<?page no="0"?> www.narr.de TBL Tübinger Beiträge zur Linguistik How do people talk about words? is work explores naming constructions such as is formation is called a “coral reef ” that are used to introduce lexicalized concepts like coral reef through metalinguistic reference. Drawing on the theoretical foundations in philosophy of language as well as recent work in semantics and pragmatics, experimental evidence from corpus analyses and psycholinguistic experiments—including self-paced reading and eye-tracking data—uncover the linguistic and cognitive processes behind these naming constructions. Special attention is given to the use and processing of quotational constructions during reading. Bridging (psycho)linguistics and philosophy of language, this interdisciplinary work o ers new insights into how people use quotational constructions to talk about language, and how readers interpret and process naming constructions with and without quotation marks. 593 Raue Reference to Names Reference to Names An empirically based study on metalinguistic uses of names of lexicalized concepts Natascha Raue ISBN 978-3-381-14641-3 <?page no="1"?> Reference to Names <?page no="2"?> Tübinger Beiträge zur Linguistik herausgegeben von Gunter Narr 593 Unbenannt-1 1 Unbenannt-1 1 20.10.2025 10: 26: 11 20.10.2025 10: 26: 11 <?page no="3"?> Natascha Raue Reference to Names An empirically based study on metalinguistic uses of names of lexicalized concepts <?page no="4"?> DOI: https: / / doi.org/ 10.24053/ 9783381146420 Dr. Natascha Raue | https: / / orcid.org/ 0000-0003-1636-5752 Universität Osnabrück | Institut für Germanistik Neuer Graben 40 49074 Osnabrück © 2025 · Natascha Raue Das Werk ist eine Open Access-Publikation. Es wird unter der Creative Commons Namensnennung - Weitergabe unter gleichen Bedingungen | CC BY-SA 4.0 (https: / / creativecommons.org/ licenses/ by-sa/ 4.0/ ) veröffentlicht, welche die Nutzung, Vervielfäl‐ tigung, Bearbeitung, Verbreitung und Wiedergabe in jeglichem Medium und Format erlaubt, solange Sie die/ den ursprünglichen Autor: innen und die Quelle ordentlich nennen, einen Link zur Creative Commons-Lizenz anfügen und angeben, ob Änderungen vorgenommen wurden. Die in diesem Werk enthaltenen Bilder und sonstiges Drittma‐ terial unterliegen ebenfalls der genannten Creative Commons Lizenz, sofern sich aus der am Material vermerkten Legende nichts anderes ergibt. In diesen Fällen ist für die oben genannten Weiterverwendungen des Materials die Einwilligung des jeweiligen Rechteinhabers einzuholen. Alle Informationen in diesem Buch wurden mit großer Sorgfalt erstellt. Fehler können dennoch nicht völlig ausgeschlossen werden. Weder Verlag noch Autor: innen oder Herausgeber: innen übernehmen deshalb eine Gewährleistung für die Korrektheit des Inhaltes und haften nicht für fehlerhafte Angaben und deren Folgen. Diese Publikation enthält gegebenenfalls Links zu externen Inhalten Dritter, auf die weder Verlag noch Autor: innen oder Herausgeber: innen Einfluss haben. Für die Inhalte der verlinkten Seiten sind stets die jeweiligen Anbieter oder Betreibenden der Seiten verantwortlich. Narr Francke Attempto Verlag GmbH + Co. KG Dischingerweg 5 · D-72070 Tübingen Internet: www.narr.de | eMail: info@narr.de Druck: Elanders Waiblingen GmbH ISSN 0564-7959 ISBN 978-3-381-14641-3 (Print) ISBN 978-3-381-14642-0 (ePDF) ISBN 978-3-381-14643-7 (ePub) Bibliografische Information der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Natio‐ nalbibliografie; detaillierte bibliografische Daten sind im Internet über http: / / dnb.dnb.de abrufbar. <?page no="6"?> 1 19 2 27 2.1 31 2.1.1 33 2.1.2 34 2.1.3 35 2.1.4 38 2.1.5 39 2.1.6 40 2.2 41 2.3 43 2.4 47 3 51 3.1 53 3.2 63 3.2.1 68 3.2.2 71 3.2.3 73 3.3 74 3.4 78 3.5 86 4 89 4.1 90 4.2 92 4.3 94 4.3.1 95 4.3.2 96 Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Approaches to defining quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Direct quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Indirect quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pure quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mixed quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emphatic quotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scare quotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paralinguistic markers of quotation . . . . . . . . . . . . . . . . . . . . . Quotation and the semantics-pragmatics debate . . . . . . . . . . . Theories of quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The concept of name-mentioning quotation . . . . . . . . . . . . . . . A copula-based approach to NMCs . . . . . . . . . . . . . . . . . . . . . . Identificational copula . . . . . . . . . . . . . . . . . . . . . . . . . . . Equative copula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drawing the boundaries between copulas . . . . . . . . . . Quotation and the use-mention distinction . . . . . . . . . . . . . . . . State of the art: Experimental and empirical research . . . . . . Research gap & objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quotation and reference to names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lexical-semantic and pragmatic properties of name-mentioning quotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . The naming status of nominal mentioned expressions . . . . . . The different readings of naming constructions . . . . . . . . . . . . Conventionalized readings . . . . . . . . . . . . . . . . . . . . . . . . Modalizing readings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . <?page no="7"?> 5 99 5.1 100 5.2 102 5.3 105 5.4 107 5.5 108 6 113 6.1 113 6.2 114 6.3 116 6.4 118 6.5 126 6.6 131 7 133 7.1 133 7.1.1 133 7.1.2 133 7.1.3 135 7.1.4 138 7.2 139 7.2.1 139 7.2.2 143 7.2.3 146 7.2.4 151 7.2.5 154 7.2.6 162 8 165 8.1 165 8.2 165 8.3 169 8.3.1 171 8.3.2 173 8.3.3 174 Characteristics of metalinguistic predicates . . . . . . . . . . . . . . . . . . . . . . . The naming predicate call . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The naming predicate nennen . . . . . . . . . . . . . . . . . . . . . . . . . . . The naming predicate refer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The naming predicate bezeichnen . . . . . . . . . . . . . . . . . . . . . . . . Additional naming predicates and their collocates . . . . . . . . . Empirical investigation I: Corpus-based construction analysis . . . . . . . . . Research questions and hypotheses . . . . . . . . . . . . . . . . . . . . . . Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Intermediate summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental investigation I: The interpretation of NMCs . . . . . . . . . . . . NI vs. MOD: A rating study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research question and hypothesis . . . . . . . . . . . . . . . . . Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis & Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NI vs. MOD: A self-paced reading study . . . . . . . . . . . . . . . . . . Accounts of verbal irony . . . . . . . . . . . . . . . . . . . . . . . . . The processing of irony: Empirical evidence . . . . . . . . . Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental investigation II: Evidence from eye-tracking . . . . . . . . . . . The processing of quotation marks in German NMCs . . . . . . . Characteristics of eye movements in reading . . . . . . . . . . . . . . Empirical evidence on punctuation marks . . . . . . . . . . . . . . . . Research question and hypotheses . . . . . . . . . . . . . . . . . Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contents 7 <?page no="8"?> 8.3.4 175 8.3.5 177 8.4 178 8.5 188 9 191 203 229 229 229 232 232 236 236 237 239 240 240 241 241 241 242 242 244 255 256 258 260 260 261 262 Materials and experimental design . . . . . . . . . . . . . . . . . Procedure and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . Data cleaning and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix A: Corpus study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Excepts from DeReKo and BNC . . . . . . . . . . . . . . . . . . . . . . . . . Appendix B: Rating Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix B1: Instructions for rating the interpretation . . . . . Appendix B2: Sample sentences with mean rating of participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ratings for bezeichnen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ratings for nennen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix B3: Total counts for the predicates (mean values by participants) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix C: Self-paced reading study . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix C1: Instructions for the SPR-study . . . . . . . . . . . . . . Appendix C2: Training Items of the SPR-Study . . . . . . . . . . . . Practice Item 1 (MODC_ns): . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practice Item 2 (NIhighC_ns): . . . . . . . . . . . . . . . . . . . . . . . . . . . Practice Item 3 (MODC_nm): . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix C3: Target words and their respective lexical frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix C4: Overview of all experimental items . . . . . . . . . . Appendix C5: Word length of the target words . . . . . . . . . . . . Appendix C6: Reading Time for the target word . . . . . . . . . . . Appendix C7: Reading Time for the spillover region . . . . . . . . Appendix D: Eye-Tracking study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix D1: Instructions for the experiment . . . . . . . . . . . . . Appendix D2: Overview of the stimuli . . . . . . . . . . . . . . . . . . . Appendix D2: Experimental items . . . . . . . . . . . . . . . . . . . . . . . 8 Contents <?page no="9"?> 265 267 269 271 Appendix D3: Participant-specific details . . . . . . . . . . . . . . . . . Appendix D4: Heat maps for target items Quotes and noQuotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contents 9 <?page no="11"?> Acknowledgements This book is a revised version of my dissertation that I submitted in April 2024 and defended on September 10, 2024 at the University of Kassel. First and foremost, I would like to express my gratitude to Holden Härtl for his supervision, support, and valuable feedback throughout the years: from the discussions in his seminars to guiding me through the process of conducting experiments and fostering my fascination for research. Thank you for your dedication and encouragement! I would also like to thank Philippe De Brabanter for serving as a second supervisor, his detailed feedback, and for joining the defence in person. I also extend my thanks to Tanja Angelovska and Lars Heiler for being part of the commission and their engaging questions as well as Marcel Schlechtweg and Florian Scharf for supporting the early stages of my PhD in writing the recommendation letter and extension assessment for my scholarship application. Many thanks go out to the University of Kassel scholarship for funding my research and conference travels. During my PhD, I had the opportunity to meet various linguists in different countries who provided useful feedback on my research and presentations. I would therefore like to thank the audience of various conferences. Participating in conferences, having discussions and exchanges about the PhD journey has been very engaging and inspiring. Most notably, I would like to thank the reviewers and audience of the 13th Humor Research conference (Texas A&M University-Commerce), the Manchester Fo‐ rum in Linguistics (The University of Manchester), Linguistic Evidence 2022 (University Paris) and 2024 (University of Potsdam), the Workshop on Syntax, Semantics and Phonology (University Barcelona), the DGfS annual meeting, and the CogSci23 (University of Sydney) reviewers for their comments and suggestions. I also enjoyed joining the CogSci mentor program and would like to thank Javier Aguado-Orea for his input. In addition, I would like to thank the team of the open access publication fund from the Kassel University Library for their advice and who made this open access publication possible. Special thanks go out to Álvaro Cortés Rodríguez - I always enjoyed our (scientific) talks and long exchanges on various matters. Additionally, I would like to thank Christine Riess, Kristina Weissbecker, Peter Hofmann, Tamara Bodden, Summer West, Katharina Zaychenko, and Anne-Kathrin Pohl for our exchanges and their (technical) support when preparing for the defence. Joining the GeKKO meetings has always been fun and a great opportunity to discuss <?page no="12"?> and share research ideas. I would also like to thank the student assistants Maria Luisa Viceconte, Lanely Lorentz and Paul Weidemann for proofreading, creating experimental items, and assisting with experiments. Heartfelt thanks go out to my parents and sister who always supported and believed in me. And thank you, Johannes, for your support, encouraging words, our PhD talks, for always listening and being there for me. I would also like to thank Katharina for our weekly chats and inspiring questions. Last, thanks to my “office companion dogs”, Coffee and Carlos, who joined my home office and work hours and reminded me to go for a walk and have some fun. 12 Acknowledgements <?page no="13"?> Abbreviations ACC Accusative AOI Area of Interest (also IA, Interest Area) AP Adjective Phrase BNC British National Corpus CP Complementizer Phrase CQL Corpus Query Language DeReKo Deutsches Referenzkorpus DP Determiner Phrase ECM Exceptional Case Marking EEG Electroencephalography ERP Event-related brain potentials FFD First Fixation Duration FRD First Run Dwell Time GEN generic KWIC Key word in context MOD Modalized reading ms milliseconds NI Name-informing reading NIC Name-informing construction NIQ Name-informing quotation NMC Name-mentioning construction NP Noun Phrase ODE Oxford Dictionary of English PP Prepositional Phrase QM Quotation marks RQ Research Question <?page no="14"?> RT Reading Time SFD Second Fixation Duration ToM Theory of Mind WMC Working Memory Capacity 14 Abbreviations <?page no="15"?> Note of the author Remarks on conventions Metalinguistic uses of words that refer to concepts and terminology are written in italics. Consider for example the following sentence: ‘The term reference encompasses a multitude of functions in linguistics.’ Quotes by other authors follow English conventions and are therefore surrounded by double upper quotes, e.g., ‘Cappelen et al. (2023) claim that quotation is “one of the most difficult and interesting topics in the philosophy of language” (1).’ In contrast, words in German sample sentences follow German writing conventions, using opening lower quotes and closing upper quotes: ‘Dieses Phänomen nennt man einen „Regenbogen”. (‘This phenomenon is called a “rainbow”.’)’. For non-English words that are used metalinguistically, a translation of the words to English is given using brackets and single quotes: ‘The predicate nennen (‘call’) is frequently found in metalinguistic constructions.’ Quotes originally written in German were translated to English by the author. Mistakes in the original source are indicated with [sic! ]. <?page no="17"?> Data availability statement All files of the experiments are uploaded to the public OSF repository (osf.io/ t4c8f). The data of the corpus study consist of the annotated corpus data for all four predicates (see Chapter 6) and can be found here: https: / / osf.io/ g3ht6 / (DOI: 10.17605/ OSF.IO/ G3HT6). The rating of the forced-choice rating study on the interpretation of NMCs (see Chapter 7.1) is available here: https: / / osf.i o/ uqkwj/ (DOI: 10.17605/ OSF.IO/ UQKWJ). The data of the self-paced reading study (see Chapter 7.2) comprises the E-Prime 3 files, i.e., the original moving window designs, an overview of the experimental items used in the study, the raw data as well as the filtered data. The files are uploaded here: https: / / osf.io/ r p7ve/ (DOI: 10.17605/ OSF.IO/ RP7VE). For the eye-tracking study (see Chapter 8) all files, more specifically of a list containing all experimental items, eye-link files, Interest Area reports, the raw and filtered data files as well as R scripts, are openly available at https: / / osf.io/ 2m76x/ (DOI: -10.17605/ OSF.IO/ 2M76X). <?page no="19"?> 1 Introduction The theory of quotation has been of interest in philosophy for the last hundred years and has recently evolved as an emerging topic in linguistics. Its roots can be traced back to the pioneering works of Gottlob Frege, with recent developments further solidifying its importance in both semantic and pragmatic domains (Cappelen et al. 2023). Quotation is a crucial device in language to serve a variety of functions, including using the words of another person, highlighting a specific term, or introducing new terminology. The use of quotational constructions not only facilitates communication by specifying the terminology used and the corresponding references but also contributes to the enrichment of our vocabu‐ lary. It reflects a common linguistic process within the evolution of language, driven by the perpetual need to introduce specialized terms and adapt to the dynamic nature of human expression. The introduction of new terminology through sentential constructions contributes to the ever-expanding demands of communication, that is the need of language to expand the lexicon by introducing specialized terms. One type of metalinguistic uses of terms is characterized by sentential constructions that make use of a variety of naming verbs such as nennen (‘call’) and bezeichnen als (‘refer to as’) that function as name-selecting predicates (see 1a-b). (1) a. This piece of writing is called a “dissertation”. - b. The term “wildlife conservation” refers to the preservation and protec‐ tion of natural habitats and species. In this dissertation, the focus is on a specific type of metalinguistic construction, which will hereafter be referred to as a name-mentioning construction (NMC), see e.g., Härtl (2015). One of the functions of quotation marks is to draw a line between object language and metalanguage. In the sentences presented above (1a-b), the expression in quotation marks is an instance of metalanguage embedded in object language. Metalinguistic sentences are not only frequently found in public discourse but also prevalent in everyday conversation in private talks (cf. Müller 2022). The study of metalanguage in natural language is a cur‐ rently emerging topic with limited empirical evidence available (Wilson 2012). Metalinguistic constructions as illustrated in (1) are essential for interlocutors <?page no="20"?> 1 Throughout this paper, I use the term constructions to cover a structured grammatical unit forming a larger unit. to use language to talk about language. Quotation itself can be characterized as a “referential device” (Recanati 2001: 681). Metalinguistic constructions 1 serve a variety of different functions and speech communities make use of a “wide variety of metalinguistic constructions” (Wilson 2012: 638). According to Searle (1990), one of the oldest debates in Western philosophy and also the philosophy of language concerns the nature of linguistic reference, and the treatment of reference can even be traced back to Plato. Implementing a pragmatic approach, Abbott (2010) defines reference as “the three-place relation that holds when a speaker x uses expression y to identify entity z” (2). Reference can either be established to the names of concrete objects in the real world or abstract objects that are not visible to the eye and have no clearly identifiable or perceivable referent. For this reason, one possibility is to distinguish abstract from concrete objects, all of which are context-dependent. Generally speaking, concrete words are processed faster than abstract words in e.g., naming and lexical decision tasks (Borghi et al. 2017). Neuropsychological research argues for a distinction between the two while other theories focus on their similarities (Borghi et al. 2017). Referring nominals, also known as referential expressions or referential nom‐ inals, are linguistic elements used to refer to specific entities or concepts in the world. These nominals allow speakers to identify and talk about particular objects or entities in the world. Within the hierarchy of referring expressions, the traditional notion claims a difference in the referent’s accessibility or activation (Gundel & Abbott 2019). Quotational theory as well as the topic of reference and referring expression have received a considerable amount of attention in the philosophy of language. Classical theories suggest that the term reference involves a mental event: A speaker uses a word to mentally recall a concept of an entity (an object or event) in order to bring it into the mental focus of an interaction (Wildt et al. 2019). In this thesis, reference to lexicalized concepts will be analysed, looking specifically at sentential constructions that mention the name of a (lexicalized) concept, see (1a-b) above. Notably, mentioned nominals in the construction under investigation do not necessarily have to be lexicalized nominals but can also constitute ad-hoc formations (see also Chapter 4). To begin with, the title of this thesis will be split up to elucidate the basic terminology. Names, in this context, are not defined as proper names, but instead as names for lexicalized concepts. The notion of name as used in the current 20 1 Introduction <?page no="21"?