swallow, and Swallow with Mrs Zapp. The same kind of gradient can be observed with be: (53) SW3 April is the cruellest month. (54) SW28 "You weren't mistaken about what you were trying to do," said Morris Zapp, relighting his cigar, "you were mistaken in trying to do it". (55) SW29 I'm working on the subject of romance for my doctorate. (56) SW29 Professor Dempsey is taking me in his car, so that will save time. The main reason for making a distinction between lexical verbs be, do and have and what would then be referred to as auxiliary verbs be, do and have could be semantic <?page no="54"?> Word classes 42 since one could argue that the lexical verb expresses meaning and the auxiliary has a purely grammatical function. Another argument in favour of such a distinction is that do-support (> 3.2.2.1), which is typical of lexical verbs, can also occur in some uses of have and do: (57) SW14 Where did you do your graduate work? (58) NW20 Gary doesn't have school today. (59) NW23 I haven't time. Nevertheless, it is clear that this is not a case of clear-cut distinctions but of gradience. For instance, while the meaning of is in (53) could be described as relating April and the cruellest month, this is equally true in the case of (54). Although most grammarians would probably tend to assign a different grammatical structure to (54) and (55), the semantic relationship between working on the subject of romance for my doctorate and I is very much the same as that between mistaken in trying to do it and you in (54). It thus seems appropriate to treat the verbs be, do and have as one class of verb, which shares many criteria with lexical verbs (A - E) but meets further criteria which do not apply to lexical verbs. 3.2.2.3 Modal verbs The class of modal verbs comprises the following verbs in English: can, may, must, shall, will, could, might, should, would. Modal verbs can occur [A] immediately before bare infinitives of lexical verbs. (60) SW20 Morris, we must get you a lapel badge so that everybody will know who you are. (42) SW27 Don't worry, if they don't know already, I'll tell them. [B] immediately before not (or in contracted forms with not) followed by another verb. (61) SW39 I can't follow an argument without an example. (62) SW22 It's a purely arbitrary relationship, and there's no reason why English speakers shouldn't decide that from tomorrow, d-o-g would signify 'cat' and c-a-t, 'dog'. [C] in front of the subject of a clause to form interrogative-'yes-no-question'-constructions or wh-interrogative-'question'-constructions (> 7.1.2.1): (63) NW326 Could you drop it here tomorrow morning before eight-thirty? (64) NW264 What shall I do about my teaching on Thursday? (65) NW307 But why shouldn't universities be in nice places rather than nasty ones? While modal verbs share criteria B and C with primary verbs (see F and G above), they are distinguished from lexical and primary verbs in their morphology in that they are <?page no="55"?> Word classes occurring in the verbal head-complex 43 [D] not morphologically marked for person and number and do not have nonfinite forms such as participles or infinitives. 17 It is remarkable that although modal verbs lack the morphological distinctions of other verbs, they must be regarded as verbs because they commute with other verbs (> 1.3.2), i.e. they can occur in the same position of the clause (see also criteria A and B): (66) NW219 Are you really? (67) SW7 Would you, indeed? (68) NW218 I know. (69) NW219 I might. 3.2.2.4 Dare, need and ought The distinction between modal verbs and lexical verbs is not totally clear-cut. Thus, the verbs dare and need can be used as lexical verbs as in (70), but they can also be used in ways which are typical of modal verbs such as (71): (70) NW353 What we need to ask ourselves is, what do we have as a department that's marketable in the outside world? (71) NW179 Well, you needn't look so apprehensive ... The verb ought is sometimes treated as a marginal modal because it does not allow complementation by a pure infinitive but by [to_INF]: 18 (72) NW85 You know, I ought to be able to order one of you to do this shadow nonsense. 3.2.3 The verbal head-complex On the basis of the word classes lexical verb, primary verb and modal verb, it is now possible to describe the structure of the English verbal head-complex (> 2.2.3). A verbal head-complex (theoretically) can consist of up to five different verbs. 19 The right-most constituent of the head-complex is called head, the others will be called pre-head. The structure of the head-complex of the following examples is given in the table below: (53) SW3 April is the cruellest month. (55) SW29 I'm working on the subject of romance for my doctorate. 17 CGEL (1985: 4.59) treats could, might, should and would as past tense forms of can, will, shall and will. See, however, also CGEL (1985: 3.30). Compare also CamG (2002: 202). 18 See CGEL (1985: 3.41-43). For used to and modal idioms compare CGEL (1985: 3.44-46). See also CamG (2002: 109-115). 19 Cf. CGEL's (1985: 2.28) examples may have been being sunk and He may have been being interviewed then, which are explicitly commented on as being very rare. A search in the BNC did not yield any such constructions. Note that possible valency complements of the head can also contain verbs. We would not analyse want as a pre-head in I don’t particularly want to emigrate NW383, however, since want lacks the properties outlined for modal and primary verbs in 3.2.2. For this kind of catenative analysis see Palmer (1974: 166-211) and Mindt (2000: 106-111). <?page no="56"?> Word classes 44 (49) SW29 "Professor Zapp, I did so enjoy your lecture," ... (45) SW27 I have listened to your paper with great interest, Morris. (17) SW3 Dismay had been already plainly written on many faces when they assembled the previous evening for the traditional sherry reception. (60) SW20 Morris, we must get you a lapel badge so that everybody will know who you are. pre-head pre-head pre-head pre-head head (53') is (55') am working (49') did enjoy (45') have listened (17') had been written (60') must get will know CGEL may have been being sunk If the verbal head-complex contains a lexical verb, this will be the head; if primary verbs occur as pre-heads, their function is to mark the verbal head-complex for person, number and tense. Otherwise, pre-heads serve to express various grammatical distinctions: have is used as a pre-head to distinguish perfective aspect forms (form of have + past participle of the head as in (45)) from non-perfective aspect forms. 20 (45) SW27 I have listened to your paper with great interest, Morris. be is used as a pre-head to distinguish: progressive aspect forms (consisting of a form of be + present participle of the head as in (55)) from simple aspect forms (55) SW29 I'm working on the subject of romance for my doctorate. passive voice forms (consisting of a form of be + past participle of the head as in (17)) from active voice forms 20 We follow CGEL (1985: 4.2-4.19) in using the category tense for the distinction between the synthetic forms present tense and past tense and subsuming the perfective - non-perfective distinction under the category of aspect, which in CamG (2002: 116) are referred to as secondary tenses. For a detailed discussion of the category of aspect in reference grammars see Schüller (2005). The notion of a future tense would seem inappropriate in English, cf. CGEL (1985: 4.3), CamG (2002: 208-210) or Herbst (2005). <?page no="57"?> Word classes occurring in the head-complex of the noun phrase 45 (17) SW3 Dismay had been already plainly written on many faces when they assembled the previous evening for the traditional sherry reception. do is used as a pre-head in interrogative-'question'-constructions of the type (50) SW28 Does your mother know you're away out listening to that sort of language? to express special emphasis (49) SW29 "Professor Zapp, I did so enjoy your lecture," she said, with a greater degree of enthusiasm than Persse had expected or could, indeed, bring himself to approve. If more than one pre-head is realised in the verbal head-complex, then the order in which these categories are expressed is as follows: modal verb or primary verb expressing number, person and tense > have for perfective aspect > be for passive > be for progressive > head. 3.3 Word classes occurring in the head-complex of the noun phrase 3.3.1 Lexical nouns 3.3.1.1 Criteria for lexical nouns In English, lexical nouns can be defined on the basis of the following grammatical and distributional criteria: 21 All lexical nouns can occur immediately [A] after a and/ or the a structuralist, the arbitrariness of the signifier [B] after 's-genitives Shakespeare's influence on T.S. Eliot [C] after adjectives cruellest month [D] before relative clauses a man who once believed in the possibility of interpretation Many lexical nouns show the following morphological properties: [E] inflected genitive forms Master's 21 For a much more detailed account see CGEL (1985: 5.51.-5.9) and CamG (2002: 326-327). See also Aarts/ Aarts (1982/ 1988: 22-28). <?page no="58"?> Word classes 46 [F] number conference - conferences Finally, a criterion concerning internal morphological structure can be given that applies to a limited number of nouns: [G] If a word ends in one of the following derivational suffixes, it is a lexical noun: 22 -ance, -dom, -ee, -eer, -ence, -ess, -ette, -hood, -ist, -ment, -ness, -ship conference, structuralist If one takes these criteria as the defining criteria for the class of lexical nouns, then it is also possible (without any circularity in the definition) to define the noun phrase as a type of phrase that can be headed by a lexical noun and to make the following statement about lexical nouns: [H] All lexical nouns can function as heads of noun phrases, which can have preheads, premodifiers and postmodifiers. 3.3.1.2 Classes of lexical nouns Lexical nouns are often subdivided into the following classes: 23 lexical nouns proper April, Shakespeare, Robyn common count structuralist, month, conference uncount dismay, gossip, structuralism The main distinguishing features between these classes concern criteria A and F in 3.3.1.1. Quirk, Greenbaum, Leech and Svartvik (CGEL 1985: 5.2) establish a table for noun classes, which can be represented as follows (* representing unacceptable forms): proper common count uncount (a) singular noun + * + (b) the + noun * + + (c) a + noun * + * (d) some + noun * * + (e) plural noun * + * This table accounts for the use of a common count noun such as conference in sentences such as 22 Compare Muthmann (2002) and Stein (2007). Note that also most words ending in the suffixes -age, -ation, -ism are nouns. 23 Cf. CGEL (1985: 5.2 and 5.60). <?page no="59"?> Word classes occurring in the head-complex of the noun phrase 47 (73) SW18 How's it going on, the Conference? (74) SW18 I've never been to a conference before, so I've no standards of comparison. (75) SW279 I told her the other day, 'If you didn't spend so much time going to conferences, Angie, you would have gotten your doctorate by now, and put all this nonsense behind you.' Proper nouns 24 can be distinguished from common nouns with respect to criterion A in that they do not occur with a or the in a way that common nouns do. Thus, with a proper noun such as April only the use without articles is possible in a sentence such as (76) SW5 Shocking weather for April. (76a) *Shocking weather for an April. (76b) *Shocking weather for the April. whereas with a common count noun such as conference, use without an article is ruled out: (76c) Shocking weather for a conference/ *conference. More precisely, one would have to say that proper nouns lack article contrast since in sentences such as (77) proper nouns such as Swallow or Rummidge cannot be used with a or the but in (78) or (79) the could not be deleted: (77) SW5 I told Swallow nobody would come to Rummidge, but he wouldn't listen. (77a) *I told a/ the Swallow nobody would come to Rummidge, but he wouldn't listen. (77b) *I told Swallow nobody would come to a/ the Rummidge, but he wouldn't listen. (78) NW305 ... when you looked up from your work, you could see the horizon of the English Channel ruled like a faint blue line between the roofs of the two neighbouring houses. (78a) ... *when you looked up from your work, you could see the horizon of English Channel ruled like a faint blue line between the roofs of the two neighbouring houses. (79) NW305 ... the photos of David Bowie and The Who and Pink Floyd had been taken off the walls ... (79a) * ... the photos of David Bowie and Who and Pink Floyd had been taken off the walls ... Similarly, uncount nouns are distinguished in CGEL (1985: 5.2) from count nouns in that the former do not occur after a and lack number contrast: (17) SW3 Dismay had already been plainly written on many faces when they assembled the previous evening for the traditional sherry reception. (17a) *A dismay had already been plainly written on many faces ... 24 CGEL (1985: 5.60) makes a distinction between proper nouns, which are single words, and names, "which may or may not consist of more than one word". <?page no="60"?> Word classes 48 (17b) *Dismays had already been plainly written on many faces ... (80) SW0 ... exchange gossip and confidences … (80a) *... exchange a gossip and confidences (80b) *... exchange gossips and confidences (37) SW15 So tell me about structuralism. (37a) *So tell me about a structuralism. (37b) *So tell me about structuralisms. The criteria given in the CGEL table do not always apply one hundred per cent, however. For instance, there are perfectly normal uses of proper nouns in the plural: (3) SW0 ... not all Aprils, for that matter, are marked by sweet showers and dulcet breezes. (81) SW21 The Swallows returned together. Similarly, count nouns can occur after some: (82) SW286 She was going to some conference or other. More importantly, proper nouns or uncount nouns can of course occur after a or the when followed by a relative clause or some other postmodifier (and with this reservation criterion A holds for all lexical nouns): 25 (83) BNC This, coupled with the gossip that for some time surrounded the couple, soon convinced the constables they had the right people. (84) BNC In some respects this method has much in common with the structuralism of the Prague School. The criteria that are used to establish the distinction between proper nouns and common nouns thus describe differences in the conditions of use with a or the rather than the possibility of the occurrence of particular forms as such. Semantic factors play a very important role here. 26 CGEL (1985: 5.2) describes count nouns as "denoting individual countable entities" and uncount (or non-count) nouns as "denoting an undifferentiated mass or continuum". It must be borne in mind, however, that such descriptions do not reflect reality as such but at best explain how a particular language classifies (or how its speakers perceive) reality. Similarly, proper nouns can be seen as names of entities in the real world, i.e. as words that do not have meaning but serve to establish reference to a particular individual entity. Thus the proper noun in a sentence such as (85) NW63 Swallow looks tired and careworn and slightly seedy. is to be interpreted as a statement referring to one particular individual (in this case a fictional character in a novel by David Lodge) and not as a statement about a class of people called Swallow or everybody carrying that name. 25 See also CGEL (1985: 5.61-5.64). 26 For uses such as a Shakespeare see CGEL (1985: 5.63). <?page no="61"?> Word classes occurring in the head-complex of the noun phrase 49 Many nouns seem to occur in both uses in different meanings. Thus, whereas confidence in (80) SW0 ... exchange gossip and confidences … can be seen as a count noun, it could be analysed as an uncount noun in (86) NW60 She has youth, she has confidence, she regrets nothing. since there is no possibility of a plural there. Since this is true of many nouns it may thus be more appropriate to speak of different uses rather than different subclasses of nouns. In fact, one could argue that (87) NW307 Well, was it a good idea to build so many universities in parks on the outskirts of cathedral cities and county towns? is an example of a use of the noun university as a count noun, that in (88) NW181 She didn't go to university, you see. it is used as an uncount noun and that in (89) NW214 She drove to and fro between her cosy little house and her warm, well-lit room at the University, ignoring the dismal weather. the University represents a proper noun similar to the City in (90) NW179 Basil's decision to go into the City, announced to an incredulous family in his last undergraduate year at Oxford, had not been an idle threat. 3.3.2 Pronouns 3.3.2.1 Word classes in the head-complex of the noun phrase: an overview Within the head-complex of a noun phrase, lexical nouns typically function as heads. Lexical nouns are not the only word class that occurs in the head-complex of a noun phrase, however. 27 On the one hand, nouns that can function as heads can be preceded by one or more pre-heads as in the case of the trouble and these conferences in (91) SW15 That's the trouble with these conferences. On the other hand, words other than nouns can function as heads of noun phrases such as that in (91) or myself in (92): (92) SW27 I wouldn't call myself a structuralist. Many of these words - like that or these - can actually function as pre-heads and heads, but a few - like the - can only function as pre-heads, while others - such as myself - can only function as heads. This has given rise to inconsistent use of terminology in linguistics, which concerns mostly the terms pronoun and determiner. The term pronoun is traditionally used to refer to words that can "take the place of a noun" and the term determiner was introduced as a cover term for all words that can take a similar function in the phrase as the article the, for instance. In order to avoid 27 Compare 2.4.2, 3.3.5 and 4.2.2 for a description of the head-complex of the noun phrase. <?page no="62"?> Word classes 50 assigning such words as that or these to two different word classes - determiner and pronoun, as is done in CGEL, for example, we suggest the following threefold distinction between: pure pronouns determiner-pronouns pure determiners. These three word classes have a number of distributional criteria in common such as that they do not usually occur after the indefinite article a 28 after adjectives. They differ in the following respects: Pure pronouns and determiner-pronouns can occur as heads of noun phrases. Determiners and determiner-pronouns can occur as pre-heads of noun phrases; so can 's-genitive forms of pronouns (and nouns). This means that a determiner-pronoun can function as pre-head and head without any formal change, whereas pronouns function as pre-heads only in an inflected form. It should be pointed out, however, that these three classes are relatively heterogeneous since some of them comprise several subclasses. Since pronouns, determinerpronouns and determiners are mostly closed classes anyway, i.e. classes whose membership is restricted to a limited number of words in the language, it is possible - and probably more appropriate - to define the respective classes in terms of an inventory of the items belonging to each class. 3.3.2.2 Subclasses of pure pronouns The class of pure pronouns comprises the following subclasses: personal pronouns (they - their) reflexive and reciprocal pronouns (themselves, each other) indefinite pronouns (somebody, everything) temporal pronouns (yesterday) wh-pronoun (who) 28 In rather special uses, the pronouns he and she can occur after a: You see, the conductor wasn't a he but a she. BNC <?page no="63"?> Word classes occurring in the head-complex of the noun phrase 51 3.3.2.3 Personal pronouns Personal pronouns can be characterized by a number of morphological criteria. Personal pronouns are marked for the grammatical categories [A] person by distinguishing between 1st, 2nd and 3rd person forms [B] number by having different singular and plural forms (1st and 3rd person only) [C] case by distinguishing between nominative, accusative (1st and 3rd person only) and genitive case forms 29 [D] gender by distinguishing between masculine, feminine and neuter (and gender-neutral) forms in the 3rd person. This can be summarized as follows: 1 sg 2 sg 3 sg 1 pl 2 pl 3 pl personal pronouns nominative I you he/ she/ it we you they 30 genitive my mine your yours his/ her/ their/ its his/ hers/ theirs/ its our ours your yours their theirs accusative me you him/ her/ them/ it us you them In this system, your, their, etc. and yours, theirs, etc., which are often referred to as possessive pronouns, are seen as genitives of the personal pronouns, the former occurring as parts of noun phrases before the noun, the latter occurring as independent noun phrases: (93) SW33 Do you know your cousin Bernadette, Persse? (94) SW33 All on account of that daughter of hers, Bernadette. With respect to phrase structure, the following criterion can be added: [E] Personal pronouns can occur as heads of noun phrases which do not have preheads or premodifiers, and, in the respective genitive-forms, as pre-heads. 31 (95) NW40 It might seem a bit bleak, a bit inhuman ('antihumanist, yes; inhuman, no', she would interject), somewhat deterministic ('not at all; the truly determined subject is he who is not aware of the discursive formations that determine him. ...) (93) SW33 Do you know your cousin Bernadette, Persse? 29 We follow CamG (2002: 456) in using the terms nominative and accusative although CGEL (1985: 5.112/ 6.2) uses the terms subjective and objective case. 30 Note that if they is used as a gender neutral pronoun referring to one person, it still takes a plural verb. Compare CGEL (1985: 6.9). 31 The range of possible postmodifiers is rather restricted: you British NW94 . Cases of premodification such as clever you or stupid me are so rare that we consider them as idiomatic here. <?page no="64"?> Word classes 52 3.3.2.4 Reflexive pronouns The class of reflexive pronouns comprises the following items: myself, yourself, himself, herself, itself, themself 32 , ourselves, yourselves and themselves. Reflexive pronouns [A] are marked for the grammatical category person as the 1st, 2nd and 3rd person forms are morphologically distinguished [B] are marked for the grammatical category number in that singular and plural are morphologically distinguished [C] often occur in clauses with an antecedent with which they concord with respect to person and number [D] are always heads of noun phrases that do not have pre-heads or modifiers. The antecedent can occur in a phrase preceding the reflexive pronoun as in (92) SW27 I wouldn't call myself a structuralist. or as head of a noun phrase in which the reflexive pronoun functions as a postmodifier as in (96) SW25 ... it isn't possible because of the nature of language itself, in which meaning is constantly being transferred from one signifier to another and can never be absolutely possessed. In certain contexts, no such antecedent is necessary: (97) SW27 To read is to surrender oneself to an endless displacement of curiosity and desire from one sentence to another, from one action to another, from one level of the text to another. 3.3.2.5 Reciprocal pronouns The class of reciprocal pronouns contains the two elements each other and one another. Reciprocal pronouns [A] are marked for the grammatical category case as they have an 's-genitiveform [B] can only occur in clauses with an antecedent with the same referent [C] can occur as heads of noun phrases which do not have pre-heads or modifiers, and, in the 's-genitive-form, as pre-heads. (98) SW4 Long before it was all over they would have sickened of each other's company ... (99) NW55 Naturally they saw less of each other ... 32 Themself as a gender-neutral reflexive pronoun is probably to be regarded as a marginal member of the class. In the BNC, 26 instances could be found as opposed to over 15000 for herself and almost 29000 for himself. LDOCE4 lists themself as a lemma but points out that many people disapprove of its use. <?page no="65"?> Word classes occurring in the head-complex of the noun phrase 53 3.3.2.6 Indefinite pronouns A further group of pronouns that ought to be identified will be referred to as indefinite pronouns here. 33 This set comprises the following items: anything, something, everything, nothing, anyone, someone, everyone, no one, none, anybody, somebody, nobody, everybody. Indefinite pronouns share a number of criteria in that they [A] cannot follow an adjective [B] can follow adverbs such as almost or hardly [C] can be followed by an of-phrase [D] can occur as heads of noun phrases which do not have pre-heads or premodifiers, and, in the 's-genitive-form, as pre-heads. 34 (100) SW0 There are conferences on almost everything these days, including the works of Geoffrey Chaucer. (101) SW4 Within a very short time they had established that none of the stars of the profession was in residence - no one, indeed, whom it would be worth travelling ten miles to meet, let alone the hundreds that many had covered. 3.3.2.7 Temporal pronouns The subclass of temporal pronouns contains the items today, tonight, tomorrow and yesterday. As is pointed out in CamG (2002: 429 and 564-565) these words, which most dictionaries and grammars subsume under the category of adverbs, can be classified as pronouns on the basis of a number of criteria. 35 Temporal pronouns [A] are marked for the grammatical category case in that they have an 's-genitive-form [B] cannot follow the (as the head of a phrase) [C] can follow be at the end of a sentence [D] can occur as heads of noun phrases which do not have pre-heads or premodifiers, and, in the 's-genitive-form, as pre-heads. (102) NW19 Magpies strut from flowerbed to flowerbed, greedily devouring the grubs that he turned up in yesterday's gardening. 33 Compare CGEL (1985: 6.45-6.49), whose category indefinite pronouns is however much wider since it includes words such as any, which will be treated as determiner-pronouns here. 34 Note, however, that not all indefinite pronouns (everything, nothing) have 's-genitive forms. 35 Classifying these words as pronouns furthermore establishes a parallel to other noun phrases expressing time relations as in Dismay had been already plainly written on many faces when they assembled the previous evening for the traditional sherry reception. SW3 <?page no="66"?> Word classes 54 (103) NW77 ... to whom every day was the same as yesterday and tomorrow, and every year the counterpart of the last and the next. [quotation from Hard Times] 3.3.2.8 who and which Finally, the words who and which can be included in the list of pronouns: who is marked for case by having a genitive form whose and the accusative forms who/ whom and gender (personal gender) (104) SW8 Swallow is the man chatting up that rather dishy girl who just came in ... (105) NW74 All the students, even those who have been staring out of the window, react to this. (106) NW62 Who will notice? which also has a genitive form whose (> 3.3.3 and 3.7): (107) NW42 Under the umbrella of a degree course in English Literature, Robyn read Freud and Marx, Kafka and Kierkegaard, which she certainly couldn't have done at Oxbridge. (108) NW85 There is a widespread feeling in the country that universities are "ivory tower" institutions, whose staff are ignorant of the realities of the modern commercial world. Who (whose, whom) and which are mostly used in interrogative constructions and relative constructions, which is why in these two uses they are often analysed as interrogative or relative pronouns (> 3.8.1). 36 3.3.3 Determiner-pronouns Determiner-pronouns can [A] be followed by a verb with which they show concord as to the categories of person and number [B] be followed by a noun (which can be followed by a verb showing concord with the noun) [C] be followed by adjectives [D] can - without any change in form - occur as pre-heads or heads in noun phrases. (109) SW3-4 ... To veterans of conferences held in British provincial universities, these were familiar discomforts ... (91) SW15 That's the trouble with these conferences. (110) SW9 I can't read these little badges without my glasses. The class of determiner-pronouns subsumes the following subclasses: 36 Compare CGEL (1985: 2.45) for wh-words. For which see also 3.3.3. <?page no="67"?> Word classes occurring in the head-complex of the noun phrase 55 demonstratives: this - these and that - those 37 (111) SW26 I have been reading books for their meaning all my life - or at least that is what I always thought I was doing. (112) SW3 For they were all well acquainted with that poem, being University Teachers of English Language and Literature, ... indefinite determiner-pronouns: any, each, some, all, both, either, neither 38 , enough, several, much - more - most, many - more - most, little - less - least, few - fewer - fewest, such (113) SW3 Some taps could not be turned on, and some could not be turned off. (112) SW3 For they were all well acquainted with that poem, being University Teachers of English Language and Literature, gathered together here, in the English Midlands, for their annual conference, and few of them were enjoying themselves. (114) SW19 Most of them are holed up at home, decorating their living-rooms or weeding their gardens, and the few with two original ideas to rub together are off somewhere at conferences in warmer, more attractive places than this. (115) SW19 I don't think we have that many lecturers over here. numerals, where a distinction can be made between 39 cardinal numbers: one, two, three, etc. ordinal numbers: first, second, third, etc. (116) SW3 "Well, there are fifty-seven people at the conference, altogether," Persse said. (117) SW3 "There must be more than fifty-seven," growled Morris Zapp. (118) SW18 The first thing I always seem to have to do when I arrive in England is buy myself a hat. wh-determiner-pronouns: what (and some uses of which 40 )(> 3.8.1) (119) NW259 You know what I mean? (120) NW306 'What letter is that? ' Robyn asked. (121) NW60 What Robyn likes to do is to deconstruct the texts, to probe the gaps and absences in them, to uncover what they are not saying, to expose their ideological bad faith, to cut a cross-section through the twisted strands of their semiotic codes and literary conventions. (122) SW8 Which one is Professor Swallow? 37 Note that the demonstratives this and that as well as a number of indefinite determiner-pronouns such as all or more also have adverb uses: I don't think we have that many lecturers over here. SW19 or ... all too eager to show off his expertise SW22 . This means that one might consider subsuming these words under a separate word class depending on whether one regards the meaning expressed by words when used in noun phrases and when used in adverb phrases as sufficiently similar. 38 On uses of the type ... neither was I or ... I wasn't either see CamG (2002: 1308). 39 For fractions (one third) and multipliers (double) see CGEL (1985: 5.18-19). 40 Note that in some uses which shows features of a pronoun and a determiner-pronoun. <?page no="68"?> Word classes 56 3.3.4 Pure determiners The class of pure determiners includes the definite article the the indefinite article a/ an the items every and no. (123) SW25 But every decoding is another encoding. (124) SW24 That is, I thought the goal of reading was to establish the meaning of texts. (125) SW28 "There is no point," said Morris Zapp. Although some grammars treat pure determiners and determiner-pronouns as members of the same word class, we regard pure determiners as a separate word class because they differ from determiner-pronouns with respect to criterion A: Pure determiners [A] cannot be followed by a verb with which they show concord as to the categories of person and number [B] can be followed by a noun (which can be followed by a verb showing concord with the noun) [C] can be followed by adjectives [D] only occur as pre-heads in noun phrases. From a phraseological point of view it would certainly make sense to analyse prefabricated chunks such as a number of, a kind of as complex determiners because semantically they have the same function as a determiner: (126) NW72 In the 1840s and 1850s," says Robyn, "a number of novels were published in England which have a certain family resemblance. 3.3.5 Lexical nouns, pronouns, determiner-pronouns and pure determiners in the noun phrase On the basis of the definitions of the classes determiner, determiner-pronoun, pronoun and lexical noun it is now possible to describe the structure of the head-complex of a noun phrase (> 4.2.2) in the following way. The head-complex consists of a head, which is the right-most constituent of the headcomplex of the noun phrase, and up to three pre-heads, where heads can be realised by lexical nouns, pronouns and determiner-pronouns pre-heads can be realised by determiner-pronouns and pure determiners as well as genitive forms of nouns or pronouns. 41 41 Note that the genitive forms my, your etc. can function as pre-heads, as in The students stir restlessly in their seats, shuffling their papers and capping their pens NW82 . However, this is not taken as an argu- <?page no="69"?> Word classes occurring in the head-complex of the noun phrase 57 (125) W28 "There is no point," said Morris Zapp. (91) SW15 That's the trouble with these conferences. (109) SW3-4 ... To veterans of conferences held in British provincial universities, these were familiar discomforts ... (127) SW28 We maintain our position in society by publicly performing a certain ritual, just like any other group of workers in the realm of discourse - lawyers, politicians, journalists. (128) NW73 They are novels in which the main characters debate topical social and economic issues as well as fall in and out of love, marry and have children, pursue careers, make or lose their fortunes, and do all the other things that characters do in more conventional novels. pre-head pre-head pre-head head (125') no point (91') these conferences (109') these (127') any other group (128') all the other things As is apparent from this table, determiner-pronouns and pure determiners can combine in certain ways. A determiner complex of this kind consists maximally of three items as in all the other. Quirk, Greenbaum, Leech and Svartvik (CGEL 1985: 5.10-25) make a distinction between predeterminers, central determiners and postdeterminers: predeterminers such as all precede central determiners such as the, which can be followed by postdeterminers such as other. 42 However, it seems that this distinction is not sufficiently discrete to account for the complexity of possible combinations. It may be more appropriate to consider it a property of the individual determiner-pronoun or determiner or at best a property of the subclasses identified above whether they can function as the first, second or last pre-head (or head) in which kinds of combinations. ment for classifying personal pronouns as determiner-pronouns because genitives of nouns also occur as pre-heads. Note, however, that not all types of pronoun have genitive forms. Genitive forms such as my, your, their are often described in terms of a word class possessive pronouns. 42 Since this terminology suggests a hierarchy in which central determiners take a more central position than preor postdeterminers, it will not be adopted here. <?page no="70"?> Word classes 58 3.4 Adjectives Adjectives are commonly defined by a number of morphological and positional criteria. 43 Adjectives [A] can occur in attributive position, i.e. as modifiers in a noun phrase (129) SW22 But there was an interesting discussion of structuralism afterwards. [B] can occur in predicative position, i.e. after verbs as heads of adjective phrases (130) SW14 I'm very interested in structuralism. [C] can be pre-modified by adverbs (131) SW23 Jakobson cites the gradation of positive, comparative and superlative forms of the adjective as evidence that language is not a totally arbitrary system. [D] have comparative and superlative forms (132) SW32 "Even the moon seems dimmer than at home," said Persse. (53) SW3 April is the cruellest month. (133) SW30 Thank you, Professor Zapp, you have been most helpful. It has to be pointed out that there is considerable overlap with other word classes since not all adjectives fulfil criteria A to D. 44 Thus, as was pointed out in 3.1.2, an adjective such as ultimate only meets criterion A but not criteria B, C or D: (18) SW26 The dancer teases the audience, as the text teases its readers, with the promise of an ultimate revelation that is infinitely postponed. Similarly, afraid occurs in predicative position (134) NW85 Yes, I'm afraid you're probably right. but not in attributive position: (134a) *an afraid person. At the same time, nouns can freely occur in attributive position as in examples (19) SW24 I think I can say in all modesty I was the Jane Austen man. (20) SW26 The tennis analogy will not do for the activity of reading - it is not a to-and-fro process, but an endless, tantalising leading on, a flirtation without consummation ... In some instances, the distinction between adjectives and participles is particularly difficult: thus one could make out a case for analysing combined or spoken, which do not meet criteria B, C and D, as participle forms of a verb in sentences such as 43 See CGEL (1985: 71.-72.) and CamG (2002: 527-531). CamG adds the criterion of possible postposition, i.e. the occurrence of an adjective after a noun or pronoun within a noun phrase as in... two people present who could agree on the simplest précis of what had been said? SW28 44 For a detailed discussion of this see CGEL (1985: 7.3-7.19) or CamG (2002: 536-542). <?page no="71"?> Adverbs and particles 59 (135) SW22 There is no absolute reason why the combined phonemes d-o-g should signify a quadruped that goes 'woof woof' rather one that goes 'miaou'. (136) SW25 Not so, because the same axiom, every decoding is another encoding, applies to literary criticism even more stringently than it does to ordinary spoken discourse. or for classifying postponed as an adjective in (18) SW26 The dancer teases the audience, as the text teases its readers, with the promise of an ultimate revelation that is infinitely postponed. 3.5 Adverbs and particles 3.5.1 Adverbs Adverbs are one of the most problematic word classes. Quirk, Greenbaum, Leech and Svartvik (CGEL 1985: 7.46) speak of "the most nebulous and puzzling of the traditional word classes" and point out "that it is tempting to say simply that an adverb is an item that does not fit the definitions for other word classes". It is for this reason that it may seem advantageous to define the class of adverbs in a more restrictive way than is done traditionally and subsume a number of words that are traditionally classified as adverbs under other word classes. This has been done by Pullum and Huddleston in the Cambridge Grammar of the English Language (2002), for instance, whose approach we will follow with some modifications. In this restricted sense of the term adverb, we classify as adverbs words that are not members of another word class and that: [A] can function as premodifiers of adjectives: (9) SW22 "Structuralism? " said Dempsey, coming up with a sherry for Angelica just in time to hear Persse's plea, and all too eager to show off his expertise. (130) SW14 I'm very interested in structuralism. (137) SW18 I went over it last night in the plane, while the movie was showing, and I was pretty pleased with it. [B] can function as premodifiers of other adverbs: (138) SW25 But if I use my words it follows that I have changed your meaning, however slightly ... (139) NW44 It was true that he had been educated at a public school, but he managed to disguise this handicap very well. [C] can occur immediately before a lexical verb (i.e. between pre-heads and the head of a verbal head complex): (140) SW22 You've really got to tell me what structuralism is all about. (141) SW22 I was just explaining structuralism to this young man. (8) SW17 Have you really and truly come all the way from London by taxi? <?page no="72"?> Word classes 60 [D] can occur at the beginnings or ends of sentences (often outside the subjectpredicate structure, i.e. as adjunct units): (142) SW26 "Actually, you know, my name is Angelica," she smiled. (143) SW37 If the words are fixed once and for all, on the page, may not their meaning be fixed also? (144) NW50 Anyway, you'll never get a job here, Robyn. Criteria A - D refer to the function of adverbs in phrases or in the structure of the clause. Morphological criteria can be used to identify some adverbs: 45 [E] Words that are derived from adjectives by means of the suffix {-ly} are adverbs. In this sense, differently and carefully in (145) SW23 Different languages divide up the world differently. (146) SW30 You have to treat these professors carefully, Persse. are clearly adverbs, although not all words ending in -ly are, since the suffix {-ly} is also a derivational suffix of adjectives in words such as friendly. [F] Some adverbs have comparative and superlative forms. (147) SW22 Lately he seems to be absent more often than he's present. The criteria listed do not apply to all members of the class. For instance, the prototypical adjective premodifier very cannot occur outside adjective or adverb phrases and thus does not meet criteria C and D: 46 (140a) *You've very got to tell me what structuralism is all about. However, the traditional concept of adverb is even broader so that classifying a word as an adverb does not really allow many generalizations about its use. What is particularly awkward about the traditional classification (as it can be found in CGEL and many dictionaries) is that it contains both very and here, which do not share any distributional features: (130) SW14 I'm very interested in structuralism. (130a) *I'm here interested in structuralism. (139) NW44 It was true that he had been educated at a public school, but he managed to disguise this handicap very well. (139a) *It was true that he had been educated at a public school, but he managed to disguise this handicap here well. It is remarkable that a word such as here, which does not meet criteria A, B or C, is traditionally classified as an adverb: 45 See Aarts/ Aarts (1982/ 1988: 31). 46 Very can occur as a premodifier in noun phrases such as the very moment NW82 , however, which must be considered a special use outside the adverb class. <?page no="73"?> Adverbs and particles 61 (138a) *But if I use my words it follows that I have changed your meaning, here slightly… (140b) *You've here got to tell me what structuralism is all about. On the other hand, here can occur at the end of a clause as a complement of be, for example, which is not possible in the case of very: 47 (148) NW104 I'll tell Mr Wilcox she's here … (148a) *I'll tell Mr Wilcox she is very ... Similarly, words such as here or where but not very can occur as complements of such trivalent verbs as place or put: (149) NW23 Where would you put it? (149a) *Very would you put it? This kind of difference provides a very good argument for assigning very and here to two different word classes, as is done by Pullum and Huddleston in CamG (2002: 612-617), for instance. In order to reduce the wide scope the adverb category traditionally has, we will subsume some words traditionally classified as adverbs under the category of particles, which will be discussed further in the next section. Within this approach, it seems appropriate to exclude a number of words from the adverb class on the basis of the following criterion: [G] Words that can function as a valency complement of verbs such as be, place or put are not adverbs unless they meet criteria A and B. (150) SW24 The moon will be up by then. Note that on the basis of this criterion now and then, which can also occur as complements of be in sentences such as (151), are nevertheless classified as adverbs because they can premodify adjectives: 48 (151) BNC I used always to look fabulous, but that was then and this is now. (152) BNC He groped in a pocket and pulled out a now useless packet of cigarettes. 3.5.2 The category particles The term particle will be used here for a word class which comprises ⌐ all words traditionally called prepositions (such as since, to, on), ⌐ all words traditionally called subordinating conjunctions 49 (such as because, since, that 50 ), and 47 See CamG (2002: 612). 48 Compare, however, the analysis in CamG (2002: 615). 49 Note that CamG (2002: 955-956) identifies a category subordinators for that, if and whether, which we include in the particle class. 50 Note that that as a determiner-pronoun will be treated as belonging to a different word class from the particle that. <?page no="74"?> Word classes 62 ⌐ some words that traditionally are classified as adverbs (such as here, on). This word class can be defined by the following properties: [A] Particles cannot take morphological endings of any kind and do not meet the criteria established for other word classes with this property (i.e. coordinating conjunctions and interjections). One advantage of this classification is that the great heterogeneity of the class of adverbs can at least be reduced to some extent by excluding words such as up on the basis of adverb criterion G above, which can be taken as a central criterion for the definition of the class in the form of criterion B: [B] Particles (or phrases headed by particles) can occur as valency complements of verbs such as be or put (also at the end of a clause): (150) SW24 The moon will be up by then. (153) SW18 They're at a lecture. (154) SW13 It was being given by the Oxford medievalist and was on the subject of Chaucerian metrics. (149) NW23 Where would you put it? [C] Particles (or phrases headed by particles) can occur as valency complements, as modifiers or as adjuncts: (37) SW15 So tell me about structuralism. (63) NW326 Could you drop it here tomorrow morning before eight-thirty? (155) SW42 Before I retire, I want to be the highest paid Professor of English in the world. (156) SW18 It says here that you're a non-resident. [D] Particles as a class have the property of valency: 51 (157) SW25 ... we can't affect the development of the text by our own words, since the text's words are already given. (158) NW44 By living in what their parents called sin, they nailed their colours to the mast of youth revolt, while enjoying the security and mutual support of old-fashioned matrimony. (15) SW24 In the event, not many people did like Morris Zapp's lecture, and several members of the audience walked out before he had finished. 