> paper relates to the lexical-conceptual representation of a common noun and the label used for it (see, e.g., Bücking, 2010; Härtl, 2016, for detailed considerations). To put it simply, names are expressions used by speakers to communicate about things or objects. Quotation serves as a tool to mention the conventionalized name of a lexicalized concept. These concepts are abstract in nature and are most commonly purely mental concepts: they are mental categories. Meaning is encoded “into memory through concepts - ideas (mental representations) to which we may attach various characteristics and with which we may connect various other ideas, such as through prepositions” (Sternberg & Ben-Zeev 2001: 202). The words speakers use are arbitrary symbols for concepts. Having a word for a concept enables humans to add information to already existing knowledge (Sternberg & Ben-Zeev 2001). Concepts evolve over time; they can change their meaning and are therefore considered as being dynamic in nature. Furthermore, concepts and their respective definitions may vary within different cultures or situations and can therefore be context-dependent. In addition, interlocutors need to decide on a specific name to refer to entities. Once these entities are established, the listener can use this name to create reference to this previously unknown concept as well (cf. Kronmüller et al. 2017). The central aim of this investigation is to explore the metalinguistic uses of names both from a theoretical and empirical angle. A prominent question for this investigation is what exactly constitutes a “metalinguistic use” of words or phrases? Generally, metalanguage is a central part of everyday language that is neither language-specific nor restricted to a specific register. The Greek loanword meta denotes something that “consciously references or comments upon its own subject” (Dictionary.com 2023). Therefore, metalanguage is lan‐ guage used to describe language: it talks or makes statements about the object language. Making use of metalanguage allows people to refer to language in order to communicate about language itself: it contributes to the understanding of the language used. Quotational constructions can be used by speakers to establish metalinguistic reference. In (2a-b), the quoted words are used metalinguistically. (2) a. The term “snaccident” has recently evolved as a new slang term in America. - b. Experts use the term “algorithm” to refer to a step-by-step set of instructions or procedures designed to perform a specific task or solve a particular problem 1 Introduction 21 <?page no="22"?> 2 I would like to thank Philippe De Brabanter for highlighting issues concerning the use of empirical studies to resolve theoretical debates. Due to the complexity of the research topic in combination with the limited scope of the empirical data and considering that empirical findings should be replicable to effectively contribute to resolving theoretical debates, I would like to note that further research is needed if one wants to draw conclusions on theoretical frameworks for quotation. Proposals for a unifying account for the different types of quotation have recently emerged (cf. Härtl & Schlechtweg 2023) and future research directions should also address the question how we can bridge the divide across quotation types based on empirical work (see Hofmann & Raue 2024). Research on quotation in general has a strong background in philosophy of language, as noted beforehand, and has considerably increased in various areas of linguistics. Broadly speaking, this discussion is located at the philosophical exploration of language in the domain of psycholinguistics. The psycholin‐ guistic perspective encompasses the notion of metacognition, which describes the ability to think about thinking - in this case, thinking about the use of language. Metacognition offers insights into the processes that govern our linguistic understanding. Metacognitive knowledge refers to one’s awareness and understanding of cognitive processes (Beran et al. 2013). Making use of quotation requires metacognitive knowledge as the speaker needs the ability to understand different perspectives, which is crucial for using quotations effectively. Discussing the appropriateness of words in a given context can be regarded as explicit metacognition, which is possibly reflected in the use of a specific intonation - a tone of voice - or the use of scare quotes when a speaker distances themselves from the words used for a specific concept (Shea 2018). For an effective conversation, the speaker needs to know the difference between using and mentioning an expression, see Chapter 3.3 for a discussion on the use-mention distinction. The research questions explored in this work are most effectively addressed through an interdisciplinary approach, wherein theoretical debates serve as the foundation for the empirical data. 2 I regard empiricism and theory to be interconnected: empirical research often uncovers new issues that require a return to theoretical considerations, which may necessitate modifications to the original theory or even the development of new theories (cf. Völkel & Kretzschmar 2021). Due to the iterative process of research, aligning theory with empirical data throughout the research process might be challenging. Further challenges in aligning theory with empirical data is the dynamic nature of research. To give just one example, newly published philosophical papers including new theoretical approaches to the theorizing on quotation have emerged over the course of experimental research. For these reasons, the current work should be seen as a starting point for collecting further experimental 22 1 Introduction <?page no="23"?> 3 Bold print as used in the original. evidence to better connect theoretical and empirical work and allow for a closer connection between empiricism and theory in general. This work draws on a broader discourse within philosophy and linguistics, two fields that I regard as foundational to an understanding and analysis of metalinguistic constructions. The notion of quotation has been examined through semantic and pragmatic approaches and can be positioned at the interface between semantics and pragmatics. The philosophical grounding of this thesis builds on classical thinkers like Gottlob Frege, whose work on semantics, logic, and the concept of reference provides a critical basis for exploring the intersection of meaning and metalanguage. This conceptual foundation facilitates a direct link between longstanding philosophical debates and the examination of a specific linguistic phenomenon, establishing a basis from which theoretical constructs can be empirically tested for. The intersection of philosophy and linguistics is relevant in studies on the semantics-pragmatics boundary to address questions about the nature of reference, meaning, and naming. This study situates itself within this interdisciplinary space, aiming to advance both the theoretical understanding of quotation and metalinguistic uses of language, as well as to contribute empirical data that might refine philosophical inquiries: here the linkage between philosophy and linguistics is operationalized through an empirical methodology that integrates a corpus analysis, self-paced reading, and eye-tracking. The primary emphasis of this thesis is the notion of linguistic reference within quotational contexts. The thesis aims to investigate referring expressions, more specifically nominals in quotational contexts. Consider the following metalinguistic sentence, which has been introduced as a name-mentioning construction in the scientific literature. (3) In linguistics, the words and expressions that people use to describe or ref er to language can be called metalanguage. 3 (Collins Dictionary 2023) The mentioned nominal, here the word metalanguage, refers to a lexicalized concept, and is introduced using the metalinguistic predicate call. Various studies have already revealed the multifaceted interplay of factors in quotational constructions. However, the precise nature of NMCs remains unclear. More specifically, this concerns the syntactic variability of NMCs, collocational patterns of metalinguistic predicates, the linguistic properties 1 Introduction 23 <?page no="24"?> of the mentioned nominal such as the lexical frequency and word type, the occurrence of quotes in NMCs, and the online processing of quotational NMCs during reading. As a consequence, further experimental and empirical research to investigate the linguistic characteristics of NMCs from a theoretical and cognitive-scientific approach is necessary. The focus is on the relationship between lexical and pragmatic factors involved in name-mentioning quotation, as well as the interpretational variation it is subject to. The aim of this thesis is to investigate the linguistic characteristics of NMCs in German and English, examine the systematicity in the use of quotation marks, and the cognitive mechanisms involved in processing literal and modalizing NMCs during read‐ ing. To investigate these research gaps, various empirical studies with different methodological approaches on NMCs will be presented, including a corpus study, a forced-choice study, a self-paced reading design as well as an eye-track‐ ing study. In order to analyse NMCs, the focus will first be on a theoretical understanding of this linguistic phenomenon before presenting the empirical investigations addressing the research questions. This thesis is structured as follows. Chapter 2 approaches the question how quotation can be defined and thus characterizes the different types of quotation, followed by a discussion of paralinguistic markers used for quotation. It further addresses the location of quotation within the debate on the semantics-pragmatics divide. The chapter closes by presenting the varieties of quotational theories. Chapter 3 explores the basic characteristics of name-mentioning quotation. It serves at providing a first overview of the topic and explores how NMCs can be analysed using a copula-based approach in considering identificational and equative copula constructions. In the following sub-chapters, the use-mention distinction will be discussed. The chapter continues with an outline of the state of the art of NMCs from an experimental and empirical perspective. It closes with a formulation of the research gap and research objectives. Chapter 4 addresses how quotation creates reference to names in investigating the lexical-semantic and pragmatic properties of name-mentioning quotation. The chapter furthermore explores the naming status of nominal expressions mentioned within the context of quotation. Additionally, it outlines the different readings effective in naming constructions: a conventionalized and a modalizing reading. Chapter 5 is dedicated to a brief sketch of the linguistic characteristics of metalinguistic predicates. It focuses on the four most prominent naming predicates in German and English, namely call, nennen (‘call’), refer and bezeichnen als (‘refer to as’). Last, collocational patterns are investigated and other relevant naming 24 1 Introduction <?page no="25"?> predicates are introduced. In Chapter 6, the first empirical investigation is presented: a large-scale corpus analysis that aims at a first examination of linguistic properties of the mentioned nominal in German and English con‐ structions involving call, nennen (‘call’), refer and bezeichnen als (‘refer to as’). Chapter 7 presents the first experimental investigation and is comprised of two studies exploring the relationship between a modalizing and name-informing interpretation of NMCs. The first study is a forced-choice study that builds on the German corpus data while the second study employs a self-paced reading design. Chapter 8 explores the processing of quotation marks in literal NMCs in German, implementing an eye-tracking study. Finally, Chapter 9 presents a general conclusion by summarizing the main findings in answering the research questions, and offering potential avenues for future research. 1 Introduction 25 <?page no="27"?> 4 Throughout this work, the term quoted material will be used to refer to any linguistic word that occurs in quotational constructions, possibly enclosed in quotation marks in the written mode. However, this quoted material does not necessarily have to be indicated by quotation marks or other types of graphemic marking. 2 Approaches to defining quotation In this section, general approaches to quotation will be considered and discussed with respect to their limitations. In the following sub-chapters, several examples will be used to illustrate the wide varieties of quotation types. To begin with, the varieties of quotation marks are one of 12 different types of punctuation marks: < . ; , : - - … ’ ? ! ( ) „“ > (see Brendel 2017). A common definition of the term quotation holds that quotation is “something that is quoted” (see Merriam Webster 2023). Another definition in the Oxford Advanced American Dictionary states that a quotation is a “group of words or a short piece of writing taken from a book, ply, speech etc. and repeater because it is interesting or useful”, adding a second definition as “the act of repeating something interesting or useful that another person has written or said” (OAAD 2023). As the reader may already guess, these definitions are neither exhaustive nor able to account for all different types of quotation, but they provide an overview of the common understanding of the phenomenon of quotation, namely as an instance of direct quotation. In the Macmillan Dictionary, another layer is added to the quotation by alluding to the use-mention distinction: “words from a book, play, film etc that you mention when you are speaking or writing” (2022) taking in account oral and written language. However, it will be shown that quotation is a complex phenomenon with a variety of definitory approaches, which gives rise to extensive philosophical discussions. From a syntactic approach, quotation is defined as everything that is written in between quotation marks, e.g., a whole sentence, a phrase, or just one word or a letter (cf. Cappelen et al. 2023). Punctuation practices can differ cross-linguistically. In Britain, it is a common practice to use single apostrophes (‘ ’) while American English prefers using double apostrophes (“ ”) instead. Even bold letters or underlined words are possible in English to indicate quoted material 4 (cf. Cappelen & Lepore 1997). In parts of Europe, quotation can be indicated by double angles (< >), and sometimes italics are used to set the quote apart. In French, double angles are referred to as guillements and date back to the 16 th century but are rarely documented in manuscripts before the 17 th -century (Issel-Dombert 2018). In written texts, quotations in German stereotypical <?page no="28"?> 5 The original work was digitalized by the Munich Digitization Center (MDZ) and is available here: https: / / www.digitale-sammlungen.de/ en/ view/ bsb00002081? page=1 6 The corpus is based on 35,021,957 sentences. preceded a quote with lower double quotes and ending with upper quotes. From a historical perspective, quotes can be traced back to the 15 th century (Schmieder 2013). Despite this, the first documentation of written quotes allegedly goes back to the work “De Vitis Sophistarum” by Flavius Philostratus 5 which was printed in the years 1516 in Strasbourg. In contrast to today’s practice of quotation marks, quotation marks were placed at the beginning of the line and continually placed there through multi-line text presentations. To put it differently, the quoted material was not simply surrounded by the marks, but rather “accompanied” by it on the left margin of the manuscript. The handwritten marks were double and closing quotation marks. Research up to the present date indicates that the marks of quotation had their origin in France some time about 1580-1590, two commas or turned commas in the margins being used to indicate cited passages (McMurtrie 1934). In the 18 th century, opening and closing quotation marks started to replace the single quotes that were used in the margin of the page. Quotation marks were more extensively used starting from the 19 th and 20 th century. In text documents from the 19 th century, the use of quotes for direct quotes was extended to the function of highlighting titles, proper names, and also for metalinguistic uses (von Polenz 1999). Despite this, online articles published by newspapers on social media for example show a tendency in recent years to adapt to the English graphemics, using only upper quotes. Examples come, among others, from online articles published by ZDF. As of now, this phenomenon and change has not been of interest in the linguistic literature. To illustrate the typographical variety of quotation marks in newspaper articles, consider the overview of punctuation marks sorted by frequency from the German News Corpora 2020. 6 28 2 Approaches to defining quotation <?page no="29"?> Punctuation mark Type of quotes Frequency " Double upper quotes; straight 4,523,836 „ Opening double lower quotes 2,104,571 “ Opening double upper quotes, curved 2,032,439 « Closing double angles 794,628 « Opening double angles 655,007 ˮ Closing double upper quotes; curved 26,772 Tab. 1: Overview of quotation marks in the German News Corpora 2020 Punctuation mark Type of quotes Frequency " Double upper quotes; straight 3,318,491 „ Opening double lower quotes 2,554,756 “ Opening double upper quotes, curved 2,410,978 « Closing double angles 645,044 » Opening double angles 581,685 ” Closing double upper quotes; curved 14,357 Tab. 2: Overview of quotation marks in the German News Corpora 2023 The two datasets from 2023 and the German News Corpora 2020 exhibit differ‐ ences in the frequency of various punctuation marks and their corresponding types of quotes. The frequency of curved closing double upper quotes sees a substantial decrease from the 2020 dataset (26,772) compared to 2023 (14,357). While there are slight variations in frequencies for other types of quotes, the overall distribution remains relatively consistent between the two datasets. Although not reflected in the corpus data, a decline of closing upper quotes seems to have manifested as a trend over the years in German news reports: upper quotes are nowadays frequently found in written text documents. This might indicate that the German convention adapts to the English typology. The repertoire of quotation marks used in the English News corpora is restricted to the following six types. 2 Approaches to defining quotation 29 <?page no="30"?> Punctuation mark Type of quotes Frequency " Double upper quotes; straight 2,210,935 „ Opening double lower quotes 8,277 “ Opening double upper quotes, curved 4,330,968 « Closing double angles 1,296 » Opening double angles 3,447 ” Closing double upper quotes; curved 531,183 Tab. 3: Overview of quotation marks in the English News Corpora 2020 Other counter arguments against a syntactic characterization of quotation are the lack of correlates in spoken language at least as far as lexical items concern, e.g., quote/ unquote (Cappelen & Lepore 2007). Above this some philosophers argue that there would be no such thing as a unified syntax for quotation and that the syntactic role of quotes varies. The great variation in the implementation of quotation marks in the written mode is not restricted to the type of quotation. Therefore, Wilson (2012) notes the following regarding the inconsistency in the practice of quotation marks: Delineating an instance of metalanguage with quotation marks is a common conven‐ tion (…). Quotation marks, italic text, and bold text—three common conventions used to highlight metalanguage—are inconsistently applied and are already “overloaded” with several distinct uses. In order to tackle the issue of the multifunctional duty of graphemic marking and inconsistency, new practices of quotation have been introduced as optional devices. Reichenbach (1947) has suggested using arrows (↘↙) for token quotes. Quine (1940) proposed an alternative notation for setting instances of quotation apart, using corner quotes (‘┌’ and ‘┐’). Most frequently, quotation marks are used for reported speech. Despite this, quotation marks are also used to indicate neologism, regionalism, foreign and colloquial lexemes, to highlight titles of books and articles, and for metalinguis‐ tic uses (Issel-Dombert 2018). Moreover, guidelines for writers would further mention another function of quotes: to create distance referred to as distance ironique. Issel-Dombrecht (2018) highlights that quotes with an ironic distancing function would not follow normative rules, but that their usage depends on the writer. Klockow (1980) summarizes the function of double angles as follows: 30 2 Approaches to defining quotation <?page no="31"?> 7 Other quotation types that have been discussed in the literature include impure quotation (Gómez-Torrente 2003), double-duty quotation (García-Carpintero 2003), sub-clausal (Potts 2007), and incorporated quotation (Clark & Gerring 1990). In the following, the terminology as presented in the examples in (4a-h) will be applied and is in line with the argumentation […] bloßes signal für die präsenz einer zusätzlichen sinnebene, eine art aufforderung, den markierten ausdruck nicht in naiver weise zu interpretieren. Wie er zu interpretieren ist, muß der leser selbst erschließen, wobei nicht nur der kontext, sondern auch die kenntnis sonstiger verwendungszusammenhänge […] eine rolle spielen kann [sic! ] (157) […] a mere signal for the presence of an additional sense level, a kind of an invitation not to interpret the marked expression in a naive way. How it is to be interpreted, the reader must discover himself, wherein not only the context, but also the knowledge of other context of use […] can play a role We will get back to the metalinguistic function of quotation marks in more detail in Chapter 3.3.3 and Chapter 4. For this work, I argue that double and single quotes do not indicate a difference in the function, such as using single quotes only for a distancing interpretation while double quotes would signal a direct quote (cf. Caduff 2009). Instead, the argumentation in this dissertation is based on the assumption that the type of quotation marks - and even italics - is not restricted to one specific function but can serve several functions. Whether quotation marks are a defining characteristic of quotation is still discussed in the literature (cf. Wang 2018; De Brabanter 2023) and will also be discussed with regards to the corpus analysis in Chapter 6. In sum, the outline on the debate on an appropriate definition of quotation has been shown to pose difficulties. The approach used for this work is one that acknowledges that not all instances of quotation are properly marked by quotation marks. To account for the inconsistency in written form, the terminology quotational construction will be used in the following to refer to instances that qualify for enclosing the expression in quotation marks, most commonly mentioned items. 2.1 Types of quotation Quotation is a phenomenon that occurs in a variety of different linguistic environ‐ ments. In the taxonomy of quotation, at least five different types of quotation are commonly distinguished (see e.g., Brendel et al. 2011 and Cappelen & Lepore 2007 for overviews). 7 The difference between the various types of quotation depends on the original’s nature and source, the degree of similarity between the original 2.1 Types of quotation 31 <?page no="32"?> presented in Härtl (2020). However, please note that scholars might adopt different names for the different types of quotation depending on their research purposes. and quotation, and the intended use of the quotation. To illustrate the variety of different types of quotation, the sample sentences from Meibauer (2014: 185) are used below. Sample sentences in (4a-f) make use of the original utterance The project is hard to understand uttered by a person called Ken. (4) Direct Quotation a Ken said, “The project is hard to under‐ stand.” - Indirect Quotation b. Ken said that the project is hard to under‐ stand. - Parenthetical Quotation c. The project, said Ken, is hard to understand. - Mixed Quotation d. Ken said that the project is “hard to under‐ stand”. - Scare Quote e. Ken said that the “project” is hard to under‐ stand. - Free (in)direct Quota‐ tion f. Ken accelerated his steps. The project is hard to understand. But he would manage it. - Empathic Quotation g). “project” [written on a door] - Pure Quotation h. “Tcejopr” is not an English word. The different types of quotations will briefly be discussed and further analysed in the following sub-chapters. For the purpose of this thesis, the focus is exclusively on the most commonly discussed types of quotation to gain a comprehensive understanding of the taxonomy. (5) a. “Kassel is a city in Hesse,” Sam said. [direct quotation] - b. Anne said she planned to travel to France. [indirect quotation] - c. “Kassel” has six letters. [pure quotation] - d. The coach declared that his team would “win the game” today. [mixed quotation] - e. The “prodigy” was in fact a master of misin‐ formation. [scare quotes] 32 2 Approaches to defining quotation <?page no="33"?> The specific characteristics of the type of quotation will be addressed in following the structure of the quotational types presented above. 2.1.1 Direct quotation Direct quotation is the most researched and well-known type of quotation. It is characterized by using the exact wording of a person or some written words, enclosing these words in quotation marks. It can be differentiated between three different types of direct quotation, all of which can be found in German and English. (6) a. “This,” said Marc, “is going to be fun.” - a’. Mark sagte: „Das wird Spaß machen! “ - b. „Das wird Spaß machen”, sagte Mark. - b’. “This is going to be fun”, said Marc. - c. Marc sagt: „Das wird Spaß machen! ” - c’. Marc said, “This is going to be fun! ” Note the difference in notation between German and English, in which German makes use of a colon while English never does. With precise reproduction of someone’s spoken or written words quotation marks are mandatory. In all three sentential constructions presented in (6), the direct quote is a syntactically independent clause that can function as a “head clause with a quotative adjunct” (D’Arcy 2015: 44). Despite this, direct quotations can also appear as syntactically dependent elements, such as when the quoted structure constitutes a prepositional phrase. Direct quotation as illustrated above are frequently found in written text. However, they are not a frequent constituent in spoken language (D’Arcy 2015). In direct quotations, words have originally been uttered in a different context. They are used to convey an utterance verbatim (cf. Brendel et al. 2011) as a reproduction of the speaker’s original words. The hearer or listener needs to interpret the direct quotation while taking into consideration the context in which the original utterance has been made. This type of quotation is consequently used to reproduce an original utterance by making use of the same wording, including all particularities of the original utterance, e.g., a dialect voice (Evans 2012). The speaker reproduced an utterance by another speaker by referring back to the original speaker. Direct quotation fulfils the 2.1 Types of quotation 33 <?page no="34"?> purpose of showcasing arguments mentioned earlier, allowing for present and/ or challenging a particular speaker’s viewpoint, and leveraging an authority appeal (Docherty & Mach 2023) as direct quotes usually mention the name of the speaker who uttered the words. 2.1.2 Indirect quotation In contrast to direct quotes, indirect quotes are not syntactically independent. The distinction between direct and indirect quotation is further characterized that pronouns need to be changed as well as tense and mood in order to report previously uttered words (cf. D’Arcy 2015). However, the distinction between direct and indirect quotes is, not evident cross-linguistically, meaning that not all languages realize this distinction grammatically (Güldemann 2008). The indirect quote in English may optionally be introduced by the complementizer that followed by the complement clause as the quoted content. In German, complementizers, verb-position, and subjunctive forms are used in indirect quotations. The sample in (7) represents a stereotypical indirect quote. (7) a. Marc said that swimming is fun. - b. Marc sagte, dass Schwimmen Spaß macht. In addition to reporting utterances, indirect quotes fulfil the function of express‐ ing beliefs and thoughts (Brendel et al. 2011). This is often contrasted with direct quotes, which are typically seen as being limited to reporting speech. Within the reported clause, uoted speakers use first-person pronouns to refer to themselves, pronouns that are coreferential to third-person constructions in the quotation frame (i.e., she said “I …”; they said “we …”); other deictic indicators of a reconstructed speaker and source context include adverbs (temporal and spatial), tense, and mood (D’Arcy 2015: 45). Shifts from direct to indirect quotes can be found in temporal deixis, personal deixis, and local deixis (underlined below). (8) a. Julia sagte gestern: „Ich werde morgen dort sein.“ - b. Julia sagte gestern, dass sie heute da sein würde. 34 2 Approaches to defining quotation <?page no="35"?> All expressions that are deictically sensitive are presented from the perspective of the speaker quoting the original speaker (Evans 2012). Despite this, deictic shifts are not a defining criterion or a test for indirect quotes (Coulmas 1986). What differentiates direct from indirect quotes is the point of view or perspective of the speaker: In direct speech, the reporter adopts the original speaker’s point of view, in a sense, by lending his voice to what they say or write. The reporter takes centre stage in indirect speech (Coulmas 1986). In direct speech the perspective of the original speaker is not changed while there is a perspective shift into the direction of the reporting speaker (De Roeck 1994). Although there is a long-standing tradition of distinguishing between direct and indirect quotation, there is currently a debate whether indirect quotation should be subsumed as a quotation type. De Brabanter (2010) for instance argues against treating indirect quotation as quotation due to the lack of the mention of a linguistic representation. We will now move on to the third type of quotation, namely pure quotation. 2.1.3 Pure quotation Consider the instances of pure quotation, also called pure quotes, as illustrated in (9). (9) a. “Cat” has three letters. - b. “N” is a letter of our alphabet. What sets pure quotation apart from direct, indirect, and mixed quotation is the lack of “attribution to any utterance or saying” (Cappelen & Lepore 2007: 14). The word in quotation marks does not refer as it commonly does. Pure quotes refer to the linguistic material of the word; it focuses on presenting the word itself for discussion. The addressee can attribute to the word cat in (9a) the property of having three letters. Instead of referring to the stereotypical denotation of the word, the quotation marks have the function of referring to the properties of the word. The quoted words in (9) are mentioned. Pure quotation is, as Searle (1969) puts it, a tool by which the “word itself is presented and then talked about, and that it is to be taken as presented rather than used conventionally to refer” (75-76). Crucially, pure quotations do not add any additional meaning to the interpretation of the sentence. Pure quotation can, in contrast to direct quotation, encompass various linguistic elements, including entire words or single letters. Even nonsensical 2.1 Types of quotation 35 <?page no="36"?> words might be implemented. Pure quotation has attracted attention by philos‐ ophers as it “creates a referentially opaque context which prevents the quoted expression from referring normally to something extralinguistic” (Wang 2018: 29). Although sentence in (9a-b) are commonly labelled as a type of “pure quotation” in the literature, Saka (2003) uses the term metalinguistic citation and Recanati (2001) word closed quotation. In instances of pure quotation, quotation marks serve as a tool that enables the quoted item to become metalinguistic, i.e., mentioned. Quine (1940) argues that mentioned nominals denote their interior. Instances of mention in pure quotes can for example serve to describe the orthographic, phonological or grammatical properties of a linguistic expression (De Brabanter 2003). Maier (2014) states that adding quotation marks to a pure quotation would turn the quoted expression into a referential expression which refers to the word itself: (10) a. “Cicero” has six letters. - b. #“Tully” has six letters. The words within the quotes cannot be replaced by a word with the same mean‐ ing because this renders the sentence illogical. This illustrates that although Cicero equals Tully in a way that they refer to the same entity in the world, the truth conditions are violated. It is not possible to preserve the quotations’ truth values when substituting synonyms or co-referential terms (Cappelen et al. 2023): The logical operation is blocked, meaning that substitution salva veritate is not possible. (11) a. Kassel is famous. - b. “Kassel” is disyllabic. Replacing Kassel with a synonymous phrasal expression in (11a) such as the largest city in North Hesse does not change the truth-value of the sentence. However, the truth value does no longer hold for (11b) in an opaque context when interchanging two expressions with the same extension (see De Brabanter 2003). Though not applicable for direct and indirect quotes, there is an ongoing debate whether quotation marks are necessary and/ or mandatory in pure quota‐ tion (cf. Gutzmann & Stei 2011, and the discussion in Section 6.5). Some scholars have argued that pure quotations without quotation marks are ungrammatical, 36 2 Approaches to defining quotation <?page no="37"?> while others argue that quotation marks are implicitly surrounding the pure quotation (Bazzoni 2016). (12) *In the garden has three words. (Bazzoni 2016: 123) Saka (2003) argues that the reading in a construction like (13a-b) changes from an object-level reading for (13a) to a meta-level reading (13b). (13) a. Arial is elegant. - b. Arial is elegant. (Saka 2003: 197) In the first sentence, the speaker indicates that the font Arial is elegant. In contrast, the second sample refers to a certain name being elegant. Furthermore, Gutzmann and Stei (2011) show that the occurrence of quotes in pure quotation has an influence on the grammaticality of the sentence. Instance of pure quotation can be tested for by adding an apposition like the word, see (14), which highlights the word’s metalinguistic status (cf. Cortés Rodríguez et al. 2022). (14) a. Dogs is a noun. - b. The word dogs is a noun. In Wang’s (2018) view, “reference is in essence mention in disguise” (27). According to Klockow (1980), there are defining characteristics of what he called an L-Zitat (logical citation). Even when the quoted material is written in plural, the item functions as a singular term: (15) a. „Katzen“ ist kein englisches Wort. “Cats” is not a word in English. - b. *„Katzen“ sind kein englisches Wort. *“Cats” are no word in English. This special type of quotation functions as a syntactically autonomous structure. Pure quotation has also been discussed with regards to slurs, questioning the impact of quotation marks on the interpretation of offensive words. Anderson & Lepore (2013: 354) use the pure quotation in (16) for their argumentation. 2.1 Types of quotation 37 <?page no="38"?> 8 I am thankful for Philippe De Brabanter pointing out that these findings are restricted to corpus data relying on journalistic texts: while the findings are true for this specific genre of written language, direct quotation should be more prevalent in spoken language. (16) “Kike” is a derogatory word. Both semantic attributions and quotation marks have the potential to cause offense. In the forthcoming analysis, NMCs will be analysed as a manifestation of pure quotation, aligning with the framework outlined in Härtl (2020). However, the categorization of NMCs solely as pure quotation remains subject to debate. There is a plausible argument for considering NMCs as an example of mixed quotation, which encompasses elements of pure quotations. The inquiry into the categorization of NMCs is tied to the classification and definition of various forms of quotation, although this specific debate will not be in focus of the present investigation. 2.1.4 Mixed quotation Mixed quotation has been investigated by philosophers and linguists as a type of quotation that functions as a complement to a speech verb. In essence, mixed quotation serves two duties which is why mixed quotes are a combination of direct and indirect quotes. On the one hand, mixed quotes refer to an original utterance while, on the other hand, they are quoting indirectly. For this reason, Maier (2007) claims that mixed quotation has a dual nature. Cappelen & Lepore (2007) argue that mixed quotation is more frequent in everyday communication than direct quotation 8 , providing a list why mixed quotation can be preferred over both direct and indirect quotation. Below the sample of mixed quotation from Stern (2022: 40) is reproduced. (17) Trump said that Obama “could not care less about terrorism”. The quoted material is both used and mentioned at the same time. As a consequence, the quoted material is a part of the sentence from a syntactic and semantic perspective. In the sample sentence above, what Trump said is quoted and the exact expressions used are repeated. Subsequently, the original utterance is partly conveyed verbatim (Brendel et al. 2011). Shan (2010: 417) 38 2 Approaches to defining quotation <?page no="39"?> illustrates the two dimensions of mixed quotation with the following sentences, using the semantic value f. (18) - Quine says quotation “has a certain anomalous feature”. - (mention) has a certain anomalous feature is used to mean some f. - (use) Quine says quotation f. The quoted words interact with both the syntactic structure and the semantics of the sentence. The terminology of mixed quotation was originally introduced by Donald Davidson in 1979. Some researchers, including among others Recanati (2001), Saka (2003), and García-Carpintero (2005), claim that mixed quotation would be semantically redundant, which is argued against by Cappelen & Lepore (2007). If the redundancy view would hold, quotation marks should not give rise to a change in meaning when removed. An interesting testing ground are indexicals in mixed quotation, which can change the reference to different people when leaving out quotation marks. For this reason, Cappelen & Lepore (2007) argue that mixed quotation indeed make a semantic difference. The reference to the words enclosed in the mixed quotation cannot be cancelled. (19) Trump said that Obama “could not care less about terrorism,” but he never uttered the words “could not care less about terrorism.” For Recanati (2008) mixed quotation is a pragmatic phenomenon as it depends on what people do with words. Due to the simultaneous use and mention in mixed quotation, it shares overlapping features with the type of pure quotation (see also Maier 2014). 2.1.5 Emphatic quotes Just like pure quotes, emphatic quotes do not refer to an original utterance. They are also discussed as greengrocer’s quotation, “noncitational” quotes by Abbott (2005) or “mystery” quotes by Saka (2013). Empathic quotes are used to highlight a specific word (Brendel et al. 2011), they are intentionally drawing attention to the word in quotation marks (Gómez-Torrente 2017). (20) a. This novel is “amazing”. - b. These fruits are “fresh”. 2.1 Types of quotation 39 <?page no="40"?> Here, the word fresh is highlighted but is not intended to give rise to an ironic interpretation. The emphasized or highlighted items are a proper part of the sentence. Notably, empathic quotes neither refer to an original utterance nor do they constitute an instance of metalanguage (Stei 2007). The quotation marks lead to a pragmatic enrichment. Nevertheless, research on empathic quotes is generally scarce, and debates continue about whether these sentences should be considered incorrect for not fulfilling the typical function of citing or expressing a distancing interpretation. 