51 This idea is also expressed in CGEL (1985: 9.65), where the category of prepositional adverbs is introduced: "A prepositional adverb is a particle which is formally identical to or related to a preposition, and which often behaves like a preposition with ellipted complement". See also CGEL (1985: 9.65n): "The relation between prepositional adverbs and prepositional phrases ... may be compared to that between intransitive and transitive use of certain verbs ...". <?page no="75"?> Adverbs and particles 63 3.5.3 Adverbs and particles versus adverbs, prepositions and subordinating conjunctions 3.5.3.1 Traditional adverbs, prepositions and subordinating conjunctions The treatment of particles as one word class largely follows the policy of CamG (2002), where some adverbs and subordinating conjunctions are included in an extended category of prepositions. 52 Many grammars and dictionaries, however, make use of the traditional categories adverb, preposition and subordinating conjunction to account for these uses. Traditional adverbs then include words such as here, since, up or where since criterion G given in 3.5.1 (the fact that they cannot occur as a clause constituent after be at the end of a clause) is not part of the traditional definition of adverb: (150) SW24 The moon will be up by then. Traditional prepositions include words such as before, for, in, of, on, to, with, subsequent to and can be defined by a number of criteria: [A] Prepositions can form clause or phrase constituents together with noun phrases. (141) SW22 I was just explaining structuralism to this young man. (130) SW14 I'm very interested in structuralism. (159) SW16 I'll see you in the bar before lunch, then. [B] Prepositions cannot occur as clause constituents of their own. (141a) *I was just explaining structuralism to. [C] Prepositions cannot be followed by what traditional grammar calls that-clauses and infinitive clauses or by nominative case forms of personal pronouns. 53 (141b) *I was just explaining structuralism to he. Traditional subordinating conjunctions (or subordinators) include words such as for, since, that, while or when and can be defined by the following criteria: 54 52 The most radical break with traditional terminology that is made in CamG (2002: 612-617) concerns the words that are traditionally labelled adverbs, which include such words as abroad, adrift, here, there, where, east, downhill, upstairs, indoors, upwards. For lists of traditional adverbs which CamG classifies as prepositions and for the reasons see CamG (2002: 612-615). Although Huddleston and Pullum (CamG 2002: 602) find reasons for doing so, we find it rather awkward to refer to a word with zerovalent uses in which it is not followed by a complement as a preposition and thus make use of the more neutral term particle for this word class (which, in any case, does not fully correspond to the category of prepositions in CamG). Compare also Eisenberg (1999: 202) for German: "... ist es sehr wohl erwägenswert, die Präpositionen mit den subordinierenden Konjunktionen (...) zu einer Kategorie zusammenzufassen und sie den koordinierenden als den eigentlichen Konjunktionen gegenüberzustellen." 53 For a definition of preposition see CGEL (1985: 9.2). 54 For a discussion of subordinators see CGEL (1985: 14.10-14.19). See also the notion of subordinators for non-finite and verbless clauses (CGEL 1985: 14.15). Compare also Aarts/ Aarts (1982/ 1988: 44-46). <?page no="76"?> Word classes 64 [A] Subordinating conjunctions can form clause or phrase constituents together with clauses. (15) SW24 In the event, not many people did like Morris Zapp's lecture, and several members of the audience walked out before he had finished. (108) NW85 There is a widespread feeling in the country that universities are 'ivory tower' institutions, whose staff are ignorant of the realities of the modern commercial world. (160) SW22 And if we are to believe language rather than our ears, English dogs go 'woof woof', French dogs go 'wouah wouah', German dogs go 'wau wau' and Italian ones 'baau baau'. [B] Subordinating conjunctions cannot occur as clause constituents of their own. (108a) *There is a widespread feeling in the country that. (160a) *And if, English dogs go 'woof woof', French dogs go 'wouah wouah', German dogs go 'wau wau' and Italian ones 'baau baau'. In this more traditional terminology, the different uses of a word such as before or since are described in terms of different word classes: (a) adverb: (161) SW18 I'm sure I've heard that name before. (162) NW42 She attended an excellent direct-grant grammar school (which has since gone independent, much to Robyn's disgust) where she was Head Girl and Captain of Games and which she left with four A grades at A-Level. (b) preposition: (159) SW16 I'll see you in the bar before lunch, then. (163) NW111 Robyn had not felt so exalted since the great women's rally at Greenham Common. (c) subordinating conjunction: (15) SW24 In the event, not many people did like Morris Zapp's lecture, and several members of the audience walked out before he had finished. (46) SW27 Your mind has lost none of its sharpness since we first met. 3.5.3.2 Reasons for the particle category The word classes that traditionally are classified as prepositions and subordinating conjunctions share important features: (1) They are uninflected words. (2) They combine with phrases or clauses to form constituents of a phrase or a clause. The main difference between prepositions and subordinating conjunctions as they are defined in traditional grammar is that prepositions are followed by phrases and subordinators by clauses. However, this distinction is blurred to a certain extent if <?page no="77"?> Adverbs and particles 65 one argues that "-ing clauses are permitted after a preposition in English" (CGEL 1985: 9.3). Up to a point, such a rule makes sense because quite a few prepositions allow complementation by [V-ing]-clauses but not by other clauses: (158) NW44 By living in what their parents called sin, they nailed their colours to the mast of youth revolt, while enjoying the security and mutual support of old-fashioned matrimony. (158a) *By they lived in what their parents called sin ... However, this leaves the problem that in the case of (164) NW27 A few weeks before, he'd returned home soon after getting to work because he'd left some important papers behind, and found the house totally silent, all three children and their mother sound asleep at 9.30 in the morning. it cannot be decided whether after should be seen as a preposition or a subordinator since it is in any case a member of both word classes: (165) SW20 Morris! It's marvellous to see you after - how many years? (166) SW26 Veil after veil, garment after garment, is removed, but it is the delay in the stripping that makes it exciting, not the stripping itself ... (167) SW14 Then after I got my Master's degree, I went home to work on the farm for two years. One of the advantages of subsuming traditional prepositions and traditional subordinating conjunctions under one word class is the avoidance of this kind of indeterminacy. If the different complementation possibilities of these words are analysed as valency phenomena, there is no need to assign them to different word classes - especially since the traditional distinction is based on the criterion of valency anyway: if a word such as before is followed by a finite clause as in (15), it is traditionally classified as a subordinator, if it is followed by a noun phrase as in (159), it is classified as a preposition and if it does not require any complementation as in (161), it is classified as an adverb. In fact, as is pointed out by Huddleston and Pullum (CamG 2002: 599-601 and 1011- 1014), valency differences such as those between the different uses of before or since are not used as an argument for assigning verbs to different classes in traditional grammar either: (32) SW21 I've forgotten. (168) SW20 How could I forget that happy face? (169) NW71 She forgot to give me your message. (170) NW59 She had forgotten how nice it was, after so long an interval. In fact, it is these parallels to verb valency that present one of the most forceful arguments in favour of the one-word-class-view. 55 It thus seems appropriate to 55 These parallels can also be taken as an argument for including into the particle category words such as here which can be analysed as having zerovalent uses and which traditionally are only <?page no="78"?> Word classes 66 assign the property of valency to members of this word class (> 6.8.2). Differences in use such as the following can then be related to the valency specifications of the respective particles: 56 (i) zerovalent use (9) SW22 "Structuralism? " said Dempsey, coming up with a sherry for Angelica just in time to hear Persse's plea, and all too eager to show off his expertise. (ii) monovalent use with noun phrase complement (171) SW17 The meetings are held on the main campus, up the road. (159) SW16 I'll see you in the bar before lunch, then. (iii) monovalent use with [V-ing]-clause complement (158) NW44 By living in what their parents called sin, they nailed their colours to the mast of youth revolt, while enjoying the security and mutual support of old-fashioned matrimony. (iv) monovalent use with finite clause complement (172) SW24 I thought that the goal of reading was to establish the meaning of texts. (46) SW27 Your mind has lost none of its sharpness since we first met. (v) monovalent use with particle phrase complement (173) NW330 I suppose you might as well leave it until after the next stage . Within the particle approach, the traditional terms preposition and subordinating conjunction can then be seen as labels for particular types of use of particles such as (ii) and (iv) above. There are of course a number of words which can be described quite well within the traditional word classes. Thus particles such as of, from, upon have the typical prepositional uses such as (ii) and (iii) (100) SW0 There are conferences on almost everything these days, including the works of Geoffrey Chaucer. but none of the other uses identified above. Similarly, although, that and whether 57 are good examples of traditional subordinators because they can be followed by a finite clause (iv) but not by a noun phrase (ii). (172) SW24 I thought that the goal of reading was to establish the meaning of texts. classified as adverbs (and not as prepositions or subordinating conjunctions): in a way, the valency properties of a word such as here can be compared to monovalent verbs such as sleep. 56 Note that this is not a complete list of all possible valency structures of particles. Compare also, for example: I spend all my time on committees arguing about how to respond to the cuts. NW324 or She began to look for a university job outside Cambridge. NW50 (> 4.5). 57 CamG treats these as subordinators as distinct from their category of prepositions. See CamG (2002: 955-956 and 1011-1014). See also 5.3. <?page no="79"?> Adverbs and particles 67 Other words, however, have to be attributed to two or three word classes in the traditional analysis: thus, for example, up or down exhibit adverbial and prepositional but not subordinating uses: (9) SW22 "Structuralism? " said Dempsey, coming up with a sherry for Angelica just in time to hear Persse's plea, and all too eager to show off his expertise. (171) SW17 The meetings are held on the main campus, up the road. and, as pointed out above, the different uses of before or since are described by assigning the words to three classes. 58 This situation can be represented as follows: particles (i) zerovalent use there (ii) monovalent use with noun phrase complement up at (iv) monovalent use with clause complement before until whether traditional label adv prep conj adv prep prep conj adv prep conj In the light of these differences in use (which do not even take into account further types of complementation to be found with particles), there seems little point in distinguishing between further subclasses since the extent of idiosyncrasy to be observed in the use of these particles is typical of valency phenomena. Rather, it seems to us that analysing these words which have traditionally been attributed to three different word classes as members of one class with different valency properties has certain advantages with respect to the description of syntactic categories. Nevertheless, it must be clear that making the distinction in this way does not resolve the considerable amount of overlap and gradience that is to be observed in the use of these words. 59 58 It is obvious that even three word class categories do not suffice to account for all the differences in the uses of these words. Thus both that and while are classified as subordinators because they can be followed by a finite clause as in (172) and in I went over it last night on the plane, while the movie was showing, ... SW18 . However, this classification does not account for the fact that that cannot be followed by a [V-ing]-clause, but while can: By living in what their parents called sin, they nailed their colours to the mast of youth revolt, while enjoying the security and mutual support of old-fashioned matrimony NW44. . 59 It might seem more appropriate to argue that the words classified here as adverbs and particles represent a rather broad spectrum which can best be described in terms of gradience and family resemblance and that adverbs and particles could thus be seen as one subclass of one larger word class. The great advantage of such an approach would be that it offers a rather elegant solution for the fact that many syntactic slots can be filled either by adverb phrases or by particle phrases (> 4.4-5). One indication of the gradient between adverbs and particles is the fact that words such as now (which we classify as an adverb on the basis of criterion G, > 3.5.1) also has zerovalent uses <?page no="80"?> Word classes 68 3.6 Conjunctions The class of conjunctions 60 comprises the following elements: and, or, nor 61 and but. A conjunction links two elements that are potential constituents of a larger linguistic unit to form one constituent at the same level. Thus in (2) SW28 ... what, with the greatest respect, is the point of our discussing your paper if, according to your own theory, we should not be discussing what you actually said at all, but discussing some imperfect memory or subjective interpretation of what you said? there are two cases of coordination, which can be represented as follows: (2a) ... we should not be [[discussing what you said at all] but [discussing some imperfect memory or subjective interpretation of what you said]]? and (2b) ... but discussing some [[imperfect memory] or [subjective interpretation]] of what you said? Coordination can take place between any type of constituent, e.g. between constituents of phrases as in (2b), or as in (3) SW0 ... not all Aprils, for that matter, are marked by [[sweet showers] and [dulcet breezes]]. (8) SW17 Have you [[really] and [truly]] come all the way from London by taxi? or clauses as in (2a) and the following examples: (15) SW24 In the event, [[not many people did like Morris Zapp's lecture,] and [several members of the audience walked out before he had finished]]. (62) SW22 [[It's a purely arbitrary relationship], and [there's no reason why English speakers shouldn't decide that from tomorrow, d-o-g would signify 'cat' and c-a-t, 'dog'.]] And, or and nor also occur in correlative constructions of the type both ... and, either ... or and neither ... nor, in which a determiner-pronoun forms a first part. Such sequences will be analysed as discontinuous conjunctions. such as Yes, I'm a professor now. SW7 and uses with clause complements such as Now she had become a pensioner she had been able to give up work as a midwife, and she spent much of her time on her allotment... BNC . For a discussion of this compare Palmer (1974: 215): "It might be plausible to argue that English does not, in fact, have two word classes adverb and preposition, but a single class 'particle' or, perhaps, 'prepositional-adverb', for there is considerable similarity in their function. Often the adverb can be replaced, with little or no change in meaning, by the preposition plus a noun phrase ...". 60 Since traditional subordinating conjunctions are subsumed under particles, what we call conjunctions here corresponds to the traditional class of coordinating conjunctions. 61 For the somewhat marginal status of nor see CGEL (1985: 13.5). See also CamG (2002: 1309). <?page no="81"?> 69 (174) NW84-5 ... make a special effort in the coming year to show themselves responsive to the needs of industry, both [[in terms of collaboration in research] and [the provision of well-trained and well-motivated graduates for recruitment to industry]]. 3.7 Interjections What is generally considered to be the word class of interjections includes the following items, for example: hey, oh, tut-tut, wow etc. Interjections can [A] occur outside the structure of the clause in that they do not enter any systematic syntactic relationship with other words in a clause [B] are typical of the spoken language [C] can contain phonological features outside the regular system of English phonology [D] can be pronounced emphatically with special lengthening and a wide pitch range 62 (175) SW9 "Oh, like in Finnegan's Wake! The Ballad of Persse O'Reilley." (176) SW17 "Ah, that figures," … (177) NW61 'Oh, hallo, Philip,' says Bob Busby. 3.8 Limits of classification 3.8.1 Relative and interrogative words One problem in the classification of word classes is that certain generalizations which are useful to make cut across the classes identified above. This concerns in particular a group of words often referred to as wh-words such as who, which, what, where, when or how. 63 Wh-words are typically associated with two particular types of syntactic construction, namely relative clauses (> 5.3.7) such as (178) NW40 According to Robyn (or, more precisely, according to the people who have influenced her thinking on these matters), there is no such thing as the 'self' on which capitalism and the classic novel are founded ... and wh-interrogative-'question'-constructions (> 7.1.2) such as (179) NW41 And who is Charles? In order to make generalizations about the function of such words within these particular types of clause, one can use the terms relative and interrogative in relation with the appropriate word class label. 62 These criteria are taken over from CGEL (1985: 11.55). 63 See CGEL (1985: 2.14, 11.14-15). Limits of classification <?page no="82"?> Word classes 70 Relative pronouns, relative particles and relative adverbs can then be defined as words [A] that form an element of the first phrase of a relative clause [B] that have the same referent as the noun or pronoun on which the relative clause depends [C] that can occur as heads of noun phrases which do not have pre-heads or modifiers. (180) SW25 Then I began a commentary on the works of Jane Austen, the aim of which was to be utterly exhaustive, to examine the novels from every conceivable angle ... The following words can be classified as relative pronouns in English the pronoun who (including the forms whose and whom) (104) SW8 Swallow is the man chatting up that rather dishy girl who just came in ... (101) SW4 Within a very short time they had established that none of the stars of the profession was in residence - no one, indeed, whom it would be worth travelling ten miles to meet, let alone the hundreds that many had covered. which (which is gender-marked since it can only refer to non-humans) (180) SW25 Then I began a commentary on the works of Jane Austen, the aim of which was to be utterly exhaustive, to examine the novels from every conceivable angle ... the pronoun that (which does not have a genitive form) (101) SW4 Within a very short time they had established that none of the stars of the profession was in residence - no one, indeed, whom it would be worth travelling ten miles to meet, let alone the hundreds that many had covered. To these can be added a number of relative particles such as where, when and why: (162) NW42 She attended an excellent direct-grant grammar school (which has since gone independent, much to Robyn's disgust) where she was Head Girl and Captain of Games and which she left with four A grades at A-Level. (21) NW82 At the very moment when they were writing about these problems, Marx and Engels were writing the seminal texts in which the political solutions were expounded. Similarly, interrogative words can be defined as words that can occur as heads or pre-heads in wh-interrogative-'question'-constructions. This class comprises the particles where, when, why: (149) NW23 Where would you put it? (181) NW323 Why don't you get her a proper drink? the adverb how: (182) NW325 How many did you sell in America? <?page no="83"?> Limits of classification 71 the pronoun who (179) NW41 And who is Charles? the determiner-pronoun what (120) NW306 'What letter is that? ' Robyn asked. (183) NW324 What shall we do without you? which (which is not gender-marked since it can also refer to humans) 64 (122) SW8 Which one is Professor Swallow? (184) SW23 I never could remember which came first, the morphemes or the phonemes. Which poses a certain problem of classification: when used in relative constructions, it has all features of a pronoun, i.e. it can function as pre-head (whose) or head (which) of a noun phrase (referring to non-human referents). When used in interrogative constructions, which can function as the head of a noun phrase (only with a nonhuman referent) and also as a pre-head (with human or non-human referents) as in which one in (122). If one takes this semantic difference as a criterion for assigning which to two different word classes, one could identify a relative pronoun which and an interrogative determiner-pronoun which. It has to be emphasized, however, that the use of such terms depends very much on the purpose of the terminology employed. If one wants to describe particular constructions, it is convenient to subsume certain words under general labels such as relative pronoun or also relative adverb or relative particle. From the point of view of the lexicon, this does not mean that one would argue, for example, that English has a relative pronoun who and an interrogative pronoun who, but rather that the word who can function both as a relative and an interrogative pronoun. 65 In the same way one can argue that the different ways in which a particle such as when can be used are part of the item-specific knowledge speakers have about that word, while at the same time it is useful for the description of relative clauses, to subsume when under a group of words such as relative particles to describe properties of the word when that are relevant to the description of this construction that it shares with other words. 3.8.2 Item-specific properties and special uses The discussion of word classes in the previous sections has shown that the behaviour of some, usually high frequency words, is relatively unsystematic. In some cases, it may make sense to leave a number of words unclassified if they do not fit into any general pattern. One such example is the use of there in sentences such as (125) SW28 "There is no point," said Morris Zapp. (135) SW22 There is no absolute reason why the combined phonemes d-o-g should signify a quadruped that goes 'woof woof' rather one that goes 'miaou'. 64 See CGEL (1985: 6.36). 65 Compare the category of wh-words in CGEL (1985: 2.45). <?page no="84"?> Word classes 72 (185) SW23 There does seem to be some iconic correlation between sound and sense across the boundaries of natural languages. which is clearly different from the use of there in (186) SW10 Nobody knows how I got there. where one would not hesitate to classify it alongside here as a particle. This is different in the case of (125), (135) and (185): firstly, it makes sense in these cases to analyse there as the subject of the clause, which for a particle would be rather untypical. Secondly, and more importantly perhaps, there cannot be replaced by another particle or a particle phrase in these clauses. It is for this reason that this existential there has to be treated as a special case although in some cases an element of the locative meaning of the particle use can also be made out: (187) SW8 There, hesitating on the threshold, was the most beautiful girl he had ever seen in his life. (188) SW8 "There's Swallow," said Dempsey. Since there in such uses does not share enough properties with members of any other word class occurring in similar positions, it would not make sense to classify there in such uses as belonging to any other word class either. 3.8.3 Problems of classification The discussion of word classes has shown that they certainly do not present a neatly structured system in the sense that we could say that if we assign a word a particular word class it necessarily shares all features of that class. In our description of word classes, we have tried to make this clear by different formulations. Criteria that apply only to some members of a class are usually given in the form of "if a word has a particular feature, then it is ..."; in other cases we can say "all members of a class have the following features." Generally, one must bear in mind that some classes can be established on a much sounder basis than others. Classes involving inflections such as nouns and lexical verbs, and also adjectives present large homogeneous classes. At the other end of the extreme, the class that is commonly identified as adverb is extremely heterogeneous. Thus a label such as adv (for adverb) in a dictionary is much less explicit than v (for verb) or n (for noun), for example. As far as word class assignment is concerned, many words are totally unproblematic. This applies to most of the words in (189) NW44-45 She spoke frequently in the Debating Society in favour of progressive causes such as abortion, animal rights, state education and nuclear disarmament. for example, where one would not hesitate to classify she as a pronoun and speak as a lexical verb because they meet all the criteria listed for these word classes. As pointed out in 3.1, problems arise in cases where a word fulfils only one or two but not all criteria given for a word class. This is the case with nuclear, for example, which only exhibits some but not all the adjective criteria since it does not allow modification by very or comparative or superlative forms and which usually occurs in attributive position. From a linguistic point of view, it still makes sense to classify words such as <?page no="85"?> Limits of classification 73 nuclear or ultimate as adjectives since they quite obviously do not belong to any other word class. 66 The usefulness of such a classification can be seen in lexicography, for instance: if a word is labelled adj in a dictionary, one can expect it to exhibit the typical properties of adjectives unless it is marked by features such as "usually before noun" (LDOCE4 in the case of nuclear). It must be emphasized that the uninflected words of English are very difficult to classify. There are a number of very small groups that can be identified rather clearly such as conjunctions, personal pronouns or pure determiners. On the other hand, for the reasons outlined we consider it advisable to subsume traditional prepositions and subordinating conjunctions under one class. It then seems consistent to include not only all subordinators in this class but also the use of to as an infinitive-marker, which CGEL (1985: 2.34) treats as a word "with unique function" and CamG (2002: 1183) as "infinitival subordinator". 67 Nevertheless, it is clearly the case that some of the words in the particle group (such as infinitival to or also whether) have a primarily grammatical function, whereas others such as before or from can be analysed as having a more obvious lexical meaning. It seems that it is in particular such high-frequency words that are difficult to classify. 68 It is psychologically plausible that the human mind is able to store a lot of item-specific information for common items and that therefore items that are used less frequently tend to be more "regular" in terms of generalizable properties. Consequently, as was pointed out, assigning word classes to words in language text is pretty straightforward in most cases, but nevertheless one should bear in mind that in some rather important cases such decisions strongly depend on the particular framework employed and that in others classifications of this kind may be of limited value in the face of the generally idiomatic nature of language. 69 3.8.4 One word or several words A final problem of word class assignment is related to the question of what one considers a word (> 3.1.1). Here the distinction between word and lexeme is relevant; for instance, give a monkey's in (190) NW217 ... doesn't it worry you that ninety-nine point nine per cent of the population couldn't give a monkey's? is to be analysed as a complex lexeme (or part of a complex lexeme), which can be classified as an idiom. As such, it could be given the word class status of a verb. At the same time, it can be analysed as consisting of different words. A slightly more difficult problem is presented by cases such as 66 Compare the distinction between central and peripheral adjectives in CGEL (1985: 7.4). 67 Compare, however, CGEL (1985: 2.34). CamG (2002: 1184-85) classifies to as a subordinator on the basis of the ungrammaticality of sentences such as *I agreed to it and go. However, this kind of problem also arises in other cases such as ? ? She decided on the ferry and booking it. 68 Compare also Sinclair's (1991: 81-98) observation that the most common preposition of the English language, of, statistically is not very typical of the use prepositions generally. 69 For word-class assignment in computational linguistics see, for instance, Garside/ Leech/ Sampson (1987). <?page no="86"?> Word classes 74 (126) NW72 In the 1840s and 1850s," says Robyn, "a number of novels were published in England which have a certain family resemblance. (191) NW46 All this of course took up a great deal of time and delayed the completion of Robyn's thesis on the nineteenth-century industrial novel, which had to be constantly revised to take the new theories into account. discussed in 3.3.4, where one might consider a number of and a great deal of as complex determiners from a phraseological point of view. Similarly, let alone could be classified as a complex conjunction: 70 (101) SW4 Within a very short time they had established that none of the stars of the profession was in residence - no one, indeed, whom it would be worth travelling ten miles to meet, let alone the hundreds that many had covered. (192) NW324 I haven't read a book in months, let alone tried to write one. 3.8.5 Item-specific knowledge and word classes All this shows that words do not seem to fall neatly into classes. Up to a point, this is a reflection of the fact that the present state of the language, which we are trying to describe as a system, must be seen as a stage in the historical development of the language. It also shows that speakers of a language must possess an enormous amount of item-specific knowledge about the use of individual words, which can only be partially captured by classifying them as members of particular word classes. 3.9 Survey of differences The following table illustrates where the word class system that was established in this chapter differs from traditional nomenclature and the terms used in the big reference grammars and learners' dictionaries. 70 Cf. CamG (2002: 1319), where let alone and not to say are described as "idioms which might be regarded as marginal coordinators". For let alone see also Fillmore/ Kay/ O'Connor (1988). <?page no="87"?> Survey of differences 75 established traditional terminology CGEL (LDOCE OALD) CamG our terminology adverb adverb furiously, very, well pronoun pronoun yesterday, today here, somewhere adverb adverb since preposition preposition since + noun phrase preposition since + finite clause conjunction (subordinating) conjunction subordinator that/ whether + finite clause infinitive marker infinitival subordinator particle to + infinitive clause article determiner/ article determinative/ article the, a determiner every determiner determinative some (as prehead of NP) pronoun pronoun pronoun determiner-pronoun some (as head of NP) <?page no="88"?> 4 Phrases 4.1 The character of the phrase 4.1.1 Syntactic units beyond the word On the basis of the word classes established in Chapter 3, it is now possible to describe in more detail the combinations which words enter to form larger units. Words make up phrases or clauses, phrases and clauses can be parts of larger phrases or clauses. Thus, for instance, a noun such as University can combine with the determiner the to form a noun phrase (1a) the University which can combine with the particle on to form a particle phrase (1b) on the University This particle phrase in turn can function as the complement of a lexical noun such as constraints to form a complex noun phrase, which can be described as follows: (1c) the financial constraints on the University In this way, larger units can be built up. Thus, in (1) NW64 Being Dean, you see, I'm very aware of the financial constraints on the University. (1c) is part of a particle phrase of the financial constraints on the University. In fact, it is a complement of the adjective aware, which functions as the head of an adjective phrase, which can be described as follows: (1d) very aware of the financial constraints on the University The reason why we consider a unit such as very aware of the financial constraints on the University a phrase is that it can function as a clause constituent in a clause such as (1). <?page no="89"?> The character of the phrase 77 4.1.2 Phrase constituents As pointed out in 2.4.2, it seems appropriate to distinguish between three different types of phrase constituents: the head (or, in some cases, a head complex consisting of a number of preheads and the head), which is an obligatory constituent of the phrase and which is realised by a lexical unit which has the property of valency, complements, which are determined by the valency of the lexical unit functioning as head, modifiers, which are optional phrase constituents that are independent of the valency properties of the head and which according to their position with respect to the head can be characterized as premodifiers, postmodifiers and discontinuous modifiers. Example (1) can then be analysed in the following way: very aware of the financial constraints on the University premodifier head complement These phrase constituents can be found in all four types of phrase that we will identify: [1] noun phrases [NP], [2] adjective phrases [AdjP], [3] adverb phrases [AdvP] and [4] particle phrases [PartP]. 4.1.3 Heads and head-complexes The head-complex can be seen as the most characteristic constituent of a phrase. This is shown by the fact that in many cases the head-complex can in fact represent the whole phrase (unless it requires an obligatory valency complement). In the case of adjective, adverb and particle phrases, the head complex always consists of a single head as in (2) NW215 The mornings were dark [adjective phrase], dusk came early [adverb phrase]. (3) NW214 Outside [particle phrase], the air was cold and clammy, thick with moisture and pollution [particle phrase]. In the case of noun phrases, the head-complex can consist of a single head as in the cases of Dean, you and I in (1) NW64 Being Dean, you see, I'm very aware of the financial constraints on the University. or a combination of pre-head(s) and head: (1a) the University <?page no="90"?> Phrases 78 (1c) the financial constraints on the University. 4.1.4 Complements and modifiers The distinction between modifiers and pre-heads on the one hand and complements on the other reflects the distinction central to valency theory between combinations of words that generally hold for a whole class of words and those that are item-specific. The fact that in the phrase the financial constraints on the University words such as university or constraint can combine with words such as the is a fact that is true of all (count) nouns of the English language; similarly, all nouns can be preceded by adjectives (as in financial constraints). These statements are statements that can be made about classes of words; therefore the fact that constraint can combine with a pre-head the and a modifier financial is not a property that is typical of the word constraint but a characteristic of a particular word class or subclass, namely that of count nouns. The fact that constraint can be followed by an on-particle phrase of the kind exemplified in (1), however, is not something that applies to all nouns in the language but a fact that has to be stated explicitly in any description of how the word constraint can be used in English. The on-phrase must thus be regarded as a construction that is specific to a particular noun. It is thus classified as a complement of constraint. Such constructions which are specific to individual words are seen as being determined by the valency of those words, whereas the fact that certain types of modifiers and preheads can occur with a particular word is seen as a feature of their word class. It must be realised, however, that - apart from the one being class-specific and the other word-specific - the type of relation holding between a modifier and its head and that holding between a complement and its governing element is very much the same. All these relations can be seen as dependency relations: the occurrence of a pre-head consisting of a determiner such as the or a modifier consisting of an adjective presupposes the existence of a noun. In this sense the determiner and the adjective are dependents of the noun. In the same way, the occurrence of a complement also presupposes the existence of a governing element, which functions as the head of the phrase so that the complement is also a dependent of the noun. 4.2 Noun phrase 4.2.1 Reference One of the most important functions of noun phrases is that they can be used by speakers to establish reference. In speech act philosophy and modern semantics, referring is looked upon as an act performed by a speaker in a particular situation of utterance in which the particular object, person, entity or class of objects etc. the speaker is talking about is identified. Thus, in slightly simplified terms, in an utterance such as (1) NW64 Being Dean, you see, I'm very aware of the financial constraints on the University. <?page no="91"?> Noun phrase 79 the speaker refers to a particular object (in this case a particular university) singling it out from all the objects of the real world that belong to the denotatum of the word university in English. He does this by using the phrase the University so that reference is established not by the single lexical noun university nor by the determiner the but by the whole phrase the University. In other cases, where the noun phrase consists only of a head, as in the case of I in (1), reference can be established by a single word, of course. There are different types of reference. For instance, reference can be generic as in (4) NW50 There is a widespread feeling in the country that universities are 'ivory tower' institutions, whose staff are ignorant of the realities of the modern commercial world. since universities in (4) refers to the class of universities as a whole. On the other hand, the University in (1) or the job at Rummidge in (5) NW76 Then, in 1984, just when Robyn was beginning to despair, the job at Rummidge came up. or a university job outside Cambridge in (6) NW50 She began to look for a university job outside Cambridge. present examples of specific reference. (5) is an example of definite reference because one particular entity is being identified, whereas (6) is an example of indefinite reference. Without going into further detail, the point to be made here is simply that it is noun phrases (and not just nouns) that can establish reference. 1 However, there are also noun phrases which are descriptive and not referential: (1) NW64 Being Dean, you see, I'm very aware of the financial constraints on the University. (7) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. 4.2.2 The head-complex of the noun phrase: pre-heads and heads The head-complex of the noun phrase is that part of a noun phrase 2 which together with obligatory complements of the head can stand for the phrase on its own in a given context that contains the elements that are relevant with respect to number which is instrumental in establishing the type of reference. 3 (7) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. 1 Compare especially Lyons (1977: 174-229), Searle (1969: 26-29), Bublitz (2001: 93-95) and also CGEL (1985: 2.30 and 5.26-5.59). 2 One might also consider analysing the structure of the nominal head-complex in terms of dependency relations and describing, for example, the difference between pure determiners and determiner-pronouns (> 3.3.3-4) as valency properties (> 6.8.1). 3 Here, the term 'type of reference' includes non-referring noun phrases such as a teacher in (7) (CGEL 1985: 5.37). <?page no="92"?> Phrases 80 The head-complex of a noun phrase consists of a head, which is the right-most constituent of the head-complex (teacher), and possibly a number of pre-heads (a). 4 All members of the following classes can function as heads of noun phrases: lexical nouns (1) NW64 Being [Dean], you see, I'm very aware of [the financial constraints on [the University]]. pure pronouns (8) NW62 Who will notice? It's not as if we're like bus drivers or air traffic controllers. I fear the general public will find they can get along quite well without universities for a day. determiner-pronouns (9) NW184 (10) SW5 (11) NW82 That's what this Government is doing to the universities: death by a thousand cuts. At Oxford or Cambridge you would expect at least a hundred and fifty. The campus clock begins to strike twelve ... Members of the following classes can function as pre-heads of noun phrases: determiners (12) SW25 (13) NW184 (14) NW64 every decoding a thousand cuts no chance of my being kept on after the three years are up determiner-pronouns (15) SW25 another encoding noun phrases headed by genitives of pronouns 5 (16) NW214 her teaching, her research, her administrative duties noun phrases headed by genitives of lexical nouns 6 (17) NW214 (18) NW214 the University's tower ten weeks' duration 4 The notion of the head-complex overcomes a number of difficulties in describing noun phrases. See also Radford's (1993) proposal of co-headedness. Quirk/ Greenbaum/ Leech/ Svartvik (CGEL 1985: 2.26) point out that the headed/ non-headed distinction is inadequate for noun phrases and verb phrases but nevertheless identify heads in noun phrases (CGEL 1985: 2.28 and 7.12). Hudson (1984: 90-2) argues for considering determiners as heads of noun phrases. For a detailed account of this discussion see Matthews (2007: 61-78). 5 Note that not all pronouns have genitives. For instance, what and that and reflexive pronouns do not have a genitive form and not all indefinite pronouns have genitive forms: compare anybody's versus *anything's. 6 Stating that noun phrases rather than nouns can occur as pre-heads etc. does not mean that all types of noun phrase can occur in that position; for instance, noun phrases with [that_CL]complements cannot occur as pre-heads or premodifiers. <?page no="93"?> Noun phrase 81 4.2.3 Complements Complements in noun phrases include: particle phrases 7 (19) NW40 (20) SW27 (21) NW85 (4) NW85 (22) NW323 nothing outside the text the mischievous influence of Continental theorizing our willingness to inform ourselves about the needs of industry. a widespread feeling in the country that universities are 'ivory tower' institutions, whose staff are ignorant of the realities of the modern commercial world. no difficulty in identifying the guest of honour wh-clauses (23) NW54 There is no reason why you shouldn't still have your own career. 4.2.4 Modifiers The elements that can occur as premodifiers of lexical nouns in noun phrases include the following (the respective modifiers are underlined, heads in normal type): 8 noun phrases (4) NW85 (24) NW42 (25) SW26 'ivory tower' institutions her long white cotton nightgown from Laura Ashley the tennis analogy adjective phrases (26) NW64 (27) NW42 (24) NW64 (28) SW0 (29) SW28 your present appointment a temporary one her long white cotton nightgown from Laura Ashley new and interesting relationships a remarkably well-read young woman ing-participle-clauses (30) NW84 its patched and disintegrating container 7 For a description of the particle phrase see 4.5: constructions traditionally called to-infinitiveclauses and that-clauses are analysed as clauses headed by the particles to and that and are thus subsumed under particle phrases. For wh-to-infinitives see 5.3.6. 8 See CGEL (1985: 17.94-114) for these and other types of premodifiers. Note that when we state that a particular type of phrase or clause can occur as a preor postmodifier this does not mean that any phrase of that kind can occur in these positions. For example, adjective phrases with postmodifications such as very interested in structuralism SW14 do not normally occur as premodifiers in a noun phrase. <?page no="94"?> Phrases 82 ed-participle-clauses (30) NW84 (31) NW43 its patched and disintegrating container its tastefully harmonized buildings in the modernist-Palladian style The elements that can occur as postmodifiers in noun phrases include: 9 particle phrases (32) NW214 (33) NW214 (24) NW42 the winter term at Rummidge her administrative duties in the Department her long white cotton nightgown from Laura Ashley adjective phrases (34) SW28 (35) VDE two people present who could agree on the simplest précis of what had been said people interested in your data as data relative clauses (> 5.3.7) (34) SW28 (36) NW40 (37) SW9 (38) NW40 two people present who could agree on the simplest précis of what had been said the writers who have influenced her thinking on these matters with a trace of something else he could not identify. he who is not aware of the discursive formations that determine him ing-participle clauses (39) SW8 the man chatting up that rather dishy girl who just came in ed-participle clauses (40) NW42 the rugs spread on the sanded and waxed pine floorboards as stepping-stones Apart from premodifiers, which precede the head, and postmodifiers, which follow the head, there are discontinuous modifiers 10 . For instance, in (41) SW42 Before I retire, I want to be the highest paid Professor of English in the world. highest paid could be analysed as a premodifier of a head Professor of English (which presumably should be taken as a compound) and in the world as a postmodifier of Professor of English. However, in the world clearly does not modify Professor of English but highest paid so that highest paid in the world has to be regarded as one constituent consisting of a participle construction which partly precedes and partly follows the head of the noun phrase. This is why such modifiers will be called discontinuous modifiers. A special type of postmodifier to be found in noun phrases is generally treated in grammars as apposition 11 - a concept that refers to cases such as 9 Compare CGEL (1985: 17.9-64) for a more detailed treatment. 