2.1.6 Scare quotes Another noteworthy type of quotation is known as scare quotes which dates back to Geach (1957). Predelli (2003) defines scare quotation as applying “a standard technical term in a nonstandard way than to invent a new term” (2). Consider the sentence replicated from Predelli (2003: 3). (21) The “debate” resulted in three cracked heads and two broken noses. The quotation marks trigger an additional pragmatic effect, i.e., pragmatic ambiguity (see Meibauer 2014), namely that the speaker considers the term debate again. The speaker does not believe the term to be appropriate in this context. Scare quotes signal an expressions’ non-standard use. Predelli (2003) further distinguishes between the message of an utterance on the one hand and the attachment, the specific message transported through the use of quotes, on the other hand. A common feature of scare quotes is that they may or may not reproduce any previously uttered words (Brendel et al. 2011; Meibauer 2014). For Predelli (2003), scare-quoting constitutes a type of conventional implicature; the quotation marks would function as so-called “attachment triggers” (7). The attachment of the sentence is the meaning conveyed by the implementation of quotation marks. The attachment is regarded as a semantic contribution, which is truth-conditional. […] the content triggered by the quotes is truth-functionally relevant, in the sense that, at least with respect to some contexts, sentences such as (1)-(8) display a truth-functionally different profile than their counterparts without quotation marks. (Predelli 2003: 18) Instead of treating scare quotes as conventional implicatures, Klockow (1980) and Gutzmann and Stei (2011) argue for analysing this type of quotation as conversational implicatures. Levinson (2000) proposed that quotation marks 40 2 Approaches to defining quotation <?page no="41"?> can be addressed by three different hypotheses. These are replicated below as presented in Meibauer (2015: 191). 1. Postsemantic pragmatics. Quotation marks in scare quotes are pragmatic indicators that trigger pragmatic inferences. They have no influence on the semantics of the construction that contains the quotation marks. 2. Presemantic pragmatics. Quotation marks in scare quotes have an influ‐ ence on the semantics of the construction that contains the quotation marks. This influence may come about indirectly, by shaping the truth conditions through pragmatic inferences. 3. Semantics. Quotation marks in scare quotes have an influence on the semantics of the construction that contains the quotation marks. This influence may come about directly, e.g., by way of a semantic operator. Meibauer (2015) acknowledges that these hypotheses are sketched too simply to adequately take into consideration the semantics-pragmatics interface. A corpus investigation by Meibauer (2015) focussed on newspaper headlines in which the focus particle was surrounded by quotation marks, e.g. Spears zahlt Mann „nur“ 750.000 Euro (‘Spears pays husband “only” 750,000 euro.’). Scare quotes were analysed as pragmatic indicators and the investigation revealed that ironic quotations frequently appear in combination with the German focus particles such as nur (‘only’): the scalar meaning of the particle was argued to trigger an ironic interpretation. Having considered the different types of quotation, paralinguistic properties used to indicate quotation will be presented in the next chapter. 2.2 Paralinguistic markers of quotation Paralinguistic markers of quotation refer to non-verbal cues in speech that signal quotation. These markers can include various vocal characteristics, intonation patterns, and non-verbal elements that help convey that the speaker is directly quoting someone else. In Ancient Greek and Latin, quoted speech was often indicated by changing verb forms or using introductory words without specific punctuation marks. These languages and many others may use a variety of methods to indicate quoted speech or text, relying on context, changes in verb forms, or other grammatical structures instead of quotation marks (Mikulová 2016). The absence of quotation marks in certain languages does not signify a lack of the concept of quoting or attributing speech. Instead, these languages 2.2 Paralinguistic markers of quotation 41 <?page no="42"?> employ distinct linguistic mechanisms within their systems to achieve the same communicative purpose within their respective linguistic systems. One exceptional phenomenon, which is a reliable indicator of quoting, is the use of air quotes, also known as “finger-dance quotes” (cf. Washington 1992). Air quotes are the non-verbal equivalent of written quotes by which speakers gesture quotation marks with their fingers in the air. This non-spoken tool of quotation can be interpreted as the counterpart of the graphemic realization of quotes (Brendel et al. 2011). In a study on the use of air quotes by Cirillo (2019) air quotes most frequently accompanied adjectives, nouns or noun phrases and were either used on top and before their referents but rarely as reparatory cues after their referent. Oral promptings of quotation include quote-unquote, and I quote, and the German construction und ich so; und er/ sie so (‘and I’m like’; ‘and he’s/ she’s like’) (see Golato 2000 for an analysis of the latter). Another possibility to highlight quoted material in German oral speech is to add the prepositional phrase in Anführungszeichen (‘in quotation marks’) as a lexical marker, which specifies the intended meaning (cf. Meibauer 2015). In addition, verbs of speaking can function as quotative markers in direct and indirect speech, most frequently say, answer, reply, speak, and ask (Aijmer 2015). Deictic pointers include “quote pro-forms like speech nouns, pronouns, deictics, and adverbs” (Güldemann 2008: 187). Ajimer (2015) further discusses pragmatic markers co-occurring with quotative verbs that serve a quotative function. Prosody is a paralinguistic marker that draws the addressee’s attention to quotation. These prosodic markers can be a pause, a falling or rising intonation (Brendel et al. 2011), and the accentuation of the quoted content. Further prosodic markers to signal quotation include a breathy voice, high pitch, slower delivery, faster delivery, a combination of a breathy voice and low pitch, combination of low pitch and slower delivery, combination of breathy voice and slower delivery, combination of higher pitch and louder volume, combination of creaky voice and lower pitch. (De Brabanter 2023: 10) Attardo et al. (2003) further discuss phonological markers of sarcasm. Given the variety of possibilities to highlight quotation in spoken language, Klewitz and Couper-Kuhlen (1999) conclude that there is no systematic marking of quotation. In sign language, role shift is used as a mechanism by sign language speakers to report utterances, i.e., for direct and indirect quotation. Quotes are an essential component in communication to convey the nuances of reported speech in sign language. Role shift is a common strategy among sign languages to report speech or thought of another person (Aristodemo et al. 2022). Role shift is characterized 42 2 Approaches to defining quotation <?page no="43"?> by the signer moving the body, head and eyes to the left or right to take over the role of another person. Quer (2011) reports that we can distinguish role shift for direct from indirect discourse: The two types of quoted discourse are identified by the lexical markers that introduce it, the syntactic location of the reported clause, and - most intriguingly - the potential for indexicals to shift between direct and indirect discourse (277). Non-manual markers of role shift may vary across sign languages but typically include body shift towards the position associated with the individual whose perspective is adopted, eye gaze disconnection from the addressee, facial expressions altered in order to mimic those of the reported signer, and head turn. (Aristodemo et al. 2022: 461) In attitude role shift, the verb say can introduce the shift to another perspective. Indexicals in the quoted sentence are also shifted to the perspective of the quoted person, as in direct quotes. Role shift is considered to be a grammatical phenomenon in the visual-gestural mode. Based on evidence of two studies, Attardo et al. (2003) list multimodal markers of irony and sarcasm including facial expressions like raised and lowered eyebrows, smiling, and blank face. The authors attribute a metacommunicative function to the above-mentioned markers of irony. Given the vide variety of facial markers, it is problematic to use these markers as a criterion for identifying irony or sarcasm. In addition, gestures and facial expressions are highly individual and should therefore not be considered a reliable criterion. Although information on gestures and facial expressions is constantly integrated into the processing of linguistic information, research has shown that gestures are processed in a distinct network from the language network, meaning that the brain regions specialized in processing linguistic signals rely on different cognitive resources than gesture processing ( Jouravlev et al. 2019). The ability to attribute and reproduce the words or thoughts of others within a conversation or text, is a prevalent and essential aspect of most languages. While the concept of quotation might be present in most languages, the typography (see Chapter 2) and (para-)linguistic mechanisms (see Chapter 3.2) can differ cross-linguistically. 2.3 Quotation and the semantics-pragmatics debate Quotation is a phenomenon that has been investigated using semantic and pragmatic approaches. To grasp the semantics-pragmatics boundary dispute, the location of quotation in this debate, and current positions on this issue, it 2.3 Quotation and the semantics-pragmatics debate 43 <?page no="44"?> is first crucial to outline the understanding and defining characteristics behind the terminology semantics and pragmatics. The relationship between semantics and pragmatics is a fundamental topic in the study of language and communication. In traditional classifications of linguistics, semantics and pragmatics are often compared against each other. Semantics is, broadly speaking, the theory of the study of meaning in (human) language (Cann et al. 2009). In contrast, pragmatics deals with the study of meaning in context. As both semantics and pragmatics are disciplines dealing with meaning, the distinction and defining characteristics have been under discussion (cf. Allan & Jaszczolt 2012; Gutzmann 2021). The discussion of the relation and differences between semantics and pragmatics has a long history. Current approaches are strongly influenced by the classical distinction initially introduced by Paul Grice (1989), who proposed the terminological difference of “what is said” and “what is implicated”. The former is concerned with how sentences compositionally determine truth conditions. Pragmatics, on the other hand, deals with the truth-conditional meaning interpreted by an addressee. Where exactly to draw the line between semantics and pragmatics is a matter of definition. For Gutzmann (2021), the defining features are summarized in the CCT view consisting of conventionality, constancy, and truth-conditionality: while semantics can be described as conventional, constant, and truth-condi‐ tional, pragmatics is characterized as conversational, context dependent, and non-truth-conditional. In classical studies, quotation is typically analysed from an interpretational semantic perspective. Conversational aspects are considered to play only a mi‐ nor role. Recanati (2001) has put forward a pragmatic, speaker-centred approach for open quotation. He argues for a distinction between c-content, which is the compositionally articulated content of an utterance, and the i-content, the intuitive truth-conditional content (Gutzmann 2007). The i-content is enriched through pragmatic processes. According to Recanati (2001), quotation and quotation marks indicate that a speaker is demonstrating linguistic content, and the quotation marks would give rise to a conventional implicature: “By using quotes, the speaker implies that the condition governing their use obtains” (663). Understanding quotation as a form of demonstration makes it a “paralinguistic phenomenon” (Recanati 2001: 680) that triggers implicatures. Gutzmann (2007) summarizes the proposal of the quotation types at the semantic-pragmatic interface, introducing a model that represents how the category of open quotes needs to be divided into several sub-categories: non-cumulative, cumulative, and autonomic quotation, in addition to closed quotation. The model highlights how these quotation types interact with both the compositional content (c-content) 44 2 Approaches to defining quotation <?page no="45"?> and the intuitive truth-conditional content (i-content), emphasizing the role of pragmatics in enriching the meaning conveyed through quotations. Overall, the model emphasizes the complexity of quotation in language and its implications at the interface of semantics and pragmatics, although other types of quotation are not elaborated upon. Within the philosophy of language, theories of quotation have been discussed as to where to draw the distinction between semantics and pragmatics (Cappelen et al. 2023). This, of course, alludes to the larger question how semantics and pragmatics are distinct from or complementary to each other. Philosophical debates have presented arguments in favour of either a semantic or a pragmatic implementation of quotation (see Chapter 3.2), although quotation was tradi‐ tionally regarded as a purely semantic phenomenon. Early semantic approaches can be traced back to Tarski and Quine who proposed the Proper Name Theory based on a truth-conditional argumentation. Generally speaking, theories of quotation share the common interest of exploring the theory of meaning (Wang 2018). For this reason, Brendel et al. (2011) argue that the study of quotation is closely linked to other linguistic phe‐ nomena located at the semantics-pragmatics boundary. Linguistic phenomena at the semantics-pragmatics interface include speech act adverbials, indexicals, demonstratives, particles, adjectives, and modality (cf. Wang). According to Cappelen & Lepore (2007), quotation can be seen as one of the linguistic phenomena that serve as an “excellent testing ground for a general theory of the interface between semantic and non-semantic content” (8). Within the debate on semantics and pragmatics, the question arises whether quotation should be regarded as a boundary dispute or an interface phenom‐ enon. This question is related to the larger question of whether it is possible to distinguish between semantics and pragmatics. In the first view, defining a clear and precise boundary between semantics and pragmatics is possible. This approach can be found in the distinction view by Bach (2007) who argues for a clear distinction between semantics and pragmatics. In contrast to the distinction view, the interface view presumes the relation‐ ship between semantics and pragmatics to be complementary in nature (Wang 2018). This means that there is no strict boundary between semantics and pragmatics, but an interaction and overlap between the two subsystems of language. The interface view is grounded in contextualism, which assumes that propositions are context-sensitive (Wang 2018). An interface dispute involves discussions about how semantics and pragmatics interact, acknowledging that there is some degree of overlap and interaction between the two domains. In other words, it claims that semantics and pragmatics cannot clearly be 2.3 Quotation and the semantics-pragmatics debate 45 <?page no="46"?> distinguished. It focuses on how content is influenced by context without explicitly stating which content is semantic or pragmatic. Although both debates deal with the relationship between content and context, they use different theoretical approaches. Wang (2018: 45) states that the only agreements can be stated out as follows: (22) a. Context influences the linguistic content to generate intuitive truth-conditional content; - b. Semantics and pragmatics cannot be clearly distinguished but their distinction is fundamental to the theorizing of meaning upon which human cognition and communication can be understood. Within pragmatic accounts, quotation marks are regarded as not being a nec‐ essary or sufficient grammatical device, instead they would serve as “practical pragmatic indicators of disambiguation” (Brendel et al. 2011: 14). Pragmaticists acknowledge that while some quotations affect can affect truth conditions, they argue that this is not universally the case, and when it occurs, it typically involves pragmatic mechanisms. In contrast, semanticists fully support the assertion that quotes affect truth conditions. (23) a. London is large. - b. “London” is large. In the pure quotation presented above, the truth-conditions differ due to the presence of quotation marks. Sentence (23a) can be considered true if London is large in terms of size or population. Regarding the second sample, the sentence is true if the word London itself, as a linguistic entity or symbol, is considered large. This might be the case with regard to the font size or visual representation. However, in a more common interpretation, this statement is likely to be consid‐ ered false because the word London does not have a physical size like an object; it is simply a linguistic representation. Consequently, pragmatic approaches use truth-conditions to produce truth-conditionally relevant content (Predelli 2003; Simchen 1999). Meibauer (2015) argues that the primary benefit of adopting a pragmatic ap‐ proach to quotation marks is that it does not require one to accept the polysemy of the punctuation or to completely disregard any of its applications. According to pragmatic approaches, quotational constructions lacking quotation marks in the written format are still considered a grammatical and meaningful sentence 46 2 Approaches to defining quotation <?page no="47"?> 9 Please note that this chapter is only expected to give an overview of the most important theories of quotation and that their discussion is beyond complete. due to the contribution of the pragmatic context. Gómez-Torrente (2003), for example, argues that omitting quotes is appropriate but less informative and might lead to unintended interpretations of the utterance. In contrast, semantic accounts claim that quotation without quotes can be materialized differently, e.g., by using acoustic markers to indicate the quote. Another option is that the contextual embedding generates a mentioning use (see Cappelen & Lepore 1999). In the following, a brief outline of the most prominent theories of quotation will be presented. 2.4 Theories of quotation In the literature on quotation, the attention has been centred around five different theories of quotation. 9 These theories vary in their approach on the status of quotation marks, questions regarding the context-dependency, as well as their role in semantic and pragmatic theories (Brendel et al. 2011). One way of classifying quotational theories has been put forward by De Brabanter (2023), who has argued that these theories can be classified in terms of the role they ascribe to marks of quotation. Type-1 theories are characterized regarding the occurrence of quotation marks as necessary while type-2 theories do not. When dividing theories of quotation into theories with and without quotation marks, the latter is supported by Recanati (2001), Washington (1992), and Saka (1998). Theories with explicit marking of quotation with quotation marks are argued for by Cappelen & Lepore (2007) and Gómez-Torrente (2003), who proposed the Use Theory claiming that Use Theories in fact postulate that the reference is created by the cited expression within a quotation expression (Bazzoni 2016). Another approach is to cluster the different theories into either semantic or pragmatic quotational theories, although most approaches to quotation are fundamentally semantic (see De Brabanter 2017). Semantic theories of quotation claim that “quotation marks have their effect in the semantics - that they alter the literal meaning of sentences without such punctuation” ( Johnson 2011: 210). Semantic theories are defended by Tarski (1956), Washington (1992), and Cappelen and Lepore (2007). Purely semantic theories of quotation include: The Name Theory, quotation as definite descriptions, and the Demonstrative Theory. First, the Proper Name Theory of Quotation has been proposed by Quine and Tarski. As the name indicates the theory claims quotations should be treated as 2.4 Theories of quotation 47 <?page no="48"?> 10 This theory has, among others, been challenged by Johnson & Lepore (2011). single words, more specifically as proper names. Tarski (1933) argues that quotes and the expressions that come before and after them serve the same purpose as the individual letters and letter complexes in single words. Then, each name wrapped in quotation marks is a constant, unique name of a definite expression, which shares the same properties with a proper name. This view is supported by Quine (1940), explaining that quotations are unstructured signs. He claims that each quotation must be viewed as a single word or sign from the perspective of logical analysis, with its individual pieces amounting to nothing more than serifs or syllables. In essence, the Proper Name Theory treats quotations similar to names in languages, that is as singular terms that are unstructured (Cappelen & Lepore 2007). As a consequence, the quoted items are not regarded as a component of the quotation. The theory can account for why co-referential terms and synonymous expressions cannot be substituted for. Second, the Description Theory was introduced to ensure that a quoted series of expressions is seen as such, posits basic units in language such as words or letters. According to this theory, basic quotations follow the Proper Name Theory, where each word has a quotation name (Geach 1957), while complex quotations describe concatenations of basic units. The Description Theory of quotation assumes that each language has a fixed set of basic units: “words, according to Geach (ibid., Ch. 18 and 1970); letters, according to Tarski (1956, p. 160) and Quine (1960, p. 143, p. 212)” (Cappelen et al. 2023). Quine and Tarski analyse statements as illustrated in (24a) as concatenations of letters, contrasting Geach’s more straightforward representation. (24) a. ‘Bachelor’ has eight letters. - b. ‘B’-‘a’-‘c’-‘h’-‘e’-‘l’-‘o’-‘r’ has eight letters. Third, the Demonstrative Theory has been put forward by Donald Davidson (1979). This theory argues that quotations are forms of demonstrations. 10 The Demonstrative Theory treats quotation as a semantic category. The quotation marks can, in this view, be replaced by the formulation “the expression of which this is a token” (Davidson 1979: 90) has eight letters. This argumentation further holds that the quoted item is not part of the syntax of the sentence and that the quoted material refers to the expression instantiated by the demonstrated token (Cappelen et al. 2023). 48 2 Approaches to defining quotation <?page no="49"?> All of the theories outlined above have been debated with respect to their application to different types of quotation, see Cappelen & Lepore (2007) and Cappelen et al. (2023) for an overview of the strengths and weaknesses. We will now move on to a characterization of pragmatic theories. Pragmatic theories have been favoured by Gutzmann & Stei (2011), claiming that quotation marks are best analysed as a device for generating conversational implicatures. They argue that all general features typically ascribed to conversational implicatures, such as context dependency, calculability, and cancelability, apply to quotational sentences. Johnson (2011), who considers exclusively direct and mixed quota‐ tion, argues that these types of quotation are best analysed pragmatically. As an opponent of the demonstrative theory of quotation, the Identity theory, also known as the Use-Theory of Quotation, attributed to Frege (1892), Washington (1992), and Searle (1969), has been proposed. It posits that the primary function of quotation marks is to indicate quotational use, where ex‐ pressions, even those not typically mentioning, become mentioning expressions in quotation (Cappelen et al. 2023). Quotation marks, according to this theory, do not refer; instead, the expression itself does the referring. Consider, for example, the quoted proper name presented below. (25) “Aristotle” Here, it is not the quotation marks that do the referring but the expression refers to itself. Washington (1992) summarized the identity Theory of Quotation as follows. The quotation as a whole is analyzed into the marks that signify quotational use of the quoted expression and the quoted expression itself used to mention an object. All expressions, even those whose standard uses are not as mentioning expressions, become mentioning expressions in quotation . . .a quoted expression is related to its value by identity: a quoted expression mentions itself. (557) The Identity Theory also suggests that a quoted expression is related to its value by identity, as it mentions itself. Wiślicki (2014) summarizes the main idea of the Identity Theory as “a name and its name are denoted by the same word” (242). Quotation marks have the function of indicating that the quoted material does not have its normal extension (Brendel et al. 2011). Both the demonstrative account and identity theory claim that quotation marks have an independent status and are not considered part of a semantic category (cf. Wang 2018). Finally, the Proper Name Theory treats quotations as single words (Tarski 1956) or “unstructured singular terms” (Cappelen & Lepore 2007). According 2.4 Theories of quotation 49 <?page no="50"?> to this view, the quoted expression is not a component of the quotation itself. Instead, quotations are considered to be referentially opaque (Brendel et al. 2011). Tarski formulates the theory as follows. Quotation-mark names may be treated like single words of a language, and thus like syntactically simple expressions. The single constituents of these names - the quotation marks and the expressions standing between them - fulfill the same function as the letters and complexes of successive letters in single words. Hence they can possess no independent meaning. Every quotation-mark name is then a constant individual name of a definite expression (the expression enclosed by the quotation marks) and is in fact a name of the same nature as the proper name of a man. New theories of quotation are constantly developed, see for example the proposal by Jan Wiślicki (2021) on quotes as modal operators and Matthew McKeever’s (2023) empirically-based theory on mixed quotation. Cappelen & Lepore (2007) reject the idea of a unified theory of quotation under the premise that quotation is defined as metalinguistic talk. For this reason, some theories have been developed exclusively for a specific kind of quotation, see e.g., the theory proposed by Bazzoni (2016) for pure quotation, and Maier (2014) on mixed quotation. An open theoretical question is whether (non-ironic) NMCs should be treated as instances of pure quotation or mixed quotation, with the latter having been addressed from a use-only perspective, mention-only theories as well as dual theories (García-Carpintero 2012). In this chapter, different types of quotation as well as a selection of quotational theories have been discussed to give a general overview of the debate. The next chapter is devoted to examining how quotation can create reference to names and the linguistic properties involved in NMCs. 50 2 Approaches to defining quotation <?page no="51"?> 3 Theoretical background Quotational constructions, as shown in (26) below, point to linguistic shapes and inform the addressee about the name of a lexical concept (cf. Härtl 2020), i.e. (26a), the compound coral reef. (26) a. This formation is called a “coral reef ”. - b. This natural phenomenon is referred to as a “polar wind”. - c. This cloud formation is known as a “lenticular cloud”. - d. Experts name this phenomenon a “sandstorm”. These types of quotational constructions have been introduced as name-inform‐ ing quotations (NIQ) in the literature (see Härtl 2018). A key characteristic of NMCs is that they provide the name of a lexical concept. The terminology is here understood as referring to the whole VP (e.g., is called a “coral reef ”) or the whole sentential constructions. The sentences in (26a-d) can be used by speakers to establish new knowledge or to make common and pre-existing information clear. The NMCs in which the theme, i.e., the expression coral reef in (26a) occurs in quotes, will be referred to as “quoted item” when it the expression is mentioned and graphemically surrounded by quotes. Expressions that appear in quotational constructions, be it quoted or unquoted, will either be called “mentioned nominal”, “names”, or “mentioned noun”. The focus of this paper is on the naming predicate taking an [NP]-complement. In this context, it is important to define and distinguish between the two terms that are central to this research project. On the one hand, the term name mentioning is a hypernym for the concept of name informing. On the other hand, name mentioning comprises a distancing interpretation, which is brought about by a modalizing use with which a speaker refers attitudinally to the shape of a certain denotatum’s name (Härtl 2018). This modalizing use is represented in a construction like And someone like that calls himself “expert”. The quoted noun expert adopts an ironic interpretation instead of a name-informing one, see Chapter 4.3 for an in-depth discussion. Before moving to the concept of name-mentioning quotation, the terminol‐ ogy including the words reference and referring is going to be classified to define what constitutes reference. The term referentiality has been used for a long time in the linguistic literature with a variety of definitions in particular <?page no="52"?> differentiating between semantic and pragmatic reference: a semantically ref‐ erential expression points inherently to an entity in the world (Abbott 2010). In contrast, pragmatic referentiality relates to how a speaker uses a linguistic expression in context (Abbott 2010). Another key term for this work is the notion of reference. Philosophers of language have traditionally been describing reference as a “semantic relation between linguistic expressions and things” (Abbott 2010: 3). It can be defined as the relation between a specific constituent of an utterance and the set of individuals that is identified by the constituent (Matthews 2014). While the terminology referent refers to a mental construction of an entity, i.e. a purely mental concept of a speaker instead of a real object in the world, referring is characterized as the “process created by a speaker who uses a linguistic expression to indicate, introduce, or identify a referent for the addressee” (Fontaine et al. 2023: 32). As a consequence, a referring expression is a linguistic representation that a speaker uses to refer to a referent. Fontaine et al. (2023) sum it up nicely: “Reference is the relationship, referring is the act, and referring expression is the linguistic expression used by a speaker to refer” (32). In addition, we need to consider the terminology metalinguistic reference (cf. Maier 2020), which is a key term for this investigation. Metalinguistic describes the “function of language in referring to itself ” (Matthews 2014) which has been coined metalingual function by the structural linguist Roman Jakobson. Metalanguage is the function of language to refer to itself. Languages are used metalinguistically e.g. when an English grammar of German uses English to describe linguistic phenomena. Here, English functions as metalanguage and German as the object language, the language that is talked about (Matthews 2014). Metalinguistic reference is linked to metacognition or metacognitive processes, i.e. the understanding as well as control of one’s own thinking processes (Sternberg & Ben-Zeev 2001: 243). People need to be aware that they can talk about language and use metalinguistic constructions to discuss terminology concerning our language use. The nature of quotation can be described as metarepresentational. When language is used to attribute properties to language or otherwise theorize about it, a device is needed that - to borrow Donald Davidson’s apt phrase - “turns language on itself ”. Quotation, by virtue of being the primary metalinguistic tool, is one such central device. (Cappelen & Lepore 2007: 1) In this famous quote, Cappelen and Lepore address the different dimensions of metarepresentation, more specifically the function of mentioning. Waugh (1980) 52 3 Theoretical background <?page no="53"?> further states that an equation is used in metalinguistic uses of language (see also Chapter 3.2.2) Various researchers have emphasized the importance of metalanguage in human communication. Anderson et al. (2006) claim that a person’s ability for meta-language would be necessary for “conversational adequacy” (973), being often used for clarifications in dialogues and also for repairs. Therefore, metalanguage has been argued to be essential for human communication. In the following section, the connection between metalanguage and the nature of name-mentioning quotation is subject to discussion. 3.1 The concept of name-mentioning quotation Name-mentioning constructions are understood to involve a naming verb like call and a mentioned nominal, which might be accompanied by quotes, see the following sample sentences in (27a-b). (27) a. This phenomenon is called “bioluminescence”. - b. One refers to this piece of furniture as a “kitchen table”. It has been argued that naming verbs like call, refer to as, stand for or speak of function as name-selecting predicates and introduce a naming convention (cf. Härtl 2020). In the above-mentioned sample (27b), the speaker introduces the conventionalized term kitchen table to refer to the piece of furniture itself, making it clear that this specific object is commonly referred to by this specific name. But what function do metalinguistic constructions like NMCs that give rise to a literal interpretation serve? According to Wilson (2012), metalanguage is used for the “introduction of new vocabulary, attribution of statement, explan‐ ations of meaning, and assignment of names” (638). In addition, metalinguistic constructions are used to explicitly mark or highlight a specific term. They can be used for clarification of a specific term, in political speeches to persuade the addressees, and to use and/ or demonstrate the knowledge of specialized terminology (Müller 2022). Müller (2022) therefore states that metalinguistic constructions have a communicative dimension. Depending on different disciplines and theoretical frameworks, NMCs can be located within different debates. From a practical standpoint, NMCs play a crucial role in language learning and acquisition, particularly in teaching the addressee the name and/ or meaning of a specific word. Definitory-like 3.1 The concept of name-mentioning quotation 53 <?page no="54"?> sentences are employed to create reference and introduce unknown words. Likewise, definitions can be subsumed as a form of meta-language with the function of facilitating the comprehension of a word’s meaning (Seres 2020). Essentially, NMCs also constitute a form of metalanguage, employing language to communicate about language. Consider, for example, a language learning situation in which a speaker introduces a new word. The use of NMCs is neither restricted to only written language nor to specific registers like news reports or academic articles. These words are often lexicalized and might be very specific terminology related to a certain topic, see (28a-b). (28) a. This chemical reaction is called “decomposition”. - b. This grammatical structure is known as a “relative clause”. Note that the sentence in (28b) has a bi-accusative structure: the name is assigned accusative case in German constructions when the name is accompanied by a determiner (cf. Härtl 2020, Cortés Rodríguez et al. 2022). In (28a-b) NMCs are, on the one hand, used in a teaching situation at school but can also be found in earlier stages of language learning. More specifically, language acquisition contexts frequently make use of NMCs. Either parents are introducing new concepts by using an NMC construction, or children themselves use them. (29) a. Child: This animal is called an elephant. - b. Parents: This is called the “greenhouse effect”. This is of no surprise as one of the first language skills of children when learning to speak a new language is learning to refer to objects: referring is a core element of human communication (Fontaine et al. 2023). In addition, these definitory-like sentences typically require speakers to possess metalinguistic abilities. This is reflected in Benelli et al. (2006), who argue that uttering definitions would require speakers to have metalinguistic abilities, in that they are based on the conception that a word (the definiendum) is a phonological unit transmitting specific semantic content and that a correct linguistic structure (the definiens) is necessary to make the content explicit, by means of an extended linguistic form linking the two terms of the definition. (75) When considering language developmental aspects, NMCs play a key role in language acquisition of young children: definitional constructions are one 54 3 Theoretical background <?page no="55"?> 11 Retrieved December 5, 2023 from https: / / bitly.ws/ 34ze8 12 Retrieved December 5, 2023 from https: / / bitly.ws/ 34zeG 13 Retrieved December 5, 2023 from https: / / bitly.ws/ 34zf8 14 Retrieved December 5, 2023 from https: / / bitly.ws/ 34zfQ option for children to acquire the lexicon (cf. Krifka 2012; Seres 2020). Perlis et al. (1998) documented metalinguistic uses of call and name in speech of a 3-year-old. With increasing age, children adapt their structure to the Aristote‐ lian format, as confirmed in a study by Benelli et al. (2006), for which lexical, phonological-semantic, and syntactic awareness is necessary. Petitta et al. (2018) differentiate between two different types if metalinguistic reference: “autonomy (i.e., the quotation of a word standing by itself) and discourse-related references (i.e., remarks on something previously said by one participant)” (52). The authors further comment on a special instance on metalinguistic reference which is used to explain a word in another language. (30) In Italian, the word for “dog” is cane. Meaning in a sentence like (8) can only be created when switching to the target language. Metalinguistic reference is characterized as a request for “information about lexical items” (Pettita et al. 2018: 50). Second, NMCs are used to draw the hearer’s attention to the quoted word. To give an example, news broadcasts frequently make use of NMCs in everyday communication to explain certain words to the audience. This becomes obvious when reading the news: (31) a. They enable AIs to learn from experience, as a person would. This is called deep learning. (BBC News 11 ) - b. This is called the “switch-cost effect”. It means that if you check your texts while trying to work, you aren’t only losing the little bursts of time you spend looking at the texts themselves - you are also losing the time it takes to refocus afterwards, which turns out to be a huge amount. (The Guardian 12 ) - c. Das nennt man Mediensucht - oder auch Computer-, Handy- oder Internetsucht. (ZDF 13 ) ‘This is called media addiction - or computer, mobile or Internet addiction.’ - d. Die Welt ist voller winzig kleiner Lebewesen, den „Mikroorganismen“. Dazu gehören etwa Bakterien, Viren, Parasiten und Pilze. Sie alle zusammen werden umgangssprachlich auch „Keime“ genannt. (Tages‐ schau 14 ) 3.1 The concept of name-mentioning quotation 55 <?page no="56"?> ‘The world is full of tiny little living creatures, the “microorganisms”. These include bacteria, viruses, parasites and fungi. All of them to‐ gether are commonly referred to as “bugs”.’ Although it is likely that NMCs are preferably used in more formal registers that are inherently suitable for explaining concepts or new pieces of information to an audience, NMCs are not restricted to a specific register. In political debates, NMCs can be used to define specialized terminology to make sure that the listeners are on the same page as the speaker. Optionally, NMCs can serve a persuasive function in public debates (see Müller 2022). Which of these functions or roles of NMCs apply to a concrete NMC construction differs with regards to the context in which the sentence is uttered, and the situation, i.e. the pragmatic information, which determines the intended meaning of the utterance. From a philosophical perspective, NMCs can be characterized as a quotational construction and located within the debate of quotational theory. This philo‐ sophical discourse, which delves into the nature of language and meaning, specifically intersects with truth-conditional semantics and pragmatics in the context of NMCs. These constructions introduce a layer of complexity to the understanding of truth conditions and the pragmatic nuances embedded in linguistic expressions. The nominal expressions in the quotational constructions in (10), can be exchanged without affecting the truth conditions of the sentence. The nominal expressions are truth-conditionally equivalent (Gutzmann 2021). (32) a. A couch is also referred to as a “sofa”. (Härtl 2020) - b. An automobile is commonly known as a car. Notably, the quoted nominal can - but does not necessarily have to - be accompanied by a determiner (see section 2.5 for a discussion of this issue). Shifting to a psycho-linguistic viewpoint, NMCs emerge as objects of partic‐ ular interest due to their involvement in metalinguistic operations, which can be encompassed within the broader concept of metacognition. In psychology, metacognition refers to the cognitive processes that involve thinking about one’s own thinking. Applied to linguistics, this implies that speakers engaging in NMCs are not merely conveying information but are cognitively aware of the act of talking explicitly about language. Metacognitive processes in language use involve considerations of linguistic form, meaning, and communicative goals. 56 3 Theoretical background <?page no="57"?