10 See Aarts/ Aarts (1982/ 1988) for a detailed description. 11 For a discussion of apposition see CGEL (1985: 17.65-17.93). Cases such as He smoked a different brand, himself, but it was as if he felt his whole philosophy of life was threatened by Robyn's analysis of the <?page no="95"?> Noun phrase 83 (42) NW60 That is to say (a favourite phrase of her own), Robyn Penrose, Temporary Lecturer in English Literature at the University of Rummidge, holds that 'character' is a bourgeois myth, an illusion created to reinforce the ideology of capitalism. Apposition is not unlike coordination, but noun phrases in apposition usually have the same referent and cannot be linked by a conjunction such as and. Thus, Temporary Lecturer in English Literature at the University of Rummidge can be regarded as an appositive postmodifier of Robyn Penrose. Similarly, the coordinated noun phrase her teaching, her research, her administrative duties in the Department can be seen as an apposition of plenty of work in (43) NW214 She had plenty of work: her teaching, her research, her administrative duties in the Department. Further cases which are sometimes treated in terms of apposition are combinations of names or of titles and names, which we see as complex proper nouns: (44) NW322 She was curious to meet Professor Zapp. (45) NW323 'I always know when I'm in England,' said Morris Zapp, as Philip Swallow went off, 'because when you go to a party, the first thing anyone says to you is, "Red or white? " ...' 4.2.5 Structure of the noun phrase The minimum realisation of a noun phrase can be a single noun, pronoun or determiner-pronoun. (1) NW64 Being Dean [lexical noun], you [pronoun] see, I [pronoun] 'm very aware of the financial constraints on the University. A noun phrase can have structures such as the following: pre-head(s) premodifier(s) head postmodifier(s) complement(s) (46) NW44 their second year (47) NW43 student demonstrations (48) SW0 conferences on almost everything (49) NW82 the seminal texts in which the political solutions were expounded advert. NW220 will be treated as appositive adjuncts and not (as in CGEL 1985: 18.39) as discontinuous noun phrases because the concept of discontinuous noun phrase is difficult to apply to cases such as Why not do it yourself? NW87 . <?page no="96"?> Phrases 84 pre-head(s) premodifier(s) head postmodifier(s) complement(s) (50) NW82 all their dismay at the squalor and exploitation generated by industrial capitalism (51) NW85 this University's contribution to IY (52) NW43 an endlessly deep stage set where apparently threedimensional objects turned out to be painted flats ... (53) SW25 ordinary spoken discourse (54) NW45 all the universities enthusiastically created or expanded in the booming sixties (55) NW52 four other equally desperate and highly qualified candidates (56) SW28 this invitation to bring the proceedings to a speedy close (4) NW85 a widespread feeling in the country that universities are 'ivory tower' institutions, whose staff are ignorant of the realities of the modern commercial world. <?page no="97"?> Noun phrase 85 (41) NW42 the highest paid Professor of English in the world pre-head(s) head complement(s) discontinuous modifier Whereas preand postmodification are usually seen as modifications of the lexical noun functioning as head, noun phrases can also take modifiers that refer to the preheads or even the pre-head and pre-modifier: Thus not in (57) NW45 Not many people perceived this quite so soon. clearly modifies many (or many people) but not just people. Similarly, almost modifies everything in (58) SW0 There are conferences on almost everything these days, including the works of Geoffrey Chaucer. While it would be possible to cover such cases as (57) and (58) by introducing a category of determiner-pronoun-phrase or pronoun-phrase which functions as a prehead in the noun phrase 12 , such a solution would not cover cases such as (59) BNC ... the largest libraries would be expected to include almost the whole range of British books, amounting to about 17,000 titles. where the premodifier almost can only occur because of the adjective whole: (59a) * ... the largest libraries would be expected to include almost the range of British books, amounting to about 17,000 titles. The essential requirement for this type of premodification seems to be that the noun phrase is generally quantifiable. Thus it seems appropriate to assume that elements such as almost, hardly etc. can occur in a premodifier slot before the pre-heads: premodifier(s) prehead(s) premodifier head postmodifier complement(s) (59) almost the whole range of British books (57) not many people 4.2.6 Elliptical noun phrases and special noun phrase construction There are a number of noun-phrase constructions in English which do not fully correspond to the description of noun phrases given above because they do not have 12 Compare the outline in CamG (2002: 431-433). <?page no="98"?> Phrases 86 heads in the defined sense. 13 The first type can be seen as elliptical noun phrases because the missing head is contextually optional in that its referent can be retrieved from the context. This is the case with examples such as: (60) BNC The green grasshoppers fittingly tended to go and settle on the green backgrounds, and the yellow grasshoppers on the yellow. (61) NW45 ... her marks, she was unofficially informed, were the highest ever achieved by a student in the School of European Studies in his short history ... (62) NW35 He hangs up his camelhair coat in the anteroom that connects his office with Shirley's and passes into the former. Secondly, there are a number of special noun phrase constructions which are subject to very special conditions of use and which can thus be said to represent formmeaning-pairings in the sense of construction grammar. One such construction is represented by (63) NW214 All the more grim, therefore, was the fate of the unemployed of Rummidge and environs, condemned to be idle in a place where there was nothing much to do, except work. where an adjective follows the to denote a group of people. A different meaning is associated with similar combinations of pre-heads and adjectives such as (64) NW85 We should do our utmost to dispel it. (65) NW179 The little he knew about Basil did not suggest that deconstruction was a likely topic of conversation in the next hour of conversation or two. It makes sense to treat these special noun phrase constructions as noun phrases since they exhibit typical features of noun phrases in that a type of pre-head that can only occur in noun phrases is followed by an adjective, which on the basis of the meaning it has in the particular construction can be treated as a head. A similar example is presented by constructions such as the following: (66) SW26 The tennis analogy will not do for the activity of reading - it is not a to-and-fro process, but an endless tantalising leading on, ... (67) SW26 Veil after veil, garment after garment is removed, but it is the delay in the stripping that makes it exciting, not the stripping itself (68) NW13 He flinches, dodges, zaps them with instant solutions, but the assault is endless: the Avco account, the Rawlinson account, the price of pig-iron, the value of the pound, the competition from Foundrax, the incompetence of his Marketing Director, the persistent breakdowns of the core blowers, the vandalizing of the toilets in the fettling shop, the pressure from his divisional boss, last month's accounts, the quarterly forecast, the annual review ... The problem here is that it seems inappropriate to classify the head of the phrase as a noun: firstly, (66) to (68) do not allow a plural; secondly, all verbs with a particular 13 For a detailed discussion of these cases in terms of fused head constructions see CamG (2002: 410- 422). CGEL (1985: 5.56) treats the unemployed as a "phrase with an adjective head referring to a group of people"; see also CGEL (1985: 7.23-26). <?page no="99"?> Adjective phrases 87 valency structure can occur in this construction. It is for this reason that we will treat this as a special type of noun phrase construction. 14 4.2.7 Typical elements of noun phrases Even if elliptical noun phrases and special noun phrase constructions do not contain a head in the sense defined here, they have as an obligatory element one of the typical pre-heads of noun phrases. A determiner or a determiner-pronoun obviously has an important function in signalling the beginning of a noun phrase in the parsing process in speech perception, whereas the noun then indicates a potential end of a noun phrase. 15 Thus one can argue that although there is no single word class that must occur in a noun phrase, all elements that can occur in the head-complex can serve as indicators of a noun phrase in the perception of speech. 4.3 Adjective phrases The defining property of an adjective phrase is that its head is an adjective. The head can stand on its own, as in (2) NW215 The mornings were dark, dusk came early. (69) SW28 I thought it was interesting. Adjective phrases allow premodification, postmodification and discontinuous modification as well as valency complementation: 16 (70) SW14 I'm very interested in structuralism. (71) NW46 Charles was not quite so committed to the new wave. (72) SW25 It is more passive in the sense that we can't interact with the text ... (73) SW27 But I am sorry to see that you have meanwhile succumbed to the virus of structuralism. (74) SW31 Oh, it's not as bad as all that. 14 Although this construction is similar to the use of the morphological form gerund in Latin, the term gerund is inappropriate for the description of English because there is no morphologically distinct form in English. See 1.4. Compare CGEL (1985: 17.54) and CamG's (2002: 80-83) use of the term gerund-participle. Cf. 3.2.1.1. 15 Compare, for instance, Clark and Clark (1977: 59 and 62) where strategies of the following kind are outlined: "Whenever you find a determiner (...) or quantifier (...), begin a new noun phrase" or "After identifying a determiner or quantifier, look for a noun, which closes out the noun phrase ...". 16 Compare CamG (2002: 542-550). <?page no="100"?> Phrases 88 premodifier(s) head postmodifier(s) complement(s) (70) very interested in structuralism (71) quite so committed to the new wave (72) more passive in the sense that we can't interact with the text (73) sorry to see that you have meanwhile succumbed to the virus of structuralism (74) SW31 as bad as all that head discontinuous modifier 4.4 Adverb phrases The structure of adverb phrases is parallel to that of adjective phrases. The defining property of an adverb phrase is that its head is an adverb. The phrase can consist of the head only: (75) SW29 "Exactly," said Morris Zapp. (76) SW29 "Professor Zapp, I did so enjoy your lecture" ... Adverb phrases allow premodification, postmodification and discontinuous modification as well as complementation: 17 (57) NW45 Not many people perceived this quite so soon. (77) SW25 But if I use my words it follows that I have changed your meaning, however slightly ... (78) VDE He assumed that native speakers could distinguish between grammatical and nongrammatical sentences independently of meaning. (79) NW64 But you never know, something may turn up later in the year. (80) BNC And a lot of people knew him very well indeed, believe me. (81) SW25 Not so, because the same axiom, every decoding is another encoding, applies to literary criticism even more stringently than it does to ordinary spoken discourse. 17 Compare CamG (2002: 570-574). <?page no="101"?> Particle phrases 89 premodifier(s) head postmodifier(s) complement (57) quite so soon (77) however slightly (78) independently of meaning (79) later in the year (80) very well indeed (81) even more stringently than it does to ordinary spoken discourse head discontinuous modifier 4.5 Particle phrases 4.5.1 Structure of the particle phrase It is a defining property of particle phrases that they contain a particle. The particle can be seen as the head of the phrase. Particle phrases allow complementation and premodification: (82) NW41 While Robyn is getting up, and getting ready for the day, thinking mostly about the nineteenth-century industrial novels on which she has to lecture this morning, I will tell you about Charles, and other salient facts of her biography. (83) NW325 What (is your book) about? (84) NW306 You don't know what the real world is like down here ... (85) NW332 You're not trying to pretend, are you, that you're genuinely interested in finding out how University Departments of English operate? (86) NW41 Unlike Vic, Robyn invariably sleeps until woken. (87) NW324 I spend all my time on committees arguing about how to respond to the cuts. (88) NW63 ... for this conversation took place in the Senior Common Room, not long after Robyn's arrival at Rummidge. (89) NW115 Besides, if it wasn't for the cuts, I'd have had a permanent job by now. (90) NW330 I suppose you might as well leave it until after the next stage. <?page no="102"?> Phrases 90 (91) NW384 I was wondering whether, in the circumstances, you would consider staying on next year ... (6) NW50 She began to look for a university job outside Cambridge. premodifier head complement (82) about the nineteenth-century industrial novels on which she has to lecture this morning [noun phrase] (82) about Charles, and other salient facts of her biography [noun phrase] (84) down here 18 (85) in finding out how University Departments of English operate [V-ing clause] (86) until woken [V-ed clause] (87) about how to respond to the cuts [wh_to_INF] (88) not long after Robyn's arrival at Rummidge [noun phrase] (89) by now [adverb phrase] (90) until after the next stage [particle phrase] (91) whether in the circumstances, you would consider staying on next year ... [finite clause] (6) to look for a university job outside Cambridge [infinitive clause] While the analysis of the constructions listed above as particle phrases reveals a structural element they have in common, it is still important to realise that particle phrases can be realised in different ways: 19 18 In the case of I went up there ... NW306 , it is unclear whether the particle there should be seen as a complement of up or whether up is a premodifer of there. 19 Types [2] and [3] roughly correspond to traditional prepositional phrases and traditional subordinate clauses introduced by a subordinating conjunction. There are two important differences between this traditional view and our approach, however: firstly, the gradient between preposi- <?page no="103"?> Particle phrases 91 [1] A particle phrase can consist only of a particle which does not have any complements (single particle construction): (82) NW41 While Robyn is getting up, and getting ready for the day, thinking mostly about the nineteenth-century industrial novels on which she has to lecture this morning, I will tell you about Charles, and other salient facts of her biography. [2] A particle phrase can be headed by a particle which has as its complement a phrase headed by a noun, adjective, adverb or another particle (particle+phrase construction): (82) NW41 While Robyn is getting up, and getting ready for the day, thinking mostly about the nineteenth-century industrial novels on which she has to lecture this morning, I will tell you about Charles, and other salient facts of her biography. 20 (89) NW115 Besides, if it wasn't for the cuts, I'd have had a permanent job by now. (90) NW330 I suppose you might as well leave it until after the next stage. [3] A particle phrase can be headed by a particle which has a clause (usually headed by a verb) as its complement (particle+clause construction): (82) NW41 While Robyn is getting up, and getting ready for the day, thinking mostly about the nineteenth-century industrial novels on which she has to lecture this morning, I will tell you about Charles, and other salient facts of her biography. (91) NW384 I was wondering whether, in the circumstances, you would consider staying on next year ... (85) NW332 You're not trying to pretend, are you, that you're genuinely interested in finding out how University Departments of English operate? 4.5.2 Headed or non-headed particle phrases It is debatable to what extent it is justified to analyse the particle as the head of a particle phrase or, indeed, whether particle phrases should be considered to be headed at all. One case in point is presented by the to-infinitive construction, which we analyse as a particle phrase with a particle to as its head and an infinitive clause complement (6) NW50 She began to look for a university job outside Cambridge. tions and subordinating conjunctions that arises in the case of non-finite clauses (> 3.5.3.2) is resolved by our categorization of constituents headed by a verb as a clause. Thus, for instance, by living in what their parents called sin NW44 is classified as a particle+clause construction. Secondly, traditional grammar does not assign any clause function to the subordinator in subordinate clauses, whereas we see it as the head of the particle phrase or the head of the particle+clause construction, (the term particle+clause construction is to be understood as a particle phrase in which a particle has a clause complement). For a different kind of analysis see Sinclair and Mauranen (2006: 13-14), who discuss the idea of treating it was during the night in I didn't see anything because it was during the night as a subordinate clause and because as a relational element that "belongs to neither clause". 20 Within these particle+phrase constructions, on which and of her biography can be identified as separate particle phrases. <?page no="104"?> Phrases 92 It is obvious that in such cases the particle is assigned head-status mainly for reasons of consistency although one could also make out a case for considering [to_INF]-constructions as single constructional units. 21 However, in this respect, as will be shown in Chapter 6, to-infinitives only present a very good example of a general feature concerning the use of particle phrases namely that the type of particle phrase occurring in a sentence very often not only depends on the valency properties of the particle itself but is specified by the valency properties of the word that governs the particle. 22 This observation could be taken as an argument for not considering particle phrases (or certain types of particle phrases) to be headed at all. Indeed, a distinction is often made between endocentric (or headed) phrases and exocentric (or non-headed) phrases. Thus Quirk, Greenbaum, Leech and Svartvik (CGEL 1985: 2.26), for instance, say: "The obligatory element of a headed construction is called its HEAD; it may function in a manner equivalent to the whole construction of which it is a part." Accordingly, the adjective phrase (very) important (indeed) is given as an example of a headed construction because the adjective important is its only obligatory element. A phrase such as to London, however, is analysed in CGEL as a prepositional phrase with "two normally obligatory elements: the preposition and the prepositional complement". Such a statement, however, depends on the definition of the category preposition as an element that cannot normally stand on its own and thus could not have the qualities describing head-status in CGEL. If one sees traditional prepositions as one (valency-based) use of words belonging to a larger word class particle, then the distinction between endocentric and exocentric phrases need not be made: if to is analysed as a particle with an obligatory valency complement London, then there are good reasons for considering particle phrases as headed. Pullum and Huddleston (CamG 2002: 598) take phrases such as to London, (which they call prepositional phrases) to be headed, but do not regard that, whether and some uses of if as heads. 23 However, if one is prepared to attribute valency or valency-like properties to these words, then it also seems possible to treat them as heads of a kind (with obligatory or contextually optional) complements. Nevertheless, it is obvious that due to the mainly grammatical function of some particles, at least in some cases there are also strong reasons for considering the combination of a particle and its complement as a single constructional unit or a constructional chunk. 24 One might consider using the term quasi-head to describe the function of the particle in such phrases, but there seems to be no need for introducing another term as long as the special nature of particle phrases is understood. 21 Nevertheless, one should not overlook the fact that the particle to can sometimes stand for [to _INF] on its own: "Do I really have to tell you? " - "Not if you don't want to." NW114 22 This is one of the arguments mentioned in CamG (2002: 955) for not considering that the head of so-called that-clauses. 23 Compare also the approach taken in X-bar theory; see, for instance, Haegeman (1991: 78-96). 24 Compare, for instance, the arguments presented by Matthews (2007: 49-55) who points out important differences between prepositional phrases and other types of phrase and argues in favour of a non-headed analysis of prepositional phrases. <?page no="105"?> 5 Clauses 5.1 Constituents of clause structure The structure of the clause can be described in a way that shows strong parallels to that of the phrase, in particular the noun phrase. So clauses can be analysed as consisting of ⌐ a verbal head-complex consisting of a head and possibly pre-heads ⌐ complements which are determined by the verb functioning as head ⌐ adjuncts, which (like the modifiers in the phrase) are independent of the valency of the lexical unit functioning as head. In the case of (1) SW18 Aren't you giving a paper this afternoon? are as pre-head and giving as head form the head-complex, you and a paper are valency complements of the verb give and not and this afternoon are adjuncts. The parallels to the noun phrase are obvious: (1) Like the noun phrase, the clause can be analysed as having a head-complex which can comprise more than one word: the verbal-head complex consists of the head and up to four pre-heads (> 3.2.3). (2) SW23 I never could remember which came first, the morphemes or the phonemes. (3) NW85 Must have been passed without discussion. In fact, the internal structure of verbal head-complexes such as could remember and must have been passed can also be analysed in terms of valency as each verb in the head-complex can be seen as a valency complement of the preceding verb. The valency of the overall head-complex is that of the verb functioning as head. 1 It should be pointed out that what we call the verbal head-complex is referred to as a verb phrase in CGEL (2.28). (2) The complements are dependent on the valency of the governing word, which in the case of the clause is the verb that functions as head of the clause, whereas with phrases it is the head of the phrase. (3) Adjuncts in clause structure correspond to the modifiers in phrase structure. They are independent of the valency properties of the head of the clause or phrase. Semantically, adjuncts and modifiers can be interpreted as either referring to the head or a larger part of the clause or phrase: 1 For a similar view see Palmer (1974: 16). See also the discussion outlined in CamG (2002: 1209- 1220). <?page no="106"?> Clauses 94 (4) SW22 You've really [adjunct] got to tell me what structuralism is all about. (5) SW22 But there was an interesting [modifier] discussion of structuralism afterwards. It is because of these parallels that some linguists regard clauses as verb-headed phrases (cf. Fillmore 1988: 43) 2 . On the other hand, there are important differences between clauses and phrases. As was pointed out in 4.2.2, the head-complex of a noun phrase - together with any obligatory complements of the noun functioning as head - can stand for the noun phrase on its own. Thus it is possible to reduce the noun phrase whose optional courses on women's writing in (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. to whose courses because optional is a premodifier and on women's writing an optional complement: (6a) She was a popular and conscientious teacher, whose courses were oversubscribed. This is not necessarily possible in the case of a clause, however. For instance, although you in (1) is an optional complement of give 3 , clause (1) cannot be represented by something such as (1a) *Aren't giving a paper this afternoon? (1b) *Aren't giving a paper? You in (1) is an obligatory clause constituent of the clause although it is neither part of the head-complex nor an obligatory valency complement of the verb functioning as head. This is an important difference between phrases and clauses because it means that the structure of a clause, unlike that of a phrase, cannot entirely be explained in terms of headedness. Rather, the more traditional view of a clause containing a subject and a predicate is equally necessary to give an account of clause structure. 4 It seems that the fact that a subject is required in clauses such as (1) SW18 Aren't you giving a paper this afternoon? (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. 5 2 Compare 2.4.3. For a discussion of such a view see Matthews (2007: 16-17). 3 For the distinction between obligatory, optional and contextually optional complements see 6.2.1. In the case of give in this sense, you is classified as an optional complement since the verb can be used in this meaning without this complement being realised: Eleven papers were given -which dealt with aspects of the history of popular culture from the 16th through to the 20th centuries in areas as diverse as early modern Germany, 19th century Russia and 20th century Mexico and India. BNC 4 For the traditional description of the difference between phrase and clause see Lyons (1968: 170- 171). See also the approach taken by Aarts/ Aarts (1982/ 1988). For a critical view of "this new term 'clause'" see Palmer (1971: 80). 5 whose optional courses on women’s writing is the subject of the relative clause (> 5.3.7). <?page no="107"?> Functions of clauses 95 (7) NW40 According to Robyn, (or, more precisely, according to the writers who have influenced her thinking on these matters), there is no such thing as the 'self' on which capitalism or the classic novel are founded ... can be explained as a property of a particular clause type since there are also clauses without subjects: (8) SW15 So tell me about structuralism. (9) NW65 Being Dean, you see, I'm very aware of the financial constraints on the University. Clauses thus differ from phrases in that the occurrence or non-occurrence of one constituent, the subject, is not determined by the valency properties of the governing verb and by general properties of clauses alone but has to be accounted for in terms of clause types. We will thus make use of the term clause rather than verb phrase to refer to these constructions. 6 5.2 Functions of clauses One important difference between phrases and clauses is that (some) clauses, socalled independent clauses (> 2.5.3), can function as sentences. Thus sentences such as (1) SW18 Aren't you giving a paper this afternoon? (10) SW3 April is the cruellest month. are sentences consisting of one clause. On the other hand, clauses cannot only function as complete sentences, they can also be constituents of larger clauses or of phrases, in which case they are referred to as dependent (or subordinate) clauses. There they can function as [1] complements of valency carriers: (2) SW23 I never could remember which came first, the morphemes or the phonemes. [complement of remember] (11) NW327 She considered the effort would be worthwhile if she could secure the endorsement of a prestigious imprint like Euphoric State University Press. [complement of consider] (12) NW64 I think you were going to say something beginning with But. [complement of think] (13) SW23 Jakobson cites the gradation of positive, comparative and superlative forms of the adjective as evidence that language is not a totally arbitrary system. [complement of that] 6 Within our terminological framework it would be inconsistent to use the term verb phrase in a way that is not identical with clause, however. Referring to all elements of a clause except the subject as a verb phrase would mean that not all valency complements of the head would be seen as belonging to the verb phrase, which would be an important difference to other phrases. As pointed out in 2.4.3, the term phrase will be restricted to constructions headed by nouns, adjectives, adverbs and particles, whereas constructions headed by verbs will be referred to as clauses. <?page no="108"?> Clauses 96 (14) NW29 ... the University day begins too late and finishes too early to inconvenience Vic himself ... [complement of to] (15) NW50 When the dust settled in Cambridge, however, it seemed that the party of reaction had triumphed. [complement of when] [2a] adjunct units in clauses: (9) NW65 Being Dean, you see, I'm very aware of the financial constraints on the University. [2b] modifiers in phrases: (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. As will be shown below, not all types of clause can fulfil all of these functions, but it is interesting to see that in English a clause such as whose optional courses on women's writing were oversubscribed cannot only be used as a relative clause as in (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. but equally well serve as a complement of a verb (6b) They didn't know whose optional courses on women's writing were oversubscribed. or as independent sentence: (6c) Whose optional courses on women's writing were oversubscribed? The same is true of a clause such as (12a) You were going to say something beginning with But. which can occur as an independent clause as in (12a), as a complement-clause of think as in (12) NW64 I think you were going to say something beginning with But. [complement of think] or as part of a particle phrase introduced by that: 7 (12b) I think that you were going to say something beginning with But. [complement of think] 7 Traditionally those particle phrases which like (12b) consist of a particle and a clause, i.e. the ones that have been called particle+clause-constructions in 4.6.1, are subsumed under subordinate clause. In this analysis, the particle is seen as a subordinator which is part of the subordinate clause. <?page no="109"?> Different types of clauses 97 5.3 Different types of clauses 5.3.1 Subject - predicate - adjunct It is the purpose of this chapter to outline different types of structural patterns that are relevant to the description of the clause in English. We will use the term clause type to distinguish different types of clause according to formal properties of morphology and position. The identification of clause types will be based on the distinction between subject, predicate and adjunct as defined in Chapter 2. Many, but not all clauses have subjects [Subj]. Subjects realise a valency complement of the predicator and in some clauses show concord with the first pre-head or the head of the predicate. Whenever a clause type contains a subject, this is specified. According to our definition of clause, all clauses have predicates [Pred]. The predicate can thus be seen as an obligatory constituent of the clause. The predicate can be a very complex constituent: Firstly, it contains the verbal head-complex, which is an obligatory constituent of the predicate. The verbal head-complex consists of a head and up to four preheads (> 3.2.3). Secondly, it provides slots for valency complements of the predicator, i.e. the verb functioning as head. As far as the clause type is concerned, there is no need to specify the number of preheads or of valency complements in the predicate. In some cases, it is necessary, however, to distinguish the first pre-head (which functions as an operator verb [op]) or other elements from the rest of the predicate [rPred]. In this way, (1) SW18 Aren't you giving a paper this afternoon? can be assigned to the type [op __ Subj __ rPred], for instance. From a structural point of view, adjuncts are optional elements of clause structure. As was pointed out in 2.2.2, their occurrence and position in the clause depend on the characteristics of the type of adjunct rather than the characteristics of the clause. This is why adjuncts play no major role in distinguishing between different kinds of clause types and will only be considered where their occurrence or position can be seen as characteristic of a particular clause type. A case in point is the wh-element in (16) NW218 Why should society pay to be told people don't mean what they say or don't say what they mean? which must be seen as characteristic of a particular clause type (> 5.3.6) and thus appears in the description of that clause type. However, nothing would be gained by establishing a separate clause type which provides for the adjunct anyway in (17) NW218 They don't pay us all that much, anyway. to distinguish it from clauses without adjunct or with the adjunct in initial position: (17a) They don't pay us all that much. <?page no="110"?> Clauses 98 (17b) Anyway, they don't pay us all that much. For this reason, (17), (17a) and (17b) are subsumed under clause type [Subj __ Pred]. The criteria to distinguish between different clause types are based on the morphology of the verbs in the verbal head-complex (5.3.2 and 5.3.3), the occurrence or non-occurrence of a subject (5.3.4), word order (5.3.5), characteristic use of particular words (5.3.6-7). 5.3.2 Finite and non-finite clauses In the description of clauses, a distinction is usually made between finite and nonfinite clauses. One important criterion on which this distinction is based is verb morphology: finite verb forms are seen as being marked for categories such as tense, mood, person and number, whereas non-finite forms are not. Thus thinks in (18) NW484 Robyn thinks. can be seen as a prototypical finite form because the {S}-affix marks it as present tense, 3rd person, singular, whereas the participle being is a prototypical non-finite form (9) NW64 Being Dean, you see, I'm very aware of the financial constraints on the University. In the case of the base form of the verb (> 3.2.1.1), however, there is no morphological reason why the infinitive in (19) NW65 It's no fun being Dean, these days, I can tell you. should be classified as a non-finite form, whereas the imperative in (8) NW15 So tell me about structuralism. should be classified as a finite form. This is why a further criterion is needed to maintain the distinction, namely that non-finite verb forms can never occur as the only verb form in an independent clause (> 2.5.3). On that basis tell in (8) can be classified as a finite form. 8 Finite clauses, then, are clauses that contain one finite verb form (i.e. a present tense, past tense or imperative form). Non-finite clauses are clauses that contain only nonfinite verb forms. They can be divided into participle clauses and infinitive clauses: [1] [V-ing]-clauses 9 (9) NW64 Being Dean, you see, I'm very aware of the financial constraints on the University. [2] [V-ed]-clauses 8 Compare Herbst/ Stoll/ Westermayr (1991). See also the discussion of the notions of finite and non-finite in CamG (2002: 88-89). Of course, one can argue that the independent clause criterion is no more than an attempt to maintain a distinction which is common in traditional grammar. 9 Abbreviations such as [V-ing] stand for a predicate which contains an ing-form as its head. <?page no="111"?> Different types of clauses 99 (20) NW215 The steady grind of intellectual work, punctuated by brief explosions of indoor physical exercise - that was the rhythm of Robyn's first winter at Rummidge. [3] [INF]-clauses (21) NW330 All I want to do at this moment is fall into bed. [INF]-clauses often function as a complement of the particle to to form particle phrases, which can be described as [4] [to_INF] (22) NW307 Well, was it a good idea to build so many new universities in parks on the outskirts of cathedral cities and country towns? 5.3.3 Active and passive clauses Both finite and non-finite clauses can be active or passive, depending on whether the governing verb occurs in an active construction or in a passive construction: (23) SW3 Some taps could not be turned on [finite passive] and some could not be turned off. [finite passive] (16) NW218 Why should society pay to be told people don't mean what they say or say what they mean? [non-finite passive] 5.3.4 Clauses with subject and without subject Both finite and non-finite clauses can occur with and without subjects: (24) NW85 I [subject of finite clause] don't remember it coming up at that meeting of the senate. (24) NW85 I don't remember it [subject of non-finite clause] coming up at that meeting of the senate. (3) NW85 Must have been passed without discussion. [finite clause without subject] (9) NW64 Being Dean [non-finite clause without subject], you see, I'm very aware of the financial constraints on the University. The use of finite clauses without subject such as (3) is restricted to very special contextual conditions (> 7.1.2.2). In some cases, it is not easy to distinguish between, for example, an [NP_V-ing]-clause and a noun phrase [NP] with an ing-clause as postmodifier: 10 (25) NW60 You would not think her unduly burdened with worries, if you watched her crossing the campus, smiling at people she knows, her eyes bright, her brow unfurrowed. (26) SW22 "Structuralism? ", said Dempsey, coming up with a sherry for Angelica just in time to hear Persse's plea, and all too eager to show off his expertise. 10 Such constructions must be kept apart from trivalent uses of verbs in which a noun phrase complement is followed by a [V-ing]-complement such as The snow adhering to the car windows created a sepulchral gloom inside, and Robyn spent several minutes brushing it off before she attempted to start the engine. NW96 <?page no="112"?> Clauses 100 5.3.5 Word order Word order can be used as a criterion to distinguish different kinds of clause: [1] [Subj__Pred]: the subject precedes the predicate (10) SW3 April is the cruellest month. [2] [op__Subj__rPred]: the operator precedes the subject, which is followed by the rest of the predicate (1) SW18 Aren't you giving a paper this afternoon? 5.3.6 Wh-clauses Clauses whose first word is a wh-word (> 3.8.1) can be referred to as wh-clauses. Whclauses can occur as independent sentences, as in the following examples: (27) NW63 Who will notice? (28) NW87 Who shall I nominate then? (29) NW342 How long are they allowed? In (27), the subject complement unit is realised by a wh-constituent (type: [wh- Subj__Pred]), whereas in (28) and (29) a predicate complement unit and an adjunct unit are realised by a wh-constituent (type [wh-A/ C__op__Subj__rPred]). The first type can also occur as a dependent clause, as in: (2) SW23 I never could remember which came first, the morphemes or the phonemes. (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. whereas in the case of predicate complement units or adjunct units word order is different in dependent clauses [wh-A/ C__Subj__rPred]: (30) NW383 In fact, I don't know how we shall manage to do without you next year. Wh-clauses can have different functions: for instance, they can be interrogative- 'question'-constructions as in (27), introduce subordinate clauses that function as complements of verbs or adjectives as in (2) or function as relative clauses as in (6). It is necessary to distinguish between two different types of construction: one could argue that in the examples given above the wh-pronouns, wh-adverbs or wh-particles introducing interrogative or relative constructions have a certain linking function. This is also true of whether in (31) NW59 Robyn is not sure whether this is wonderfully modern and liberated of them, or rather depraved. The difference between whether in (31) and how in (30) and what in (32) (30) NW383 In fact, I don't know how we shall manage to do without you next year. (32) NW87 I'm sure I don't know what you are talking about. <?page no="113"?> Different types of clauses 101 is that whether in (31) does not, for example, function as a complement unit or adjunct unit of the subordinate clause, which is, however, the case with how or what in (30) and what in (32). It is for this reason that whether this is wonderfully modern and liberated of them, or rather depraved in (31) has to be analysed as a particle phrase (headed by whether). What you are talking about in (32), however, can be analysed as a wh-clause of the type identified above, in which about what functions as a predicate complement unit: 11 (31) whether this is wonderfully modern and liberated of them, or rather depraved. h: particle c: clause SCU VHC PCU (32) what you are talking about PCU (discontinuous) SCU VHC PCU (discontinuous) Wh-clauses such as (32) find a parallel in wh-to-infinitive constructions of the following kind: 12 (33) NW59-60 It is not that she does not know what to say, it is that there is not enough time to say all she knows. (34) NW61 She is always uncertain how to address her Head of Department. 5.3.7 Relative clauses Clauses are often given labels according to their function. One type of clause that ought to be mentioned here is the relative clause. Relative clauses are postmodifiers of heads in noun phrases, which contain as (part of) a first constituent a wh-word or that functioning as a relative pronoun or relative particle 13 : (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. 11 Note that where, when etc. can often function as complements in the same valency slots as [wh__CL]-complements: I don't know where. BNC . 12 Here, the wh-element is part of the infinitive clause. The whole construction is a particle phrase headed by to. 13 Since the particle in such clauses represents a clause constituent of the subordinate clause and does not introduce it, the term relative adverb is equally justified. Note the difference between He left the previous evening to drive to Ipswich, where his own term is due to begin today at the University of Suffolk NW41 , in which where functions as an adjunct in the clause underlined, and When the dust settled in Cambridge, however, it seemed that the party of reaction had triumphed NW50 , in which when introduces the subordinate clause but is not a clause constituent of that clause. <?page no="114"?> Clauses 102 (7) NW40 According to Robyn, (or, more precisely, according to the writers who have influenced her thinking on these matters), there is no such thing as the 'self' on which capitalism or the classic novel are founded ... (35) NW41 He left the previous evening to drive to Ipswich, where his own term is due to begin today at the University of Suffolk. Relative clause constructions without a relative pronoun or relative particle are possible if the head of the noun phrase corresponds to a noun-complement of the predicate of the relative clause: (36) NW215 The Shadow Scheme contributed nothing to that - on the contrary, it interfered with it, taking up the precious one day a week she had kept free from Departmental duties. 5.4 Clause types 5.4.1 The notion of clause type Using the categories established above, the following clause types will be established. As was pointed out above, neither adjuncts nor complements that occur in the predicate will be specified unless they are characteristic of a particular clause type. The columns on the right indicate whether a clause can occur as an independent clause, i.e. a sentence, a valency complement of a verb, noun or adjective, an adjunct in a clause, a modifier in a phrase and as a complement of a particle to form a particle phrase. Subj subject Pred predicate op operator verb wh-Subj subject containing wh-word wh-A adjunct containing wh-word wh-C complement belonging to the predicate containing wh-word rPred remaining predicate (without op or other element specified separately) <?page no="115"?> Clause types 103 5.4.2 Finite clauses Sentence Complement Adjunct Modifier C. of particle 1 [Subj__Pred] (10) SW3 April is the cruellest month. (37) NW84 Sometimes I think the VC takes his economy drive a little too far. (15) NW50 When the dust settled in Cambridge, however, it seemed that the party of reaction had triumphed. 1a [wh-Subj__Pred] (27) NW63 Who will notice? (2) SW23 I never could remember which came first, the morphemes or the phonemes. (6) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. (38) BNC Anyway, I thought you wanted to be sure about who was organizing it. 1b [wh-A/ C__Subj__rPred] (39) NW110 I can't imagine where that would be (40) NW327 ... it was nearly eight-fifteen in the morning when she finished the task. (41) NW305 She set up her word-processor on the scratched and stained desk where she had swotted for her A-levels ... (42) NW96 Robyn was well aware that clothes do not merely serve the practical purpose of covering our bodies, but also convey messages about who we are, what we are doing, and how we feel. <?page no="116"?> Clauses 104 Sentence Complement Adjunct Modifier C. of particle 2 [op__Subj__rPred] (1) SW18 Aren't you giving a paper this afternoon? (43) NW124 'Can't you do something about these pictures? ' she asked Wilcox. 14 2a [wh-A/ C__op__Subj__rPred] (28) NW87 Who shall I nominate then? (44) NW23 'Where would you put it? ' Vic inquires, ... 14 3 [___Pred] (8) NW15 So tell me about structuralism. (45) NW104 'Take a seat,' he said, ... 14 5.4.3 Non-finite clauses without subject Complement Adjunct Modifier C. of particle [V-ing] (46) NW314 'You shit,' Robyn said aloud, when she had finished reading the letter. (47) NW214 She drove to and fro between her cosy little house and her warm, well-lit room at the University, ignoring the dismal weather. (48) SW8 Swallow is the man chatting up that rather dishy girl who just came in ... (49) SW60 ... and even after publishing her book she went on accumulating ideas and insights about the subject. 14 Note that these clause types only occur as complements of the type QUOTE (> 6.4.2.7). <?page no="117"?> Clause types 105 Complement Adjunct Modifier C. of particle [V-ed] (50) VDE What was Laura doing married to this kind man? (51) NW43 Cut off from normal social intercourse with the adult world, relieved of inhibition by the ethos of the Permissive Society, the students were apt to run wild ... (52) NW43 The Sussex campus, with its tastefully harmonized buildings in the modernist-Palladian style, arranged in elegant perspective at the foot of the South Downs a few miles outside Brighton, was much admired by architects ... (53) NW15 FOUR toilets? he said, when first shown over the house. [INF] (54) NW214 No other reason would make anyone come here, or having come, stay. (22) NW307 Well, was it a good idea to build so many new universities in parks on the outskirts of cathedral cities and country towns? 5.4.4 Non-finite clauses with subject [NP_V-ing] (25) NW60 You would not think her unduly burdened with worries, if you watched her crossing the campus, smiling at people she knows, her eyes bright, her brow unfurrowed. (55) NW70 Robyn Penrose is making her way to Lecture Room A, along corridors and down staircases thronged with students changing classes. [NP_V-ed] (56) SW44 I would like to get my poems published … <?page no="118"?> Clauses 106 Complement Adjunct Modifier C. of particle [NP_INF] 15 (57) NW149 After a moment or two she heard him call, "Turn the ignition key," and when she did so, the engine fired. 5.5 Quasi-clauses Many grammars identify so-called verbless clauses. If one defines the clause as being headed by a verb as we have done, this is a rather unfortunate choice of term. We thus use the term quasi-clauses for constructions which do not have a verbal head-complex which function in a way similar to clauses but cannot be analysed as phrases. Quasi-clauses can be of two types. Firstly, there are quasi-clauses that do not contain a verbal head-complex but consist of more than one clause constituent. (58) NW45 Robyn, a dressing-gown over her underclothes and slippers on her feet, descends the short dark staircase to the groundfloor ... Such constructions can be introduced by with or without (CamG, 2002: 1267). (59) NW96 With the Rummidge A to Z open on the passenger seat beside her, she set off to find J. Pringle & Sons, somewhere on the other side of the city: the dark side of Rummidge, as foreign to her as the dark side of the moon. (60) SW24 Morris Zapp delivered it striding up and down the platform with his notes in one hand and a fat cigar in the other. Secondly, there are constructions such as (61) BNC In other words, if in doubt, don't take a chance. (62) BNC When in London, he stayed with the Fabers ... which are introduced by words that are usually followed by a clause. Such "verbless clauses" can be introduced by although, as if, as though, if, though, when, whenever, while and whilst (CamG 2002: 1267). 16 In the framework introduced here, cases such as (59- 15 Note that an [NP_to INF]-construction, which we analyse as a particle phrase, can occur as a complement of a verb or a particle: What the students want her to do is to give them some basic facts ... NW60 , You're a third of the way into your second year, with another full year still to run from next September. NW64 . 16 It is often seen as a characteristic of verbless clauses, that a "missing" form of be can be inserted: With the Rummidge A to Z being open on the passenger seat beside her... or When he was in London .... This is not a particularly reliable criterion, however, since it results in a very broad concept of <?page no="119"?> Quasi-clauses 107 62) are analysed as particle phrases (> 4.5 and 6.8.2). (61) and (62) represent monovalent uses of the particles if and when, whereas (59) and (60) are examples of divalent uses of the particle with. quasi-clauses. While CGEL (1985: 10.16) treats cases such as He waited, anxious for a reply or She was standing, a picture of innocence as a typical example of "a verbless clause with adverbial status", we do not classify them as quasi-clauses but as adjective phrases and noun phrases used as adjuncts that are co-referential with the subject. <?page no="120"?> 6 Valency 6.1 Basic concepts of the valency model 6.1.1 Valency slots This chapter is concerned with the different aspects of the description of the valency properties of words that have valency - above all, verbs, adjectives and nouns. Before we delineate the categories for the description of valency in detail, we would like to outline the basic principles of the model employed here: A lexical unit has the property of valency if it opens up one or more valency slots which can or must be realised by a complement. 1 At the formal level, valency slots will be described in terms of the complements which can fill them. A complement is any formal realisation of a valency slot for instance, a phrase or a clause. At the semantic level, valency slots can be characterized in terms of participants or participant roles, which characterize the semantic function of the complement in the clause. With respect to optionality, valency slots will be characterized as to whether a slot must or can be realised by a complement. Adjuncts are not part of the valency description. In the case of (1) NW51 Robyn considered herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave. the lexical unit consider (meaning 'judge') opens up three valency slots, which are realised by the complements Robyn, herself and lucky to get a job for one term at one of the London colleges … and whose participant roles can be described tentatively as " JUDGE ", " THING JUDGED " and " JUDGEMENT ". Within the third complement, the adjective lucky opens up a separate valency slot, which is realised by a [to_INF]-complement and which can be attributed a participant role such as ' CAUSE '. Valency is a property of lexical units, i.e. combinations of a lexical form with a particular meaning (Cruse 1986: 80). This can be shown by the fact that in (2) NW384 But I was wondering whether, in the circumstances, you would consider staying on next year, and see what happens. 1 Compare Helbig/ Schenkel's ( 2 1973: 49) definition of syntactic valency: "... verstehen wir unter syntaktischer Valenz die Fähigkeit des Verbs, bestimmte Leerstellen um sich herum zu eröffnen, die durch obligatorische oder fakultative Mitspieler zu besetzen sind." <?page no="121"?> Basic concepts of the valency model 109 the verb consider has a different valency structure and a different meaning ('think about'). This lexical unit opens up two valency slots, which are realised by the complements you and the [V-ing]-clause complement staying on next year, whose participant roles can be described as " THINKER " and " THING THOUGHT ABOUT ". 6.1.2 Complements and adjuncts The concept of valency is based on the idea of distinguishing between those constituents of a clause whose occurrence is in some way or other typically associated with the occurrence of a particular item, the valency carrier, and those elements which are not. The former are called complements (actants, Ergänzungen, Komplemente), the latter adjuncts (circonstants, freie Angaben). Although the notion of valency is often associated with Tesnière's dependency grammar (1959), the valency approach was developed to a large extent in German linguistics (Helbig/ Schenkel (1969), Helbig (1992), Heringer (1970, 1996), Engel (1977), Schumacher et al. (2004) and many others) 2 , where it can now be regarded as the standard model of syntactic description. 3 It is interesting to see that much of the work on valency was carried out in the context of foreign language teaching (Helbig/ Schenkel 1969). This is hardly surprising since the distinction between complements and adjuncts is particularly relevant in a foreign language context. If one wants to describe the typical uses of a word in, say, a dictionary for foreign learners of English, it is important to specify that a verb such as read can be used with a noun phrase such as a book (3) NW324 I haven't read a book in months, let alone tried to write one. but not that it can co-occur with an adjunct such as in months. Similarly, it is far more important to indicate that adjectives such as desperate can be used with a [to_INF]construction as in (4) NW64 The University is desperate to save on salaries. than it is to indicate that it can be premodified by an adverb as in (5) NW52 Thus a three-year lectureship in English Literature was advertised, Robyn applied, was interviewed along with four other equally desperate and highly qualified candidates, and was appointed. This is simply because not all adjectives can take [to_INF]-complements, and not all verbs can take [NP]-complements such as a book in (3), whereas it is a general property of the word class of adjectives that they can be premodified by adverbs (> 3.4) and all verbs can co-occur with adjuncts such as in months. 2 For early indications of ideas related to valency see Bühler (1934/ 1999: esp. 173, 251, 295) and de Groot (1949/ 1964: 114-115). 3 For early applications of the valency approach to English see Emons (1974, 1978), Allerton (1982), Herbst (1983) or Somers (1987). <?page no="122"?> Valency 110 From the foreign learner's perspective, it is almost equally important to know which constructions are not possible with a particular word although they occur with another word: (6) NW93 I don't understand why you agreed to do it. (6a) *I don't understand why you agreed doing it. (2) NW384 But I was wondering whether, in the circumstances, you would consider staying on next year, and see what happens. (2a) *But I was wondering whether, in the circumstances, you would consider to stay on next year, and see what happens. (7) NW307 This Government is capable of anything. (7a) *This Government is capable to do anything. (8) NW307 You know, I ought to be able to order one of you to do this shadow nonsense. (8a) *You know, I ought to be able of ordering one of you to do this shadow nonsense. It is safe to assume that their status as item-specific constructions also affects the representation of complements in the mental lexicon in the sense that they are stored together with or linked with the item in question. Nevertheless, the distinction between the item-specific constituents (complements) and constituents such as adjuncts (in clauses) and modifiers (in phrases), which are not related in such a specific way to a governing element, is by no means clear-cut. 6.2 Different degrees of optionality 6.2.1 Obligatory, optional and contextually optional complement slots One aspect concerning the relationship between a valency carrier and a particular complement concerns the question of whether the respective valency slot has to be realised syntactically whenever the valency carrier is used or not. Thus, for instance, a book in (3) NW324 I haven't read a book in months, let alone tried to write one. can be seen as a typical example of what is commonly called an optional complement because the verb read can equally well be used without it: (9) NW215 At home she read, she took notes, she distilled her notes into continuous prose on her word-processor, she marked essays; at the University she lectured, she gave seminars and tutorials, she counselled students, interviewed applicants, drew up reading-lists, attended committee meetings, and marked essays. The verb take in (9), however, could not be used without the [NP]-complement represented here by notes, which can thus be classified as an obligatory complement of take. It is important to realise that our definition of obligatory and optional focuses on possible uses of the lexical unit in question and not on whether a complement can be deleted from an actual sentence or not. Thus the fact that the first she cannot be de- <?page no="123"?> Different degrees of optionality 111 leted in (9) is regarded as a structural requirement of the sentence (> 7.1.2.1). From a valency point of view, she in (9) is seen as representing an optional complement of take and read because both verbs can be used without that complement in passive clauses, clauses with imperative forms and many types of subordinate clause: 4 (9a) At home, books were read, notes were taken ... (10) NW215 There were always so many books, so many articles in so many journals, waiting to be read, digested, distilled and synthesized with all the other books and articles she had read, digested, distilled and synthesized. The distinction between optional and obligatory valency slots represents a gradient, in which different degrees of optionality or obligatoriness must be distinguished. Thus, the use of read without the [NP]-complement in (9) NW215 At home she read ... seems perfectly normal and it does not seem unusual or odd that the reader of the novel is not told what it is that she reads. A sentence such as (11) NW306 'I shall be extremely sorry,' said Robyn. on the other hand, can only be fully perceived if the reader is able to identify from the context what the person is sorry about. In the case of (11), this information is provided in the immediately preceding sentence: (11) NW306 'I don't suppose you'll be sorry to leave Rummidge, will you, dear? ' said her mother. 'I shall be extremely sorry,' said Robyn. In the same way, the first use of understand in the following example is only possible because the information expected can be retrieved from the context: (12) NW216 Of course, they don't know what we do, but even if one tried to explain it to them they wouldn't understand, and even if they understood what we were doing they wouldn't understand why we are doing it, or why anybody should pay us to do it. The kind of optionality observed with sorry or understand is subject to severe contextual conditions, which do not hold in cases such as read. We will thus divide up the optionality scale and make the following distinctions: The valency slots that a verb, adjective or noun opens up can be obligatory, i.e. the complement slot must be filled whenever the valency carrier is used contextually optional, i.e. the complement slot need not be filled if the information required can be obtained from the context optional, i.e. the complement slot can but need not be filled. 4 Cf. also: The notes were taken early on in the research when I was chiefly concerned with documenting pollution control work ... BNC <?page no="124"?> Valency 112 6.2.2 Three levels of necessity It has to be noted that we classify the valency slots provided by the valency carrier as optional or contextually-optional but not the complements as such. Such a careful formulation is necessary because it does not imply that any elements that are present in existing language text could be deleted. This is not only true of subjects, which can be realised by complements filling an optional valency slot of the verb but cannot be deleted on the basis of the structural requirements of the clause, but also of any other elements which from a structural or valency point of view can be considered as optional. For instance, it is obvious that in the case of (13) NW328 Great, I'll read it as soon as I can. the complement it or the adjunct as soon as I can could not be deleted if (13) were to serve as an answer to (13a) What will you read as soon as you can? or (13b) When will you read it? respectively. This means that one can distinguish between three types of necessity: 5 In the case of structural necessity a clause constituent is required by the syntactic construction: thus, clauses always require a predicate and many of the sentence types identified (> 7.1) require the presence of a subject, for instance. Necessity at the level of valency refers to the fact that certain verbs etc. may require certain complements to occur whenever they are used in a particular meaning. Communicative necessity refers to the information that needs to be provided in a particular situation. Very often, communicative necessity arises from the immediate context of the utterance. Thus, although in parks on the outskirts of cathedral cities and county towns is an adjunct in (14) NW307 Well, was it a good idea to build so many new universities in parks on the outskirts of cathedral cities and county towns? the context shows very clearly that the sentence would hardly make sense without it: (14) NW307 ... but don't you think, in retrospect, that the way Robbins was implemented was a mistake? ' Professor Penrose laid down his knife and fork and look at Robyn over his spectacles. 'What do you mean? ' 'Well, was it a good idea to build so many new universities in parks on the outskirts of cathedral cities and county towns? ' 5 Compare VDE (2004: xxx) for this distinction. See also Herbst/ Roe (1996) and Herbst (1999). <?page no="125"?> Distinguishing between complements and adjuncts 113 Similarly, in America, although an adjunct, is an essential part of the question (15) NW325 How many did you sell in America? It is important to bear in mind that necessity at the levels of structure and of valency abstracts from actual utterances as they occur in actual language use in that they describe generally which elements can or must be used if a particular construction (such as an interrogative-'question'-construction) or a particular valency carrier (such as the verb put) is used. Thus in (16) NW324 'Ah, Robyn,' said Philip Swallow, 'you put us all to shame. ...' (17) NW61 'Put me down for picketing, Bob,' says Robyn. one could argue that to shame and down fill obligatory valency slots of put, which can thus be attributed to necessity at the level of valency. You in (16), however, is optional from a valency point of view but required for structural reasons, whereas the imperative-'directive'-construction (> 7.1.3) in (17) does not require a subject. Communicative necessity, on the other hand, refers to the communicative weight of elements in a particular situation of utterance. In fact, it could be argued that communicative necessity is dominant anyway since presumably the choice of a particular construction and the choice of a particular verb are driven by what a speaker wants to say. It is precisely for this reason, however, that elements that are classified as optional from a structural or valency point of view could not necessarily be deleted from actual language text. 6 6.3 Distinguishing between complements and adjuncts As indicated in 6.1.2, the distinction between complements and adjuncts is not always clear-cut. It is obvious that a simple deletion test will not be sufficient to distinguish complements from adjuncts: firstly, because, as was shown above, deletion tests are not entirely unproblematic methodologically anyway, but, secondly, because not all elements that are potentially deletable are adjuncts. What makes distinguishing complements from adjuncts difficult is that the distinction does not coincide with formal categories of any kind. Thus, for instance, noun phrases can realise complements or adjuncts: (18) NW65 It's no fun being Dean, these days, I can tell you. (19) NW46 The more conservative dons viewed these ideas and their proponents with alarm, seeing in them a threat to the traditional values and methods of literary scholarship. The same applies to [V-ing]-clauses, which, as in (19), can be adjuncts, or, complements, as in 6 This view of the optionality of valency complements is very much in line with a surface-oriented approach and in particular the position described by Sinclair and Mauranen (2006: 150). Compare Herbst (2008 forthcoming). <?page no="126"?> Valency 114 (2) NW384 But I was wondering whether, in the circumstances, you would consider staying on next year, and see what happens. German linguistics has developed a number of test criteria to support the distinction between complements and adjuncts. 7 For English, one test that could be applied is that adjuncts can sometimes be separated off in a clause of their own: (20) SW18 I bought this hat at Heathrow this morning. (20a) I bought this hat. This happened at Heathrow this morning. (21) NW46 So Robyn and Charles went to Cambridge to do their PhDs. (21a) Robyn and Charles went to Cambridge. They did this to do their PhDs. While such criteria can be quite useful to decide on borderline cases, it has to be said that all tests of this kind are of limited value. In cases such as (18) or (19) this test produces very awkward results. Furthermore, while one could find arguments for considering on Romantic landscape poetry a complement of lecture in (22) NW130 She thought of Charles at the University of Suffolk, giving a lecture, perhaps, on Romantic landscape poetry, illustrated with slides. it would still be possible to isolate it in a separate clause: (22a) She thought of Charles at the University of Suffolk, giving a lecture. It was, perhaps, on Romantic landscape poetry, illustrated with slides. A more reliable indication of complement status is that adjuncts generally have more positional mobility than complements: (19) NW46 The more conservative dons viewed these ideas and their proponents with alarm, seeing in them a threat to the traditional values and methods of literary scholarship. (19a) Seeing in them a threat to the traditional values and methods of literary scholarship, the more conservative dons viewed these ideas and their proponents with alarm. (23) NW329 They have given him such a rough time in the past, at conferences and in reviews. (23a) At conferences and in reviews they have given him such a rough time in the past. (23b) In the past they have given him such a rough time, at conferences and in reviews. (24) NW111 It was a bit chaotic at the University last Wednesday. (24a) Last Wednesday, it was a bit chaotic at the University. (24b) At the University it was a bit chaotic last Wednesday. (25) NW218 They don't pay us all that much, anyway. (25a) Anyway, they don't pay us all that much. 7 Cf., for example, Helbig/ Schenkel ( 2 1973: 31-40), Helbig (1992: 72-87), Zifonun/ Hoffmann/ Strecker et al. (1997: 1043-1064). See also Herbst/ Heath/ Dederding (1980). For English, see especially Emons (1974: 65-105), Herbst (1983 and 1988) and Somers (1987: esp. 11-18). <?page no="127"?> Distinguishing between complements and adjuncts 115 While the sentences with the adjuncts in front position seem perfectly normal, front position of complements which belong to the predicate is either only marginally acceptable or entails strong emphasis: (26) NW47 These she reads with the kind of pure, trance-like attention that she used to give, as a child, to the stories of Enid Blyton. (25b) ? Us, they don't pay all that much, anyway. (25c) *All that much, they don't pay us, anyway. (20c) This hat I bought at Heathrow this morning. ['strong emphasis'] (20d) This morning I bought this hat at Heathrow. (20e) At Heathrow I bought this hat this morning. This also applies to clauses: thus a [to_INF]-adjunct expressing ' PURPOSE ' as in (21) can also occur at the beginning of a sentence: (21) NW46 So Robyn and Charles went to Cambridge to do their PhDs. (21b) To do their PhDs, Robyn and Charles went to Cambridge. This is not possible in the case of the [to_INF]-complement of desperate, for example. (4) NW64 The University is desperate to save on salaries. (4a) *To save on salaries, the University is desperate. A further criterion which has been suggested to distinguish between valency-bound complements and adjuncts is that of question forms. It concerns particle phrases, which can realise adjuncts as in (27) NW70 Robyn herself would disapprove of it on ideological grounds ... (28) NW71 Robyn arranges her notes on the lectern … or complements as of it in (27) or also on Romantic landscape poetry in (22) NW130 She thought of Charles at the University of Suffolk, giving a lecture, perhaps, on Romantic landscape poetry, illustrated with slides. Adjuncts very often allow adverbial or particle question forms such as (27a) Why would Robyn disapprove of it? (28a) Where does Robyn arrange her notes? whereas question forms with who or what can be taken as indicators of complement status: (27b) What would Robyn disapprove of? (22b) What was Charles giving a lecture on? Nevertheless, it has to be said that none of these tests are wholly satisfactory since not all adjuncts have the same positional mobility and not all elements that allow whoor what-questions are necessarily item-specific. Such tests (in some more refined <?page no="128"?> Valency 116 form than has been indicated here) can nevertheless be regarded as a useful methodological device that can contribute towards achieving consistency in the analysis of borderline cases. 8 As far as the distinction between adjuncts and complements is concerned, it seems appropriate to make it on the basis of the two criteria outlined in 2.3, where a complement was defined as a constituent that either fills an obligatory valency slot and/ or is determined in its form by the governing element. The latter criterion is the obvious one since it shows that a particular constituent is specifically related to a governing word. This is illustrated by the fact that some verbs in English allow [to_INF]-complements and others [in_V-ing]-complements: (6) NW93 I don't understand why you agreed to do it. (29) NW44 It was true that he had been educated at a public school, but he managed to disguise this handicap very well. (30) VDE She had finally succeeded in winning the title. The first criterion is necessary because some types of complement cannot really be distinguished from adjuncts as far as their formal realisation is concerned. A case in point is the verb live (the lexical unit meaning 'reside'). It would be difficult to argue that the constituents underlined are determined in their form by the verb: (31) NW48 Charles continued to live with her in Cambridge, commuting by car to Ipswich to teach his classes, and staying there for a night or two each week. (32) NW28 Though he lives on its doorstep, so to speak, Vic has never been inside the place. (33) NW47 ... and it crosses Robyn's mind, not for the first time, that it is a pity she lives so far from the metropolis where such exciting events are always happening. (34) NW97 This part of the city lacks the individual character of Robyn's own suburb, where healthfood stores and sportswear boutiques and alternative bookshops have sprung up to cater for the students and liberal-minded yuppies who live there; and still more does it lack the green amenities of the residential streets around the University. Since the formal realisations of these complements are the same as for adjuncts, it is the fact that they fill an obligatory valency slot of the verb that justifies their complement status. 9 8 A further test that has been suggested is that of free addibility. Adjuncts can be added to most sentences relatively freely (provided they are semantically compatible). Thus one could easily add adjuncts such as now, at the university, whenever I am at the university, etc. to a sentence such as I spend all my time on committees arguing about how to respond to the cuts. NW324 . Compare Ágel (2000: 185) and Heringer (1996: 159). 9 For a discussion of these complements see 6.4.2.4. <?page no="129"?> Formal realisations of complements 117 6.4 Formal realisations of complements 6.4.1 Formal and functional properties in the characterization of complements Complements are formal realisations of valency slots. One important component in the description of complements concerns their formal realisation in terms of a particular type of phrase or clause. Thus we can say that all three complements of the following use of consider are noun phrases: (35) NW332 I consider this an underhand trick. However, this purely formal classification is not sufficient to distinguish the three complements from one another, which is why two further criteria will have to be used: namely their ability to occur as subject and their order of occurrence in the clause. Since the formal realisation of a complement slot in some cases differs between active and passive clauses, the following distinctions will be introduced: Complements that can occur as subjects of active clauses will be marked by a subscript of the type act-subj , complements that can occur as subjects of passive clauses will be marked by a subscript of the type pass-subj , 10 complements which can neither serve as a subject of an active nor of a passive clause will not receive any such subscript. It is obvious that complements that do not function as subjects occur in the predicate of the clause. If more than one complement can occur in the predicate, their order is very often governed by general principles such as that noun phrases precede particle phrases etc. Should the order of complements in the predicate not be clear, it can be indicated by subscripts such as 1 or 2 . Following these conventions, which are similar to those used in the Valency Dictionary of English (2004), the complements of consider can be described as follows to cover uses such as 11 (1) NW51 Robyn considered herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave. (35) NW332 I consider this an underhand trick. (36) NW58-59 Robyn continued to give much of her time and emotional energy to the Women's Group, but Charles did not object; after all, he considered himself a feminist too. (37) NW42 Though urged by the school to apply for a place at Oxbridge, she chose instead to go to Sussex University, as bright young people often did in the 1970s, because the new universities were considered exciting and innovatory places to study at. 10 This kind of complement corresponds to what in traditional grammar is often referred to as an object (see 8.2). 11 For a complete valency description of this use of consider, which is rather more complex, see VDE. <?page no="130"?> Valency 118 valency slot (35) (1) complements I I Robyn [NP] act-subj [by_NP] noun phrase that functions as subject in an active clause by-phrase in passive clauses consider considered II this herself [NP] pass-subj noun phrase that occurs as subject of a passive clause or otherwise in the predicate in active clauses III an underhand trick lucky to get... [NP] [AdjP] noun phrase in the predicate adjective phrase in the predicate 6.4.2 Types of complement 6.4.2.1 Phrases and clauses Valency complements can be described as formal units. In some cases these units can be described in terms of particular phrases. Thus the complements underlined in (25) are noun phrases, (38) contains an example of an adjective phrase complement and (39) an example of an adverb phrase complement: [NP] (25) NW218 They don't pay us all that much, anyway. [AdjP] (38) NW218 But doesn't that make us rather marginal? [AdvP] 12 (39) VDE His favourite wood is pine, because it’s soft, perfect for carving and cuts easily. In other cases, complements are represented by clauses without subjects such as [INF] (40) NW214 No other reason would make anyone come here, or having come, stay. [V-ing] (41) NW314 'You shit,' Robyn said aloud, when she had finished reading the letter. or by clauses containing subjects: [CL fin ] (42) SW7 I suppose everyone says that to you. [NP_V-ing] (43) NW60 You would not think her unduly burdened with worries, if you watched her crossing the campus, smiling at people she knows, her eyes bright, her brow unfurrowed. 12 This category is referred to as [ADV: QUALITY] in VDE. <?page no="131"?> Formal realisations of complements 119 [NP_V-ed] (44) NW325 If by any chance there should be a job advertised in the next twelve months ... [NP_INF] (45) NW149 After a moment or two she heard him call, "Turn the ignition key," and when she did so, the engine fired. [wh_CL] (46) SW23 I never could remember which came first, the morphemes or the phonemes. 6.4.2.2 Particle complements Some types of complement are introduced by a particle. As pointed out in 3.5.2 and 6.8.2, particles themselves can be considered valency carriers of a kind in that they differ in their complementation possibilities. Some particles can have [NP]-complements, others [AdvP]-complements or [V-ing]-complements etc. Thus one could argue that complementation with particles involves two choices: first a valency carrier such as a verb can provide for a particle-complement and then the particle opens a valency slot for a particular complement of the particle. It seems, however, that, in very many cases, the valency of the verb not only determines the choice of particle but also influences the range of possible choices concerning the valency potential of the particle. It is for this reason that particle complements (or, more precisely, particle phrase complements) of verbs, nouns, adjectives and adverbs will be specified as combinations of a particle and the possible complements when used with that particular verb, noun, adjective or adverb. Thus in (47) NW331 I wrote to you about it. the verb write is used with two particle complements, which can be described as [to_ NP] and [about_NP]. Similarly, (48) NW60 After all, she worked on the nineteenth-century industrial novel for something like ten years ... represents a particle complement [on_NP]. This complement consists of the particle on and a noun phrase which is the complement of the particle. Similarly, on replacing signalling cables damaged in Saturday’s fire in (49) VDE Birmingham New Street station was closed for 10 hours yesterday to allow engineers to work on replacing signalling cables damaged in Saturday’s fire. can be analysed as representing a particle complement of the type [on_V-ing]. 13 In some cases, the choice of construction that can follow a particular particle is extremely limited: (50) NW328 Haven't you noticed that in the modern world good news comes by telephone and bad news by mail? (51) NW326 Professor Zapp has kindly offered to look at my work-in-progress. 13 In terms of a simplification of the description, these can be subsumed under the label [on_N/ Ving]. <?page no="132"?> Valency 120 In such cases - and this is perhaps especially true of [to_INF]-complements - it is psychologically plausible to analyse the particle complement as a unified whole, i.e. a single choice in the sense of Sinclair (2004a). 6.4.2.3 Particle complements or prepositional verbs? 14 It should be pointed out that the analysis of sentences such as (52) SW28 Be honest, have you ever been to a lecture or seminar at the end of which you could have found two people present who could agree on the simplest précis of what had been said? presents one of the major points of difference between the valency approach and the account of verb complementation as it is given in CGEL (1985: 16.2-10). CGEL gives priority to the combination of verb and (in their terminology) preposition and introduces the category prepositional verb and postulates units such as agree on, explain to or write on. This has a certain attraction with respect to sentences such as (53) NW218 Whose side are you on, Robyn? (54) NW217 ... things like that, which we argue about and read about and write about endlessly... However, it seems rather awkward to treat agree in (52) differently from (55) NW95 But, like Penny, he was surprised Robyn had agreed to be nominated as Arts Faculty Shadow. or to analyse the uses of write in (56) NW361 I'll write you a glowing reference … (47) NW331 I wrote to you about it. (57) NW82 At the very moment when they were writing about these problems, Marx and Engels were writing the seminal texts in which the political solutions were expounded. as representing three different lexemes: write (in 56 and second occurrence in 57), write about (first use in 57) and write about to (in 47). If the prepositional verb analysis is applied consistently in dictionaries, it results in agree on or decide on being given separate entries set apart from the verbs agree and decide, which is not particularly helpful either. prepositional-verb-analysis valency analysis verb complement verb complement agree to be nominated ... to be nominated ... agree on the simplest précis of what had been said agree on the simplest précis of what had been said 14 For phrasal verbs see 3.8.3.2. <?page no="133"?> Formal realisations of complements 121 Furthermore, the prepositional verb analysis should find a parallel in prepositional adjectives or prepositional nouns such as 15 (58) NW240 The universities are open to everyone. (59) NW59 Indeed, her uneasiness about Charles scarcely counts as a conscious worry at all; while the worry about the lecture is, she is well aware, a trivial and mechanical one. However, such concepts are never used in the analysis: not even those learner's dictionaries which treat prepositional verbs as separate entries would treat open to or uneasiness about as different lexemes from open or uneasiness. It is for these reasons that we consider an analysis of sentences such as (52) in terms of particle complements far more appropriate, where a particle complement is defined as a complement where a specific particle and its possible complement is specified by the valency of the governing word. 6.4.2.4 Unspecified particle phrase complements It is important to realise that particle complements as defined above are not the only complements that can be realised by particle phrases. Thus in Cambridge in (31) NW48 Charles continued to live with her in Cambridge, commuting by car to Ipswich to teach his classes, and staying there for a night or two each week. is also a particle phrase, but the particle in is not part of the valency specification of the verb live since the same valency slot can also be realised by particle phrases with other particles such as (33) NW47 ... and it crosses Robyn's mind, not for the first time, that it is a pity she lives so far from the metropolis where such exciting events are always happening. (60) NW64 I just wanted to say that, we would of course be very sorry to lose you, you've been a tremendous asset to the Department, even in the short time you've been here. It seems appropriate to distinguish this type of valency complement from the particle complements identified above (in which a specific particle is required). We will use the label [PartP] to refer to this category of unspecified particle phrases, which allows a number of formal realisations. Such particle phrases are attributed complement status, however, because they fill an obligatory valency slot of the verb. In other respects, they resemble adjuncts since their form is not determined by the governing verb. In fact, the range of formal realisations of these complements is the same as when the same semantic role (> 6.5) is realised in an adjunct unit. 16 15 See CGEL (1985: 16.68 note). For a discussion of the prepositional verb analysis versus the particle complement analysis see VDE (2004: xxvi), Herbst (1999), Götz/ Herbst (1989). Compare also Emons (1974, 1978). Since these uses (overwhelmingly) correspond to traditional prepositional phrases, the term prepositional complement is also justified. 16 The category of unspecified particle phrase is also required to account for obligatory complements of verbs such as behave. Compare: Both sides behaved as if they were sounding each other out. VDE <?page no="134"?> Valency 122 6.4.2.5 Further complements There are a number of complement types which cannot easily be described in terms of a particular type of phrase or clause but which can take a number of different forms. For instance, say has an obligatory valency slot which can be filled by either a sentence or a single word: (61) NW62 'I couldn't afford the time, even if I had the money,' says Robyn. (62) NW62 'Agreed,' says Philip Swallow. This type of complement will be referred to as [QUOTE]. 17 In VDE, [QUOTE] is distinguished from the category [SENTENCE] which covers cases such as 18 (63) NW62 I was one of them, I don't mind admitting … Here, the clause I was one of them can syntactically be analysed as a complement of admit. In such cases, the remaining clause which contains the governing element fulfils a function similar to an adjunct so that one could argue that such constructions border on the area of phraseology. This is particularly apparent in (64) SW6 That was ten years ago, Darlington was small in those days, still is, I suppose, but it was a challenge, and the students are quite good, you'd be surprised. The label [DESCRIPTION] can be used to describe another category rather similar to [QUOTE], but which need not be marked off by quotation marks: 19 (65) NW52 Entitled The Industrious Muse: Narrativity and Contradiction in the Industrial Novel (the title was foisted on her by the publishers, the subtitle was her own) it received enthusiastic if sparse reviews, and the publishers commissioned another book provisionally entitled Domestic Angels and Unfortunate Females: Woman as Sign and Commodity in Victorian Fiction. 6.4.2.6 Pro-forms A construction which is rather similar to [SENTENCE] involves the use of a pro-form as in cases such as: (66) NW351 Language is a system of differences, as Saussure said. (67) NW352 One upon a time, as Rupert remembers, there was a single syllabus, essentially a survey course on Eng.Lit. From Beowulf to Virginia Woolf, which all the students followed in common, through lectures and a weekly tutorial, and life was very simple and comfortable, if a little dull. Here, as can be seen as a pro-form referring to the clauses which are separated off from the clauses containing the verbs say or remember. where the same valency slot can be realized by [AdvP] as in But if they behave badly it is their problem, not mine. VDE . Compare the category [ADV] in VDE. 17 Compare the annotation "Quotation" in FrameNet, see Fillmore (2007: 150). 18 Such cases as (61) and (62) cannot simply be treated as variations of a pattern with a [that_CL]-complement since not all verbs allowing [that_CL]-complements occur in this construction (cf. accept). 19 [DESCRIPTION] allows realizations other than noun phrases: A column entitled "Bulletin" informs her that ... NW47 <?page no="135"?> Formal realisations of complements 123 Other pro-forms that can function as complements are so and not: (68) NW217 I suppose so. 6.4.2.7 List of complements On the basis of these considerations, the following labels can be used to describe valency complements in English. (Specifications as to the ability of a complement to occur as the subject of an active or a passive clause and to the position have to be added where appropriate): 20 [NP] (25) NW218 They don't pay us all that much, anyway. [AdjP] (38) NW218 But doesn't that make us rather marginal? [AdvP] 21 (39) VDE His favourite wood is pine, because it’s soft, perfect for carving and cuts easily. [PartP] 22 (60) NW64 (69) NW45 ... you've been a tremendous asset to the Department, even in the short time you've been here. Go to Oxbridge. [to_NP] etc. 23 (60) NW64 (70) NW307 ... you've been a tremendous asset to the Department, even in the short time you've been here. Whatever happened to the spirit of the Robbins Report? [V-ing] (41) NW314 'You shit,' Robyn said aloud, when she had finished reading the letter. [INF] (40) NW214 No other reason would make anyone come here, or having come, stay. [to_INF] (51) NW326 Professor Zapp has kindly offered to look at my work-in-progress. [CL fin ] (42) SW7 I suppose everyone says that to you. [that_CL] 24 (50) NW328 Haven't you noticed that in the modern world good news comes by telephone and bad news by mail? [NP_V-ing] (43) NW60 You would not think her unduly burdened with worries, if you watched her crossing the campus, smiling at people she knows, her eyes bright, her brow unfurrowed. 20 Similarly, if a category only applies to very specific lexical items, this can be indicated accordingly: [if_CL], [whether_CL] etc. For further categories such as [ORDINAL], [NUM] or [SCORE] (as in The Welshmen lost three-nil... VDE ). See VDE (2004: xvii-xx). 21 This category is referred to as [ADV: QUALITY] in VDE. 22 For a definition of this category see 6.4.2.4. 23 [Part P] can be used for a particle phrase complement where the particle is not dependent on the governing word, [to_NP], [on_NP], [with_NP], etc. is the notation used for particle phrases where the valency carrier demands a certain particle (see 6.4.2.2 and 6.4.2.4). 24 The label [(that_) CL] is used to express the fact that a complement can either be realized by a that-clause (which is a particle phrase) or by a finite clause that is not introduced by that. <?page no="136"?> Valency 124 [NP_V-ed] (44) NW325 If by any chance there should be a job advertised in the next twelve months ... [NP_to_INF] 25 (46) NW60 What the students want her to do is to give them some basic facts ... [NP_INF] (45) NW149 After a moment or two she heard him call, "Turn the ignition key," and when she did so, the engine fired. [wh_to_INF] 26 (71) NW61 She is always uncertain how to address her Head of Department. [wh_CL] 27 (46) SW23 I never could remember which came first, the morphemes or the phonemes. [QUOTE] (61) NW62 'I couldn't afford the time, even if I had the money,' says Robyn. [SENTENCE] (63) NW62 I was one of them, I don't mind admitting … 6.4.2.8 Discontinuous realisations of complements Complements that consist of a clause with a subject can have a discontinuous realisation in the passive: 28 (72) VDE Even yesterday she was heard to dismiss the whole exercise bitterly as state gambling. This is by no means unusual since discontinuous realisation can also occur with postmodifiers as in (73) NW72 In the 1840s and 1850s," says Robyn, "a number of novels were published in England which have a certain family resemblance. Cases which require or allow what is generally described as extraposition can also be seen as discontinuous realisations of complements. Thus in (74) NW58 It was impossible to commute from Rummidge to Ipswich or vice versa. (75) NW113 She did not think it necessary to tell him that the University administration had been obliged to circulate all members of staff with a memorandum asking them, if they had been on strike, to volunteer the information (since there was no other way of finding out) so that their pay could be docked. it and [to_INF] can be seen as realising the same valency slot and, as such, we treat them as discontinuous complements. 29 25 This complement type represents a particle phrase that has an infinitive clause with subject as the complement of the particle. 26 This label covers (a) particle phrases introduced by to with an infinitive clause complement which has a wh-element as an adjunct unit or a complement unit in initial position and (b) particle phrases with the head whether followed by a to-infinitive complement. See 5.3.6. 