> Another perspective is to locate NMCs within the taxonomy of definitory utterances, more specifically as meta-terminological definitions: Grundsätzlich bestehen Definitionen aus einem zu definierenden Sachverhalt (Defi‐ niendum) und einem definierenden Konzept (Definiens), die durch eine Verbalkon‐ struktion (Definitor) zueinander ins gesetzt werden […]. Roelcke (2005: 54-61) gibt eine Synthese der Literatur zu Definitionen und stellt Typen von Definitionen zusammen. Im Hinblick auf den semiotischen Status des Definiendum unterscheidet er Nominaldefinitionen, welche sich auf die sprachliche Repräsentation eines Sach‐ verhalts, also die Bedeutung des Terminus im Sinne eines sprachkonstruktivistischen Modells, beziehen, von Realdefinitionen, welche sich auf den Begriff des Sachverhalt im Sinne eines erkenntnisrealistischen Modells beziehen. (Müller 2022: 100) ‘In principle, definitions consist of a something that needs to be defined (definiendum) and a defining concept (definiens), which are placed together by means of a verbal construction (definitor). Roelcke (2005: 54-61) gives a synthesis of literature to definitions and compiles types of definitions. With regard to the semiotic status of the definition, he distinguishes nominal definitions, which relate to the linguistic representation of a fact, i.e. the meaning of the term in the sense of a linguistic constructivist model, from real definitions that refer to the concept of the fact in the senses of a cognitive realistic model.’ As indicated in this quote, many definitions in the logical and philosophical literature can be analysed as consisting of three different parts: the definiendum, the term that is defined, the definiens, the expression which provides the meaning of the term, and an additional expression that is equated (Gupta & Mackereth 2023). Definitions create semantic equivalence between two terms (Benelli et al. 2006). They are thus characterized as containing the most relevant information about a specific term. Stereotypical ways of requesting a definition can be communicationally be done by posing the question “What ix X? ” or “What does X mean/ refer to? ” (Burton-Roberts 1986). As one would expect when classifying NMCs as a definitional sentence, they can be found in encyclopaedias, dictionaries, and educational books. An Aristotelian format of definitions has been described as ‘An X is a Y that Z’, where ‘X’ is a given object or concept, ‘Y’ represents the genus proximum (the superordinate category), and Z is the differentia specifica, i.e. information that allows the specific object or concept to be identified. (Benelli et al. 2006: 73) While the traditional definition does not contain a pure quotation and also no metalinguistic predicate, which sets it apart from NMCs, the words Y and Z 3.1 The concept of name-mentioning quotation 57 <?page no="58"?> share characteristics with the name(s) in NMCs, such that a superordinate term establishes a copular-like relation with a specific concept. In mathematics and logic, the distinction between real and nominal definitions has been introduced by John Locke (1689). A real definition, according to Locke, is a definition that provides the essence or nature of a thing. It goes beyond mere linguistic conventions and aims to capture the underlying reality or nature of the object or concept being defined ( Jones 2023). Nominal definitions, on the other hand, are definitions that assign names to things based on their observable properties or characteristics. These definitions are more concerned with the use of language and how terms are conventionally employed rather than with the essential nature of the object or concept. Nominal definitions are essentially linguistic conventions that help communication but may not necessarily capture the true nature of the thing being defined. (33) A triangle is a three-sided polygon with internal angles that always add up to 180 degrees, and its sides may have different lengths or be of equal length. [real definition] - In geometry, we call any three-sided figure a “triangle”. [nominal definition] In the first example, the essential characteristics and properties that define the nature of a triangle, such as the number of sides, internal angles, and the relationship between these elements is emphasized. In the nominal definition, the term triangle functions as a linguistic label or name assigned to any shape with three sides. The definition is more concerned with the conventional use of the term rather than the underlying nature of the geometric figure. Crucially, definitions can establish a copula relation: Semantic equivalence takes the form of a COPULA (verbs, such as ‘is’, ‘means’, ‘refers to’) and essentially consists in the categorical identity of a stimulus-item (e.g. ‘a dog is a domestic animal’; ‘mean is an adjective meaning nasty’, ‘running refers to the action of moving forward very quickly’, etc.). (Benelli et al. 2006: 72). This observation will be further addressed in Chapter 3.2. Definitions have been argued to be metalinguistic in nature (Watson 1985; Snow 1990; McGhee-Bidlack 1991; Seres & Espinal 2019; Benelli et al. 2006). Metalinguistic skills that involve reflection on language, more specifically knowledge about the lexical inventory of a language, are required to create a definition: speakers need to know about appropriate linguistic structures to convey ideas and get to the semantic core of language objects. 58 3 Theoretical background <?page no="59"?> An open research question that still needs to be addressed concerns the location of NMCs within the taxonomy of types of quotation. While name-in‐ forming sentences have been argued to be a type of pure quotation (Härtl 2020), they might also qualify for an analysis as a type of mixed quotation (see also Section 2.1.3). The debate over the classification of NMCs within this taxonomy affects how empirical data on NMCs is collected, analyzed and interpreted and could consequently influence the conclusions drawn from experimental studies. If NMCs are treated as pure quotations, the quoted word is treated as a syntactically isolated element that functions outside the typical referential structure. Conversely, if NMCs are treated as mixed quotations, they should be more syntactically embedded in the sentential structure. This distinction directly impacts assumptions on language processing stages, calling for a closer examination in further research. NMCs have only recently sparked interest and are therefore understud‐ ied, especially with respect to the lexical and pragmatic factors involved in name-mentioning quotation. The notion of name-mentioning quotation is a concept that is closely linked to two hyponyms: name informing and distancing (Härtl 2018). In general, name-mentioning quotation appears in a variety of different constructions that differ in interpretation. To illustrate, consider the following sample sentences: (34) a. Der sogenannte „Baum“ ist eine Yoga Übung zur Verbesserung des Gleichgewichts. - - (The so-called “tree” yoga exercise is used to improve balance.) - b. ? ? Der sogenannte „Baum“ ist eine mehrjährige Pflanze mit Blättern und Ästen. - - ( ? ? The so-called “tree” is a perennial plant with leaves and branches.) - c. Der sogenannte „Baum“ ist eigentlich eher ein Busch. - - (The so-called tree is rather a bush.) While (34a) and (34c) are perfectly acceptable with a distancing interpretation of the word, the sentence in (34b) is less acceptable with a conventionalized interpretation of the noun. As illustrated above, NMCs appear in a variety of different constructions that differ in interpretation. When considering (34a) the quoted expression tree adopts a non-conventionalized reading. In this sentence, the quoted nominal refers to the name of a yoga pose rather than a plant. The quotation marks draw attention to the fact that the semantics deviate from the common understanding of a tree being a type of plant. The quotation marks 3.1 The concept of name-mentioning quotation 59 <?page no="60"?> highlight that the semantics depart from the stereotypical interpretation. In contrast, the highly conventionalized noun tree in (34b) fails in this construction. Tree is a conventionalized noun that can occur in a so-called construction, i.e. highly conventionalized nouns are unlikely to appear in NMCs in general (cf. Härtl 2020). The term is, however, acceptable in (12c) in which the name of the object is questioned and an alternative, better-fitting term is given instead. The acceptability of the sentences consequently depends on the conventionality of the term when used in combination with sogenannt (‘so-called’). Generally speaking, compounds differ from simple nouns in their acceptabil‐ ity in NMCs. (35) a. Der sogenannte „Klappstuhl“ ist ein Möbelstück zum Campen. (The so-called “camp-chair” is a piece of furniture for camping.) - b. ? Der sogenannte „Stuhl“ ist ein Möbelstück. NMCs involving the participle form sogenannt (‘so-called’) of the naming verb nennen (‘call’) in combination with the low-frequent noun camp-chair is perfectly acceptable. Words with a low lexical frequency are expected to not be well-established in the addressee’s mental lexicon which makes it compatible with a being-called construction. In contrast, highly familiar words are less acceptable as the addressee is familiar with the term which makes its introduction redundant and would not be an economical use of language. In the paper by Wilson (2012), the phenomenon of “mentioned language”, which can be used as an umbrella term for a variety of sentential constructions involving metalinguistic reference, is described. In order to further specify and narrow down the definition of name-mentioning constructions important for this thesis, the definition of mentioned language will now be reproduced and further developed. The definition on metalanguage provided by Wilson (2012: 639): For T a token or a set of tokens in a sentence, if T is produced to draw attention to a property of the token T or the type of T, then T is an instance of mentioned language. My working definition of name-mentioning constructions holds the following: For T a token or a set of tokens in a sentence, if T is produced to draw attention to a property of the token T or the type of T, then T is an instance of mentioned language. The sentential construction containing T is introduced 60 3 Theoretical background <?page no="61"?> 15 I am thankful for the comment of a member at the GeKKO biannual meeting who noted that NMCs might be genre-specific, especially be found in medical texts in which the name of a disease is introduced to the addressee. The empirical analyses presented in this work did not focus on a specific text genre but I share the assumption that some genre might prefer a use of NMCs. by a metalinguistic naming predicate such as call, refer to as, nennen (‘call’) or bezeichnen als (‘refer to as’). 1. Token: defined as a linguistic entity which is either a word or a phrase that creates reference to a lexicalized concept; T can but does not necessarily have to be accompanied by quotation marks and might also be highlighted with e.g. bold letters, in italics etc. 2. The meaning of the linguistic entity can either be literal or non-literal, i.e., modalizing 3. The predicate functions to establish a link between two concepts, either creating an equation or an identification between the two linguistic units 4. The superordinate term clarifies the meaning of the linguistic entity: the concept is defined or further specified in this sentential structure, typically representing a sub-concept of the other concept. This working definition is applied for the theoretical and empirical exploration of NMCs in this thesis. Another noticeable fact is that NMCs frequently occur in formal registers and should not be expected to occur extensively in colloquial conversations. 15 Yet, they are anticipated to be present in both spoken and written communication. NMC appear in a variety of syntactic constructions. Mentioned expressions are frequently nominals, i.e., taking the role as a noun phrase (NP henceforth). In German and English NMCs, the quoted nominal can function as the subject (36a), predicative complement (36b), or appear in apposition with the name (36c). (36) a. “Copper” is the scientific name for the chemical element with the symbol Cu. - b. Strong winds are named “hurricanes” when wind speeds reach 74 miles per hour. - c. The name “hurricane” is shared by various severe tropical cyclones that form in different parts of the world. In addition, quoted expressions can be embedded in a relative clause (see 37a-b). 3.1 The concept of name-mentioning quotation 61 <?page no="62"?> 16 Bold print as used in the original. (37) a. This phenomenon, known as globalisation, is commonly often used in journalism. - b. This practice, known as mindfulness, has gained popularity in contem‐ porary discussions about mental health. Possibly, NMC sentences can be split into two different sentences. (38) a. In the world of chemistry, “Copper” holds a significant place. It is the scientific name for the chemical element with the symbol Cu. - b. Scientists study the phenomenon in which a region in space where gravity is so intense that nothing, not even light, can escape its gravitational pull. This is commonly known as a “black hole”. This great variety indicates the flexibility of language users to include men‐ tioned expressions in everyday language. A downside of this variation is that NMCs might be difficult to be extracted from corpora, not only because there is no systematic way of indicating mentioned language but also because of the syntactic flexibility. General characteristics of NMC constructions include the feature of clarity. Due to its function of introducing terminology, NMCs are expected to be written in a clear and precise language to offer a straightforward understanding of the quoted material. Just like definitions, NMC sentences are typically concise, and written in a neutral language, meaning without using an emotional or persuasive tone. Moreover, NMCs can be used to comment on (new) words, expressing an evaluation of the term. Svanlund (2018: 1) presents samples from Swedish, in which an unofficial term is metalinguistically commented on. (39) Göran Tägtström värjer sig inför uttrycket “stafettläkare” som används påmånga håll i landet. - Stafettläkare låter lite nedsättande. Jag föredrar att kalla mig konsultläkare. - Göran Tägtström opposes the expression “relay doctor” which is used in many places around the country.---Relay doctor sounds a bit disparaging. I prefer to call myself consulting doctor. 16 62 3 Theoretical background <?page no="63"?> These metalinguistic comments could address a variety of linguistic properties, such as their “pronunciation, spelling, meaning, usage patterns, etymology, history of usage, stylistic value, etc.” (Svanlund 2018: 2). Coming back to the distinction of definiendum and definiens, Seres & Espinal (2019) state that canonical definitions typically follow the structure of copular sentences. In a “definitional mode of speaking” (Seres & Espinal 2019: 4), the sentential patterns follow a logical structure making use of is or mean to create a link between the term that needs to be defined and the defining characteristics of the term. In the next section, NMCs will be analysed using a copula-based approach. 3.2 A copula-based approach to NMCs In this section, a look into the nature of copular verbs and copular constructions is presented, which has been of interest in syntactic and semantic discussions. The taxonomy of copular clauses was originally proposed by Higgins (1979) and includes four different types of copular clauses. Before taking a look at this taxonomy, the general characteristics of copula verbs will be reviewed as an overview to determine the copular sentences underlying NMC construction. The term copula has a Latin origin and refers to “sth. that “connects” or “binds” or “links” things together” (cf. OED 2023), which describes the fact that copulas are elements that create a link between the predicate and subject. Copula words have the syntactic function of expressing a relation between the subject and the predicate. It can therefore be defined as an intransitivity verb that links a subject either to (i) a noun phrase, (ii) an adjective, (iii) or another constituent that expresses the predicate in the sentences (Glossary of Linguistic Terms 2023). (40) a. Dinner-is-in the kitchen. - b. The weather-seems-good. Copula relations are commonly encoded across using languages using verbs that are similar to the English copula verb be (Curnow 1999). In German, the copula verbs sein, werden, and bleiben form the predicate only in combination with a predicative complement (cf. Schneider & Lang 2022). Copula construction refers to the fundamental structures employed by language to convey the concepts of (a) establishing the equivalence of two participants typically expressed by an NP (e.g., ‘that man is my teacher’, ‘that woman is Louisa’) and (b) indicating 3.2 A copula-based approach to NMCs 63 <?page no="64"?> 17 Arche et al. (2019) comment on cross-linguistic difference regarding these prototypical features of copula. group affiliation or classification through NPs (e.g., ‘that woman is a doctor’, ‘that man is a teacher’) (Curnow 1999). (41) a. Mary is nice - b. Mary: λx[nice’(x)] The predicating adjective nice is linked to the referring expression Mary. Arche et al. (2019) note that it is generally challenging to properly define the concept copula. The prototypical behaviour of copulas includes them carrying verbal inflection, and their appearance in “contexts where the predicate is nonverbal” (Arche et al. 2019: 3). Copulas by themselves are unable to define a predicate (42a) in the languages German and English, which is the reason why the sentences are rendered illogical. 17 Similarly, copulas are required for the adjective to be able to define the predicate (42b). (42) a. *Mary is - b. *Mary happy In the linguistic literature, a common distinction is made between predicational and non-predicational types of copula. To differentiate between predication and equative copula, the referential status of the second NP is crucial. (43) a. Charles is a writer. predicational - b. Charles Dodgson is Lewis Carroll. equative - c. That’s Charles. identificational - d. The problem is Charles. specificational - - - (Seres 2020: 87) The subject referent of the first NP is assigned a property or characteristic that is referred to by the postcopular phrase in predicational sentences: Die Erdkugel ist nicht ganz rund (‘The earth is not entirely round’) and Peter ist Lehrer (‘Peter is a teacher’). In contrast to non-predicational sentences (43b-d), the second NP 64 3 Theoretical background <?page no="65"?> in predicational ones (43a) is non-referential (Seres & Espinal 2019). Crucially, unlike auxiliaries, the copula does not contribute to the semantic content of the phrase (Pustet 2003). Further, Geist (2006) notes that copula verbs have an outstanding character‐ istic as they do not allow for modifications with VP-adverbials. She uses the following two sentences to illustrate the difference: (44) a. Heidi schläft (gerade) in der Hängematte. ‘Heidi is (currently) sleeping in the hammock.’ stative verb - b. *Paul ist (gerade) unter der Straßenlaterne betrunken. ‘*Paul is (currently) drunk under the street light.’ copula verb - - - (Geist 2006: 92) The same observation is true for NMCs, which supports the idea of an underspe‐ cified copula relation. Consider the examples in (45) which can also be applied to German NMCs. (45) a. This piece of writing is called a “short story”. - b. *This piece of writing is right now called a “short story”. NMCs are used to describe facts and a stative form of language use. Depending on the referent of the lexicalized item, it is commonly differentiated between situative and static verbs, with the latter one having a special referent that has been referred to as stative state argument (Geist 2006). Copula sentences exhibit the same behaviour as stative verbs when it comes to applying eventuality test: they both fail these tests (Maienborn 2003). Stern (2022) provides more evidence for this claim, demonstrating that the cancellation of the metalinguistic reference, i.e., negation of the metalinguistic implicature, is impossible. (46) a. *He is presenting his dissertation at the ‘paper’ session but no one ever called it that. - b. *The ‘debate’ resulted in three cracked heads and two broken noses, but no one ever described it in those terms. - - - (Stern 2022: 132) 3.2 A copula-based approach to NMCs 65 <?page no="66"?