27 [wh_CL] comprises both clauses with an initial wh-element which functions as a complement unit or an adjunct unit within that clause or a clause introduced by a wh − particle such as whether; see 5.3.6. Strictly speaking, these could be distinguished as [wh-CL] and [wh_CL]. 28 In (72), she + to_INF is one complement which corresponds to Someone heard her dismiss the whole exercise bitterly. <?page no="137"?> Formal realisations of complements 125 6.4.3 One complement or two? It is obvious that the analysis of valency structures poses a number of problems for which there is no obvious solution. Thus one could question the wisdom of a trivalent analysis of (1) NW51 Robyn considered herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave. (36) NW58-59 Robyn continued to give much of her time and emotional energy to the Women's Group, but Charles did not object; after all, he considered himself a feminist too. (43) NW60 You would not think her unduly burdened with worries, if you watched her crossing the campus, smiling at people she knows, her eyes bright, her brow unfurrowed. In view of the fact that her crossing the campus was analysed as an [NP_V-ing]-complement in 6.4.2.1, one could consider analysing the uses of consider in (1) and (36) and of think in (43) as divalent (see 6.6). This would involve identifying quasi-clause complements of the following types [NP_AdjP] herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave [NP_NP] himself a feminist However, there is an important difference between watch and consider with respect to question forms: a question form such as (43a) What are you watching? seems acceptable (if marginally so) in cases such as (43) or (76) NW125 Robyn peered through a Perspex window and watched things moving round and going in and out in sudden spasms, lubricated by spurts of a liquid that looked like milky coffee. (77) VDE Today you can watch the catch come in (approx 7: 30-8 am) at Hayle, St Ives, Newlyn and Mousehole, as fishing boats bring back dover sole and haddock. A question of the type (1a) What did Robyn consider? can be ruled out for cases such as (1) or (36), but be appropriate for (2) NW384 But I was wondering whether, in the circumstances, you would consider staying on next year, and see what happens. On the other hand, such a question might be possible in the case of think as in (43), which shows that structural tests of this kind can only be used sensibly up to a certain point. Both in the case of the divalent analysis of the watch-construction and the 29 For a very similar analysis in FrameNet see Fillmore (2007: 146). VDE introduces the category [it + pattern of N/ ADJ] for the respective valency slot of verbs such as consider to make clear that the type of clausal complement that can occur depends on the valency of the noun or adjective. <?page no="138"?> Valency 126 trivalent analysis of the uses of consider and think arguments for an alternative analysis can be found. Even if in terms of a valency pattern analysis (> 6.7) this is perhaps less central, it has to be pointed out that in some cases patterns which have the same formal structure have different semantic interpretations: thus it is probably appropriate to analyse examples such as (78) BNC Some say it is good for the Prime Minister to mix with the crowd and others that for a Prime Minister it is less than dignified. (79) VDE "In principle, it would not be good for the national military to get involved," he said. in terms of a [for_ NP_ to_ INF]-complement, whereas in cases such as (80) VDE I think it was good for her to be able to talk it all out. an analysis with [for_NP] and [to_INF] as two separate complements seems more appropriate. The latter construction can be interpreted as having a for-complement as a kind of ' BENEFICIARY ' of the action expressed by the infinitive complement, whereas the former construction contains a judgement about the action described by the [for_ NP_to_INF]-complement. Although the two are clearly distinct, in many cases it will not be possible to decide whether the one or the other reading is intended. 6.5 Semantic valency: participants 30 6.5.1 Participant roles Complements have been defined as syntactic constituents that are governed by a valency carrier such as a verb, adjective or noun. The semantic function of a complement with respect to the valency carrier (or in the clause) can be described in terms of participants or participant roles. The idea of semantic roles was developed by linguists such as Halliday (1967/ 1968, 1970) and Fillmore (1968). In his famous and influential model of case grammar, Fillmore introduced the concept of semantic cases to account for the relationship of sentences such as: (81) QE John ('AGENTIVE') opened the door ('OBJECTIVE'). (82) QE The door ('OBJECTIVE') was opened by John ('AGENTIVE'). (83) QE The key ('INSTRUMENTAL') opened the door ('OBJECTIVE'). 30 Halliday (1970: 146-147) makes a distinction between "the process itself", "participant functions" and "circumstantial functions". Since the distinction between verb-specific and non-verb-specific elements is reflected in the distinction between complements and adjuncts anyway and because there is a considerable gradient between what one considers to be central to a particular "process" and what is not, we will use the term participant in a more neutral sense so that at the level of semantics, adjuncts will also be seen as representing participants of the clause (which are not determined by the valency of the verb). The term argument is also often used in this context to refer to valency-bound participants or to what we call complements. <?page no="139"?> Semantic valency: participants 127 (84) QE John ('AGENTIVE') opened the door ('OBJECTIVE') with the key ('INSTRU- MENTAL'). A model of this kind is very attractive for a number of reasons: it shows that different complements - such as John in (81) and (84) and by John in (82) or the key and with the key in (83) and (84) - represent the same participant (which is why they can be subsumed under one valency slot in our model) semantic cases give an indication of the semantic function of the participants in the sentences in which they occur. Although Fillmore's approach was taken over by many schools of linguistics ranging from valency theory to generative grammar (under different labels such as semantic roles or theta-roles), 31 it soon became clear that the original idea of establishing a very small set of possibly universal semantic roles would be difficult to put into practice. Since it is far less problematic to describe the roles of individual verb uses, an alternative is to make use of relatively specific participant roles in the description of valency complements. 32 These can be developed in a relatively ad-hoc fashion since their only purpose is to describe the semantic function of the complements in valency structures. Thus for write in (85) SW24 I wrote five books on Jane Austen, every one of which was trying to establish what her novels meant and, naturally, to prove that no one had understood what they meant before. (86) NW78 It is interesting how many of the industrial novels were written by women. the participant roles " WRITER " and " THING WRITTEN " can be identified. 33 The disadvantage of such very specific roles is, of course, that apart from distinguishing the different participants of one valency carrier from one another they hardly fulfil any other purpose. In particular, they do not systematically allow any generalizations that would account for the fact that complements of different verbs may have the same or a very similar semantic potential. In some cases at least, such parallels undoubtedly exist: thus the [on_NP]-complement on Jane Austen in (85) can 31 See, for example, Dirven/ Radden (1977) or Helbig (1992). See also Haegeman's (1991: 41-2) discussion of theta-roles (the latter without any reference to Fillmore). 32 For problems of the case grammar model see Palmer (1981: 146-149). Compare also Fillmore's FrameNet approach (2007: 131): "With respect to the naming of frame elements, we learned early that for many of the complex frames there is no non-arbitrary way of fitting them into the standard sets of case roles or thematic roles in recent literature." Compare also Götz-Votteler (2007). See Herbst (1983) for an attempt to describe the semantic functions of adjective valency. 33 To give a rough indication of how item-specific the roles indicated are we shall use double inverted commas for more specific and single inverted commas for more general roles. Compare also the role labels used by Lamb (2001: 178-182) and Stefanowitsch and Gries (2003: 213). <?page no="140"?> Valency 128 be described by a participant role ' TOPIC ' 34 , which at a certain level of discreteness 35 , can also be used for the [about_NP]-complement as in (47) NW77 I wrote to you about it. Similarly, the labels introduced for the participant roles of (1) NW51 Robyn considered herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave. - " JUDGE ", " THING JUDGED " and " JUDGEMENT " - can equally be applied to verbs such as regard, judge or think: 36 (87) SW27 ... and I suppose that most of us in this room must be regarded as compulsive readers ... (88) VDE Any success it might have runs the risk of being judged a merely formal solution. (75) NW113 She did not think it necessary to tell him that the University administration had been obliged to circulate all members of staff with a memorandum asking them, if they had been on strike, to volunteer the information (since there was no other way of finding out) so that their pay could be docked. This also applies to cases such as buy, purchase or sell, which can be used to describe the same action from different perspectives (> 7.2.2.2) and, which, quite obviously, involve the same participants of the action: (89) NW53 It was formerly her parents' second car, sold to Robyn at a bargain price when her mother replaced it. (90) NW148 I didn't buy it, my parents gave it to me, when they changed it." (91) NW41 Her alarm clock, a replica of an old-fashioned instrument purchased from Habitat, with an analogue dial and a little brass bell on the top, rouses her from a deep sleep at 7.30. One way of accounting for such relations was developed by Fillmore in terms of a frame-semantic approach by identifying so-called frame elements (Fillmore 2007). In FrameNet (framenet.icsi.berkeley.edu), which is a large-scale online lexicographical project, labels such as Buyer, Goods, Seller, and Money are used to describe frame elements that occur in both the Commerce__buy frame and the Commerce__sell frame. 37 Using the same descriptive categories such as "goods" or "money" in the 34 This is true irrespective of whether one considers on Jane Austen to be a complement of write or of books. 35 One could also subsume uses of about under a more general role ' REFERENCE ' (Herbst 1983) or introduce sub-roles (Schüller forthcoming). 36 The complement marked by double underlining is a discontinuous constituent with it as an extraposed element which refers to the clause that follows the second complement in the predicate. This element has the participant role " THING JUDGED ", necessary has the role of " JUDGEMENT ". 37 Buyer and Goods are core frame elements in both frames, Seller also in the Commerce__sell frame; other frame elements are classified as non-core. It is interesting to see that both the labels used in FrameNet and in VDE do not always represent roles but rather descriptions of the participants involved in the action. <?page no="141"?> Semantic valency: participants 129 description of different verbs is obviously desirable because it contributes to the recognition of semantic relations across individual verbs. 38 In a similar, though perhaps less systematic way, this is expressed by the semantic information given in the notes of VDE: 39 buy Generally buy means 'to acquire something in exchange for money'. (i) A person I can buy something II (a) from a shop VIII (b) from or off a person VIII selling it to them (c) by a particular method V , e.g. by mail order or by the kilo etc. (d) with something that one has IX , esp. money (e) for (1) another person, (2) a special occasion such as Christmas, someone‘s birthday, (3) a particular purpose III (f) for or at a certain price IV . ... sell The basic meaning of sell is 'hand over in return for money'. (i) A person, shop, company, etc. I can sell goods II to somebody III for or at a certain price IV . ... The main function of the notes in VDE is to characterize the semantic function of the complements and to indicate which lexical items can be used in particular complement positions. At this level of specificity, as demonstrated in the descriptions provided by the VDE, it often seems appropriate - both for reasons of pedagogic presentation and linguistic accuracy - to give open or closed lists of collocates or rather specific labels such as "goods" or "money". 40 In other cases and for other purposes of description, broader generalizations seem possible and desirable: 41 thus it is worth considering whether the participant roles expressed by the complements with double underlining in the following examples could be subsumed under one common semantic role. (92) NW323 Are you going to get this person a drink, Philip? 38 Compare Fillmore/ Johnson/ Petruck (2003: 237): "In FrameNet, the semantic valence properties of a word are expressed in terms of the kinds of entities that can participate in frames of the type evoked by the word. We call these roles frame elements (FEs)." Since it can be seen as one of the main purposes of the FrameNet descriptions to point out how particular facts of the world can be verbalized, frame elements might be seen as "participants" of certain extralinguistic actions or facts rather than as categories that arise from the description of the valency of particular verbs. 39 The Roman numerals refer to the participants identified in the complement inventory of the entries. 40 See VDE (2004: xxxvii-xxxviii). 41 More general semantic roles can be used for instance to describe semantic functions of constructions. Compare, for instance, Goldberg (2006: 21): "... the verb and its own arguments are integrated (unified, "fused") with the predicate and arguments of the construction. ... the argument role of the construction must fuse with an independently existing participant role of the verb." For these more general roles Goldberg (2006: 5) uses labels such as "agent", "recipient" and "theme", but equally argues "no assumption is made that these thematic role labels are drawn from a universal or limited set." <?page no="142"?> Valency 130 (93) NW45 She lights the gas stove, and makes herself a breakfast of muesli, wholemeal toast and decaffeinated coffee. (94) NW45 He ... always had a cup of cocoa or a bowl of soup ready for her when she returned home, tired but invariably triumphant. (95) NW44 She organized an all-night telephone counselling service for students in despair about their grades or love-lives. (96) NW75 And in any case his degree was good enough to earn him, as Robyn's did for her, a Major State Studentship to do postgraduate research. In all these cases, the complement in question refers to a person who benefits from the action described in some way or other. There is thus no need to use a more specific role label, but the participant role can be described as ' BENEFICIARY ' in these cases. A very similar role is expressed in the following examples, where the action is directed towards a person and which therefore will be labelled ' RECIPIENT ': (97) NW325 Send me your manuscript and I'll have a look at it. (46) NW60 What the students want her to do is to give them some basic facts ... Both the role ' BENEFICIARY ' and the role ' RECIPIENT ' can be expressed by noun phrase complements, but while ' BENEFICIARY ' is also often expressed by [for_NP]-complements, ' RECIPIENT ' can be expressed by [to_NP]-complements: (98) NW355 The point is, it means Charles was wrong to say that we shouldn't teach theory to students who haven't read anything. (90) NW148 I didn't buy it, my parents gave it to me, when they changed it. How problematic such general semantic roles are is shown by examples such as (36) and (99), where the [to_NP]-complements do not represent a human or animate recipient, which - depending on the definition of the category - can be taken as a reason for not subsuming it under ' RECIPIENT ': (36) NW58-59 Robyn continued to give much of her time and emotional energy to the Women's Group ... (99) NW41 This morning, she gives priority to the fact that it is the first day of the winter term ... A further complication can be found in the case of a verb such as send, where a construction with a to-phrase can be interpreted as either a realisation of a [to_NP ' RECIPIENT ']-complement as in (100) NW382 I want to send a Gorillagram to somebody in London … or as a realisation of a [PartP]-complement, which could be labelled ' ADDRESSEE ' or ' LOCATIVE : GOAL '. The latter allows a different range of lexical items from ' RECIPIENT ': 42 42 Compare Goldberg's (2006: 26, 33-34) discussion of send. <?page no="143"?> Semantic valency: participants 131 (101) NW45-6 She thinks about the structure of Disraeli's Sybil: or, the Two Nations (1845), until the sound of the Guardian dropping on to the doormat sends her scurrying to the front door. (102) BNC He was a prosperous farmer who had sent his son to Balliol ... (103) NW21 They're just sitting around in their warm staff-rooms, chewing the fat, while the kids are sent home to get into mischief. 6.5.2 Selected semantic roles As the discussion in the previous section has shown, developing a model of semantic roles is like steering a course between Scylla and Charybdis: either one designs roles that are very specific and therefore serve only a very limited purpose or one makes use of more general semantic roles, which inevitably results in considerable amount of overlap between categories and gradience in the analysis. Nevertheless, such characterizations of participants 43 are useful since they give at least an indication of the meaning expressed by particular complements. It is for this reason that we would like to introduce a short list of roles which can be used for the characterization of participants or adjuncts without too many problems: 44 ' BENEFICIARY ': someone or something that benefits from the action expressed by the verb: (56) NW361 I'll write you a glowing reference … ' RECIPIENT ': someone or something who is the intended receiver of the action described by the verb (90) NW148 I didn't buy it, my parents gave it to me, when they changed it. ' EFFECTED ': something that is the outcome or result of the action described by the verb (56) NW361 I'll write you a glowing reference … (86) NW78 It is interesting how many of the industrial novels were written by women. ' AFFECTED ': something that is affected by the action described by the verb 43 Perhaps a term such as participant characterization would be more appropriate since the categories used in FrameNet or VDE cannot always be called roles in any strict sense of the word. One could argue, however, that the labels used for example in VDE serve as indicators for the role of the participant in the action referred to. For a discussion of the character of roles see Tarvainen (1985: 14). 44 This list is not to be seen as a comprehensive list in the sense that it would cover all possible cases but rather as a list of role labels that can usefully be applied in many of the more prototypical cases. On the whole, we would agree with CamG (2002: 228) in this respect: "It is by no means clear that (as assumed in some modern work) a small number of general roles can be established (perhaps in the order of a dozen or so) such that all arguments can be assigned to one or other of these roles, with no two arguments in the same clause having the same role." For a more comprehensive account than the one given in this section see, for instance, the lists of semantic roles given in CGEL (1985: 8.2, 10.5-8 and 10.18-32) or CamG (2002: 228-233). <?page no="144"?> Valency 132 (3) NW324 I haven't read a book in months, let alone tried to write one. In cases where there is no contrast between ' BENEFICIARY ' and ' RECIPIENT ', a more general role ' BENREC ' can be used, for instance for cases such as lend as in (104) NW52 The three years stretching ahead seemed like a long time, then, worth buying a little house in Rummidge for (Robyn's father lent her the money for the deposit) rather than paying rent. Similarly, ' EFFECTED ' and ' AFFECTED ' can be subsumed under the combined role of ' ÆFFECTED '. ' PREDICATIVE ': A role ' PREDICATIVE ' will be introduced to cover cases where the role of a complement is that of identification, equation, characterization or description (which could be seen as subroles of ' PREDICATIVE '): (105) NW65 Being Dean, you see, I'm very aware of the financial constraints on the University. (66) NW351 Language is a system of differences, as Saussure said. (106) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. (107) SW27 I wouldn't call myself a structuralist. The labels given so far can be applied without too much difficulty to a large number of predicate complement units, especially but not only to those realised by noun phrases. It is obvious that there is a certain amount of overlap between these categories since the ' RECIPIENT ' or a ' BENEFICIARY ' of an action can normally also be considered to be ' AFFECTED ' by that action so that the label ' AFFECTED ' must be understood as meaning 'affected' in a way other than being ' BENEFICIARY ' or ' RECIPIENT '. Up to a point, syntactic tests may help support one's intuitions about role assignment: thus the fact that ' BENEFICIARY ' can be expressed by [for_NP]-complements and [NP]-complements and ' RECIPIENT ' by [to_NP]-complements and [NP]-complements may provide an indication for classifying a particular [NP] as either ' BENEFICIARY ' or ' RECIPIENT '. In the case of the label ' ÆFFECTED ', it seems that passivization can serve as an indicator in the sense that if this test produces acceptable sentences, the label ' ÆFFECTED ' can be applied: (90) NW148 I didn't buy it ... (90a) ... it was bought by my parents ... Nevertheless, the labels ' BENREC ', ' ÆFFECTED ' and ' PREDICATIVE ' will not suffice to characterize all predicate complement units. A label that can also be used for subject complement units is: ' AGENT ': someone that carries out an action (cf. CamG 2002: 230): 45 45 The label ' AGENT ' is not entirely unproblematic. If ' AGENT ' is defined in terms of a conscious and intended action, it does not cover cases such as An extremely public row about the denial of tenure to a young lecturer associated with the progressive party opened old wounds and inflicted new ones on this <?page no="145"?> Semantic valency: participants 133 (13) NW328 Great, I'll read it as soon as I can. (92) NW323 Are you going to get this person a drink, Philip? (86) NW78 It is interesting how many of the industrial novels were written by women. Further roles that can be used to characterize complements and adjuncts without too many problems include: ' LOCATIVE : STATIVE ': (31) NW48 Charles continued to live with her in Cambridge .... ' LOCATIVE : GOAL ': (21) NW46 So Robyn and Charles went to Cambridge to do their PhDs. (108) NW307 Every autumn there's this absurd migration of well-heeled youth going from Norwich to Brighton or from Brighton to York. ' LOCATIVE : SOURCE ': (108) NW307 Every autumn there's this absurd migration of well-heeled youth going from Norwich to Brighton or from Brighton to York. ' TIME ': (108) NW307 Every autumn there's this absurd migration of well-heeled youth going from Norwich to Brighton or from Brighton to York. ' PURPOSE ': (21) NW46 So Robyn and Charles went to Cambridge to do their PhDs. ' TOPIC ': (109) SW0 There are conferences on almost everything these days ... While some of these roles may be relatively unproblematic, others can only be seen as prototypical categories which may be useful in the analysis for certain purposes, whereas for other purposes more specific participant roles (or participant characterizations) are more appropriate. It must be emphasized that the fact that we have listed a number of more general roles in this section should not in any way be taken to mean that all occurrences of complements or adjuncts could be described in these terms. Since there are considerable differences between the categories developed in different approaches, there seems little point in outlining yet another system in this always thin-skinned community. NW48 . CamG (2002: 230-1) treats 'agent' as a "subtype of causer" to distinguish between My uncle returned ('agent') and My headache returned ('causer') or I coughed to remind them I was waiting ('agent') and I coughed [involuntarily] ('non-agent'). CamG (2002: 231-2) further distinguishes between 'theme' as an "entity that moves or is located" and 'experiencer' as "the one who feels or perceives", where it is slightly unclear why She fell off the balcony and She was in a happy frame of mind should be 'theme' together with her in It made her angry, whereas he in He hates me is classified as 'experiencer'. We will not make any attempt here to introduce a general label for subject complement units that are not ' AGENT ' but would rather consider this a case for a more specific characterization of participants. In combination with the role ' PREDICATIVE ' many subject complement units could be described as providing a point of reference within the sentence, which is a description that might be extended to other verb meanings. <?page no="146"?> Valency 134 book. What is important is that the attractions and limitations of such labelling systems for participants are understood and that in any analysis or description a framework is used that is appropriate to the overall purpose of that analysis or description. In this book, we will thus only occasionally refer to participant roles at all, making use of more general and more specific labels as it seems appropriate. 6.5.3 Participants and complements 6.5.3.1 Participant and complement sharing Valency slots have been characterized (> 6.1) in terms of complements (at the formal level) and in terms of participants (at the level of semantics). In very many cases, participants and complements coincide. Thus in (110) NW45 He kept the flat tidy while Robyn was out doing good works, and always had a cup of cocoa or a bowl of soup ready for her when she returned home, tired but invariably triumphant. there are two complements, which correspond to two participants whose roles can be described as ' AGENT ' and ' LOCATIVE : GOAL '. However, in (111) NW53 Returning to the kitchen, Robyn turns down the thermostat of the central heating and checks that the back door of the house is locked and bolted. (112) NW42 Instead of returning to Australia, he took a post at a university on the south coast of England, where he has been ever since, now occupying a personal Chair. the verb return is used with only one complement, realising the role ' LOCATIVE : GOAL '. Nevertheless the other participant can also be identified, Robyn in (111) and he in (112). The same is true in the case of constructions such as (37) NW42 Though urged by the school to apply for a place at Oxbridge, she chose instead to go to Sussex University, as bright young people often did in the 1970s, because the new universities were considered exciting and innovatory places to study at. where she, which is a complement of chose can also be analysed as representing not only a participant of chose, but also a participant of urged. The same kind of analysis can be applied to the use of nominate in (55) NW95 But, like Penny, he was surprised Robyn had agreed to be nominated as Arts Faculty Shadow. where Robyn syntactically is a complement of agree, but semantically represents a participant both of agree and of nominate. Such cases can be described as participant sharing. Note that in cases of coordination it may be more appropriate to speak of complement sharing: thus in (113) SW1 Today's conferees have an additional advantage over the pilgrims of old in that their expenses are usually paid, or at least subsidised, by an institution to which they belong, be it a government department, a commercial firm, or, most commonly perhaps, a university. <?page no="147"?> Semantic valency: participants 135 both the [NP]-complement their expenses and the [by_NP]-complement can be analysed as complements (and participants) of the verb pay as well as of the verb subsidise. In the case of particle phrase complements, this can also apply to the [NP]-complement of the particle, as in (114) NW39 Both are expressions of a secularized Protestant ethic, both dependent on the idea of an autonomous individual self who is responsible for and in control of his/ her own destiny, seeking happiness and fortune in competition with other autonomous selves. (54) NW217 ... things like that, which we argue about and read about and write about endlessly... 6.5.3.2 Participant mergers and impersonal uses The number of participants and the number of complements can also differ for a number of other reasons. Thus it is probably appropriate to identify two valency slots with verbs such as meet or kiss on the basis of uses such as: 46 (115) NW60 Outside the Department Office she meets Bob Busby, the Department's representative on the local committee of the Association of University Teachers, pinning a sheet of paper to the AUT noticeboard. Here, the verb meet has two complements, which seem to correspond to two participants. In (116) NW110 Haven't we met before? (117) SW4 More than customary aggravation was generated by the discovery that the conference would be sleeping in one building, eating in another, and meeting for lectures and discussions on the main campus ... one could however argue that the two participants are expressed by only one complement. Since this complement must either be plural as in (116) or semantically represent a group as in (117), it is probably appropriate to explain this in terms of a merger of participants, although it is not easy to decide on the participant role of this complement. 47 A related problem concerns the fact that some constructions with clause complements can be equivalent semantically to constructions involving two complements. For instance, we analysed the use of think in (75) NW113 She did not think it necessary to tell him that the University administration had been obliged to circulate all members of staff with a memorandum asking them, if they had been on strike, to volunteer the information (since there was no other way of finding out) so that their pay could be docked. 46 See Herbst/ Klotz (2002). 47 VDE specifies such complements as [N pl/ group ]. See Herbst/ Klotz (2002) for a more detailed discussion of these problems. See also the distinction between logical, semantic and syntactic valency discussed by Helbig (1992: 7-10). It may be appropriate to account for the different uses of verbs such as meet in terms of clausal roles (> 7.2.2.3). <?page no="148"?> Valency 136 as containing three complements with the participant roles " JUDGE ", " THING JUDGED " and " JUDGEMENT ". One could argue, however, that the [that_CL]-complement in (14) NW307 ... but don't you think, in retrospect, that the way Robbins was implemented was a mistake? ' contains both the participant roles of " THING JUDGED " and " JUDGEMENT ", which would then also be a case of a merger of participants. 48 In any case, these examples show that there may be differences between what one could term semantic valency and syntactic valency. A very interesting case in point is provided by so-called meteorological verbs such as rain, which typically occur in constructions such as (118) NW141 It's snowing again. where it certainly cannot be attributed any participant role but is generally analysed as an impersonal it. It seems appropriate to classify it as a complement, however, (a) because it is a clause constituent, and (b) because it contrasts with impersonal there. 49 (119) NW87 I know, and there isn't time. 6.6 Quantitative valency At a relatively superficial level, one can distinguish between different uses of a valency carrier in purely quantitative terms. At the syntactic level quantitative valency describes the number of complements realised in a particular clause. One can distinguish between [0] zerovalent uses (120) SW25 Reading, of course, is different from conversation. [1] monovalent uses (9) NW215 At home she read, ... [2] divalent uses (3) NW324 I haven't read a book in months ... (121) SW3 April is the cruellest month. [3] trivalent uses (99) NW41 This morning, she gives priority to the fact that it is the first day of the winter term ... 48 This could also be seen as an argument for analysing (75) as representing a divalent construction with it necessary to tell him ... as a complex clause complement. 49 See Emons (1974: 118-122) for a discussion of the complement status of it in such cases. Compare Fillmore's (2007: 145-6) treatment of there as a syntactic valent with no assigned semantic role. <?page no="149"?> Complement inventories or valency patterns 137 [4] quadrovalent uses (122) VDE The court could impose conditions that, for example, access be limited to certain hours or that you pay compensation to your neighbours for any loss of privacy or inconvenience suffered. On the basis of all possible uses of a verb (or a semantic unit), one can also classify verbs according to their maximum valency and distinguish accordingly between mono-, di-, triand quadrovalent verbs. 50 Similarly, statements about quantitative valency can also be made at the semantic level referring to the number of participants involved. Semantic and syntactic valency need not always coincide: for instance (116) NW110 Haven't we met before? represents a use of meet, which, for the reasons given above, can be analysed as semantically divalent but syntactically monovalent. 6.7 Complement inventories or valency patterns 6.7.1 Complement inventory On this basis, it is possible to draw up a complement inventory for a verb, adjective or noun by making the following specifications for each valency slot: ⌐ participant role ⌐ list of all possible complement types realising this valency slot ⌐ specification of degree of obligatoriness. Modifying the approach taken in VDE slightly, a complement inventory for the verb apply in the sense in which it is used in the examples given below can take the following form: I optional ' AGENT ' [NP] act-subj / [by_NP] II optional ' ÆFFECTED ' [to_INF] [for_NP] pass-subj III optional ' BENREC ' [to_NP] pass-subj 50 Statements about the maximum valency of verbs can be made without referring to a particular sentence type. In the case of adjectives and nouns, terminology is less clear. Thus one could argue that in The University is desperate to save on salaries NW64 not only to save on salaries should count as a complement: syntactically, desperate is a complement of is, but semantically it refers to the University. Generally, the fact that adjectives either require a verb (to function as complement of that verb) or a noun (to function as its modifier in a noun phrase) could be taken as an argument for providing a corresponding valency slot. Since this is mainly a terminological matter, we will treat such uses as monovalent. <?page no="150"?> Valency 138 (5) NW52 Thus a three-year lectureship in English Literature was advertised, Robyn I applied, was interviewed along with four other equally desperate and highly qualified candidates, and was appointed. (37) NW42 Though urged by the school to apply for a place at Oxbridge II , she chose instead to go to Sussex University, as bright young people often did in the 1970s, because the new universities were considered exciting and innovatory places to study at. (123) SW383 'You did say, back in January,' says Robyn, 'that if a job came up, I I should apply for it II . (124) VDE When I I applied to do those subjects II , I still wanted to become an architect. (125) VDE She I applied to a lot of universities III but this was the particular course she wanted. (126) VDE You I have to apply in writing to Dyfed County Council III for a permit II . • 6.7.2 Valency patterns and valency constructions 6.7.2.1 Valency patterns Complement inventories are a way of describing valency relations from the point of view of the valency carrier. Alternatively, one can describe the different uses of a verb (such as apply above) as occurrences in different types of valency pattern, where valency patterns are identified as sequences of complements and the valency carrier such as 51 [NP] act-subj + verb act (5) [NP] act-subj + verb act + [to_INF] (124) [NP] act-subj + verb act + [for_NP] (37) (123) [NP] act-subj + verb act + [to_NP] (125) [NP] act-subj + verb act + [to_NP] + [for_NP] (126) In the projected online valency pattern bank for English (www.patternbank.unierlangen.de), which is based on the Valency Dictionary of English (2004), valency patterns are specified with respect to the following elements: 52 the subject complement unit (which is a contextually optional constituent) (> 7.1.2.2), the valency carrier (with a specification as to active/ passive morphology), the predicate complement units that are part of the pattern. This means that a valency pattern such as 51 Compare the notion of Satzmuster as defined by Engel (1977: 180): "Die Struktur des Satzes wird zwar entscheidend durch die Struktur des Verbs bestimmt; dabei spielt aber die A r t der jeweiligen Ergänzung eine wenigstens ebenso große Rolle wie ihre Anzahl. Ein Überblick über die Kombinationsmöglichkeiten von Ergänzungen hat also deren Zahl u n d A r t zu berücksichtigen. Solchermaßen festgelegte Kombinationsmöglichkeiten werden S a t z m u s t e r genannt." For Satzbaupläne see Engel (1977: 181). 52 In valency patterns and valency constructions the order of the elements is specified in terms of the order of the elements in a thematically unmarked declarative clause (7.1.2.1). <?page no="151"?> Complement inventories or valency patterns 139 [NP] actsubj + verb act + [NP] covers the following uses of read, write and publish: 53 (3) NW324 I haven't read a book in months, let alone tried to write one. (127) NW52 When she arrived she had already published several articles and reviews in academic journals, and shortly afterwards her much-revised thesis appeared under the imprint of Lecky, Windrush and Bernstein. whereas (86) NW78 It is interesting how many of the industrial novels were written by women. would be attributed to the corresponding passive pattern [NP] pass-subj + verb pass + [by_NP]. In the description of valency patterns, it is thus not necessary to specify whether a complement can occur as the subject of a passive clause (as is done in a complement inventory) since this information concerns the use of the verb in a different pattern. 54 The idea of the valency pattern bank is to list all verbs that can occur in a particular valency pattern independently of any other patterns in which they may occur. Thus (56) NW361 I'll write you a glowing reference … (107) SW27 I wouldn't call myself a structuralist. can both be attributed to the valency pattern [NP] act-subj + verb act + [NP] + [NP]. If the semantic roles of the participants of the verb are also to be taken into account, the description must be extended to valency constructions. 6.7.2.2 Valency constructions The term valency construction 55 can be used to refer to abstract representations that comprise valency patterns (i.e. the lexical unit of the valency carrier and the valency complements) and specifications of the participant roles of the complements. In terms of valency constructions, (56) and (107) would be distinguished as (56) NW361 I'll write you a glowing reference … [NP act-subj ' AGENT ']__write act __[NP ' BENREC ']__[NP ' ÆFFECTED '] (107) SW27 I wouldn't call myself a structuralist. [NP act-subj ' AGENT ']__write act __[NP ' ÆFFECTED ']__[NP ' PREDICATIVE '] 53 According to the principles outlined in 6.5.3.1, I in (3) is analysed as the subject of read and of try but not of write. 54 In the valency pattern bank verbs that have corresponding passive uses will be given in a different colour from verbs that only show active uses. 55 For the similar concept of argument structure constructions in construction grammar see e.g. Goldberg (2006) or Stefanowitsch and Gries (2003). <?page no="152"?> Valency 140 Otherwise, the identification of valency constructions follows the same principles as that of valency patterns. From the point of view of valency constructions the statement that the subject complement unit of the pattern is contextually optional (> 7.1.2.2) means that the respective participants will be considered as inherent: their respective participant roles are implied irrespective of whether the clause contains a subject or not. This means that the two uses of publish in (127) NW52 When she arrived she had already published several articles and reviews in academic journals, and shortly afterwards her much-revised thesis appeared under the imprint of Lecky, Windrush and Bernstein. (48) NW60 After all, she worked on the nineteenth-century industrial novel for something like ten years, and even after publishing her book she went on accumulating ideas and insights about the subject. are subsumed under one valency construction, which can be described as [NP act-subj " PUBLISHER "]__publish act __[NP " ITEM PUBLISHED "] although only in (127) is the " PUBLISHER " syntactically expressed. On the other hand, (128) SW18 Aren't you giving a paper this afternoon? (129) SW13 It was being given by the Oxford medievalist and was on the subject of Chaucerian metrics. will be seen as representing two different valency constructions, namely: [NP subj " GIVER "]__give act __[NP " ITEM GIVEN "] [NP subj " ITEM GIVEN "]__give pass __[by NP " GIVER "] Obviously, the use of give in (46) NW60 What the students want her to do is to give them some basic facts that will enable them to read the novels as simple straight-forward reflections of 'reality', and to write simple, straightforward, exam-passing essays about them. represents yet another valency construction of the verb (though not necessarily of the same lexical unit): [NP subj " GIVER "]__give act __[NP " GIVEE "]__[NP " ITEM GIVEN "]. 6.7.2.3 Valency constructions or complement inventories It is obvious that complement inventories, valency patterns and valency constructions represent different kinds of abstractions from the actual uses of a verb, adjective or noun. Although we do not know to what extent either of these abstractions is psychologically "real" in that it plays a role in the way speakers store valency information in the mental lexicon, from a descriptive linguistic point of view both complement each other: on the one hand, speakers are able to identify that one in (3) NW324 I haven't read a book in months, let alone tried to write one. and a glowing reference in (56) NW361 I'll write you a glowing reference … <?page no="153"?> Possible valency carriers 141 represent the same participant (with the same participant role), although in (3) the complement immediately follows the verb and in (56) it does not. Similarly, speakers are able to realise that him and to me in (130) NW26 Didn't we give him one a few years ago? (90) NW148 I didn't buy it, my parents gave it to me, when they changed it. represent the same participant despite different morphological forms of the complements. This means that assuming some sort of generalization by identifying individual valency slots in terms of a complement inventory as outlined above is rather plausible. On the other hand, there is a lot to be said in favour of a representation of valency relations as valency constructions: firstly, the order of complements in the clause will have to be accounted for anyway. Secondly, it seems that certain lexical realisations of complements (Klotz 2000, Herbst 2007) are related to patterns and not just to complements, compare: 56 (23) NW329 They have given him such a rough time in the past, at conferences and in reviews. (131) NW118 Give me a couple of minutes, will you? (23c) *They have given such a rough time to him in the past, at conferences and in reviews. (131a) *Give a couple of minutes to me, will you? 6.8 Possible valency carriers 6.8.1 Verbs, adjectives and nouns - lexical units Since the valency of verbs largely determines the structure of clauses, it is obvious that verbs are valency carriers. However, as has been shown, the notion of valency has been extended to adjectives, which, after all, exhibit rather similar complementation structures to verbs. In fact, it is quite tempting to analyse a combination of, say, a verb and an adjective such as be sorry in (132) NW64 ... we would of course be very sorry to lose you ... as having a function equivalent to that of a verb: (133) NW383 I don't particularly want to emigrate. Furthermore, it is interesting to see that when adjectives occur in constructions without a verb such as be, a parallel [to_INF]-construction with be is also possible: (1) NW51 Robyn considered herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave. 56 For the lexical aspect see especially Klotz (2000). For a discussion of valency patterns versus valency complements see Herbst (2007). See also Stefanowitsch/ Gries (2003). <?page no="154"?> Valency 142 (134) BNC Despite this many bankers consider it to be more important than the aforementioned economic factors. The combination of be and adjective can thus be seen as representing a predicator in very much the same way as the verb. In noun phrases, the head noun can have a similar function with respect to the relationship of pre-head and complement: thus (135) NW47 Charles's success in landing a job provoked in Robyn the first twinge of jealousy, the first spasm of pique, to mar their relationship. finds a parallel in (135a) Charles succeeded in landing a job. This means that the pre-head Charles can be seen as the expression of a participant of the noun success. Since genitives are not specific to particular nouns, there is no formal reason for classifying them as complements but this is definitely an area in which the distinction between complements and non-noun-specific constituents is very difficult to draw. 57 In any case, the distinction between item-specific complements and non-item-specific modifiers can be drawn in the case of nouns and adjectives in the same way as that between complements and adjuncts for verbs: thus the to-phrase and the on-phrase in (136) and (22) clearly are dependent on the valency of the noun lecture (since, for example, they could not occur with any noun), whereas the form of the particle phrase at eleven in (137) (which could be replaced by then or tomorrow) is not determined by the valency of lecture and thus a postmodifier: (136) NW60 How can all that knowledge be condensed into a fifty-minute lecture to students who know almost nothing about it? (22) NW130 She thought of Charles at the University of Suffolk, giving a lecture, perhaps, on Romantic landscape poetry, illustrated with slides. (137) NW69 Eventually they are all dealt with, and Robyn is free to prepare for her lecture at eleven. Similarly, in (132) NW64 ... we would of course be very sorry to lose you ... very is a premodifier, which is not provided for by a valency slot of the adjective, whereas the [to_INF] must be regarded as a valency complement of sorry. In the same way the property of valency can be ascribed to adverbs such as independently since of meaning in (138) clearly has the status of a valency complement: (138) VDE He assumed that native speakers could distinguish between grammatical and nongrammatical sentences independently of meaning. Strictly speaking, it is of course not a verb, an adjective or a noun that is the valency carrier but a particular lexical unit. Thus, while the monovalent use of sleep in (139) NW41 Unlike Vic, Robyn invariably sleeps until woken. 