> Matushansky (2008) argues that naming predicates exhibit the same syntactic behaviour in both languages like German, English, French etc. She describes the syntax of naming constructions as involving a small clause. This approach to naming constructions involving proper names as in They named their daughter Luisa is supported by Fara (2015). In following Stowell (1981), a small clause is a minimal syntactic structure containing a subject, either a DP or CP, and a non-verbal predicate, as an AP, PP, NP or DP. Small clauses lack the full syntactic complexity of an independent clause as the verbs take a single constituent as their complement. This subject-predicate construction does not contain a finite verb (Kerstens et al. 2023). Rothstein (2004) argues for a small clause analysis for exceptional case marking verbs (ECM) like consider and make. ECM involves a construction where a non-subject argument of one verb appears to be marked with the case that is expected for the subject of another verb (Kerstens et al. 2023). In English, this often involves using the pronoun it or there as a syntactic placeholder. (47) Tine believes [Luis to have won]. Here, the case-marking of the subject Luis by the verb believe is an instance of an ECM. In the following, a copula-based approach to name-informing quotation will be outlined. Härtl (2020) argues that NMCs involving call introduce a copular relationship, either an identificational copula or equative copula. When analyzing NMCs as small clauses, they involve either a subject or an object (cf. Ágel 2017). The predicate nennen (‘call’) functions as a transitive copula verb, which involves an object predicative in NMC constructions in active voice: (48) Experten nennen diese Diskussion eine „Grundsatzdebatte“. - experts call this discus‐ sion.acc a fundamental debate.acc. The object predicative assigns the first object DP, i.e., diese Diskussion, with accusative case. Passive voice in German is built using a periphrasis, specifically the so-called “werden-passiv”, which consists of the auxiliary verb werden and the past participle of the main verb used with the metalinguistic predicate nennen (see 49). 66 3 Theoretical background <?page no="67"?> (49) Diese Diskussion wird eine „Grundsatzdebatte“ genannt. - this discussion.nom is a fundamental debate.nom called In contrast to active voice, the preand post-copula NPs in sentences involving the predicate nennen are marked with nominative case in passive voice. The argumentation in favour of a copula-based approach for NMCs is backed up from an analysis on definitional sentences, which have been discussed as being similar to equative, identificational, and specificational sentences (see Seres & Espinal 2019). Referential NPs in NMC constructions have been argued to have kind reference (Härtl 2018, Seres 2020). In addition, based on the development of the Wangeroorge Frisian naming verb heit, which translates back to call, Hoekstra (2023) even claims that heit are basically copular verbs (BE) with an additional naming component (the name quali‐ fier), both in their attributive-predicational use with nonreferential NPs (proper names or bare kind names) and in their identificational-predicational use with kind-referring NumPs (127-128). The classical division of copular sentences typically distinguishes between four different types of copular sentences, which dates back to Higgins (1979): predicational, specificational, equative, and identificational. (50) a. The winner is a man with a red beard. - b. The winner is Charlie Brown. - c. Britt Reid is the Green Hornet. - d. That woman is Susan. (Arche et al. 2019: 6) The post-copular NP in (50a) is predicative, ascribing certain attributes to the subject. Meanwhile, the subject NP is referential. In (50b), the post-copular NP identifies the referent. Both of the NPs in the identity statement (50c) are equally referential. In contrast, while (50d) has a referential subject without providing the identity of the referent, the identity is provided by the post-copular NP. Name-informing constructions have been argued to either contain an iden‐ tificational copula or an equative copula. These two types of copula will be discussed in the next two sections. 3.2 A copula-based approach to NMCs 67 <?page no="68"?> 18 I would like to thank Philippe De Brabanter for pointing out that not all of the sentences provided here, e.g. sample sentences referring to an individual as in (43c) and (50d), fit neatly to the argumentation that the post-copular NP denotes a sort. One possibility to resolve this issue might be to regard ‘sorts’ as a category that groups entities based on shared characteristics or fundamental properties, meaning that there exist underlying sorts (e.g. types of individuals). This argumentation requires further investigations in future research. 3.2.1 Identificational copula The first type of copula in NMCs is the so-called identificational copula which are used to communicate the identity of a referent (see e.g. Scott 2019). Identificational copulas have been described as containing a pre-copular phrase as “a demonstrative and the post-copular phrase is a nominal” (Heller & Wolter 2008: 226). Identificational copula clauses are characterized by expressing an identity between two denotations: the preand the post-copular (noun) phrase. (51) a. The morning star is Venus. - b. That is a man. Mikkelsen (2011) states that identificational clauses generally have a demon‐ strative as their subject. Crucially, identificational clauses “are typically used for teaching the names of people or of things” (Higgins 1979: 237). This argumentation is supported by Kiss (2006) who states that identificational sentences have a referential subject, and that their predicate has the function of teaching the name of this item. Frege (1892) has analysed identity statements in terms of concept functions. Consider the semantic representation in (52). (52) λx. λy. [y is no other than x] When applied to a sentence like The whale is a mammal, an identity relationship between two expressions is established, e.g. between whale and mammal. The two NPs enclosing the copula are syntactically similar noun phrases while other types of copula allow for a greater variance in the subject and complement position (Geist 2006). Heller & Wolter (2008) argue that the post-copular phrase in identificational sentences denotes a sort. 18 Gupta (1980) argues that predicational sentences 68 3 Theoretical background <?page no="69"?> 19 Interestingly, the sentence seems to be more acceptable when accompanied by a determiner This process is a “dehydration”. as Mary is a woman involves identity: Mary = a woman. In the covert copula sentence down below, the sort of the subject referent is specified. (53) a. This phenomenon is called a “nuclear fission”. 19 - b. This phenomenon is a nuclear fission. - c. λy GENs … [NUCLEAR FISSION(s, y)] (the phenomenon) In canonical definitions, a bi-nominal copular sentence is commonly used. These definitions typically “consists of a name of the kind to be defined followed by BE and a list of the most salient characteristics of the prototypical object” (Seres 2020: 84), i.e., following the structure “NP1 is NP2”. The first NP can be preceded by both a definite and indefinite determiner and serves as a hyponym for the hyperonym of the second NP. This becomes apparent in the next sentence. (54) a. The dog is a mammal. - b. This type of flower is a rose. In definitional sentences, the second NP is often more specific and provides the defining characteristics of the term. This specificity of the second NP is a common feature, like in the following two sentences represented in (55). (55) a. The term cracker was used in Elizabethan times to describe braggarts. - b. Confire comes from the Latin word conficere. - - (Wilson 2012: 112) Wilson (2012) states that specific structures of the sentence are commonly found in corpus data on mentioned sentences. Noun appositions, likely used to provide additional details to the first NP, can be found in the data. This includes appositive constructions like the name, the term, the title, and the phrase. Hoekstra (2023) analyses the development from the naming verb heit ‘to be called’ of the Wangerooge East Frisian dialect into the copular verb to be. This functional change was initially present in identificational copular clauses. Together with the German verb heißen, heit was originally a naming verb 3.2 A copula-based approach to NMCs 69 <?page no="70"?> 20 Jones (2018) argues that the definite article before the second NP can be omitted in English. cognate. The naming verb heit can both mean ‘to call’ and ‘to be called’ and was particularly common in copular clauses. Sentences in which the Wangerooge Frisian heit is taking kind names as complements are repeated below from Hoekstra (2023: 123). (56) a. daa híiten wii fon múurwettel those called we of-QUOT parsnip - b. ‘We called those parsnips’ (Mitth. II, 57 cited in Hoekstra 2023: 115) - c. dait fin hat den püünmillii - - the fine is-called then pound-flour - d. un dat grof hat grant - - and the coarse is-called groats - - ‘The fine one (i.e. wheat flour) is called ‘pound flour’ then and the coarse one is called ‘groats’” - - (Mitth. II, 46) The first sample sentences are with causative while the last ones are unaccu‐ sative. Heit was in competition with the copula wízze ‘to be’ and Hoekstra (2023) concludes that the “copular heit is probably best analysed as a suppletive allomorph of wízze ‘to be’ used with identificational small clauses” (141). Undergoing a grammaticalization process, heit developed into a copular verb. Please note that NMCs need to be distinguished from copula constructions such as Giscard d’Estaing is the President of France  20 , in which the post-copula phrase the President of France can be interpreted as referring to a role, post, or office, providing context and information about the subject rather than a specific person ( Jones 2018: 24). In other words, these types of constructions offer an ambiguous interpretation which does not appear in NMCs. Even when it comes to NMCs following the definition sketched in Chapter 2.1, not all NMC constructions allow for an identificational copula. Consider, for example, the following sample which does not allow for substituting the predicate with is. (57) a. This process is called “oxidation” - b. *This process is “oxidation”. 70 3 Theoretical background <?page no="71"?> 21 Note that some scholars argue that equative copulas do not exist and should be regarded as instances of predication instead (see Mikkelsen 2011). For these sentences, another type of copula needs to be considered, namely an equative copula approach. We will therefore move on to the second type of copula in NMCs. 3.2.2 Equative copula The second type of copula relation to be found in NMCs are equative copular sentences which are also termed identity or equational clause in the literature, cf. Mikkelsen (2011). A bona fide instance of an equative sentence is illustrated underneath, in which two NPs are equated. (58) a. The morning star is the evening star. Equative copula constructions equate two expressions of the same type flanking the copula, for example two proper nouns (59a) or two adjectives (59b). (59) a. Cicero is Tully - - b. Honest is honest. (Mikkelsen 2011: 1806; 1808) The two expressions, NP1 and NP2, that are equated represent the same referent (Mikkelsen 2011: 1807). The referential term in argument position introduces a referent. Both NPs in equative copulas are referential (The morning star is the evening star). In (37a), the second NP provides information on the name of a person and can roughly be understood as the person known as Cicero is also referred to as Tully (Mikkelsen 2011). 21 Equative copula clauses “have verb-second order — the finite copula occurs in second position (…), and is flanked by two phrases, giving [XP + Verb + XP]” ( Jones 2018: 61). When considering copula clauses of the form NP be NP what distinguishes equative copulas from other types of copula is their referential status. In equative clauses, the NP preceding and the NP following the copula are referential. In contrast, only the second NP in predicational and the subject, i.e., the first NP, in specificational clauses is non-referential (Mikkelsen 2011). In equative copula, the two nominals in NMCs that fill the space before and after the copular refer to the same expression. 3.2 A copula-based approach to NMCs 71 <?page no="72"?> 22 There remains debate over whether both NPs refer in a strict sense. If the left-hand NP in (60b) evokes an expression, it would be more appropriate to analyze the sentence as a hybrid between object-language and metalanguage, as De Brabanter points out. This issue deserves further attention in future research. (60) a. A hospital is also called a “clinic”. - b. A job is also referred to as an occupation. In example (60a) the propositional synonyms hospital and clinic, are equated. 22 Switching the two synonyms leaves the truth-conditions of a sentence un‐ touched. The felicity of the utterance would not be affected (Gutzmann 2021; Cruse 2004). Importantly, constructions that are built with refer to as require also to express an equative relationship between the two nominals. Copular clauses (61a) as well as NMCs (61c) cannot be reversed (Horton 1996: 321). (61) a. Joe is a teacher. - b. *A teacher is Joe. - c. This phenomenon is a moonbow. - d. *A moonbow is this phenomenon. I argue in favour of the argumentation outlined in Härtl (2020) that the sentence in (39c) is uttered by a speaker who “veridically commits” (Giannakidou & Mari 2019) to the truth of the utterance. This argumentation is also in line with the general characteristics of definitional sentences. As Seres (2020) puts it: “Such sentences are not bound to specific times or worlds, they are atemporal and non-accidental (…), they represent general truths and are law-like” (80). Despite this, as Meibauer (2014) notes, non-truthful or false quotations are quite common in everyday communication. Yet, it is crucial to draw a distinction between the speaker lying and intentionally using a false quotation. In the philosophical paper by Buekens (2011), the author elaborates on the assertive dimension. In following Williamson (2000), the Rule of Assertion can be sum‐ marized as follows: (62) One must ((assert p) only if one knows that p) The content of the speaker’s assertion is an enriched and/ or modified version of the utterance’s compositionally determined semantic content (Buekens 2011). 72 3 Theoretical background <?page no="73"?> 23 Bold letters are my own to highlight the possible properties of NMCs. Coming back to the sample sentence presented in (39c), the asserted content is that according to the speaker, the phenomenon is called a moonbow. Following this logic, the conversation in (63) does not hold and would give rise to a faultless disagreement. (63) A: This phenomenon is called a “moonbow”. - B: No, this phenomenon is not a moonbow. In this exchange, both interlocutors A and B are making claims about the phenomenon in question, asserting differing viewpoints without either being factually wrong. As Buekens (2011) notes, “if X asserts that p and Y asserts that not-p, they disagree as to whether p, and at least one of them must be at fault” (642). Although both speakers arrive at divergent conclusions about the appropriate label for the phenomenon, neither of them can be deemed at fault in their assertions. Each speaker’s perspective is valid based on their own definitions and interpretations, leading to a faultless disagreement despite the apparent contradiction in the dialogue. This section has investigated how a link is established between two linguistic entities in name-mentioning quotation: through identification or equation of the linguistic entities. 3.2.3 Drawing the boundaries between copulas In the previous two sections, I have outlined the distinction between equative and identificational copulas. Coming back to the initially proposed distinction on copula clauses as outlined by Higgins (1979), it is crucial to note that predi‐ cational copula do not occur in NMCs. Predicational copula are characterized by a referential NP1 and a predicational, non-referential NP2: the NP “predicate[s] some property (characteristic, attribute, quality) of the referent of the subject NP” (Declerck 1988: 3). By definition, NMCs contain a referring nominal (see Härtl 2020), and further linguistic analyses have revealed that definitional sentences that can be classified as NMCs are non-predicational (Seres & Espinal 2019; Seres 2020). Similarly, NMCs do not fit into the category of attributive-predicational copula, which exhibit the structure of DP/ NumP Copula AP/ NP/ PP 23 (see Dobrovie-Sorin & Beyssade 2012): “the small clause subject is referential (DP/ 3.2 A copula-based approach to NMCs 73 <?page no="74"?> NumP) and the small clause predicate nonreferential (AP/ NP/ PP)” (Hoekstra 2023: 136). Hoekstra (2023) outlines the argumentation that Wangerooge Frisian heit does not occur in predicational copular clauses but only in classificational, specificational, identificational, and equative copular clauses. Summing up, name-informing constructions involve a referring use of the post-copula nominal (see also Härtl 2020), contradicting with the definition of (attributive-)predicational copula. To the best of my knowledge, no NMCs exist that fall outside the framework discussed in this chapter, although empir‐ ical evidence distinguishing which NMCs should be classified as equative or identificational is still lacking. While an empirical classification of the types of copula clauses commonly represented in NMCs remains an open research question, I expect NMCs to predominantly be classified as identificational due to the fact that equative copula often link two nominal elements together, such as the synonyms in (60), which is not frequently used in communications. The preference for identificational copular clauses in naming constructions is compatible with corpus data presented in Hoekstra (2023). Additionally, the observations on NMCs can further serve as a testing ground for the taxonomy of copula clauses. In the next chapter we will move on to the philosophical distinction between using and mentioning expression, which is central to our understanding of NMCs. 3.3 Quotation and the use-mention distinction The use-mention distinction is a fundamental concept in the philosophy of language. One of the phenomena located at the boundary between use and mention is quotation. The terminology, known as the use-mention distinction in the literature, was introduced by Quine in his work Mathematical Logic (1940). It originally aimed at differentiating between metalanguage and object language. Essentially, this distinction clarifies the difference between using linguistic expressions to refer to entities or concepts, and mentioning those expressions to discuss them as objects of language itself. In general, words can, on the one hand, refer to their customary reference or, on the other hand, refer to themselves. To grasp the use-mention distinction, consider the following sample sentences taken from Wilson (2012: 638). (64) a. I watch football on weekends. - b. Football may refer to one of several sports. 74 3 Theoretical background <?page no="75"?> In the first sample sentence, the word football is used to refer to a sporting activity. In contrast, the second sentence refers to the term football itself, discussing it as a linguistic term. Similarly, we can illustrate the distinction with the following sentences. (65) a. Kassel is a city in Hesse. - b. “Kassel” has six letters. In (65a), the word Kassel is denotationally used with its customary reference, i.e., as a city located in Germany. Sentence (65b), on the contrary, does not denote a city but refers to the linguistic setup of the word Kassel. Here, the word is mentioned instead of used. When a word is mentioned, it is not used to refer to a specific entity in the word, here a specific city in Hesse. Instead, the term itself is explicitly discussed. As illustrated in (66a), quotation marks or italics can be used to highlight the convention of mentioning a linguistic entity. Importantly, it is not possible to replace the referential expression with a synonymous term in (66b) without contradicting the sentence’s truth-conditions. (66) a. “Sofa” has 4 letters. - b. *“Couch” has 4 letters. Mentioning an expression, as opposed to using it, can affect the truth-conditions of a sentence. When an expression is mentioned, its truth-conditions need to be considered in a meta-linguistic context. Therefore, truth conditions vary depending on whether an expression is used or mentioned. The truth-value of a sentence is typically derived from the compositional elements of the sentence, i.e. depending on the semantic features of the sentence (Davidson 1977). The same difficulty of non-replaceability applies for co-referential terms. Cappelen et al. (2023) provide the following two sample sentences in (67) to address this issue. (67) a. “Bachelor” has eight letters. - b. “Unmarried man” has eight letters. Some philosophers, including Quine (1940), have argued that quotation marks are necessary for a sentence as in (67a) and that the quotation “denotes its 3.3 Quotation and the use-mention distinction 75 <?page no="76"?