57 Cf. Herbst (1988) and Herbst/ Roe (1996). For aspects of noun valency cf. also Schmid (2000). <?page no="155"?> Possible valency carriers 143 and the divalent use in (140) NW59 However, on these occasions they invariably slept together. clearly represent different senses of the verb sleep, i.e. different lexical units. In the same way, one can use valency properties to distinguish between the uses of have (which we classified as a primary verb in 3.2.2.2): it may make sense to analyse uses of have with an [NP]-complement such as (141) NW377 Basil has more money than is good for him already. (142) NW214 She had plenty of work: her teaching, her research, her administrative duties in the Department. as representing a different lexical unit from its uses as a pre-head 58 as in (143) NW214 She had survived the previous winter by surrendering herself to work. Certainly, have with an obligatory [to_INF]-complement must be seen as a separate lexical unit: (144) NW214 Now you won't have to go to America. (145) NW66 I have to get ready for a lecture. The problem is, of course, that in many cases the distinction between different lexical units is by no means as clear. Thus, for instance, there are good reasons for distinguishing the uses of consider exemplified by (1) NW51 Robyn considered herself lucky to get a job for one term at one of the London colleges, deputizing for a woman lecturer on maternity leave. (2) NW384 But I was wondering whether, in the circumstances, you would consider staying on next year, and see what happens. but one could also find reasons for subsuming them under one sense. To a certain extent at least, this problem can be overcome in valency descriptions since complement inventories or valency patterns are always specified with respect to the particular lexical unit identified so that the result is a description which is consistent. It is of course debatable to what extent senses of lexical items can be identified independently of their valency structures, which is one of the attractions of the concept of valency constructions as combinations of formal valency patterns, the respective participant roles of the complements and the meaning of the respective lexical units as outlined above. 6.8.2 Particles as valency carriers In 3.5.3.2 it was seen as one of the reasons for establishing the word class particle that there are parallels between different valency structures of verbs and the complementation possibilities of words such as before that provide an important argument 58 This means that the properties of primary verbs and modal verbs and their uses as heads and preheads in the verbal-head complex can be described in terms of valency. Similarly, one could consider taking a similar approach towards the description of pure determiners, determiner-pronouns and pure pronouns and extending the notion of valency to these word classes. <?page no="156"?> Valency 144 for subsuming uninflected words that occur in the traditional word classes of adverb, preposition and subordinating conjunction under one word class: (146) SW18 I'm sure I've heard that name before. (147) SW16 I'll see you in the bar before lunch, then. (148) SW24 In the event, not many people did like Morris Zapp's lecture, and several members of the audience walked out before he had finished. As in verb valency, we could argue that since has a contextually-optional valency slot whose participant is expressed by a complement in (149) and has to be identified from the context in (150): (149) NW111 Robyn had not felt so exalted since the great women's rally at Greenham Common. (150) NW42 She attended an excellent direct-grant grammar school (which has since gone independent, much to Robyn's disgust) where she was Head Girl and Captain of Games and which she left with four A grades at A-Level. In the case of up or down, the valency slot is purely optional: (151) SW24 Morris Zapp delivered it striding up and down the platform with his notes in one hand and a cigar in the other. (151a) Morris Zapp delivered it striding up and down with his notes in one hand and a cigar in the other. Nevertheless, the case for considering particles valency carriers is perhaps less obvious than it is for verbs, adjectives or nouns, which is why Matthews (2007), with respect to traditional prepositions, speaks of "properties at least 'akin' ... to valencies". One complication is that in many cases where zerovalent uses (traditionally classified as adverbs) and monovalent uses co-exist, they very often do not correspond to each other: thus (129) SW13 It was being given by the Oxford medievalist and was on the subject of Chaucerian metrics. could not be replaced by (129a) *It was being given by and was on the subject of Chaucerian metrics. (129b) *It was being given by the Oxford medievalist and was on. despite the existence of (152) NW308 There was never anything "on", Mummy, if you mean, as I presume, getting married. In principle, this situation is no different from that with verbs and other valency carriers, where valency also is a property of lexical units and not of words. It needs to be emphasized in the case of particles, however, because we are perhaps not as aware of meaning differences in the case of particles as with verbs or nouns. In particular, as a comparison of monovalent uses with zerovalent uses shows, the latter very often are restricted to very special kinds of meanings: <?page no="157"?> Possible valency carriers 145 (153) NW58 They were both so obsessed with the place, its gossip and rumours and intrigues, that they wanted to spend every possible moment together there, comparing notes, exchanging opinions: who was in, who out, what X said about Y's review of P's book about Q. (154) NW330 Will you come in for a cup of coffee? While in in (154) resembles the use of since in (150) in that it is deictic in character and could be viewed in terms of a contextually optional slot, it does not allow any such complementation (unless one wishes to treat into as an allomorph of in). Similarly, there is a zerovalent use of outside, which can also be interpreted as being deictic, as in (155) NW49 Now she rubs moisturizer into her facial skin as protection against the raw wintry air outside, coats her lips with lipsalve, and brushes some green eyeshadow on her eyelids, pondering shifts of point of view in Charles Dickens's Hard Times (1854). (156) NW53 Outside, in the street, her car is parked, a red six-year-old Renault Five with a yellow sticker in its rear window, "BRITAIN NEEDS ITS UNIVERSITIES." alongside with monovalent uses such as the following: (157) NW43 The Sussex campus, with its tastefully harmonized buildings in the modernist- Palladian style, arranged in elegant perspective at the foot of the South Downs a few miles outside Brighton, was much admired by architects, but had a somewhat disorienting effect on the young people who came to study there. (158) NW258 ... in fact, she had not been anywhere outside England for the past two years ... (159) NW40 Every text is a product of intertextuality, a tissue of allusions to and citations of other texts; and, in the famous words of Jacques Derrida (famous to people like Robyn, anyway), "il n'y a pas de hors-texte ", there is nothing outside the text. (160) NW51 Of course she was aware, cognitively, that there was a life outside universities, but she knew nothing about it, nor did Charles, or her parents. Although the zerovalent uses of outside in (155) and (156) can be attributed to a lexical unit 'not inside a building' (LDOCE4) and thus be distinguished from (157) - (160), it is probably more difficult to distinguish (160) from a sentence such as 59 (161) NW37 You'd do better to buy in from outside. However, even if a description of the uses of at least some of the items we call particles (and for which CamG (2002) uses the word preposition) will have to allow for a rather strong idiomatic component and take into account different lexical units in the case of some particles, there are very good reasons for describing these items in terms of valency. 60 59 Compare, however, LDOCE4: "if someone is outside a group of people, an organization etc., they do not belong to it: ... The university administrators ignored criticism from outside." 60 Nevertheless, such an analysis is by no means generally accepted since scholars such as Matthews (2007) reject it on the grounds that the relationship established between the constituents linked by a particle is radically different in character from other cases of valency. <?page no="158"?> Valency 146 6.8.3 Idioms and other item-based constructions 6.8.3.1 Idiomatic combinations One must not overlook the fact that the property of valency does not only apply to single words but that combinations of several words can also have valency. Thus, for instance, a phrase such as at a loss opens up a number of different valency patterns, which the noun loss on its own does not have: (162) VDE Not only didn’t we know the flight number, we were at a loss as to which airport he was landing at. (163) VDE She found herself walking, talking about nothing much, all at a loss what to say or do about this assault upon her daily routine. This means that the notion of valency carrier must be extended beyond the individual word. Another case in point is presented by had better: 61 (164) NW116 Well, I suppose I'd better show you round the estate. While in the case of at a loss or had better a combination opens up a new valency structure, a case such as (165) NW218 But the idea of a centre is precisely what poststructuralism calls into question. is not quite as straightforward: although formally it could be subsumed under the trivalent use of call with an [PartP]-complement as in (166) VDE He had called him into his office on Tuesday. it would be difficult to do so semantically, where (165) is more like (107) SW27 I wouldn't call myself a structuralist. Furthermore, question seems to be by far the most frequent noun used after into in a non-locative sense 62 , and into question is used in this sense not only with the verb call, but also with bring and throw. Despite these possibilities of variation, it is tempting to look at call NP into question as an idiomatic use of call in a special valency structure. 63 6.8.3.2 Phrasal verbs A very obvious case of certain valency patterns of verbs achieving idiom status is that of phrasal verbs, i.e. combinations of verbs with particles such as in, on, up, which even in traditional accounts are often referred to as adverbial particles: 64 (17) NW61 Put me down for picketing, Bob … 61 CGEL (1985: 3.45) classifies had better as a modal idiom. 62 In a search for called into BNC also yields action, default and effect, for instance. 63 Compare LDOCE 4 (entry: question). 64 Phrasal verbs can thus be seen as item-specific valency constructions of the respective verb, which, however, can have valency slots for further complements. It was already pointed out in 6.4.2.2 that cases of particle phrase complementation are not analysed as multi-word verbs in valency theory. In other approaches, the term phrasal verbs is often used to refer to such constructions as well. For adverb particles and prepositional adverbs see CGEL (1985: 9.5). <?page no="159"?> Possible valency carriers 147 (167) NW308 Oh dear, is it off again? One of the characteristics of such phrasal verbs is that in many cases the particle can either precede or follow a noun phrase complement of the verb: (168) NW359 Robyn put down the telephone receiver and laughed aloud. (169) NW37 It was probably a foolish thing to say, but Vic shrugs off any regret as he puts the receiver down. In the case of pronouns, the particle follows the [NP]-complement: (170) NW150 But he lingered by the window, inhibiting her from winding it up. (171) NW83 As Robyn Penrose is winding up her lecture ... Not all phrasal verbs allow shifting of the particle, however: (172) NW50 The man himself departed to take up a more remunerative and prestigious post elsewhere, and his friends and supporters fell silent, or retired, or resigned and took jobs in America. It is important to realise that the particle in these cases realises a whole particle phrase. Thus, from the point of view of the verb, (17) and (168) represent three-valent structures (17) NW61 Put me down for picketing, Bob … (168) NW359 Robyn put down the telephone receiver and laughed aloud. which differ from a use of put as in (173) NW47 She puts her soiled breakfast things into the sink .... only in terms of the shiftability of the particle and the idiomatic status of the combination. 65 65 It must be obvious that the telephone receiver in (168) is a complement of the verb put but not a complement of the particle down. This is different, of course, in the case of Morris Zapp delivered it striding up and down the platform with his notes in one hand and a cigar in the other SW24 , where the platform is a complement of down (and of up) so that up and down the platform can be analysed as a (coordinated) particle phrase. <?page no="160"?> 7 The meaning of sentences 7.1 Sentence types 7.1.1 Sentence types and meaning It is generally assumed that the structure of sentences has a certain impact on their meaning. Thus it is obvious that sentences that contain the same words in a different order such as (1) SW14 She had been for a walk. (1a) Had she been for a walk? or (2) SW25 The door is open. (2a) Is the door open? differ in meaning in that the first in each pair is a 'statement' and the second a 'question'. Quirk, Greenbaum, Leech and Svartvik (CGEL 1985: 11.2) distinguish between the following four classes of discourse function: (a) statements ("primarily used to convey information") (b) questions ("primarily used to seek information on a specific point") (c) directives ("primarily used to instruct somebody to do something") (d) exclamations ("primarily used for expressing the extent to which the speaker is impressed by something"). This kind of meaning carried by the sentence type has to be distinguished from the illocutionary force of a particular utterance in a given context of situation. Thus a statement such as (2) can be taken to represent the type of speech act which is often referred to as directive in the pragmatics literature 1 , as is also obvious from the source of the example: (2b) SW25-26 ... when for instance I say, 'The door is open,' and you say, 'Do you mean you would like me to shut it? ' ... In this sense, sentence (2), which is a statement, would have the same speech act function as a question of the type (2c) Can you shut the door? 1 Cf. Bublitz (2001). For a distinction between the levels of syntax, semantics and pragmatics see Leech (1983: 114). Compare also CamG (2002: 61-62 and Chapter 10). <?page no="161"?> Sentence types 149 Similarly, sentences such as (3) SW30 You have to treat these professors carefully, Persse. (4) SW27 The floor is now open for discussion … semantically are statements, but pragmatically (3) can be an explanation, a piece of advice etc., whereas (4) must be seen as a kind of directive. 7.1.2 Statements and questions 7.1.2.1 Major types of statements and questions The most common sentence type is represented by (5) NW484 Robyn thinks. (6) SW3 April is the cruellest month. (7) NW84 Sometimes I think the VC takes his economy drive a little too far. Syntactically, these are instances of clause type [Subj__Pred] (usually referred to as declarative); semantically they represent statements. It seems appropriate to describe sentence types in terms of constructions, i.e. pairings of form and meaning. 2 In order to make this clear, the sentence types are given combined names, which indicate formal properties and semantic properties (the semantic labels are given in single inverted commas to distinguish them from labels referring to formal aspects). Sentences such as (5)-(7) can then be described as representing a declarative-'statement'construction, which can be defined as follows: 2 For a definition of construction see Goldberg (2006: 5). One reason for not discussing formal syntactic and semantic properties of sentence types separately is that the terminology used in traditional grammars is rather confusing. Thus a term such as interrogative is generally used as a formal term which is related to two different formal properties - wh-words and subject-operatorinversion. Cf. CGEL (1985: 803) and CamG (2002: 856). <?page no="162"?> The meaning of sentences 150 Declarative-'statement'-construction Meaning: 'statement' Formal criteria: [Subj__Pred] or [Pred] I. Typical constituents: (a) Declarative 'statements' usually have a subject (which is not a wh-word). (b) Declarative 'statements' always have a predicate. II. Word order: The subject precedes the predicate. III. Intonation: falling 3 On the basis of the three formal criteria - typical constituents, word order and intonation - it is possible to distinguish this construction from other major sentence types. For instance, the only difference between this type of construction and questions of the type (8) SW8 Which one is Professor Swallow? (9) NW41 And who is Charles? is a lexical one, namely that the subject is realised by a phrase which contains a question word such as which or who. They belong to the following construction: Wh-subject-'question'-construction Meaning: 'question' Formal criteria: [wh-Subj__Pred] I. Typical constituents: (a) a subject which contains a wh-word as head or pre-head. (b) a predicate. II. Word order: The subject precedes the predicate. III. Intonation: falling 4 If constituents other than the subject are the focus of the question, there is also a change in word order so that the corresponding construction can be described in the following form: 3 It is extremely difficult to make general statements about intonation and thus it must be obvious that the labels given here have to be regarded as generalizations. Leech and Svartvik (1994: 365) point out that yes-no-questions "usually have rising intonation" and that wh-questions "normally have falling intonation". See also CGEL (1985: 817). For details concerning intonation patterns compare Gimson/ Cruttenden (2001: 268-271). 4 See Leech/ Svartvik (1994: 366). <?page no="163"?> Sentence types 151 Wh-interrogative-'question'-construction Meaning: 'question' Formal criteria: [wh-A/ C__op_Subj_rPred] I. Typical constituents: (a) a wh-constituent (i.e. a phrase which contains a question word as head or pre-head) (b) a subject (c) a predicate containing a pre-head functioning as operator II. Word order: wh + operator + subject + (rest of predicate) III. Intonation: falling 5 (10) NW87 Who shall I nominate then? (11) NW88 Why don't you ask Robyn Penrose? (12) SW18 Now, where do I check in? (13) SW18 What size of audience am I likely to get? The fact that the operator precedes the subject (sometimes called subject-operatorinversion) is also typical of other kinds of question, which however do not contain a wh-word: Interrogative 'yes-no-question'-construction Meaning: 'question' Formal criteria: [op__Subj__rPred] I. Typical constituents: (a) Interrogative 'yes-no-questions' usually have a subject (b) Interrogative 'yes-no-questions' usually have a predicate containing a prehead functioning as operator II. Word order: operator + subject + (rest of predicate) III. Intonation: rising 6 Since such questions can often be answered by yes or no, they are often referred to as yes-no-questions: (14) SW18 Have you really flown all the way from America for this conference, Professor Zapp? (15) SW5 Was it Professor Swallow who was giving out the little badges? A different type of question is represented by the following type: 5 See Leech/ Svartvik (1994: 365-6). 6 See Leech/ Svartvik (1994: 365). <?page no="164"?> The meaning of sentences 152 Interrogative 'alternative-question'-construction Meaning: 'question' Formal criteria: I. Typical constituents: (a) Interrogative 'alternative questions' usually have a subject (b) Interrogative 'alternative questions' usually have a predicate containing a pre-head functioning as operator (c) Interrogative 'alternative questions' contain a phrase with at least two elements coordinated by or II. Word order: operator + subject + (rest of predicate) III. Intonation: rise on each question item, fall on last question item 7 (16) NW98 ... was it Women in Love or Lady Chatterley? ... A further type of question is identical with the declarative-'statement'-construction in terms of constituents and word order but differs in intonation: Declarative-'question'-construction Meaning: question Formal criteria: [Subj__Pred] and [Pred] I. Typical constituents: (a) Declarative 'questions' usually have a subject (which is not a question word). (b) Declarative 'questions' always have a predicate. II. Word order: The subject precedes the predicate. III. Intonation: rising 8 (17) SW11 Then you are no doubt familiar with Miss Weston's book, From Ritual to Romance, on which Mr Eliot drew for much of the imagery and allusion in The Waste Land? (18) SW18 You like it? The six constructions outlined above are of course not the only constructions expressing statements or questions in English. For instance, a sentence such as (19) SW18 Morris! It's marvellous to see you after - how many years? does not fall into any of the categories listed above but could be identified as a whdeclarative-'question'-construction. 9 A full list of all sentence types in English is beyond the scope of this book, however. 7 See CGEL (1985: 11.20). 8 See CGEL (1985: 11.12). <?page no="165"?> Sentence types 153 7.1.2.2 Context-bound types of statements and questions Some of the constituents of the constructions listed above are not obligatory, but contextually optional, i.e. they do not have to be realised if their referent or meaning is clear from the context. This applies in particular to (a) subjects of clauses as in (20) NW85 'I don't remember it coming up at that meeting of Senate,' said Philip. 'Must have been passed without discussion. ...' (b) (in the spoken language) operators and other pre-heads of the verbal core as in (21) NW324 You leaving Rummidge, Robyn? According to the definitions given above, (20) can still be classified as a declarative- 'statement'-construction since we said that such constructions "usually" have a subject. Likewise, (21) is an example of an interrogative yes-no-'question'. We refer to such sentences as context-bound declarative 'statements' and context-bound interrogative 'yes-no-questions' since they can only be interpreted in the context in which they occur. Such context-bound constructions are often dealt with in terms of ellipsis. 10 Note that readers are not always able to retrieve the "missing words" from the linguistic context in such cases. Thus if one wanted to expand the underlined sentence in (22) SW5 No, that's Bob Busby, he's just as bad. Worse, if anything. Been beside himself with excitement for weeks, organizing outings and so forth. to a "full" sentence such as (22a) He has been beside himself with excitement for weeks ... only the he can be found in the preceding clause, but not has. The same applies to the sentences underlined in (23) NW324 '... I haven't read a book in months, let alone tried to write one.' 'Well, I have,' said Robyn. 'Read one or written one? ' said Morris Zapp. 'Written one,' said Robyn. where the change between speakers implies a change between you and I, which is not immediately recoverable from the text. For these reasons, we will not make use of the concept of ellipsis here, but simply consider such sentences as context-bound constructions. 11 9 Cf. CGEL (1985: 11.14 note f) for cases such as (19). For a more detailed descriptions of sentence types see CGEL (1985) or CamG (2002). See also Cobuild Grammar (1990: 195-206). 10 Compare, for instance, Quirk, Greenbaum, Leech and Svartvik (1985: 12.47-51), where ellipsis is described in terms of "omission" and "verbatim recoverability" (CGEL 1985: 12.32). 11 Compare Sinclair/ Mauranen (2006: 150): "Earlier grammars used categories like 'ellipsis' and 'words understood' to explain this kind of phenomenon, but now that many linguists respect the actual wordings of corpora, and are committed to describing the text and not some rewritten ver- <?page no="166"?> The meaning of sentences 154 7.1.3 Directives Sentences such as (24) SW18 Never go to lectures. (25) NW87 Don't be absurd. can be considered to be prototypical examples of the imperative-'directive'-construction. Imperative-'directive'-construction Meaning: 'directive' Formal criteria: I. Typical constituents: (a) Imperative 'directives' contain a predicate with the verb in the base form (which in negation needs do-support, also for the verb be) (b) Imperative 'directives' can have a subject. II. Word order: If present, the subject precedes the predicate. III. Intonation: falling Although, like (24) and (25), most imperative-'directive'-constructions do not have a subject, most grammars analyse you or someone in sentences such as (26) QE You be QÙIET! (27) QE Someone open this door. as the subject of the verb, although, of course, there is no concord between subject and verb. According to Quirk, Greenbaum, Leech and Svartvik (1985: 11.25) 12 , such cases have to be distinguished from (28) QE YÒU | come HÈRE in which you represents a separate tone unit and thus is analysed as a vocative, i.e. a name, a pronoun or other type of noun phrase identifying the person at whom the imperative is directed, which functions as an adjunct in sentence structure: (29) NW94 Don't be ridiculous, Penny. (30) NW61 Put me down for picketing, Bob. Subject and adjuncts with the role ' VOCATIVE ' can also be combined: (31) BNC You be quiet, Hughie. sion of it, notions that there are some words missing or that the text cannot be understood as it stands are no longer tenable." 12 For a very similar approach compare Biber/ Conrad/ Leech (2002: 254-255) and the discussion in CamG (2002: 924-928). One criterion supporting such an analysis is the negative: Don't you be so damn patronizing BNC . <?page no="167"?> Sentence types 155 Apart from the imperative-'directive'-construction, there are of course other constructions expressing the meaning of 'directive'. Both CGEL (1985: 829-830) and CamG (2002: 924-925) mention let-constructions in this context: 13 (32) SW44 Let's catch her up. (33) NW107 Let me find that letter first. The let-'directive'-construction, if we want to call it such, is a good example of the view, put forward by Sinclair (2004a: 164), for example, "that the fundamental distinction between grammar, on the one hand, and lexis, on the other hand, is not as fundamental as it is usually held to be". The same applies to exclamative constructions which also depend on the presence of particular lexical items (what and how) (see below). 7.1.4 Minor sentence types The 'question'-constructions and 'directive'-constructions identified above differ from the declarative-'statement'-construction in terms of the meaning of the whole sentence. Other constructions differ from the declarative-'statement'-construction in giving special prominence to one of the elements within the sentence. Such prominence can be expressed by prosodic means (or, in the written language, by a special typeface) without changing the word order of the declarative- 'statement'-construction as in (34) NW94 To be fair, he doesn't seem to have (italics in original) much power. Another means of attributing prominence to an item is fronting, which means that a clause constituent occurs in initial position in the sentence: 14 (35) BNC Mark I think his name is. (36) SW3 To this element of the developing story Wordsworth soon added several more ... (37) SW28 Tea, I'm afraid it will have to be. In some cases, fronting may be accompanied by inversion of subject and verb: (38) BNC Now comes the dodgy part. (39) BNC Most remarkable was the language used by some senior members of the judiciary. Two constructions that involve more than a mere shift of elements must be mentioned in this context - cleft and pseudo-cleft constructions. 15 Cleft-constructions have the following structure: 13 Both CGEL (1985: 829-830) and CamG (2002: 924-925) refer to these constructions as a type of imperative, which seems inappropriate since the term imperative is generally used to refer to a morphological property of the predicate (represented by the base form for singular and plural uses). What distinguishes this construction from other uses of let as a trivalent verb is that the latter cannot be abbreviated to let's (CGEL) and that question tag constructions of the type Let's go, shall we? are only possible in this use. 14 For fronting and inversion see CGEL (1985: 1377-1381). <?page no="168"?> The meaning of sentences 156 it + (form of) be + complement/ adjunct of clause + (relative pronouns) + (rest of) clause (40) NW238 It's my kids who should be worrying about getting qualifications. (41) NW133 It was the most terrible place she had ever been in in her life. Pseudo-cleft-constructions have the following structure: clause with what as a first element + (form of) be + rest of clause. (42) BNC What they have done is to look in detail at the impact of two severe floods which hit the little Araglin river in Co Cork in August 1986 and how it has recovered since. (43) BNC What you are trying to do is to develop an employee who thinks for himself. (44) NW60 What Robyn likes to do is to deconstruct the texts, to probe the gaps and absences in them, to uncover what they are not saying, to expose their ideological bad faith, to cut a cross-section through the twisted strands of their semiotic codes and literary conventions. A further type of construction that involves fronting is presented by the so-called exclamatives, which are related in grammars such as CGEL (1985: 833-835) and CamG (2002: 918-924) to a separate discourse function exclamation or "exclamatory statement" (CamG 2002: 853). We consider exclamations to be a special kind of 'statement' and thus refer to the constructions in question as exclamative-'statement'constructions. [what/ how-A/ C_Subj_Pred] (45) NW99 How strange it is, strange and sad, to see all these tropical faces amid the slush and dirty snow, the grey gritty hopelessness of an English industrial city in the middle of winter. (46) NW111 What fun it had been! [what/ how-Subj_rPred] (47) BNC What a curious and deep shaft into English society is opened by the reflection that when reputations and perhaps a million pounds in costs alone are at stake, we revert to the patterns of upper-class education. Exclamative-'statement'-constructions, and to a certain extent, also pseudo-cleft and cleft-constructions, are further examples of how constructions can be constituted by a combination of structural-grammatical and lexical elements in the sense of Sinclair (2004a). 15 Compare CGEL (1985: 1383-1389) for a more detailed description. <?page no="169"?> Sentence meaning 157 7.2 Sentence meaning 7.2.1 Conceptual or propositional meaning The sentence types identified above constitute an important dimension of the meaning of a sentence. At the same time, the meaning of a sentence or clause arises from the meanings of the words, or, to be more precise, the lexical units that make up the constituents of the sentence, and from the relationship of the constituents to one another, or, in other words, from the participant roles expressed by these constituents. Thus a sentence such as (48) NW218 They don't pay us all that much, anyway. receives a certain interpretation because its words have certain meanings and because words such as they and us refer to certain people that have been mentioned before 16 . At this level, the meaning of (48) contrasts very obviously with the meaning of (48a) They pay us quite a lot, actually. What contributes further to the meaning of the sentence is the fact that they expresses the participant role of ' AGENT ', that much that of ' ÆFFECTED ' and us that of ' BENREC '. Here, (48) obviously contrasts with (48b) We don't pay them all that much, anyway. This kind of meaning is often called conceptual meaning (Leech 1981: 9) or propositional meaning (Brown/ Yule 1983: 107). 17 7.2.2 Perspective 7.2.2.1 Thematic meaning A further level of meaning that can be used in the description of sentences is that of thematic meaning, which Leech (1981: 19) describes as "what is communicated by the way in which a speaker or writer organizes the message, in terms of ordering, focus, and emphasis". The information structure of sentences was discussed extensively by linguists of the Prague School, who made use of the terms theme and rheme in this context: The theme is the left-most constituent of a sentence, the starting-point of a message, in other words, what is being talked about. The rheme is what follows the theme, i.e. what is said about it. 16 For reference in texts compare Sinclair (2004a: 15). 17 The term proposition is not used in the strict sense of logic here, but in the way that Brown and Yule (1983: 107) describe it as being typical of text-analysis literature, namely "to represent a 'once-off' interpretation of a text-sentence as it is used in a context." <?page no="170"?> The meaning of sentences 158 It is obvious that constructions such as pseudo-cleft sentences and cleft sentences (> 7.1.4) can be seen in terms of thematic meaning, but up to a point this also applies to differences between an active sentence such as (48) and the corresponding passive (48c) We aren't paid all that much, anyway. (48d) We don't get paid all that much, anyway. Such differences in meaning do not concern the propositional content of the sentence but are rather a matter of the perspective from which the action is described. 18 7.2.2.2 Perspectivization and lexical choice In some cases, the choice of perspective entails a lexical choice, as is the case with verbs such as buy and sell: (49) NW53 It was formerly her parents' second car, sold to Robyn at a bargain price when her mother replaced it. (50) NW148 I didn't buy it, my parents gave it to me, when they changed it." Here, I in (50) represents the same participant of the action as to Robyn in (49) but the way in which they are presented is rather different. 19 One could argue that buy and sell belong to the same semantic frame because X buys Y from Z is conceptually equivalent to Z sells Y to X. 7.2.2.3 Participant roles and clausal roles If one wants to describe the different levels of sentence meaning it may thus make sense to investigate alternative ways of expressing the same propositional content in a language. This includes lexical choices such as the one between buy and sell or grammatical choices such as the one between active clauses and passive clauses. A further case in point is presented by cases where the same participants of a verb are realised in different valency patterns: (51) NW325 Send me your manuscript ... (51a) Send your manuscript to me. (52) NW323 Are you going to get this person a drink, Philip? (52a) Are you going to get a drink for this person, Philip? 18 Theme and rheme must not be confused with the terms "given" and "new". Given information is the information that the speaker assumes the hearer to be familiar with, while new information is not assumed to be familiar to the hearer or reader. For this use of theme and rheme see Halliday (1970: 160-164) and Brown/ Yule (1983: 126-133). Compare also Esser (1979). For the notions of given and new see Brown/ Yule (1983). Compare Halliday (1967/ 1968) for a detailed analysis of clause structure involving the notions of theme and different participant roles. 19 Compare the discussion of this example by Fillmore (1977: 16-18) and Palmer (1981: 148). Compare the descriptions in FrameNet. <?page no="171"?> Sentence meaning 159 The question is whether there is any difference in meaning between (51) and (51a) or (52) and (52a). On the one hand, the choice of pattern is quite clearly dependent on structural factors such as the length of the complement in question: (53) NW41 This morning, she gives priority to the fact that it is the first day of the winter term, and that she has a lecture to deliver and two tutorials to conduct. (53a) ? ? This morning, she gives the fact that it is the first day of the winter term, and that she has a lecture to deliver and two tutorials to conduct priority. (54) BNC Public policy will give it priority. On the other hand, one would not necessarily want to argue that (52) and (52a) are totally synonymous either. One indication of this is that the BNC yields only one instance of get/ got a drink for, but more than forty for get me/ us/ you/ him/ her/ them a drink. A case that is often discussed in the literature (e.g. Palmer 1981: 146-149, Goldberg 2006) concerns verbs such as smear, paint or load, where alternatives of the following kind exist (Roman numerals indicate the participants of the verb): (55) NW53 She lifts from the floor a leather Gladstone bag, and begins to load it II with the things she will need for the day: well-thumbed, much underlined and annotated copies of Shirley, Mary Barton, North and South, Sybil, Alton Locke, Felix Holt, Hard Times; her lecture notes - a palimpsest of holograph revisions in different-coloured inks, beneath which the original typescript is scarcely legible; and a thick sheaf of student essays marked over the Christmas vacation III . (56) NW25 He strides through the kitchen, where Marjorie I is listlessly loading his soiled breakfast things III into the dishwasher II , and runs up the stairs. (57) QE Pat I loaded the wagon II with hay III . 20 (58) QE Pat I loaded the hay III onto the wagon II . Apart from the fact that in (55) the length of the [with_NP]-complement may be one reason for the choice of construction, one could argue that the participant expressed by the noun phrase complement is presented as being more immediately affected by the action than the particle complement. If this is the case, then a description of the role of the complements purely in terms of participant roles as defined in Chapter 6.5 would not be sufficient since the participant roles are identical for each participant in all cases. It could indeed be argued that the element of 'affectedness' that characterizes the noun phrase complements is a property of the syntactic structure that combines with the participant roles, which are properties of the valency of the verb. 21 It thus seems that one has to make a distinction between different levels and aspects of description: A first dimension of the semantic relations between the constituents of a sentence (or the complements of a verb) arises from the participants. As was 20 For a discussion of these examples see Goldberg (2006: 34-37) (Roman numerals added by us). For a detailed analysis see Dowty (1991). 21 Compare Palmer's (1994) distinction between grammatical roles and notional roles. <?page no="172"?> The meaning of sentences 160 shown in 6.5.1, this can be expressed in terms of relatively specific roles such as " ITEM LOADED " or in terms of more general roles such as ' ÆFFECTED '. A second dimension arises from the fact that sentence structures seem to be able to express more general relations between the constituents in terms of who does what to whom or what affects what etc. Thus it can be seen as a property of the particular construction chosen in (57) that the wagon is presented as an element to which something is done, i.e. as being 'affected' by the action described. In (58), however, this is the case with the hay. The problem for the analysis of such relations is that while in (57) the hay is being presented as being particularly 'affected' by the action, and in (58) the wagon, when analysing the process of loading, it would be difficult to argue that the hay would not be 'affected' in (58) or the wagon would not be 'affected' by the action in (57). It seems that by choosing (57) or (58), however, what the speaker does is not to say that participant II is 'affected' and participant III is not or vice versa, but to put special emphasis on this 'affectedness' of one particular participant. It thus seems appropriate to treat this kind of phenomenon as part of the process of perspectivization. 22 This can be expressed by making a distinction between: 23 participant roles, which are properties of verb valency and remain constant for each participant such as " ITEM LOADED " or ' ÆFFECTED ' clausal roles, which arise from sentence structure, and add a further dimension to the interpretation of the sentence such as FOCUSSED AFFECTEDNESS . If we describe the participant roles of load as I " LOADER ", II " CONTAINER " and III " ITEM LOADED " and describe the role of the noun phrase complements in the above examples as FOCUSSED AFFECTEDNESS , then the uses of load in (55) and (56) can be characterised as follows: 24 (55) NW53 She lifts from the floor a leather Gladstone bag, and begins to load it II with the things she will need for the day: well-thumbed, much underlined and annotated copies of Shirley, Mary Barton, North and South, Sybil, Alton Locke, Felix Holt, Hard Times; her lecture notes - a palimpsest of holograph revisions in different-coloured inks, beneath which the original typescript is scarcely legible; and a thick sheaf of student essays marked over the Christmas vacation III . 22 Compare Fillmore (1977). 23 Compare Palmer's (1994: 4-11) distinction between notional roles and grammatical roles. 24 One reason for assigning these roles only to noun phrases is that there does not seem to be an equally perceptible change of perspective in the case of clausal complements. Compare also CamG (2002: 710): "NPs functioning as core complements are related directly to the verb, while those functioning within PPs are related to the verb only indirectly, via the preposition. An NP related to the verb by a preposition in this way is referred to as an oblique." Compare also Goldberg (2006: 40). <?page no="173"?> Sentence meaning 161 (55') I load II III NP NP with NP " LOADER " " CONTAINER " " ITEM LOADED " FOCUSSED AFFECTEDNESS (56) NW25 He strides through the kitchen, where Marjorie I is listlessly loading his soiled breakfast things III into the dishwasher II , and runs up the stairs. (56') I load III II NP NP into NP " LOADER " " ITEM LOADED " " CONTAINER " FOCUSSED AFFECTEDNESS The difference between participant roles and clausal roles also becomes apparent when analysing the following uses of close down: (59) NW113 The division I could close us II down. (60) NW33 Since the election of the Tory Government of 1979, which allowed the pound to rise on the back of North Sea oil in the early eighties and left British industry defenceless in the face of foreign competition, or (according to your point of view) exposed its inefficiency (Vic inclines to the first view, but in certain moods will admit the force of the second), one-third of all the engineering companies in the West Midlands II have closed down. The phrasal verb close down can be seen as having two participants (I and II), which can be described as I ' AGENT ' and II ' ÆFFECTED '. Again, one could argue that in (59) the 'affectedness' of II is presented more prominently than in (60) and attribute the clausal role FOCUSSED AFFECTEDNESS to (59) but not to (60). (59') I close II down NP NP ' AGENT ' ' ÆFFECTED ' FOCUSSED AFFECTEDNESS (60') II have closed down NP ' ÆFFECTED ' <?page no="174"?> The meaning of sentences 162 The effect syntactic structure has on the semantic interpretation of participants also shows with respect to subject complement units. In (60) NW33 ... one-third of all the engineering companies in the West Midlands II have closed down. participant II is presented as being more actively and causally involved in the process than in (59) NW113 The division I could close us II down. Again, we get an interesting discrepancy between participant roles looked at in isolation and semantic roles at the level of the clause. Participant I is a relatively clear case of the category ' AGENT ' and clearly will be interpreted like that in a sentence such as (59). However, II, whose participant role can be characterized as ' ÆFFECTED ', seems to receive an 'agentive' kind of interpretation when it occurs as the subject of an active clause as in (60). Here, (60) contrasts with the passive construction, in which the subject complement unit does not express 'agentivity' of any kind: (61) NW307 There are rumours that one or two universities II will be closed down completely. It thus seems to be a characteristic of subject complement units of active clauses that they may add a dimension of AGENTIVITY to the respective participant (which could then be seen as a clausal role). 25 This property of subjects of active clauses becomes particularly apparent with a verb such as fly, where different kinds of participants receive an AGENTIVITY -interpretation when realised by a subject complement unit: 26 (62) CP4 It is obvious from his stiff, upright posture, and fulsome gratitude to the stewardess serving him a glass of orange juice, that Philip Swallow, flying westward, is unaccustomed to air travel; while to Morris Zapp, slouched in the seat of his eastbound aircraft ... the experience of long-distance air travel is tediously familiar. (63) BNC Then a helicopter pilot flew them to an area of bare ice studded with dark rocks near the Allan Hills (76°S; 156°E). (64) VDE The airline flew 89 flights to Sweden in 1941. (65) SW232 Which airline did you fly? (66) BNC The plane flew up the fjord, which seemed so narrow that the mountains were on both wing tips at the same time. (67) VDE He flew by helicopter from the US naval air station in Bermuda to the grounds of Government House where he was welcomed by Mrs Thatcher. 25 Compare Palmer's (1994: 25) view of the "Agent being essentially the cause or 'initiator' of the action and the Patient the one directly affected by it or its 'endpoint'". See also Goldberg (2006: 184-186). 26 For a discussion of semantic roles in connection with the verb fly see Götz-Votteler (2007). In these examples, the participant role of Philip Swallow in (62) like that of them in (63) could be subsumed under ' AFFECTED ', but Philip Swallow still receives an AGENTIVE interpretation in a similar way to a helicopter pilot in (63). The latter can also be attributed a participant role ' AGENT ', which is also exhibited by its ability to occur as a [by_NP]-complement in a corresponding passive. <?page no="175"?> Sentence meaning 163 (68) NW132 She couldn't see their features, because they wore facemasks and goggles -not without reason, for when they tipped the ladles to pour, white-hot metal splashed like pancake batter and sparks flew through the air. The clausal roles of AGENTIVITY and FOCUSSED AFFECTEDNESS can also be used to explain the difference between such uses of meet as exemplified by (69) NW60 Outside the Department Office she meets Bob Busby, the Department's representative on the local committee of the Association of University Teachers, pinning a sheet of paper to the AUT noticeboard. (70) NW110 Haven't we met before? On the basis of (70) it does not seem possible to differentiate between the two participants of meet by attributing different characterizations or roles to them. However, the elements of AGENTIVITY and FOCUSSED AFFECTEDNESS that she and, respectively, Bob Busby have in (69) can be accounted for on the basis of different clausal roles. However, it is important to bear in mind that these clausal roles do not hold generally, but are dependent on certain verb meanings: thus the role of FOCUSSED AF - FECTEDNESS only seems to apply to noun phrase predicate complement units in combination with verbs referring to activities such as (71) NW307 They are systematically destroying the finest university system in the world. but not in combination with verbs expressing states such as (6) SW3 April is the cruellest month. (72) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. Similarly, as far as subject complement units are concerned, the clausal role of AGENTIVITY also only applies in combination with verbs expressing activities, i.e. to they in (71) or (73) NW322 Robyn arrived a little late at the Swallows' modernized Victorian villa ... but not with verbs expressing states as in (6) SW3 April is the cruellest month. (72) NW52 She was a popular and conscientious teacher, whose optional courses on women's writing were oversubscribed. (74) NW323 'I always know when I'm in England,' said Morris Zapp ... These observations show that the meaning of a sentence is not solely determined by the participant roles that are part of the valency specifications of the governing verb, but also by aspects of the structure of the clause, which interact and combine with the participant roles and with the meaning of the verb. <?page no="176"?> 8 Analysis of sentences 8.1 Clause units 8.1.1 Clause units as a combination of structural and valency criteria The approach towards the description of sentences outlined here attempts to combine two different aspects: the aspect of grammatical structure, which manifests itself in the identification of formal clause types and sentence types, which are constructions in the sense of a combination of form and meaning: in terms of grammatical structure, a clause contains three types of clause elements, namely subject, predicate and adjunct(s) the aspect of the lexical property of valency, which is based on the distinction between item-specific complements and non-item-specific adjuncts and which is made apparent by identifying complement inventories, formal valency patterns and valency constructions. Both aspects are closely related: on the one hand, the structure of the predicate depends on the number of valency complements of the governing verb. On the other hand, the description of valency phenomena is necessarily related to clause structure: firstly, it is part of the characterisation of valency complements as to whether they can occur as the subject of an active or a passive clause or not, and, secondly, particular valency constructions include information on the linear order of the complements. There is considerable overlap between the two aspects, but not complete identity: in a sentence such as (1) NW361 You will write me a reference, then? the analysis of clause structure yields you as the subject of the clause and will write me a reference as the predicate. From a valency point of view, write is the governing verb and you, me and a reference are valency complements of write. From both perspectives, then is seen as an adjunct. As was pointed out in Chapter 1, this results in conflicting hierarchies: the structural analysis takes the sentence as the highest unit in the hierarchy and establishes constituents so that the subject and the predicate are at the same level of analysis. The complements occurring in the predicate are established in the next step of the segmentation process: <?page no="177"?> Clause units 165 (1') S Subject Predicate Adjunct You then will write me a reference The valency analysis does not establish part-whole relations, but takes the verb as the highest element in the hierarchy and sees the complements as dependent upon it. As a result, all complements are to be found at the same level in the hierarchy: (1'') (will) write Complement Complement Complement Adjunct You me a reference then We would like to combine both perspectives in a linear representation at the level of clause units: (1''') subject predicate adjunct You will write me a reference then complement verbal head complex complement complement adjunct The names for some clause units combine both aspects: 1 subject complement unit (SCU): the clause unit that functions as the subject of the clause and is realised by a complement of the governing verb (you) 2 1 The clause units identified here are not identical with the units of linear unit grammar as outlined by Sinclair and Mauranen (2006). 2 The term subject complement unit is to be taken in the sense in which the German term Subjektergänzung is often used, i.e. as referring to a complement that functions as the subject. We use the <?page no="178"?> Analysis of sentences 166 predicate complement unit (PCU): a clause unit that is a constituent of the predicate and is realised by a complement of the governing verb (me and a reference) predicate head unit (PHU): the clause unit that contains the verbal head-complex, i.e. the governing verb which is considered the head of the clause, and possible pre-heads (will write) adjunct unit (AU): a clause unit that functions as an adjunct and is independent of the valency of the governing verb (then) One further type of clause unit can be identified: linking unit (LU): a clause unit that coordinates clauses or sentences. Thus, for sentences such as (1) NW361 You will write me a reference, then? (2) NW215 But this year the winter term was different. (3) NW325 Finally he asked her if she had made arrangements to publish her book in America. the following analysis can be provided: (1'''') You will write me a reference then SCU PHU PCU1 PCU2 AU (2'''') But this year the winter term was different LU AU SCU PHU PCU (3'''') Finally he asked her if she ... America AU SCU PHU PCU1 PCU2 8.1.2 Description of SCUs and PCUs In the light of the discussion of valency relations in the previous chapters it seems appropriate to describe SCUs and PCUs in terms of a semantic role of the corresponding participant (see 6.5) the type of complement they represent (> 6.4), i.e. using such categories as [AdjP], [NP] pass-subj or [to_INF]. term subject complement unit in order to make quite clear that this category is not identical with the term subject complement in CGEL (1985: 10.8), where it is used for an element of clause structure such as different in sentence (2), which refers to the subject. Most other accounts do not use the term subject complement and object complement in the sense of CGEL, but rather terms such as subject attribute and object attribute (Aarts/ Aarts 1982/ 1988), subject predicative and object predicative (Biber, Johansson, Leech, Conrad and Finegan 1999: 126 and 130) or predicative complement (CamG 2002: 53-54). In our terminology, these correspond to PCUs with the semantic role ' PREDICATIVE '. Compare also Jespersen’s (1927: 355-404) use of the term predicative. <?page no="179"?> Alternative accounts of clause structure 167 (1'''') You will write me a reference then SCU PHU PCU1 PCU2 AU [NP] act-subj [NP] pass-subj [NP] pass-subj ' AGENT ' ' BENREC ' ' ÆFFECTED ' 8.2 Alternative accounts of clause structure 8.2.1 PCUs versus objects, predicatives and adverbials The model outlined above contains elements of dependency and valency grammar in that it takes the head of the predicate head unit (PHU) as the governing verb of the whole clause and the subject complement unit and any predicate complement units as dependent on that head. At the same time, it maintains a linear presentation of the clause units in a way similar to the elements of clause structure of CGEL, for instance. One fundamental difference to many other approaches is that with the exception of the distinction between SCUs and PCUs, the complements of the verb are not subdivided into further classes at this level. CGEL (1985: 2.16) uses terms such as "indirect object" (O i ) and "direct object" (O d ), "subject complement" (C s ) and "object complement" (C o ) (meaning elements referring to the subject or object of a clause), and "adverbial" (A) to establish a distinction between seven clause types. (4) NW384 Robyn thinks S V CGEL SCU PHU this book (5) NW244 ... Robyn telephoned Charles S V O CGEL SCU PHU PCU ' ÆFFECTED ' this book (6) SW3 April is the cruellest month ... S V C s CGEL SCU PHU PCU ' PREDICATIVE ' this book (7) NW323 ... I am in England S V A CGEL SCU PHU PCU ' LOCATIVE ' this book <?page no="180"?> Analysis of sentences 168 (8) NW223 I told you that S V O i O d CGEL SCU PHU PCU ' BENREC ' PCU ' ÆFFECTED ' this book (9) SW27 I wouldn't call myself a structuralist S V O C o CGEL SCU PHU PCU ' ÆFFECTED ' PCU ' PREDICATIVE ' this book (10) NW47 She puts her soiled breakfast things into the sink ... S V O A CGEL SCU PHU PCU ' ÆFFECTED ' PCU ' LOCATIVE ' this book These clause types are then taken in CGEL and related approaches as the basis for distinguishing between different verb classes: intransitive: SV (4) NW384 Robyn thinks. transitive: monotransitive: SVO (5) NW244 When she got home, Robyn telephoned Charles. ditransitive: SVOO (8) NW223 I told you that. complex transitive: SVOC and SVOA (9) SW27 I wouldn't call myself a structuralist. (10) NW47 She puts her soiled breakfast things into the sink ... copular: SVC and SVA (6) SW3 April is the cruellest month ... (7) NW323 I always know when I 'm in England ... Two objections to this classification immediately spring to mind: firstly, it is odd that seven clause types should be reflected by five types of verbs, and it will be shown in the next section that this may indeed have to do with major problems of classification. Secondly, in view of the enormous complexity of valency relations and the <?page no="181"?> Alternative accounts of clause structure 169 optionality of complements, it is at best possible to talk about different verb uses in this context but not about different classes of verbs. 3 8.2.2 Advantages and disadvantages of traditional terminology One major problem of the CGEL terminology is that it does not keep the levels of formal and semantic analysis apart, which tends to lead to problems of classification. Thus CGEL (1985: 10.8) defines subject and object complements by a number of criteria such as (a) FORM The complement is normally a noun phrase or an adjective phrase, but it may also be a nominal clause ... ... (d) SEMANTIC PROPERTIES The complement typically identifies or characterizes the referent of the clause element to which it is related. ... This means that while the adjective phrase unemployed in (11) NW214 Robyn Penrose was not unemployed - yet. can be classified as a "subject complement" (in the sense of CGEL), out of work in (12) NW113 They look around this area, they see the factories that have closed in the past few years, they know how many people are out of work. is not covered by the criteria listed in CGEL (1985: 10.8) for "subject complements". On the other hand, classifying out of work as an adverbial is unsatisfactory, too, because it does not meet the semantic criteria given for adverbials in CGEL (1985: 10.9) since it hardly "refers to the circumstances of the situation" etc. 4 The same applies to cases such as (13) NW50 Universities everywhere were in disarray, faced with swingeing cuts in their funding. A particularly problematical category is that of the object, which is of course a term that is used widely in linguistics. Some consider the criterion of passivization as the defining criterion of the category, others take semantic characteristics as the defining property. CGEL (1985: 10.7) includes both criteria in its description: 3 This may be a matter of interpretation, of course. We would not want to treat read in I haven't read a book in months, let alone tried to write one NW324 as a different verb from the read in At home she read ... NW215 . CGEL (1985: I.54) treats changes from transitive to intransitive or from intransitive to transitive as changes of secondary word class under the heading of conversion in the chapter on word formation. 4 This is discussed in CGEL (1985: 10.11) under the heading of "prepositional phrases and adjuncts as complement": "Some clause structures and clause elements can be analysed in more than one way". It is mainly for semantic reasons and because "such prepositional phrases can be coordinated with, or placed in apposition with, adjective phrases that undoubtedly function as complement" that such cases are then classified as subject complements in CGEL. <?page no="182"?> Analysis of sentences 170 (c) SYNTACTIC FUNCTION ... The object of an active clause may generally become the subject of the corresponding passive clause (...) ... (d) SEMANTIC PROPERTIES (i) The direct object typically refers to an entity that is affected by the action denoted in the clause .... (ii) The indirect object typically refers to an animate being that is the recipient of the action... Here, it is particularly the conflict between the passive criterion and semantic properties that is problematic: first of all, there are so-called middle verbs (CGEL 1985: 16.27), where the PCU can be attributed the role of ' ÆFFECTED ' but where no passive is possible: (14) NW60 She has boxes full of notes and file cards on it. Secondly, the word "generally" used in the CGEL description of objects above is a major qualification since quite a number of cases where elements that would presumably be classified as objects cannot easily become subjects of passive clauses: 5 (15) NW80 If we rationalize production as you say, will that mean redundancies? (15a) *... will redundancies be meant? (16) NW30 But never mind common sense for the moment ... (16a) *But common sense should not be minded for the moment ... (17) VDE The president closed his speech today with remarks directed to the Russian people. (17a) *The speech was closed with remarks to the Russian people. These examples may suffice to show that definitions of categories which draw upon both formal and semantic criteria run into difficulties. In construction grammar and similar frameworks, the term object seems to receive a purely semantic interpretation: 6 5 Compare, for instance, the category of benefactive objects introduced by Aarts and Aarts (1982/ 1988: 140) that cannot usually be passivized. For problems of the passive criterion see also Jespersen (1927: 232). 6 See Croft and Cruse (2004: 279-280): "Cognitive Grammar argues that fundamental syntactic categories such as Noun, Verb, Subject and Object are abstract (schematic) semantic construals of the conceptual content of their denotations. Thus, fundamental syntactic categories have an essentially semantic basis, but in terms of the construal of experience, not in terms of semantic classes." For passivizability see also Croft (2001: esp. 46 and 48-49): "I contend that if one takes all distributional patterns seriously, then one would end up with a set of features each of which is in essence indexed whether or not the word (or phrase) in question occurs in each role in each construction. For each mismatch of distributions across constructions, the mismatching words would somehow have to be indexed with a feature that in effect blocks its occurrence in one construction or permits its occurrence in the other construction. For instance, one might attach a feature to the representation of the Object of the English verb weigh that permits it to occur without a Preposition (on the assumption that it is an anomalous Oblique), or prevents it from being passivized (on the assumption that it is an anomalous Direct Object). ... For example, the Object of weigh would, <?page no="183"?> Alternative accounts of clause structure 171 Matthews (1981: 98) arrives at a purely semantic definition of the three verb classes featuring prominently in traditional accounts - transitive, intransitive and copula: 7 The predicates are of three types, in each of which a verb (...) is an essential element. In the first the verb has an object ... In that case the construction is described as transitive: the act referred to by the verb (...) is said to pass across (Latin 'transire') from an actor to a goal. ... The term is also applied to the verbal lexeme. Thus KILL is an inherently transitive verb, its sense implying a person or thing to whom the killing is done as well as a person or thing who is doing it. In the second type of predicate the verb is a copula accompanied by a noun or an adjective. ... In this case no action is referred to, and the verb is merely a linking element (the meaning of the Latin 'copula') between this second element and the subject. The noun or adjective is then described as predicative, or is said to stand in predicative position. ... In the third type of predicate the verb may again refer to an action; but there is no goal to which it is directed. ... In this case the construction is described as intransitive, the verb being accompanied at most by adverbials. However, if the distinction between objects, predicatives and adverbials is one that is made at the level of semantics, then this is not entirely unproblematic since the subject can be defined in purely formal terms as an element of clause structure like predicate and adjuncts. Furthermore, if the distinction is a semantic one, there is no need to distinguish between different types of predicate clause units since they receive a semantic characterization in terms of a role specification anyway. It is for this reason that we like, for instance, Halliday (1994: 26) - will not make use of the term object at all. However, our system of describing PCUs can easily be translated into such terminology: If one takes passivization as a defining criterion for the category object, all complements with the subscript pass-subj will correspond to objects. If one chooses a semantic definition of object, all PCUs marked ' BENREC ' that are realised by noun phrases correspond to indirect objects, whereas those marked ' ÆFFECTED ' correspond to direct objects. 8 PCUs with the semantic role ' PREDICATIVE ' correspond to what many grammars call predicatives and what is called "subject complement" and "object complement" in CGEL. 9 among other things, have the features which essentially reduce to <weigh <obj = +ActiveDirectObject, - Passive Subject>>." 7 For a discussion of the terms direct and indirect object see Matthews (1981: 124-130). For the role of the passive criterion see also CamG (2002: 246). 8 This of course depends on the definition of these roles. Depending on the role inventory, there might be further roles that would fall under the category of direct object. This also concerns the classification of cases such as She had plenty of work... NW214 , in which it is debatable whether plenty of work should be classified as ' ÆFFECTED ' or not and whether despite the lack of a corresponding passive it should be considered a direct object. Semantic roles can also be used to translate this system into the terminology used by Aarts/ Aarts (1982/ 1988: 136-143), who distinguish between direct object, indirect object and benefactive object and introduce a category such as predicator complement for cases such as The boy resembles his father. - <?page no="184"?> Analysis of sentences 172 Whether one finds terms such as object and predicate useful or not depends, to a certain extent at least, on whether one takes a grammatical or a lexical perspective. For the grammarian interested in generalizations, it is worth noting that the two classes of complement that are traditionally distinguished as object and predicative share not only semantic, but also certain syntactic properties: 10 (a) Objects and predicatives can be realised by [NP]-complements, but only predicatives can be realised by [AdjP]-complements. (b) Most objects can be subjects of corresponding passive clauses, predicatives can never be subjects of corresponding passive clauses. When the individual lexical unit used in actual language text is considered, such generalizations are relatively irrelevant, however. Valency descriptions such as those provided by VDE indicate in detail whether a particular valency slot of a verb can be realised by an [NP]-complement and an [AdjP]-complement or whether or not parallel passive constructions exist. Since in English there does not seem to be a cluster of formal properties shared by all direct objects, indirect objects or predicatives, the terms have no predictive power in the sense that if something is classified as an object it will necessarily also be able to occur as the subject of a passive clause in a corresponding pattern. It seems to us that the semantic and syntactic properties can be covered more appropriately by distinguishing systematically between the formal and semantic properties of the complements. Even if criteria at these levels are also subject to gradience - this applies to the acceptability of passive constructions as well as to the identification of semantic roles - keeping the levels apart at least reduces the vagueness in the analysis by not making use of prototypical categories that combine the problems caused by gradience at different levels of description. 9 The same applies to the terms subject attribute and object attribute as defined by Aarts/ Aarts (1988: 140-142 and 158-161). 10 Compare, for instance, the list of defining properties given in CGEL (1985: 10.7-8) or CamG (2002: 244-263). See also Huddleston (1984: 181-194), in particular also for a discussion of the problem of number concord between subject and predicative. <?page no="185"?> 9 Analytic framework 9.1 Steps in the analysis of sentences On the basis of the framework developed in the previous chapters of this book, we would now like to outline an approach to the syntactic analysis of sentences, which will consist of eight steps. We will illustrate these steps using the following sentences: (1) NW215 But this year the winter term was different. (2) NW325 Send me your manuscript and I'll have a look at it. (3) NW215 Her only chance of staying in academic life was to build up an irresistibly impressive record of research and publication. (4) NW361 You will write me a reference, then? (5) NW41 This morning, she gives priority to the fact that it is the first day of the winter term .... (6) NW47 She puts her soiled breakfast things into the sink, already crammed with the relics of last night's supper, and hurries upstairs. (7) NW325 Finally he asked her if she had made arrangements to publish her book in America. Step I Identification of sentences The first step in the analysis of a sentence is to determine whether the sentence can be analysed as consisting of two or more sentences (> 2.5) or whether it is a single sentence. Each of the examples (1-7), with the exception of (2), represents a single independent clause, i.e. a sentence. In (2), however, we have a case of coordination of two sentences linked by the conjunction and. sentence (1) NW325 But this year the winter term was different. sentence <I> sentence <II> (2) NW325 Send me your manuscript and I'll have a look at it. sentence (3) NW315 Her only chance of staying in academic life was to build up an irresistibly impressive record of research and publication. Note on coordination: Roman numerals will be used to indicate cases of coordination: thus sentence <I> + conjunction + sentence <II> indicates a sentence that contains two sentences in coordination. Similarly, NP <I> + conjunction + NP <II> indicates a <?page no="186"?> Analytic framework 174 noun phrase that consists of two coordinated noun phrases. Coordinated elements will have to be identified at every level of analysis. 1 Step II Identification of sentence types The second step is to identify the sentence type for each sentence. Most of the sentences (1-7) represent declarative-'statement'-constructions, (2) consists of two sentences, one of which is an imperative-'directive'-construction, and (4) is an example of a declarative-'question'-construction: sentence: declarative-'statement' (1) NW325 But this year the winter term was different. sentence <I> : imperative-'directive' sentence <II> : declarative-'statement' (2) NW325 Send me your manuscript and I'll have a look at it. sentence: declarative-'statement' (3) NW315 Her only chance of staying in academic life was to build up an irresistibly impressive record of research and publication. sentence: declarative-'question' (4) NW361 You'll write me a reference, then? Step III Identification of clause units The next step is to identify the clause units. We suggest a linear analysis, in which the constituents we have called clause units are identified, namely the subject complement unit: SCU the predicate head unit: PHU predicate complement units: PCU adjunct units: AU linking units: LU This means that the level of clause elements - subject, predicate and adjuncts - is only indirectly represented: since it is clear that PHU and PCUs are to be seen as constituents of the predicate, however, no separate hierarchical level is indicated in the analysis. 1 Thus, in (3) the particle phrase of research and publication consists of the particle of and a coordinated noun phrase complement [NP <I> ] research + conjunction and + [NP <II> ] publication. <?page no="187"?> Steps in the analysis of sentences 175 sentence: declarative-'statement' (1) NW325 But this year the winter term was different. LU AU SCU PHU PCU sentence <I> : imp.-'directive' sentence <II> : decl.-'statement' (2) NW325 Send me your manuscript and I ll' have a look at it. PHU PCU1 PCU2 LU SCU PHU PCU Step IV Valency analysis: description of valency patterns Step IV consists of an identification of the valency pattern (> 6.7) exemplified by the sentence. This means describing the verbal head complex and the complement units. The verbal head complex will be specified with respect to active and passive voice: VHC act or VHC pass the number of valency complements of the head realized in the clause: VHC act: 1 , VHC pass: 2 , VHC act: 3 , etc. 2 The description of complements involves in particular: indicating the formal properties of the complement [NP], [AdjP], [that_CL] etc. 3 indicating whether a particle phrase represents a specific particle phrase of the type [to_NP], [about_NP], etc., where the particle is determined by the valency of the verb or an unspecified particle phrase [PartP], in which no specific particle is specified in the valency characterization of the verb. sentence: declarative-'statement' (5) NW41 This morning she gives priority to the fact that it is the first day of the winter term .... AU SCU: NP PHU: VHC act: 3 PCU1: NP ÆFFECTED PCU2: to_NP 2 Note that in the case of dependent clauses, the quantitative valency of the VHC will be specified with respects to the number of complements actually realized in the dependent clause itself. Thus, in the case of (3), to build up an irresistibly impressive record of research and publication is a [to_INF]-complement of was; the valency of build is thus analysed as 2 on the grounds that two complements - up and an irresistibly impressive record of research and publication - depend on it. 3 One could indicate further whether complements can function as the subjects of active or passive clauses, which (for reasons of simplicity) will not be done in this framework although it would be a requirement of a full valency description of the verb. <?page no="188"?> Analytic framework 176 sentence: declarative-'statement' (6) NW41 She puts her soiled breakfast things into the sink ... SCU: NP PHU: VHC act: 3 PCU1: NP ÆFFECTED PCU2: PartP The valency description of complement units also includes the indication of participant roles. Although, in principle, participant roles can be attributed to any valency complement, in this kind of analysis we suggest a simplified approach: firstly, participant roles will only be indicated for predicate complement units realized by noun phrases secondly, only a limited number of role labels will be used. These simplifications seem justified if we consider it the main purpose of the indication of these roles that they make clear semantic differences between sentences such as 4 (4) NW361 You will write me a reference, then? (8) SW27 I wouldn't call myself a structuralist. For this purpose it seems sufficient to make use of the following role labels: ' PREDICATIVE ' (9) SW0 Rummidge is not Birmingham, although it owes something to popular prejudices about that city. Of course, the role ' PREDICATVE ' also applies to PCUs such as (3) NW215 Her only chance of staying in academic life was to build up an irresistibly impressive record of research and publication. but since this is not a noun phrase realization, it will not be indicated in the analysis. ' BENREC ' (2) NW325 Send me your manuscript and I'll have a look at it. (7) NW325 Finally he asked her if she had made arrangements to publish her book in America. ' ÆFFECTED ' (4) NW361 You will write me a reference, then? (2) NW325 Send me your manuscript and I'll have a look at it. (3) NW215 Her only chance of staying in academic life was to build up an irresistibly impressive record of research and publication. 4 This can be justified on the grounds that the semantic function of subject complement units can be described generally in the terms outlined in Chapter 7. See, for instance, CamG (2002: 229, 235- 236). Similarly, one can argue that [PartP]-complements and specific particle phrase complements very often have a limited repertoire of semantic roles they can express so that their specification is not necessary here. <?page no="189"?> Steps in the analysis of sentences 177 (10) NW215 At some deeper level of feeling and reflection she derived a subtle satisfaction from her association with the factory, and a certain sense of superiority over her friends. If a predicate complement unit realized by a noun phrase cannot be assigned to any of the roles above, it will be labelled ' OTHER ' to make clear that a more specific semantic role must be used. (11) SW0 The modern conference resembles the pilgrimage of medieval Christendom in that it allows the participants to indulge themselves in all the pleasures and diversions of travel while appearing to be austerely bent on self-improvement. sentence: declarative-'question' (4) NW361 You will write me a reference then? SCU: NP PHU: VHC act: 3 PCU1: NP BENREC PCU2: NP ÆFFECTED AU sentence: declarative-'statement' (7) NW325 Finally he asked her if she made arrangements to publish her book in America. AU SCU: NP PHU: VHC act: 3 PCU1: NP BENREC PCU2: if_CL Step V Formal classification of adjunct units Step V, which can be carried out simultaneously with step IV, consists of identifying the formal constituents of the adjunct units in terms of the type of phrase or clause in which they are realized in the sentence, i.e. particle phrases will be classified as PartP. sentence: declarative-'question' (4) NW361 You will write me a reference then? SCU: NP PHU: VHC act: 3 PCU1: NP BENREC PCU2: NP ÆFFECTED AU: AdvP sentence: declarative-'statement' (5) NW41 This morning she gives priority to the fact that it is the first day of the winter term .... AU: NP SCU: NP PHU: VHC act: 3 PCU1: NP ÆFFECTED PCU2: to_NP Step VI Identification of constituents The result of applying steps I to V is that all SCUs and PCUs of a sentence will have received a description in terms of a phrase or a clause. In the case of clauses, steps III to V will have to be repeated until all the resulting constituents are phrases. <?page no="190"?> Analytic framework 178 Step VII Description of phrase structure Phrases and verbal head complexes will be described in terms of the categories established in Chapter 4: head (h) pre-head (preh) premodifier (prem) postmodifier (postm) valency complement of head (c) sentence: declarative-'statement' (1) NW325 But this year the winter term was different. LU AU: NP SCU: NP PHU: VHC act: 2 PCU: AdjP preh h preh h h h Step VIII Indication of word class The categories identified in VII will then have to be described in terms of the formal constituents by which they are realized. If this is a single word, the word class will be indicated. If it is a phrase or clause, the analytical process has to be repeated from step III or step VI onwards. 5 sentence: declarative-'statement' (1) NW325 But this year the winter term was different. LU AU: NP SCU: NP PHU: VHC act: 2 PCU: AdjP conj preh: detpron h: ln preh: det h: ln h: pv h: adj 5 If a premodifier or a postmodifier consists of only one word, only the word class will be indicated (i.e. we will not explicitly indicate that this word represents a particular type of phrase of which it is the head). <?page no="191"?> Examples of syntactic analysis 179 9.2 Examples of syntactic analysis 6 (1) NW215 But this year the winter term was different. sentence: declarative-'statement' LU: conj But AU: NP preh: detpron this h: ln year SCU: NP preh: det the h: ln winter term PHU: VHC act: 2 h: pv was PCU: AdjP h: adj different. (2) NW325 Send me your manuscript and I'll have a look at it. sentence <I> : imperative-'directive' PHU: VHC act: 2 h: lv Send PCU1: NP BENREC h: pron me PCU2: NP ÆFFECTED preh: pron your h: ln manuscript LU: conj and sentence <II> : declarative-'statement' SCU: NP h: pron I PHU: VHC act: 2 preh: mv 'll h: pv have PCU: NP ÆFFECTED preh: det a h: ln look c: at_NP h: part at c: NP h: pron it. 6 This section contains suggestions for the analysis of 25 sentences on the basis of the principles outlined in 9.1. Compare the schema for the structural description of clauses used by Aarts/ Aarts (1982/ 1988). <?page no="192"?> Analytic framework 180 (3) NW215 Her only chance of staying in academic life was to build up an irresistibly impressive record of research and publication. sentence: declarative-'statement' SCU: NP preh: pron Her prem: adj/ special use only 7 h: ln chance c: of_V-ing h: part of c: V-ing PHU: VHC act: 1 h: lv staying PCU: PartP h: part in c: NP prem: adj academic h: ln life PHU: VHC act: 2 h: pv was PCU: to_INF h: part to c: INF PHU: VHC act: 2 h: lv build phr. PCU1: up h: part up 8 verb PCU2: NP ÆFFECTED preh: det an prem: AdjP prem: adv irresistibly h: adj impressive h: ln record c: of_NP h: part of NP <I> h: ln research c: NP conj and NP <II> h: ln publication. (4) NW361 You will write me a reference, then? sentence: declarative-'question' SCU: NP h: pron You PHU: VHC act: 3 preh: mv will h: lv write PCU1: NP BENREC h: pron me PCU2: NP ÆFFECTED preh: det a h: ln reference, AU: AdvP h: adv then? 7 This can be regarded either as an adjective or as a special use of only, which is typically an adverb. Only does not meet any of the defining properties of adjectives except criterion A (predicative use). 8 This presentation is meant to indicate that the phrasal verb build up is a phraseological unit, which consists of the verb build and the particle up, which fills one of the valency slots of build (> 6.8.3.2). <?page no="193"?> Examples of syntactic analysis 181 (5) NW41 This morning, she gives priority to the fact that it is the first day of the winter term. sentence: declarative-'statement' AU: NP preh: detpron This h: ln morning, SCU: NP h: pron she PHU: VHC act: 3 h: lv gives PCU1: NP ÆFFECTED h: ln priority PCU2: to_NP h: part to c: NP preh: det the h: ln fact c: that_CL h: part that c: CL fin SCU: NP h: pron it 9 PHU: VHC act: 2 h: pv is PCU: NP PREDICATIVE preh: det the preh: detpron first h: ln day c: of_NP h: part of 10 c: NP preh: det the h: ln winter term. (7) NW325 Finally he asked her if she had made arrangements to publish her book in America. sentence: declarative-'statement' AU: AdvP h: adv Finally SCU: NP h: pron he PHU: VHC act: 3 h: lv asked PCU1: NP BENREC h: pron her PCU2: if_CL h: part if c: CL fin SCU: NP h: pron she PHU: VHC act: 2 preh: pv had h: lv made PCU: NP ÆFFECTED h: ln arrangements c: to_INF h: part to c: INF PHU: VHC act: 1 h: lv publish PCU: NP ÆFFECTED preh: pron her h: ln book AU: PartP h: part in c: NP h: ln America. 9 Impersonal uses of it are regarded as complements because they fill a valency slot of the verb. 10 Of the winter term modifies first rather than day and can thus also be seen as a complement of first. <?page no="194"?> Analytic framework 182 (6) NW47 She puts her soiled breakfast things into the sink, already crammed with the relics of last night's supper, and hurries upstairs. sentence: declarative-'statement' SCU: NP h: pron She PHU: VHC act: 3 h: lv puts PCU1: NP ÆFFECTED preh: pron her <I> prem: adj soiled prem: ln breakfast h: ln things PCU2: PartP h: part into c: NP preh: det the h: ln sink, postm: AdjP 11 prem: adv already h: adj crammed c: with_NP h: part with c: NP preh: det the h: ln relics c: of_NP h: part of c: NP preh: NP preh: detpron last 12 h: ln night's h: ln supper, LU: conj and <II> PHU: VHC act: 2 h: lv hurries PCU: PartP h: part upstairs. 11 Note that the postmodifier already crammed with the relics of last night’s supper could not be analysed as a [V-ed]-clause with crammed as predicate head unit (PHU) and the particle phrase headed by with as a predicate complement unit (PCU), since cram does not take a [with_NP]complement when used as a verb in an active construction. Compare ? She crammed the sink with the relics of last night’s supper and She crammed the relics of last night’s supper into the sink. 12 This use of last is classified as a determiner-pronoun because it can also be used in phrases such as the week before last. <?page no="195"?> Examples of syntactic analysis 183 (8) SW27 I wouldn't call myself a structuralist. sentence: declarative-'statement' SCU: NP h: pron I PHU: VHC act: 3 preh: mv would AU: AdvP h: adv n't h: lv call PCU1: NP ÆFFECTED h: pron myself PCU2: NP PREDICATIVE preh: det a h: ln structuralist. (9) SW0 Rummidge is not Birmingham, though it owes something to popular prejudices about that city. sentence: declarative-'statement' SCU: NP h: ln Rummidge PHU: VHC act: 2 h: pv is AU: AdvP h: adv not PCU: NP PREDICATIVE h: ln Birmingham, AU: PartP h: part though c: CL fin SCU: NP h: pron it PHU: VHC act: 3 h: lv owes PCU1: NP OTHER h: pron something PCU2: to_NP h: part to c: NP prem: adj popular h: ln prejudices c: about_NP h: part about c: NP preh: detpron that h: ln city. <?page no="196"?> Analytic framework 184 (10) NW215 At some deeper level of feeling and reflection she derived a subtle satisfaction from her association with the factory, and a certain sense of superiority over her friends. sentence: declarative-'statement' AU: PartP h: part At c: NP preh: detpron some prem: adj deeper h: ln level c: of_NP h: part of c: NP NP <I> h: ln feeling conj and NP <II> h: ln reflection SCU: NP h: pron she PHU: VHC act: 3 h: lv derived PCU1 <I> : NP ÆFFECTED preh: det a prem: adj subtle h: ln satisfaction PCU2: from_NP h: part from c: NP preh: pron her h: ln association c: with_NP h: part with c: NP preh: det the h: ln factory, LU: conj and PCU1 <II> : NP ÆFFECTED preh: det a prem: adj certain h: ln sense c: of_NP h: part of c: NP h: ln superiority c: over_NP h: part over c: NP preh: pron her h: ln friends. <?page no="197"?> Examples of syntactic analysis 185 (11) SW0 The modern conference resembles the pilgrimage of medieval Christendom in that it allows the participants to indulge themselves in all the pleasures and diversions of travel while appearing to be austerely bent on self-improvement. sentence: declarative-'statement' SCU: NP preh: det The prem: adj modern h: ln conference PHU: VHC act: 2 h: lv resembles PCU: NP OTHER13 preh: det the h: ln pilgrimage c: of_NP h: part of c: NP prem: adj medieval h: ln Christendom AU: PartP h: part in that 14 c: CL fin SCU: NP h: pron it PHU: VHC act: 3 h: lv allows PCU1: NP BENREC preh: det the h: ln participants PCU2: to_INF h: part to c: INF PHU: VHC act: 2 h: lv indulge PCU1: NP BENREC h: pron themselves PCU2: in_NP h: part in c: NP preh: detpron all preh: det the h <I> : ln pleasures h conj and h <II> : ln diversions c: of_NP h: part of c: NP h: ln travel AU: PartP h: part while c: V-ing PHU: VHC act: 1 h: lv appearing PCU: to_INF h: part to c: INF PHU: VHC act: 1 h: pv be PCU: AdjP prem: adv austerely h: adj bent 15 c: on_NP h: part on c: NP h: ln selfimprovement. 13 Although the participant role of the pilgrimage of medieval Christendom is similar to that of ' PREDICATIVE ' it is probably more appropriate to describe in terms of " RESEMBLEE " or " ITEM OF COMPARISON " (> 6.5 and 9.1). 14 Note that in that is analysed as a complex particle (> 3.8.4). 15 Note that it is not possible to analyse be...bent as a passive verb form constituting a predicate head unit (PHU) of [INF]-clause since there is no corresponding active pattern: *They bend him on selfimprovement. CGEL refers to such constructions as pseudo-passives (cf. CGEL: 3.77). <?page no="198"?> Analytic framework 186 (12) NW326 "Could you drop it here tomorrow morning before eight-thirty? " sentence: interrogative-'yes-no-question' PHU: VHC act: 3 preh: mv "Could SCU: NP h: pron you h: lv drop PCU1: NP ÆFFECTED h: pron it PCU2: PartP h: part here AU: NP h: pron tomorrow 16 postm: ln morning postm: PartP h: part before c: NP h: detpron eight-thirty? " (13) NW327 She considered the effort would be worthwhile if she could secure the endorsement of a prestigious imprint like Euphoric State University Press. sentence: declarative-'statement' SCU: NP h: pron She PHU: VHC act: 2 h: lv considered PCU: CL fin SCU: NP preh: det the h: ln effort PHU: VHC act: 2 preh: mv would h: pv be PCU: AdjP h: adj worthwhile AU: PartP h: part if c: CL fin SCU: NP h: pron she PHU: VHC act: 2 preh: mv could h: lv secure PCU: NP ÆFFECTED preh: det the h: n endorsement c: of_NP h: part of c: NP preh: det a prem: adj prestigious h: ln imprint postm: PartP h: part like c: NP h: ln Euphoric State University Press. 17 16 This analysis reflects the fact that tomorrow can realise AU on its own here. Alternatively, one could consider tomorrow morning to be a phrase similar to this morning. 17 Euphoric State University Press is analysed as a complex proper noun (> 4.2.4). <?page no="199"?> Examples of syntactic analysis 187 (14) NW327 Robyn sat up all night printing out her book. sentence: declarative-'statement' SCU: NP h: pn Robyn PHU: VHC act: 2 h: lv sat phrasal PCU: up h: part up verb AU: NP preh: detpron all h: ln night AU: V-ing PHU: VHC act: 2 h: lv printing phrasal PCU1: out h: part out verb PCU2: NP ÆFFECTED preh: pron her h: ln book. (15) NW324 I spend all my time on committees arguing about how to respond to the cuts. sentence: declarative-'statement' SCU : NP h: pron I PHU: VHC act: 4 h: lv spend PCU1: NP ÆFFECTED preh: detpron all preh: pron my h: ln time PCU2: on_NP h: part on c: NP h: ln committees PCU3: V-ing PHU: VHC act: 1 h: lv arguing 18 PCU: about_wh_to_INF h: part about c: wh_to_INF AU: AdvP h: adv how 19 h: part to c: INF PHU: VHC act: 1 h: lv respond PCU: to_NP h: part to c: NP preh: det the h: ln cuts. 18 Since spend can occur both in the valency pattern [NP + verb + NP + on_NP] and [NP + verb + NP + V-ing] in trivalent uses both [on_NP] and [V-ing] are analysed as complements (> Chapter 6). 19 Note that despite its position in front of the particle to, the wh-adverb how functions as an adjunct unit (AU) within the [INF]-clause respond to the cuts, which serves as a complement to the particle to. The [to_INF] (particle phrase headed by to) is again a complement of the particle about. (> 5.3.6). <?page no="200"?> Analytic framework 188 (16) NW50 Universities everywhere were in disarray, faced with swingeing cuts in their funding. sentence: declarative-'statement' SCU: NP h: ln Universities postm: part everywhere PHU: VHC act: 2 h: pv were PCU: PartP h: part in c: NP h: ln disarray, AU: V-ed PHU: VHC act: 1 h: lv faced 20 PCU: with_NP h: part with c: NP prem: adj swingeing h: ln cuts c: in_NP h: part in c: NP preh: pron their h: ln funding. (17) NW184 I want to teach women's studies and post-structuralism and the nineteenth-century novel and write books about them. sentence: declarative-'statement' SCU: NP h: pron I PHU: VHC act: 2 h: lv want PCU: to_INF h: part to c <I> : INF PHU: VHC act: 1 h: lv teach PCU: NP ÆFFECTED NP <I> prem: ln women's h: ln studies conj and NP <II> h: ln post-structuralism conj and NP <III> preh: det the prem: ln nineteenthcentury 21 h: ln novel conj and c <II> : INF PHU: VHC act: 2 h: lv write PCU1: NP ÆFFECTED h: ln books PCU2: about_NP h: part about 22 c: NP h: pron them. 20 Alternatively, the [V-ed]-clause faced with swingeing cuts in their funding could be seen as a predicate complement unit (PCU) in this sentence, coordinated with the PCU in disarray. 21 Nineteenth-century is here analysed as a compound noun because of its orthography. Alternatively, one could take it as an NP headed by the lexical noun century with nineteenth (word-class detpron) as its pre-head. 22 Alternatively, about them could also be analysed as a complement of books. <?page no="201"?> Examples of syntactic analysis 189 (18) NW56 So Charles moved his books and other possessions to Ipswich, which Robyn found rather inconvenient, since she was in the habit of borrowing his books, and occasionally his sweaters. sentence: declarative-'statement' AU: AdvP h: adv So SCU: NP h: ln Charles PHU: VHC act: 3 h: lv moved PCU1: NP ÆFFECTED preh: pron his h <I> : ln books conj and preh <II> : detpron other h <II> : ln possessions PCU2: to_NP h: part to c: NP h: ln Ipswich, AU: CL fin23 PCU1: NP ÆFFECTED h: pron which SCU: NP h: ln Robyn PHU: VHC act: 3 h: lv found PCU2: AdjP prem: adv rather h: adj inconvenient, AU: PartP h: part since c: CL fin SCU: NP h: pron she PHU: VHC act: 2 h: pv was PCU: PartP h: part in c: NP preh: det the h: ln habit c: of_V-ing h: part of c: V-ing PHU: VHC act: 1 h: lv borrowing PCU <I> : NP ÆFFECTED preh: pron his h: ln books, conj and AU <II> : AdvP h: adv occasionally PCU <II> : NP ÆFFECTED preh: pron his h: ln sweaters. 23 Note that this clause is a relative clause (> 5.3.7). <?page no="202"?> Analytic framework 190 (19) SW0 Only, these days, professional people call them conferences. sentence: declarative-'statement' AU: AdvP h: adv Only, AU: NP preh: detpron these h: ln days, SCU: NP prem: adj professional h: ln people PHU: VHC act: 3 h: lv call PCU1: NP ÆFFECTED h: pron them PCU2: NP PREDICATIVE h: ln conferences. (20) NW326 How shall I get the manuscript to you? sentence: wh-interrogative-'question' AU: AdvP h: adv How PHU: VHC act: 3 preh: mv shall SCU: NP h: pron I h: lv get PCU1: NP ÆFFECTED preh: det the h: ln manuscript PCU2: to_NP h: part to c: NP h: pron you? (21) NW218 They don't pay us all that much, anyway. sentence: declarative-'statement' SCU: NP h: pron They PHU: VHC act: 3 preh: pv do AU: AdvP h: adv n't h: lv pay PCU1: NP BENREC h: pron us PCU2: NP ÆFFECTED preh: detpron all preh: adv that 24 h: detpron much, AU: AdvP h: adv anyway. 24 Note that that in the NP all that much has a premodifying function and is thus analysed as an adverb (> 3.3.3). <?page no="203"?> Examples of syntactic analysis 191 (22) NW52 Thus a three-year lectureship in English literature was advertised, Robyn applied, was interviewed along with four other equally desperate and highly qualified candidates, and was appointed. sentence <I> : declarative-'statement' AU: AdvP h: adv Thus SCU: NP preh: det a prem: ln three-year 25 h: ln lectureship c: in_NP h: part in prem: adj English h: ln literature PHU: VHC pass: 1 preh: pv was h: lv advertised, sentence <II> : declarative-'statement' SCU: NP h: ln Robyn <I> PHU: VHC act: 1 h: lv applied, <II> PHU: VHC pass: 1 preh: pv was h: lv interviewed AU: PartP h: part along with 26 c: NP preh: detpron four preh: detpron other prem: AdjP AdjP <I> prem: adv equally h: adj desperate conj and AdjP <II> prem: adv highly h: adj qualified h: ln candidates, LU: conj and <III> PHU: VHC pass: 1 preh: pv was h: lv appointed. 25 Three-year is analysed as a compound noun because of its orthography. Alternatively one could take it as an NP headed by the lexical noun year with three (word-class detpron) as its prehead. 26 We take along with as a complex particle (> 3.8.4). <?page no="204"?> Analytic framework 192 (23) NW325 If by any chance there should be a job advertised in the next twelve months, I ought to stand a fair chance of getting it, with two books on my record. sentence: declarative-'statement' AU: PartP h: part If c: CL fin AU: PartP h: part by c: NP preh: detpron any h: ln chance SCU: PartP h: part there PHU: VHC act: 2 preh: mv should h: pv be PCU: NP_V-ed SCU: NP preh: det a h: ln job PHU: VHC pass: 1 h: lv advertised AU: PartP h: part in c: NP preh: det the preh: detpron next preh: detpron twelve h: ln months, SCU: NP h: pron I PHU: VHC act: 2 preh: v ought to 27 h: lv stand PCU: NP ÆFFECTED preh: det a prem: adj fair h: ln chance c: of_V-ing h: part of c: V-ing PHU: VHC act: 1 h: lv getting PCU: NP ÆFFECTED h: pron it, AU: PartP h: part with 28 c: NP preh: detpron two h: ln books c: PartP h: part on c: NP preh: pron my h: ln record. 27 Since ought does not meet all criteria of modal verbs (> 3.2.2.4), it is only classified as verb here; ought to is seen as an idiomatic unit functioning as pre-head. Alternatively, ought could be analysed as a verb with a [to_INF]-complement. 28 With is classified as a divalent particle here. Otherwise, one could see this as a quasi-clause (> 5.5). <?page no="205"?> Examples of syntactic analysis 193 (24) NW324 ‘That’s what life is like in British universities these days, Morris,’ said Philip Swallow, presenting Robyn with a glass of rather warm Soave. sentence: declarative-'statement' PCU: QUOTE 29 SCU: NP h: detpron 'That PHU: VHC act: 2 h: pv 's PCU: wh_CL c: NP h: detpron what SCU: NP h: ln life PHU: VHC act: 2 h: pv is PCU 30 : PartP h: part like AU: PartP h: part in c: NP prem: adj British h: ln universities AU: NP preh: detpron these h: ln days, AU: NP h: ln Morris,' 31 PHU: VHC act: 2 h: lv said SCU: NP h: ln Philip Swallow 32 , AU: V-ing PHU: VHC act: 2 h: lv presenting PCU1: NP BENREC h: ln Robyn PCU2: with_NP h: part with c: NP preh: det a h: ln glass c: of_NP h: part of c: NP prem: AdjP prem: adv rather h: adj warm h: ln Soave. (25) NW324 ‘You leaving Rummidge, Robyn? ’ said Morris Zapp. sentence: declarative-'statement' PCU: QUOTE 33 SCU: NP h: pron 'You PHU: VHC act: 2 h: lv leaving PCU: NP ÆFFECTED h: ln Rummidge, AU: NP h: ln Robyn? 34 PHU: VHC act: 2 h: lv said SCU: NP h: ln Morris Zapp. 35 29 In cases where complements described as QUOTE are sentences, the sentence type can be identified. ‘That’s what life is like in British universities these days, Morris,’ is a declarative- 'statement'-construction. 30 This is a case of a discontinuous realization of a predicate complement unit. Compare: Life is like this in British Universities (> 6.4.2.8). 31 The adjunct unit (AU) Morris has the semantic function of a ' VOCATIVE ' (> 7). 32 Philip Swallow is taken to be a complex proper noun (> 4.2.4). 33 ‘You leaving Rummidge, Robyn? is a context-bound interrogative 'yes-no-question'-construction (> 7.1.2.2). 34 The participant role of the adjunct unit Robyn can be described as ' VOCATIVE ' (> 7). 35 Morris Zapp is analysed as a complex proper noun (> 4.2.4). <?page no="206"?> Bibliography Aarts, Flor/ Jan Aarts (1982/ 1988): English Syntactic Structures. New York/ Leyden: Prentice Hall/ Martinus Nijhoff. Ágel, Vilmos (2000): Valenztheorie. Tübingen: Narr. Allerton, David J. (1982): Valency and the English Verb. London: Academic Press. Allerton, David J. (1990): 'Language as form and pattern: grammar and its categories', in: Neville E. Collinge (ed.): An Encyclopaedia of Language. London/ New York: Routledge, 68-111. Behrens, Heike (2005): 'Wortarten-Erwerb durch Induktion', in: Clemens Knobloch/ Burkhard Schaeder (eds.): Wortarten und Grammatikalisierung. Perspektiven in System und Erwerb. Berlin/ New York: Walter de Gruyter. Biber, Douglas/ Stig Johansson/ Geoffrey Leech/ Susan Conrad/ Edward Finegan (1999): Longman Grammar of Spoken and Written English. Harlow: Pearson. Biber, Douglas/ Susan Conrad/ Geoffrey Leech (2002): Longman Student Grammar of Spoken and Written English. Harlow: Pearson. Bloomfield (1933/ 2 1935): Language. New York: Holt/ London: Allen & Unwin Ltd. Brown, Gillian/ George Yule (1983): Discourse analysis. Cambridge: Cambridge University Press. Bublitz, Wolfram (2001): Englische Pragmatik. Berlin: Schmidt. Bühler, Karl (1934/ 1999): Sprachtheorie. Stuttgart: Lucius & Lucius (UTB). Burgschmidt, Ernst/ Dieter Götz (1974): Kontrastive Linguistik Deutsch/ Englisch. München: Hueber. Clark, Herbert H./ Eve V. Clark (1977): Psychology and Language. An Introduction to Psycholinguistics. San Diego: Harcourt Brace Jovanovich. Croft, William (2001): Radical Construction Grammar. Syntactic Theory in Typological Perspective. Oxford: Oxford University Press. Croft, William/ D. Alan Cruse (2004): Cognitive Linguistics. Cambridge: Cambridge University Press. Cruse, David A. (1986): Lexical Semantics. Cambridge: Cambridge University Press. Culler, Jonathan (1976): Saussure. Glasgow: Fontana. de Groot, Albert W. (1949/ 1964): Structurele Syntaxis. Den Haag: Servire. Dirven, René/ Günter Radden (1977): Semantische Syntax des Englischen. Wiesbaden: Athenaion. Dowty, David (1991): 'Thematic proto-roles and argument selection', Language 67, 547-619. <?page no="207"?> Bibliography 195 Eisenberg, Peter (1999): Grundriß der deutschen Grammatik. Band 2. Der Satz. Stuttgart/ Weimar: Metzler. Emons, Rudolf (1974): Valenzen englischer Prädikatsverben. Tübingen: Niemeyer. Emons, Rudolf (1978): Valenzgrammatik für das Englische. Eine Einführung. Tübingen: Niemeyer. Engel, Ulrich (1977): Syntax der deutschen Gegenwartssprache. Berlin: Schmidt. Esser, Jürgen (1979): Englische Prosodie. Tübingen: Gunter Narr Verlag. Esser, Jürgen (1992): 'Neuere Tendenzen in der Grammatikschreibung des Englischen', Zeitschrift für Anglistik und Amerikanistik 40, 112-123. Fillmore, Charles (1968): 'The case for case', in: Emmon Bach/ Robert T. Harms (eds.): Universals in Linguistic Theory. New York: Holt, Rinehart and Winston, 1-88. Fillmore, Charles (1977): 'The case for case reopened', in: Klaus Heger/ János S. Petöfi (eds.): Kasustheorie, Klassifikation, semantische Interpretation. Hamburg: Buske, 3-26. Fillmore, Charles (1988): 'The mechanisms of "construction grammar"', in: Shelley Axmaker/ Annie Jassier/ Helen Singmaster (eds): General Session and Parasession on Grammaticalization. Berkeley Linguistics Society. Proceedings of the Fourteenth Annual Meeting. Berkeley, CA: Berkeley Linguistics Society, 35-55. Fillmore, Charles (2007): 'Valency issues in FrameNet', in: Thomas Herbst/ Katrin Götz- Votteler (eds.): Valency - Theoretical, Descriptive and Cognitive Issues. Berlin/ New York: Mouton de Gruyter, 129-160. Fillmore, Charles/ Christopher R. Johnson/ Miriam R.L. Petruck (2003): 'Background to Framenet', International Journal of Lexicography 16.3, 236-250. Fillmore, Charles/ Paul Kay/ Mary O'Connor (1988): 'Regularity and idiomaticity in grammatical constructions: The case of let alone', Language 64, 501-538. Fischer, Kerstin/ Anatol Stefanowitsch (2006): 'Konstruktionsgrammatik: Ein Überblick', in: Kerstin Fischer/ Anatol Stefanowitsch (eds.): Konstruktionsgrammatik. Von der Anwendung zur Theorie. Tübingen: Stauffenburg, 3-17. Garside, Roger/ Geoffrey Leech/ Geoffrey Sampson (1987): The Computational Analysis of English. London: Longman. Gimson, Alfred C. (revised by Alan Cruttenden) ( 6 2001): Gimson’s Pronunciation of English. London: Hodder Arnold. Goldberg, Adele (2006): Constructions at Work. The Nature of Generalization in Language. Oxford: Oxford University Press. Götz, Dieter/ Thomas Herbst (1989): 'Language description for linguistics and language teaching: The London School and its latest grammar', Die Neueren Sprachen 88.3, 220- 235. Götz-Votteler, Katrin (2007): 'Describing semantic valency', in: Thomas Herbst/ Katrin Götz-Votteler (eds.): Valency. Theoretical, descriptive and cognitive issues. Berlin/ New York: Mouton de Gruyter, 37-49. <?page no="208"?> Bibliography 196 Haegeman, Liliane (1991): Introduction to Government and Binding Theory. Oxford/ Cambridge, Mass: Blackwell. Halliday, Michael A.K. (1967/ 1968): 'Notes on transitivity and theme in English', Journal of Linguistics, Part 1: 37-81, Part 2: 199-244, Part 3: 179-215. Halliday, Michael A.K. (1970): 'Language Structure and Language Function', in: John Lyons (ed.): New Horizons in Linguistics. Harmondsworth: Penguin, 140-165. Halliday, Michael A.K. ( 2 1994): An Introduction to Functional Grammar. London/ Melbourne/ Auckland: Edward Arnold. Hausser, Roland (2006): A Computational Model of Natural Language Communication. Berlin/ Heidelberg: Springer. Helbig, Gerhard (1992): Probleme der Valenz- und Kasustheorie. Tübingen: Niemeyer. Helbig, Gerhard/ Wolfgang Schenkel (1969/ 2 1973): Wörterbuch zur Valenz und Distribution deutscher Verben. Leipzig: Enzyklopädie. Herbst, Thomas (1983): Untersuchungen zur Valenz englischer Adjektive und ihrer Nominalisierungen. Tübingen: Narr. Herbst, Thomas (1988): 'A valency model for nouns in English', Journal of Linguistics 24.2, 265-301. Herbst, Thomas (1999): 'Valency between syntax and lexicology', in: Wolfgang Falkner/ Hans-Jörg Schmid (eds.): Words, Lexemes, Concepts. Approaches to the Lexicon. Studies in Honour of Leonhard Lipka. Tübingen: Narr, 167-173. Herbst, Thomas (2005): 'Englische Grammatik ist nicht so kompliziert: Pro Minimalismus, Lücke und Polysemie - Kontra Prototypik und Semantik in grammatischer Terminologie', in: Thomas Herbst (ed.): Linguistische Dimensionen des Fremdsprachenunterrichts. Würzburg: Königshausen und Neumann, 11-28. Herbst, Thomas (2007): 'Valency complements or valency patterns? ', in: Thomas Herbst/ Katrin Götz-Votteler (eds.): Valency. Theoretical, Descriptive and Cognitive Issues. Berlin/ New York: Mouton de Gruyter, 15-35. Herbst, Thomas (forthcoming 2008): 'Valency - item-specificity and idiom principle', in: Ute Römer/ Rainer Schulze (eds.): Exploring the Grammar-Lexis Interface. Amsterdam/ Philadelphia: John Benjamins. Herbst, Thomas/ Ian Roe (1996): 'How obligatory are obligatory complements? An alternative approach to the categorization of subjects and other complements in valency grammar', English Studies 2, 179-199. Herbst, Thomas/ David Heath/ Hans-Martin Dederding (1980): Grimm's Grandchildren. Current Topics in German Linguistics. London/ New York: Longman. Herbst, Thomas/ David Heath/ Ian Roe/ Dieter Götz (2004): A Valency Dictionary of English. A Corpus-Based Analysis of the Complementation Patterns of English Verbs, Nouns and Adjectives. Berlin/ New York: Mouton de Gruyter. [VDE] Herbst, Thomas/ Michael Klotz (2002): 'Meeting and kissing as valency problems - Some remarks on the treatment of reciprocity and reflexivity in a valency description of English', in: Ekkehard König/ Volker Gast (eds.): Reflexives and Intensifiers: The Use of self-Forms in English. Zeitschrift für Anglistik und Amerikanistik, 3.3, 239-249. <?page no="209"?> Bibliography 197 Herbst, Thomas/ Rita Stoll/ Rudolf Westermayr (1991): Terminologie der Sprachbeschreibung. Ein Lernwörterbuch für das Anglistikstudium. Ismaning: Hueber. Heringer, Hans Joachim (1970): Deutsche Syntax. Berlin: Walter de Gruyter. Heringer, Hans Joachim (1993): 'Basic ideas and the classical model', in: Joachim Jacobs/ Arnim von Stechow/ Wolfgang Sternefeld/ Theo Vennemann (eds.): Syntax. Ein internationales Handbuch zeitgenössischer Forschung. Berlin/ New York: Walter de Gruyter, 298-316. Heringer, Hans Joachim (1996): Deutsche Syntax dependentiell. Tübingen: Stauffenburg. Huddleston, Rodney (1984): Introduction to the Grammar of English. Cambridge: Cambridge University Press. Huddleston, Rodney/ Geoffrey K. Pullum (2002): The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. [CamG] Hudson, Richard (1984): Word Grammar. Oxford: Blackwell. Hülser, Karlheinz (1996): 'Stoa (Beginn ca. 300 v. Chr.)', in: Tilman Borsche (ed.): Klassiker der Sprachphilosophie. München: Beck, 49-62. Jespersen, Otto (1927): Modern English Grammar. On Historical Principles. Part III. Syntax. Second Volume. Heidelberg: Carl Winter Universitätsbuchhandlung. Kastovsky, Dieter (1982): Wortbildung und Semantik. Düsseldorf/ Bern/ München: Basgel/ Francke. Klotz, Michael (2000): Grammatik und Lexik. Studien zur Syntagmatik englischer Verben. Tübingen: Stauffenburg. Kortmann, Bernd (2005): English Linguistics: Essentials. Berlin: Cornelsen. Lamb, Sidney (2001): 'Learning syntax - a neurocognitive approach', in: Martin Pütz/ Susanne Niemeier/ René Dirven (eds.): Applied Cognitive Linguistics I: Theory and Language Acquisition. Berlin/ New York: Mouton de Gruyter, 167-191. Leech, Geoffrey ( 2 1981): Semantics. Harmondsworth: Penguin. Leech, Geoffrey (1983): Principles of Pragmatics. Harlow: Longman. Leech, Geoffrey/ Jan Svartvik ( 2 1994): A Communicative Grammar of English. London/ New York: Longman. Leisi, Ernst/ Christian Mair: ( 8 1999): Das heutige Englisch: Wesenszüge und Probleme. Heidelberg: Winter. Lyons, John (1968): Introduction to Theoretical Linguistics. Cambridge: Cambridge University Press. Lyons, John (1977): Semantics 1. Cambridge: Cambridge University Press. Matthews, Peter (1981): Syntax. Cambridge: Cambridge University Press. Matthews, Peter (1993): Grammatical theory in the United States from Bloomfield to Chomsky. Cambridge: Cambridge University Press. <?page no="210"?> Bibliography 198 Matthews, Peter (2007): 'The scope of valency in grammar', in: Thomas Herbst/ Katrin Götz-Votteler (eds.): Valency. Theoretical, Descriptive and Cognitive Issues. Berlin/ New York: Mouton de Gruyter, 3-14. Mindt, Dieter (2000): An Empirical Grammar of the English Verb System. Berlin: Cornelsen. Muthmann, Gustav (2002): Reverse English Dictionary. Berlin/ New York: Mouton de Gruyter. Nida, Eugene A. (1964): A Synopsis of English Syntax. Norman: Summer Institute of Linguistics of the University of Oklahoma. Palmer, Frank (1971): Grammar. Harmondsworth: Penguin. Palmer, Frank ( 2 1974): The English Verb. London: Longman. Palmer, Frank ( 2 1981): Semantics. Cambridge: Cambridge University Press. Palmer, Frank (1994): Grammatical roles and relations. Cambridge: Cambridge University Press. Paul, Hermann (1919/ 2000): Deutsche Grammatik, Band III. Tübingen: Niemeyer [Auszüge in: Ludger Hoffmann (ed.): Sprachwissenschaft. Ein Reader. 2. Auflage, 2000, 494-502]. Quirk, Randolph/ Sidney Greenbaum/ Geoffrey Leech/ Jan Svartvik (1985): A Comprehensive Grammar of the English Language. London/ New York: Longman. [CGEL] Radford, Andrew (1993): 'Head-hunting: on the trail of the nominal Janus', in: Greville G. Corbett/ Norman M. Fraser/ Scott McClashan (eds.): Heads in Grammatical Theory. Cambridge: Cambridge University Press, 73-113. Robins, Robert H. (1966/ 2000): 'The development of the word class system of the European grammatical tradition', in: Ludger Hoffmann (ed.): Sprachwissenschaft. Berlin/ New York: Walter de Gruyter, 452-469 (also: Foundations of Language 2). Robins, Robert H. (1967): A Short History of Linguistics. London/ Harlow: Longmans, Green and Co. Robins, Robert H. ( 2 1971): General Linguistics. An Introductory Survey. London: Longman. Saussure, Ferdinand de (1894/ 1964): 'Lettres de Ferdinand de Saussure à Antoine Meillet (4 janvier 1894) ', in: Société Genevoise de Linguistique, Cahiers Ferdinand de Saussure, 21, 1964, Genève : Librairie Droz, 93-130. Schmid, Hans-Jörg (2000): English abstract nouns as conceptual shells: from corpus to cognition. Berlin/ New York: Mouton de Gruyter. Schüller, Susen (2005): 'Problematik der Darstellung von Aspektunterschieden in englischen Referenzgrammatiken', in: Thomas Herbst (ed.): Linguistische Dimensionen des Fremdsprachenunterrichts. Würzburg: Königshausen & Neumann, 29-42. Schüller, Susen (forthcoming): Semantic Aspects of Verb Valency. (working title). Schumacher, Helmut/ Jacqueline Kubczak/ Renate Schmidt/ Vera de Ruiter (2004): VALBU - Valenzwörterbuch deutscher Verben. Tübingen: Narr. Searle, John R. (1969): Speech Acts. An Essay in the Philosophy of Language. London/ New York: Cambridge University Press. <?page no="211"?> Bibliography 199 Sinclair, John (1991): Corpus. Concordance. Collocation. Oxford: Oxford University Press. Sinclair, John (2004a): Trust the Text. Language, corpus and discourse. (ed. with Ronald Carter). London/ New York: Routledge. Sinclair, John (2004b): 'New evidence, new priorities, new attitudes', in: John Sinclair (ed.): How to Use Corpora in Language Teaching. Amsterdam/ Philadelphia: John Benjamins, 271-299. Sinclair, John (ed.) (1990): Collins Cobuild English Grammar. London/ Glasgow: Collins. Sinclair, John/ Anna Mauranen (2006): Linear Unit Grammar. Integrating Speech and Writing. Amsterdam/ Philadelphia: Benjamins. Somers, Harold L. (1987): Valency and Case in Computational Linguistics. Edinburgh: Edinburgh University Press. Stefanowitsch, Anatol/ Stefan Th. Gries (2003): 'Collostructions. Investigating the interaction of words and constructions', International Journal of Corpus Linguistics 8: 2, 209-243. Stein, Gabriele (2007): A Dictionary of English Affixes. Their Function and Meaning. München: Lincom. Tarvainen, Kalevi (1985): Semantic Cases in the Framework of Dependency Theory. Trier: LAUT. Tesnière, Lucien (1959): Éléments de syntaxe structurale. Paris: Klincksieck. Tesnière, Lucien (1959/ 2000): 'Grundzüge der strukturalen Syntax (Auszüge)', in: Ludger Hoffmann (ed.): Sprachwissenschaft. Ein Reader. Berlin/ New York: Walter de Gruyter, 517-542. Thümmel, Wolf (1993): 'Westliche Entwicklungen', in: Joachim Jacobs/ Arnim von Stechow/ Wolfgang Sternefeld/ Theo Vennemann (eds.): Syntax. HSK, Berlin/ New York: de Gruyter, 130-199. Tomasello, Michael (2003): Constructing a Language. A Usage-Based Theory of Language Acquisition. Cambridge, Mass./ London: Harvard University Press. Zifonun, Gisela/ Ludger Hoffmann/ Bruno Strecker et al. (1997): Grammatik der Deutschen Sprache Band 2. Berlin/ New York: Walter de Gruyter. Dictionaries Oxford Advanced Learner’s Dictionary (2007) (chief editor: Sally Wehmeier), Oxford: Oxford University Press. [OALD7] Longman Dictionary of Contemporary English (2003)(director: Della Summers), Harlow: Pearson, [LDOCE4] The Oxford English Dictionary ( 2 1989) (J.A. Simpson and E.S.C. Weiner), Oxford: Oxford University Press. [OED2] Langenscheidt Collins Großwörterbuch Englisch (2004), Berlin: Langenscheidt Collins Cobuild English Dictionary ( 2 1995) (editor-in-chief John Sinclair), London: HarperCollins. <?page no="212"?> 200 Further acronyms used CamG Huddleston/ Pullum (2002) CGEL Quirk/ Greenbaum/ Leech/ Svartvik (1985) VDE Herbst/ Heath/ Roe/ Götz (2004) Sources of examples BNC British National Corpus VDE Valency Dictionary of English (Cobuild Corpus) QE quoted example NW David Lodge: Nice Work SW David Lodge: Small World CP David Lodge: Changing Places David Lodge (1976): Changing Places. London: Secker & Warburg. (first published 1975). David Lodge (1984): Small World. London: Secker & Warburg. David Lodge (1989): Nice Work. Harmondsworth: Penguin (first published 1988). Bibliography <?page no="213"?> Index Aarts 14, 45, 60, 63, 82, 94, 166, 170, 171, 172 acceptability 172 acceptable 3, 40, 115, 125, 132 actant 109 active 14, 21, 23, 44, 99, 117, 118, 123, 138, 139, 155, 158, 162, 164, 170, 175 adjective 22, 23, 25, 26, 28, 34-37, 45, 50, 53, 54, 56, 58-61, 72, 73, 76-78, 81, 82, 85-88, 91, 92, 95, 100, 102, 107-109, 111, 118, 119, 121, 125-127, 137, 140-142, 144, 169, 171 AdjP 77, 118, 119, 123, 167, 172, 175, 178 adjunct 8, 17, 19, 20-24, 27, 60, 62, 83, 93, 94, 96, 97, 100-110, 112-116, 121, 122, 124, 126, 131, 133, 142, 154, 156, 164-166, 169, 171, 174, 177 adjunct unit 24, 25, 30, 60, 96, 100, 101, 121, 124, 166, 167, 174, 175, 177, 178 adverb 14, 15, 28, 31, 36, 37, 53, 55, 58-67, 70-72, 75, 77, 88, 90, 91, 95, 100, 101, 109, 118, 119, 142, 144, 146 adverbial 15, 23, 67, 107, 115, 146, 167, 169, 171 AdvP 77, 118, 122, 123, 177 AGENT 129, 132, 134, 137, 139, 157, 161, 162, 167 AGENTIVE 126, 127, 162 AGENTIVITY 162, 163 Allerton 34, 109 allomorph 145 antecedent 52 apposition 82, 83, 169 appositive 83 Aristotle 1 attribute 11, 15, 24, 92, 161, 166, 172 attributive 35, 58, 72 base form 38, 39, 41, 98, 154, 155 Behrens 32, 34 benefactive object 170, 171 Biber 154, 166 Bloomfield 9 BNC 40, 43, 48, 50, 52, 61, 68, 85, 86, 88, 94, 101, 103, 106, 111, 126, 131, 142, 146, 154, 155, 156, 159, 162 Brown 157, 158 Bublitz 79, 148 Bühler 109 Burgschmidt 13 CamG 4, 13, 27, 29, 37, 39, 40, 43-45, 51, 53, 55, 58, 59, 61, 63, 65, 66, 68, 73-75, 85-88, 92, 93, 98, 106, 107, 131-133, 145, 148, 149, 153-156, 160, 166, 171, 172, 176 CGEL 4, 8, 9, 11, 13, 14, 17, 23, 24, 29, 36, 37, 39, 40, 43-48, 50, 51, 53-55, 57-60, 62, 63, 65, 68, 69, 71, 73, 75, 79-82, 86, 87, 92, 93, 107, 120, 121, 131, 146, 148-150, 152, 153, 155, 156, 166-170, 172 Clark 5, 87 clausal role 135, 158, 160-163 clause 1, 3-6, 8-14, 16-29, 34, 35, 40-43, 45, 48, 52, 60-72, 75, 76, 80-82, 90-115, 117- 119, 122-126, 128, 131, 135, 136, 138-141, 149, 153, 155-158, 162-178 cleft 155, 156, 158 CL fin 118, 123 cognitive 2, 12, 13, 32, 33, 145, 170 collocation 17 comparative 13, 28, 58, 60, 72, 95 complement 17, 21-27, 43, 61-63, 66-68, 76- 79, 81, 83-85, 88-97, 99-106, 108-147, 156, 159, 160, 162, 164-166, 167, 169, 171, 172, 174-176, 178 complement inventory 129, 137, 139, 140, 141, 143, 164 complementation 43, 65, 67, 87, 88, 89, 119, 120, 141, 143, 145, 146 compound 29, 82 concord 18-20, 52, 54, 56, 97, 154, 172 conjunction 15, 36, 37, 61-68, 73-75, 83, 90, 144, 173, 174, 178 Conrad 154, 166 constituency 11 constituent 4-12, 16, 17, 19-29, 43, 56, 63, 64, 68, 76, 77, 80, 82, 91, 93-95, 97, 100, <?page no="214"?> Index 202 101, 106, 109, 110, 112, 116, 126, 128, 136, 138, 142, 145, 150-155, 157, 159, 160, 164, 166, 174, 177, 178 construction 1, 2, 4, 6, 17-19, 22, 24, 36, 38- 43, 45, 54, 68-71, 78, 82, 85-87, 90, 91, 92, 95, 96, 99-102, 106, 109, 110, 112, 113, 119, 122, 125, 126, 129, 130, 134-136, 138- 141, 143, 146, 149-156, 158-160, 162, 164, 170-172, 174 contextually optional 86, 92, 94, 110, 111, 138, 140, 145, 153 coordinated 27, 28, 62, 68, 83, 147, 152, 166, 169, 174 coordination 27-29, 68, 83, 134, 173 copula 171 copular 171 corpus 1, 153 count 46-49, 78, 137 countable 48 Croft 1, 170 Cruse 1, 108, 170 Cruttenden 150 Culler 13 declarative 17-19, 21, 138, 149, 150, 152, 153, 155, 174-178 Dederding 11, 114 de Groot 109 deletable 113 deleted 22, 25, 47, 110, 112, 113 deletion 22, 113 demonstrative 55 dependency 9-11, 19, 21, 78, 79, 109, 167 dependent clause 29, 30, 100, 175 determinative 75 determiner 14, 36, 49, 50, 56, 57, 61, 73-76, 78-80, 87 determiner-pronoun 36, 50, 53, 54, 55, 56, 57, 61, 68, 71, 75, 79, 80, 83, 85, 87, 143, 178 dictionary 1, 4, 14, 23, 31, 34, 53, 60, 63, 72- 74, 109, 117, 120, 121, 138 direct object 14, 167, 170-172 directive 38, 113, 148, 149, 154, 155, 174, 175 Dirven 127 ditransitive 168 divalent 23, 24, 125, 136, 137, 143 Dowty 159 Eisenberg 63 emergentism 32 Emons 6, 109, 114, 121, 136 endocentric 92 Engel 109, 138 Esser 37, 158 exclamation 148, 156 exclamative 155, 156 existential 72 exocentric 92 extralinguistic 129 extraposition 124, 128 Fillmore 1, 19, 27, 36, 74, 94, 122, 125-129, 136, 158, 160 Fischer 17 frame 19, 127, 128, 129, 133, 158 FrameNet 122, 125, 127-129, 131, 158 frequency 71, 73 frequent 40, 146 fronting 155, 156 function 12, 14, 16, 19, 20, 24, 27, 35, 36, 40- 42, 44, 46, 49, 50, 52, 56, 57, 59, 60, 61, 68, 69, 71, 73, 76, 78, 80, 85, 87, 91-93, 95-97, 99-101, 106, 108, 117, 118, 122-124, 126, 127, 129, 137, 141, 142, 148, 154, 156, 165, 166, 169, 170, 175, 176 functional 2, 8, 14, 117 Garside 73 gender 51, 52, 54, 70, 71 generative 9, 32, 127 generic 79 genitive 45, 50-54, 57, 70, 80, 142 gerund 13, 37, 87 Gimson 150 Goldberg 1, 2, 17, 24, 129, 130, 139, 149, 159, 160, 162 Götz 13, 14, 23, 121, 127, 162 Götz-Votteler 127, 162 gradience 42, 67, 131, 172 gradient 41, 67, 90, 111, 126 grammar 1, 2, 4, 9, 10, 11, 13, 16, 17, 19, 24, 32-34, 64, 70, 86, 109, 126, 127, 139, 144, 153, 155, 165, 167, 170 grammatical 1, 4, 15, 21, 22, 37, 40-42, 44, 45, 51-53, 73, 88, 92, 142, 156, 158-160, 164, 172 Greenbaum 4, 8, 11, 13, 14, 46, 57, 59, 80, 92, 148, 153, 154 <?page no="215"?> Index 203 Gries 127, 139, 141 Haegeman 8, 9, 92, 127 Halliday 19, 126, 158, 171 Hausser 28 head 20, 21, 40, 41, 43-46, 49-54, 56-59, 71, 75-94, 97, 98, 102, 142, 143, 150-153, 166, 167, 175, 178 headed 25, 27, 46, 62, 80, 81, 91, 92, 94, 95, 101, 106 headedness 80, 94 Heath 11, 14, 23, 114 Helbig 108, 109, 114, 127, 135 Herbst 11, 13, 14, 23, 37, 44, 98, 109, 112- 114, 121, 127, 128, 135, 141, 142 Heringer 11, 109, 116 Hoffmann 114 Huddleston 4, 59, 61, 63, 65, 92, 172 Hudson 80 idiom 43, 73, 74, 146 idiomatic 51, 73, 145-147 illocutionary 148 imperative 38, 39, 41, 98, 111, 113, 154, 155, 174 impersonal 135, 136 indefinite 50, 53, 55, 56, 79, 80 indirect object 14, 167, 170-172 infinitive 1, 36, 38, 39, 41-43, 63, 73, 75, 81, 90, 91, 92, 98, 99, 101, 105, 106, 108, 109, 115, 116, 118-120, 123, 124, 126, 137, 138, 141-143, 167, 175 inflection 13, 33, 38, 45, 50, 72 interjection 36, 37, 62, 69 interrogative 6, 18, 40-42, 45, 54, 69, 70, 71, 100, 113, 149, 151-153 intonation 3, 150-152, 154 intransitive 62, 168, 169, 171 inversion 39, 149, 151, 155 irregular 37, 38 Jespersen 166, 170 Johansson 166 Johnson 129 Kay 36, 74 Klotz 135, 141 Kortmann 4, 8, 9 Lamb 127 LDOCE 31, 75, 146 Leech 4, 8, 11, 13, 14, 46, 57, 59, 73, 80, 92, 148, 150, 151, 153, 154, 157, 166 Leisi 33 lemma 40, 52 lexeme 31, 73, 120, 121, 171 lexical item 2, 17, 21, 24, 123, 129, 130, 143, 155 lexical noun 36, 45, 46, 48, 49, 56, 76, 79-81, 83, 85 lexical unit 23, 77, 93, 108-110, 116, 139, 140-145, 157, 172 lexical verb 36-44, 59, 72 lexicographical 128 lexicography 73 lexicological 31 lexicon 32, 71, 110, 140 lexis 1, 2, 155 linear 9, 11, 16, 164, 165, 167, 174 linking 100, 166, 171, 174 linking unit 29, 166, 174, 175, 178 LOCATIVE 72, 130, 133, 134, 146, 168 Lodge 3, 48 Lyons 23, 79, 94 Matthews 4, 9, 11, 80, 92, 94, 144, 145, 171 Mauranen 5, 16, 91, 113, 153, 165 meaning 15-17, 32, 34, 42, 48, 49, 68, 72, 73, 86, 94, 108, 109, 112, 129, 131, 133, 143, 144, 148, 149, 150-155, 157-159, 163, 164, 167, 171 merger of participants 135, 136 Mindt 43 modal verb 36, 39, 40, 42, 43, 45, 143 modification 14, 59, 72, 85, 87, 88 modifier 25-27, 52, 58, 62, 70, 77, 78, 81, 82, 85, 88, 89, 93, 94, 96, 102-106, 110, 137, 142 monotransitive 168 monovalent 23, 24, 66, 67, 136, 137, 142, 144, 145 morpheme 5, 38 morphology 1, 9, 12, 13, 34, 35, 37-40, 42, 43, 45, 46, 50, 52, 58, 60, 62, 87, 97, 98, 138, 141, 155 Muthmann 39, 46 negation 7, 39, 154 neuter 51 nominal 79, 169 <?page no="216"?> Index 204 nominative 19, 51, 63 noun 1, 8, 18, 22, 23, 26, 28, 33-35, 37, 45-58, 60, 63, 65-68, 70-73, 75-83, 85-87, 90, 91, 93-95, 99, 101, 102, 107-109, 111, 113, 117- 119, 121, 122, 125, 126, 130, 132, 137, 140- 142, 144, 146, 147, 154, 159, 160, 163, 169, 170, 171, 174, 176, 177 NP 75, 77, 99, 105, 106, 109-111, 118, 119, 123-127, 130, 132, 135, 137-140, 143, 146, 147, 159-162, 167, 172-178 nuclear 23, 72 number 37, 40, 43-45, 54-56 numeral 37, 55, 129, 159, 173 OALD 31, 75 object 2, 14, 78, 79, 84, 117, 125, 166, 167, 169-172 obligatoriness 111, 137 obligatory 77, 79, 87, 92, 94, 97, 110, 111, 113, 116, 121, 122, 143, 153 oblique 160, 171 operator 18, 97, 100, 102, 149, 151-153 optional 19, 23, 26, 27, 77, 79, 94-97, 100, 101, 103, 110-113, 132, 137, 144, 163 optionality 108, 110, 111, 113, 169 ordinal 55 Palmer 4, 9, 39, 43, 68, 93, 94, 127, 158-160, 162 part of speech 36 participant 19, 108, 109, 126-131, 133-137, 139-144, 157-163, 166, 176, 177 participle 13, 15, 37-41, 43, 44, 58, 81, 82, 87, 98 particle 29, 36, 37, 59, 61-68, 70-73, 75-78, 81, 82, 89-92, 95, 96, 99-107, 115, 117, 119- 121, 123, 124, 135, 142-147, 159, 174-177 particle complement 119, 120, 121, 159 particle phrase 66, 67, 72, 76, 77, 78, 81, 82, 89-92, 96, 99, 101, 102, 106, 107, 115, 117, 119, 121, 123, 124, 130, 135, 142, 146, 147, 174-177 passive 1, 14, 23, 44, 45, 87, 88, 99, 111, 117, 118, 123, 124, 138, 139, 158, 162, 164, 170- 72, 175 passivizability 170 passivization 132, 169, 171 Patient 162 pattern 12, 17, 71, 97, 122, 125, 126, 137-141, 143, 146, 150, 158, 159, 164, 170, 172, 175 patternbank 138 PCU 24, 25, 29, 30, 101, 166-168, 170-172, 174-178 peripheral 19, 73 periphery 23 person 37-40, 43, 44, 45, 52, 54, 56, 98 personal 50, 51, 54, 57, 63, 73, 134 perspective 12, 22, 24, 105, 110, 128, 145, 157, 158, 160, 164, 165, 172 perspectivization 158, 160 Petruck 129 phrasal verb 120, 146, 147, 161 phrase 1, 5, 7, 8, 9, 10, 17, 18, 24-29, 35, 45, 46, 49, 50-56, 58, 60, 62-68, 70, 71, 75-83, 85-99, 101, 102, 106, 107, 108, 109, 110, 113, 117-119, 121-124, 130, 132, 137, 142, 146, 147, 150-152, 154, 159, 160, 163, 169- 171, 174-178 phraseological 35, 56, 74 phraseology 122 plural 18, 46, 48, 49, 51, 52, 86, 135, 155 possessive 51, 57 postdeterminer 57 postmodification 81, 85, 87, 88 postmodifier 26, 46, 48, 51, 52, 77, 80, 81, 82, 83, 84, 85, 88, 89, 99, 101, 124, 142, 178 postposition 58 pragmatics 148 Prague School 157 predeterminer 57 predicate 4, 8, 9, 13, 17-24, 27, 60, 94, 97, 98, 100-102, 112, 115, 117, 118, 128, 129, 132, 138, 150-152, 154, 155, 163-167, 171, 172, 174, 176, 177 predicate complement unit 24, 25, 29, 30, 100, 101, 132, 138, 163, 166-168, 170-172, 174-178 predicate head unit 24, 29, 30, 166-168, 174- 178 predicative 58, 132, 133, 139, 166-168, 171, 172, 176 predicator 21-24, 97, 142, 171 premodification 51, 85, 87-89 premodifier 26, 46, 51, 53, 59, 60, 77, 80-84, 85, 88, 89, 90, 94, 109, 142, 178 preposition 15, 37, 61-66, 68, 73, 75, 92, 120, 144, 145, 160, 171 prepositional 62, 66-68, 90, 92, 120, 121, 146, 169 <?page no="217"?> Index 205 pronoun 14, 37, 49, 50, 52, 56, 70-72, 75, 83, 85, 101, 102, 154 proper noun 47-49, 83 proposition 157 propositional 157, 158 prosodic 3, 155 psychological 2, 5, 12, 32, 73, 120, 140 Pullum 4, 59, 61, 63, 65, 92 punctuation 3 quadrovalent 137 question 6-8, 18, 40-42, 45, 69, 70, 100, 113, 115, 125, 150-153, 155, 156, 174, 177 Quirk 4, 8, 11, 13, 14, 46, 57, 59, 80, 92, 148, 153, 154 QUOTE 4, 104, 122, 124 Radden 127 Radford 80 reciprocal 50, 52 reference 48, 78, 79, 127, 128 referent 52, 70, 71, 83, 86, 153, 169 referential 79, 107 reflexive 50, 52, 80 relative 45, 48, 54, 69, 70, 71, 82, 95, 96, 100, 101, 102, 156 rheme 157, 158 Robins 1, 8, 34 Roe 14, 23, 112, 142 rPred 97, 100, 102-104, 151, 156 Sampson 73 Satzbauplan 138 Satzmuster 138 Saussure 13 Schenkel 108, 109, 114 Schmid 142 Schüller 44, 128 Schumacher 109 Searle 79 semantic 2, 13, 16, 19, 20, 23, 31, 34, 35, 41, 42, 48, 56, 71, 94, 108, 116, 121, 126-131, 134-137, 139, 146, 149, 158, 159, 162, 166, 169-172, 176, 177 semantics 78, 126, 134, 148, 171 sense 14, 143, 146, 172 sentence 1-12, 14, 16-2, 24, 27-31, 32, 41, 46- 48, 53, 58, 60, 61, 71, 73, 92, 95, 96, 100, 102-104, 110-112, 115, 116, 120-122, 124, 126, 127, 132, 133, 137, 145, 148-150, 152- 155, 157-160, 162-164, 166, 173-178 spoken 3, 153 statement 17-19, 148-150, 152, 153, 155, 156, 174-178 Stefanowitsch 17, 127, 139, 141 Stein 39, 46 Stoic 1, 8 Stoll 98 stored 1, 2, 33, 110 Strecker 114 subject 1, 2, 4, 8, 14, 17-24, 27, 28, 40-44, 51, 60, 62, 72, 86, 94, 95, 97-100, 102-105, 107, 111-113, 117, 118, 123, 124, 132, 133, 137- 140, 144, 149-156, 162-167, 169-172, 174- 176 subject complement unit 24, 25, 29, 30, 100, 101, 132, 133, 138, 140, 162, 163, 165, 166- 168, 174-178 subjunctive 39 subordinate 5, 28, 29, 90, 95, 96, 100, 101, 111 subordinating 37, 61, 63-68, 73, 75, 90, 144 subordination 27-29 subordinator 61, 63-67, 73, 75, 91, 96 Svartvik 4, 8, 11, 13, 14, 46, 57, 59, 80, 92, 148, 150, 151, 153, 154 syntactic 1, 2, 4, 5, 7, 9, 10, 12, 13, 15-17, 24, 31, 32, 67, 69, 76, 108, 109, 112, 126, 132, 135-137, 149, 159, 162, 170, 172, 173 Tarvainen 131 tense 13, 37-41, 43-45, 98 terminology 2-4, 8, 9, 12-14, 24, 25, 27, 34, 35, 49, 57, 63, 64, 71, 75, 95, 120, 137, 149, 166, 169, 171 Tesnière 11, 19, 21, 109 thematic 127, 129, 138, 157, 158 theme 129, 133, 157, 158 theta role 13, 127 Thümmel 1 Tomasello 2, 19 traditional grammar 29, 34, 64, 65, 91, 98, 117, 149 transitive 62, 168, 169, 171 trivalent 10, 23, 24, 61, 99, 125, 126, 136, 146, 155 uncount 46-49 uninflected 64, 73, 144 <?page no="218"?> Index 206 unspecified particle phrase 121, 175 utterance 1, 2, 78, 112, 113, 148 valency 4, 12, 14, 17, 18, 20-27, 43, 61, 62, 65-67, 77-79, 87, 92-95, 97, 101, 102, 108- 113, 115-127, 129, 134-146, 158-160, 163- 167, 168, 172, 175, 176, 178 valency slot 24, 101, 108-113, 116-119, 121, 122, 124, 125, 127, 135, 137, 141, 142, 144, 146, 172 valent 136, 147 Valenz 108 VDE 4, 14, 23, 82, 88, 105, 112, 116-119, 121- 126, 128, 129, 131, 135, 137, 138, 142, 146, 162, 170, 172 verb 1, 8-13, 17-25, 27, 33, 34, 36-45, 51, 54, 56, 58, 59, 61, 62, 65, 66, 72, 73, 80, 86, 91, 93-100, 102, 106, 108-113, 116, 119-122, 125-131, 133-144, 146, 147, 153-155, 158- 160, 162-172, 175, 178 verbal head complex 20, 21, 24, 37, 40, 43- 45, 59, 93, 97, 98, 101, 106, 165, 166, 175, 178 verbless 63, 106, 107 vocative 154 Westermayr 98 word 1, 2, 4-6, 10-12, 14, 21, 25-27, 31-41, 43, 45-50, 53-58, 62-67, 69-74, 76, 78, 79, 87, 92, 93, 98, 100-102, 108, 109, 116, 121- 123, 129, 131, 143, 144, 146, 148-154, 157, 169, 170, 178 word class 2, 14, 27, 31-38, 40, 41, 43, 45, 49, 50, 55-59, 61-67, 69, 71-74, 76, 78, 87, 92, 109, 143, 169, 178 zerovalent 63, 65, 66, 67, 136, 144, 145 Zifonun 114 <?page no="219"?> Glossary adjective: word class comprising words such as good, clever, old (> 3.4) adjunct: an element in the structure of a clause that is independent of the valency of the governing verb (> 2.2.2, 2.3.1) adjunct unit [AU]: one of the five units of clause structure; a unit that is realized by an adjunct and thus independent of the valency of the verb that functions as the head of the clause (> 8.1) adverb: word class comprising words such as beautifully or however that can function as adjuncts or modifiers (Traditionally words such as yesterday, which we classify as a temporal pronoun, and here, there, then, when, which we classify as particles, are also referred to as adverbs.) (> 3.5.1) clause/ phrase: a construction consisting of a group of words which are headed by a verb (clause), a noun or pronoun (noun phrase NP), an adjective (adjective phrase AdjP), an adverb (adverb phrase AdvP) or a particle (particle phrase PartP) (> 2.4, 4, 5) complement: an element in the structure of a clause or a phrase that is determined in its form by the word functioning as the head of the clause or phrase or/ and that fills one of its obligatory valency slots (> 2.3.1, 6.4) conjunction/ coordinating conjunction: word class comprising words such as and and but that can coordinate elements of a sentence, clause or phrase (> 3.6) declarative 'question' construction: > sentence type declarative 'statement' construction: > sentence type determiner/ pure determiner: word class comprising the words a, an, the, no, every that can function as pre-heads in a noun phrase (> 3.3.4) determiner-pronoun: word class comprising words such as these that can function both as pre-heads and heads in a noun phrase (> 3.3.3) head-complex: that part of a noun phrase or a clause that contains the head and possibly pre-heads (> 3.3.5) imperative 'directive' construction: > sentence type interjection: word class comprising words such as oh or hello (> 3.7) interrogative 'alternative question' construction: > sentence type interrogative 'yes-no-question' construction: > sentence type lexical verb: a word such as read that can function as the head of the verbal headcomplex (> 3.2.1) linking unit [LU]: one of the five units of clause structure; a unit that is realized by a conjunction that coordinates other units of clause structure (> 8.1) <?page no="220"?> Glossary 208 modal verb: word class comprising words such as can or might that can function as heads or pre-heads in the verbal head-complex (> 3.2.2.3) modifier: an element in the structure of a phrase that is independent of the valency of the word that functions as the head of the phrase (> 4.1.4) noun/ lexical noun: word class comprising words such as university that can function as the head of a noun phrase (> 3.3.1) operator verb: the verb functioning as the first pre-head in the verbal head-complex (> 5.3.1) particle: word class that subsumes traditional prepositions (about, in, on), traditional subordinating conjunctions (that, whether) and some traditional adverbs (here) (> 3.5.2) particle complement (or particle phrase complement): a valency complement that is realized by a particle phrase; in the case of specific particle complements a particular particle is part of the valency specification of the valency carrier (write about it); in the case of unspecified particle phrase complements no particular particle is required by the valency carrier (ie.: live in Cambridge, here etc.) phrase > clause predicate complement unit [PCU]: one of the five units of clause structure; a unit that is part of the predicate of the clause and is realized by a complement of the unit that functions as the head of the clause (> 8.1) predicational head unit [PHU]: one of the five units of clause structure; the unit that contains the verb functioning as the head of the clause and the verbs functioning as pre-heads (> 8.1) primary verb: word class comprising the words be, do and have that can function as heads or pre-heads in the verbal head-complex (> 3.2.2.2) pronoun/ pure pronoun: word class comprising words such as they that can function as the head of a noun phrase; genitives of pronouns (their) can function as preheads (> 3.3.2) sentence type: a particular type of sentence characterised by formal (syntactic) and semantic properties; for example: (> 7.1) declarative 'statement' construction (April is the cruellest month. SW1 ) wh-subject 'question' construction (And who is Charles? NW41 ) wh-interrogative 'question' construction (Why don't you ask Robyn Penrose? NW88 ) interrogative 'yes-no-question' construction (Have you really flown all the way from America for this conference, Professor Zapp SW18 ) interrogative 'alternative question' construction (... was it Women in Love or Lady Chatterley? ... NW98 ) declarative 'question' construction (You like it? SW18 ) <?page no="221"?> 209 imperative 'directive' construction (Don't be absurd NW87 ) subject complement unit [SCU]: one of the five units of clause structures; the unit which functions as the subject of the clause and is realized by a complement of the verb that functions as the head of the clause (> 8.1) [that_CL]: a particle phrase in which that (which is regarded as the head of the phrase) is followed by a finite clause (traditionally referred to as a that-clause) (> 4.5, 6.4.2) [to_INF]: a particle phrase in which to (which is regarded as the head of the phrase) is followed by an infinitive clause (traditionally called a to-infinitive clause) (> 4.5, 6.4.2) unit > adjunct unit, linking unit, predicate complement unit, predicate head unit, subject complement unit (> 8.1) valency: ability of a verb, noun, adjective, adverb or particle to open up valency slots to be filled by complements (> 2.3, 6) verbal head-complex: the formal realisation of the (>) predicational head unit of a clause (> 3.2.3) [wh_CL]: a construction which contains a wh-element as its first element: either (a) a wh-particle is the head of a particle phrase or (b) the wh-element is a clause unit of a wh-clause (> 5.3.6) wh-subject 'question' construction > sentence type wh-interrogative 'question' construction > sentence type Glossary <?page no="222"?> Abbreviations used in the analysis of sentences Clause units AU adjunct unit LU linking unit PCU predicate complement unit PHU predicate head unit SCU subject complement unit Formal realisations of adjunct units and complement units Simple phrase labels AdjP adjective phrase: But doesn't that make us rather marginal? NW218 AdvP adverb phrase: Not many people perceived this quite so soon. NW45 NP noun phrase: They don't pay us all that much, anyway. NW218 PartP unspecified particle phrase: ... you've been a tremendous asset to the Department, even in the short time you've been here. NW64 Go to Oxbridge. NW45 Simple clause labels CL fin finite clause (with subject): I suppose everyone says that to you. SW7 INF non-finite clause (without subject) headed by an infinitive form of a verb: No other reason would make anyone come here, or having come, stay. NW214 V-ing non-finite clause (without subject) headed by a verb in the ing-participle form: Robyn sat up all night printing out her book. NW327 V-ed non-finite clause (without subject) headed by a verb in the -ed-participleform: The steady grind of intellectual work, punctuated by brief explosions of indoor physical exercise - that was the rhythm of Robyn's first winter at Rummidge. NW215 Combined labels NP_INF After a moment or two she heard him call, "Turn the ignition key," and when she did so, the engine fired. NW149 NP_to_INF What the students want her to do is to give them some basic facts ... NW60 NP_V-ed If by any chance there should be a job advertised in the next twelve months ... NW325 NP_V-ing You would not think her unduly burdened with worries, if you watched her crossing the campus, smiling at people she knows, her eyes bright, her brow unfurrowed. NW60 <?page no="223"?> 211 to_NP etc. specified particle phrase (headed by the particle indicated and followed by the type of complement indicated): ... you've been a tremendous asset to the Department, even in the short time you've been here. NW64 Whatever happened to the spirit of the Robbins Report? NW307 that_CL (specified) particle phrase headed by that followed by a finite-clause complement: Haven't you noticed that in the modern world good news comes by telephone and bad news by mail? NW328 to_INF specified particle phrase headed by to followed by an infinitive-clause complement: Professor Zapp has kindly offered to look at my work-in-progress. NW326 wh_to_INF She is always uncertain how to address her Head of Department. NW61 wh_CL I never could remember which came first, the morphemes or the phonemes. W23 Further labels 1 QUOTE 'I couldn't afford the time, even if I had the money,' says Robyn. NW62 SENTENCE I was one of them, I don't mind admitting … NW62 Formal realisation of predicate head units VHC act: 0, 1, 2, etc. verbal head complex (active + number of valency complements occurring in clause) VHC pass: 0, 1, 2, etc. verbal head complex (passive + number of valency complements occurring in clause) Labels for the description of phrases h head preh pre-head prem premodifier postm postmodifier c valency complement of head Word class labels adj adjective adv adverb conj conjunction det pure determiner detpron determiner-pronoun 1 For further labels see 6.4.2.7. Abbreviations used in the analysis of sentences <?page no="224"?> 212 interj interjection ln lexical noun lv lexical verb mv modal verb part particle pron pure pronoun pv primary verb v verb (unclassified) Abbreviations used in the analysis of sentences <?page no="225"?> This book provides an introduction to the analysis of sentences for students of English. It outlines principles of syntactic analysis and develops the categories used for a framework in which the concept of valency plays a major role. The basic categories of syntactic description are elucidated in great detail: word classes, phrases, clauses. One major chapter deals with the notion of valency and how it can be implemented in a description of the English language. The main purpose of this book is to describe a framework which will enable students to scrutinise English sentences. The final chapter provides a practically oriented outline of how the categories identified in the preceding parts of the book can be adopted in an analysis of English sentences. The method outlined consists of eight clearly defined steps and is demonstrated in detail in the analysis of a number of authentic English sentences. ISBN 978-3-8233-6390-3