> interior” (23). Therefore, mentioning is characterized by “using a word to talk either about itself or about another word” (Stern 2022: 21). Stern (2022) further argues that Frege (1892) has been the first philosopher to distinguish between using a sign and speaking about the sign itself by enclosing the latter in inverted commas in order to disambiguate use and mention. However, mentioning an expression is also possible without using quotation marks (Riegelnik 2011) and quotation marks are not the only device that enables speaker to mention expressions (Davidson 1984), including for example italics. Another option is by adding the word or the expression to mention an expression (Moore 2019). Mentioned entities can be letters, words, phrases, sentences, and also symbols. Although it has been a standard practice on philosophy to differentiate between use and mention, it is still controversial where to draw the boundary (see Cappelen et al. 2023). Within the framework of the Ostension Theory, Saka (2003) proposed a differentiation between use and mention in which the intention of the speaker plays a crucial role. Thus, he connects use with the extension of a quoted expression, while speakers mentioning words would never refer to the extension: „S intends to refer to the extension of x” (Saka 2003: 190). Mention, on the other hand, is characterized by the speaker intending to refer to something that is associated with x but that is not the extension of x. The metalinguistic nature of quotation has been summarized by Cappelen & Lepore (2007), who argue that quotations are used to talk about a very particular part of the word, namely, “language itself ” (5). Cappelen et al. (2023) further provide a definition that characterizes the use-mention distinction as a basic feature of quotation. Therefore, their definition is based on the idea that a certain expression has to be quoted: “Expression E is mentioned in sentence S just in case it is used to refer to itself in S” (para. 2.2). As the researchers themselves note, this characterization is limited to theories considering only mentioned expressions being quoted. The first problem is according to this description E that must be used in order to mention it. More importantly, a critique questions whether standard meta-linguistic devices including quotation can be considered as referring expressions. Theories have debated and addressed the question whether quotation can be used and mentioned. Saka (2003) defends the view that simultaneous use and mention is possible, adding the following definition: “S produces a token of x, thereby ostending an open-ended number of items associated with x”. Wang (2018) addresses this issue and puts it as follows: The reason (…) is that the intuition of pure quotation being mentioned is so strong that no theoretical space is left for it to possibly be used. Behind this is the entrenched idea that use and mention are exclusive of each other. When an expression is mentioned, 76 3 Theoretical background <?page no="77"?> it refers to something linguistic being itself, or its type; when it is used, it refers to something extralinguistic. (30) First, Wang refers to the question whether pure quotations can be used. In some theoretical frameworks, use and mention are mutually exclusive within the same linguistic act or utterance: in a specific context, a word or expression can either be used or mentioned, but not both would be possible simultaneously within the same sentence or statement. Counter evidence is, among others, presented by Davidson (1979). He claims that mixed quotation simultaneously involves use and mention. (68) a. Alice said that life “is difficult to understand”. - b. Alice said that life is difficult to understand. - c. Alice uttered the words “is difficult to under‐ stand”. (Wang-2018: -30) In the mixed quotation (68), Alice is not merely mentioning the phrase but the quotes reveal that the phrase is being cited verbatim as part of Alice’s speech. This is contrasted with (68a), which does not involve mention, while the last sample sentence (68c) exemplifies explicit mention. According to Wilson & Sperber (1992) the long-lasting philosophical debate on the distinction between use and mention is illustrated when speakers are using irony. The use-mention distinction offers possibility to make reference to some state of affairs that was predicted, expected or desired, either because of some explicit prediction or based upon a mutually shared domain of knowledge. (Colston & Gibbs 2007: 5) Broadly speaking, mentioned language has several functions. These functions encompass introducing new terms, the “attribution of statements, explanation of meaning, and assignment of names” (Wilson 2011). Additionally, mention‐ ing serves various other functions: “for indicating titles, explaining meaning, introducing new words, attributing exact utterances to others” (Perlis et al. 1998), and fulfilling common functions of metalanguage. Perlis et al. (1998) further note that use and mention is a crucial aspect of human conversation and communication due to its high frequency. Employing the use-mention distinction enables a more refined and precise utilization of language, enhancing its capacity for nuanced communication and metalinguistic discourse. In fact, Perlis et al. (1998) take this one step further and claim that meta discourse is “the very essence of conversation and communication”. This is in line 3.3 Quotation and the use-mention distinction 77 <?page no="78"?> with Evans (2012), who states that our capacity to coordinate, discern, and calibrate the competing interpretations of a partially shared social environment is fundamentally based on our ability to represent the words and thoughts of others and relate them to our own and our interlocutors' perspectives. Although metalanguage is crucial for communicating metalinguistic proper‐ ties and explicit information about the properties of language itself, it is no surprise corpus data reveals that we generally use words way more frequently than mentioning them (cf. Wilson 2011). 3.4 State of the art: Experimental and empirical research This section reviews studies investigating the notion of quotation from an em‐ pirical angle. The aim is to offer a comprehensive overview of the current state of experimental and empirical research on NMCs. Studies that are more relevant to the individual specifications of the various methodological approaches used in this thesis will be discussed before presenting the respective study. In an analysis of the NMC „X“ ist gut (‘“X” is good’), Finkbeiner (2015) em‐ ployed qualitative corpus data to determine formal features of the construction and detect considerable variation in the use of quotes. The meta-communica‐ tively used construction was analysed using online discussions, employing a pragmatic approach from Gutzmann & Stei (2011). X was considered as a mentioned expression irrespectively of occurring with or without quotation marks, arguing in favour of quotes being optional. She argues that the quotes are pragmatic indicators that block the canonical interpretation of X. The only large-scale corpus investigation can be found in Quaßdorf (2016), which examines quotations from a phraseological angle in Hamlet. The details are not subject to discussion here due to the lack of applicability to NMCs. Gutzmann & Stei (2011) conducted a corpus study to bolster their claim that the use of quotes is not compulsory in name-mentioning contexts and that mentioned expressions regularly occur without quotes. They report a corpus-study performed in the Archiv der geschriebenen Sprache of the COSMAS II corpus, looking at quotation marks in meta-linguistic contexts, which has been defined as metalinguistically used nominals, following the key words Wort ‘word’ and Begriff ‘expression’. In total, only 61.1% of the nominals have been marked by quotation marks. For a corpus study focussing on adjective-noun phrases in naming contexts involving sogenannt (‘so-called’), adjective noun phrases and compounds were extracted from the DeReKo corpus (Härtl 2016). The data revealed a preference 78 3 Theoretical background <?page no="79"?> 24 The Enhanced Cues Corpus is available here: https: / / shomir.net/ data/ um_corpus/ um_ corpus.html; latest access 01.08.2023 25 Note that these mention-significant words consisted of nouns and verbs form the pilot study. 26 Bold print as used in the original, here used to highlight mentioned language. for phrases in these naming constructions as compared to compounds. This finding was interpreted as a higher pressure for phrases to highlight their naming status. The aforementioned studies lack systematic accounts of the correlation between quotation (and quotes) and the linguistic properties of the quoted content. Schrader (2017) studied the interaction between the lexical frequency of a mentioned nominal involved in sogenannt-constructions and its interpretation as either literal or modalizing, making systematic use of corpus data (DeReKo). The corpus data was rated with regards to giving rise to a literal or modalizing reading, revealing that highly frequent nominals are more likely to link to ironic uses of the construction. She reports an interaction to hold between the lexical frequency of a mentioned nominal involved in sogenannt-constructions and its interpretation as either literal or modalizing, such that highly frequent nominals are more likely to link to ironic uses of the construction, concluding that compounds are less likely to appear in a modalized reading. Further empirical evidence comes from Wilson (2012) who presents the first corpus collection on metalanguage in English. The Enhanced Cues Corpus  24 is based on 5,000 Wikipedia articles and makes use of a variety of instances of metalanguage based on a total of 23 - what he labels as - mention-significant words. 25 The corpus contained sentences as presented in (69). (69) a. This is sometimes called tough love. - b. I wrote “meet outside” on the chalkboard. - c. Has is a conjugation of the verb have. - d. The button labeled go was illuminated. - e. That bus, was its name 61C? - f. Mississippi is fun to spell. 26 (Wilson 2012: 638) Notably, sentence (48a) represents and instance of a name-informing construc‐ tion involving the predicate call, which assigns the “label tough love as an attribute of the sentence subject” (Wilson 2012). When looking at the metalan‐ guage corpus, the corpus data confirms that 80 % of appearances with the 3.4 State of the art: Experimental and empirical research 79 <?page no="80"?> 27 For the purpose of this investigation, the corpus presented in Wilson (2011) could not be taken into consideration as it has several shortcomings for the current purpose such as not taking ironical sentences and instances of mixed quotation, for which he uses the term partial quotation, into consideration. naming verb call were preceded by an instance labelled as mentioned language (Wilson 2012). This in turn means that NMCs involving call are frequently used for metalinguistic reference even when compared to other metalinguistic constructions in English. It needs to be highlighted that the corpus collection is based on different categories for mentioned language so that the category Names as Names, names referring to e.g. proper names or titles, have also been taken into consideration. The most common type of category found in the data belonged to the Words as Words category which was defined as the target item referring either to the word or phrase itself. Although the corpus contains a wide variety of mentioned language, the disadvantage for the present investigation is that constructions involving quotation marks are regarded as stylistic cues that are informal and would be used as a limitation for selecting mentioned language (Wilson 2011). 27 When considering metalinguistic phenomena, studies have been presented on metalinguistic negation and metalinguistic comparison. Metalinguistic neg‐ ation concerns for example the pronunciation of words: (70) He didn’t order ‘[eI]pricots’; he ordered ‘[æ]pricots’. - - (Morzycki-2011: -76) In a corpus study on neologisms, Svanlund (2018) examined which metalinguis‐ tic instances appeared with quotes and contained an explicit metalinguistic comment, which he summarizes with the term metafocusing. More than 96 % of the neologisms were meta-focused and commonly implemented into the lexical process of becoming a conventionalized term. The number of constructions with quotation marks made up 47 % and more than half of the neologies were morphologically complex words, more specifically compounds. Further empirical evidence has looked into the presence and absence of a determiner in NMCs. These investigations are crucial to the study of NMCs as they raise the question whether the determiner is used or mentioned. As argued previously, NMCs are best analyzed as instances of pure quotation. However, pure quotations as in the sentence „Kassel” has 2 syllables do not allow for a determiner. Despite this, if quotes are present in NMCs, the determiner is usually not enclosed in quotation marks. In this case, the determiner is not mentioned; 80 3 Theoretical background <?page no="81"?> the determiner is syntactically not part of the metalinguistic argument and hence not of the mentioned nominal. In a pilot study on NMCs, Härtl (2020) presents evidence from two corpus studies. The investigation focussed on sentences with nennen (‘call’) in combination with auch (‘also’). 500 sentences following the pattern Man nennt X (auch) Y were selected for analysis from the DeReKo corpus. It revealed that naming constructions involving a nominal compound appeared more likely with quotation marks when the noun-noun compound was preceded by a determiner. The hypothesis that quotes are used more frequently in NMCs when preceded by an article was confirmed. The argumentation states that quotes highlight the metalinguistic use of the construction when preceded by a determiner. To investigate this issue further, a set of follow-up studies were conducted. In contrast to the corpus study, a forced-choice survey with German partici‐ pants (Raue 2020, see also Cortés Rodríguez et al. 2021) could not confirm this finding. Participants were asked to read sentences with and without quotes, varying the presence and absence of quotes around the target item as presented below. (71) a. Die Wanderer beobachten, dass ein vorüberge‐ hender, starker Niederschlag gefallen ist. Dieses Phänomen nennt man einen Wolkenbruch/ einen „Wolkenbruch“. ‘Hikers observe a temporary, heavy precipitation. This phenomenon is called a cloudburst/ “cloud‐ burst”.’ [withDET] - b. Der Architekt schlägt vor, auf das neue Haus ein Dach mit abgewinkelten Seiten zu setzen. Man nennt die Konstruktion Satteldach/ „Satteldach“. ‘The architect suggests putting a roof with angled sides on the new house. The construction is called a pitched roof/ “saddle roof ”.’ [withoutDET] One finding of the survey presented in Raue (2020) was that there was no statistically significant difference between naming constructions involving the three different naming verbs call, refer to and name. For this reason, the relation in the naming construction is assumed to be a be a copular one too: to refer to Y as “N” entails that Y is an N; to name Y “N” entails that Y is an N. In contrast to the corpus study in Härtl (2020), the rating study showed that there was no correlation between the occurrence of a determiner and the use of quotation marks in name-informing constructions, suggesting that nominal compounds in NMCs do not prefer quotation marking and that further empirical and experimental research is needed. In other words, the presence of a determiner 3.4 State of the art: Experimental and empirical research 81 <?page no="82"?> was not a predictor for the use of quotes in name-informing sentences, as both conditions tended to not be accompanied by quotes. It was argued that there was no need to highlight the construction’s metalinguistic use with quotes. In contrast to the corpus study in Härtl (2020), the rating study showed that there was no correlation between the occurrence of a determiner and the use of quotation marks in name-informing constructions, suggesting that nominal compounds in NMCs do not prefer quotation marking and that further empirical and experimental research is needed. In an acceptability judgement in German, NMCs were used in which the presence and absence of a determiner was manipulated (Cortes Rodríguez et al. 2022). The sentences made use of the predicates nennen (‘call’) and bezeichnen als (‘refer to as’). (72) a. Seit einigen Jahren nennt man dieses Möbelstück Bücherregal. ‘For several years, this piece of furniture has been called bookshelf. [-determiner, nennen] - b. Seit einigen Jahren nennt man dieses Möbelstück ein Bücherregal. 'For several years, this piece of furniture has been called a bookshelf.’ [+determiner, nennen] The acceptability judgement revealed no significant difference between the ratings of the constructions with and without determiner, indicating that both conditions were perceived as acceptable constructions by the participants. Third, a self-paced reading study was conducted to investigate the anaphoric resolution of nominal expressions in NMCs. The experimental items are repro‐ duced from Cortés Rodríguez et al. (2022: 21). (73) Das Frühstücksei war letztes Jahr in dieser Fam‐ ilie so beliebt, dass man ein besonderes Gefäß, dass es vor dem Wegrollen sichert, benutzte. ‘Last year, the breakfast egg was so popular in this family that one used a special container, which keeps it from rolling away.’ - [Context] - a. / Man / nennt / die Erfindung / einen Eierbecher. ‘One calls this invention an egg cup.’ [+determiner] - - Wochenlang / half / dieser / immer / auch / beim Frühstück. ‘For weeks, this always has helped at breakfast, too.’ [Target (demonstra‐ tive)] 82 3 Theoretical background <?page no="83"?> b. / Man / nennt / die Erfindung / einen Eierbecher. ‘One calls this invention an egg cup.’ [+determiner] - - Wochenlang / half / er / immer / auch / beim Früh‐ stück. ‘For weeks, it always has helped at breakfast, too.’ [Target (pronoun)] - c. / Man / nennt / die Erfindung / Eierbecher. ‘One calls this invention egg cup.’ [-determiner] - - Wochenlang / half / dieser / immer / auch / beim Frühstück. ‘For weeks, this always has helped at breakfast, too.’ [Target (demonstra‐ tive)] - d. / Man / nennt / die Erfindung / Eierbecher. ‘One calls this invention egg cup.’ [-determiner] - - Wochenlang / half / er / immer / auch / beim Früh‐ stück. ‘For weeks, it always has helped at breakfast, too.’ [Target (pronoun)] The results for the critical reading region, i.e., the target word, showed no significant effects, neither for the anaphora type nor determiner factors. Following this series of experiments, the researchers conclude that because none of the results point to differences in the behavioral treatment, names used in NMCs with determiner do not differ referentially from uses without determiner. Further experimental research is related to the interpretation of NMCs. In a reading time study reported in Schlechtweg & Härtl (2023), the processing of quotation marks in ironic sentences in English was tested for using the self-paced reading paradigm. They used six different conditions: literal with quotes (a), literal without quotes (b), ironic with quotes (c), ironic without quotes (d), unrelated meaning without quotes, unrelated meaning with quotes. The conditions of interest are repeated in (74) with sample sentences from the experimental study. (74) a. Context: Yesterday, a young woman won the jackpot of this month’s lottery. She decided to donate ten million of the thirty million dollars to cancer research. Target: We all hope that the generous lady uses the rest of the money to fulfill her own dreams. 3.4 State of the art: Experimental and empirical research 83 <?page no="84"?> b. Context: Yesterday, a young woman won the jackpot of this month’s lottery. She decided to donate ten million of the thirty million dollars to cancer research. Target: We all hope that the “generous” lady uses the rest of the money to fulfill her own dreams. - c. Context: Yesterday, a young woman won the jackpot of this month’s lottery. She decided to donate sixteen cents of the thirty million dollars to cancer research. Target: We all hope that the generous lady uses the rest of the money to fulfill her own dreams. - d. Context: Yesterday, a young woman won the jackpot of this month’s lottery. She decided to donate sixteen cents of the thirty million dollars to cancer research. Target: We all hope that the “generous” lady uses the rest of the money to fulfill her own dreams. For the target reading region ironic parts were processed more quickly when being accompanied by quotes in contrast to no quotes. Literal meanings containing quotes were processed less quickly than those without quotes. They conclude that - independently of the sentences’ interpretation being literal or non-literal - quotes first increase the processing burden, but would facilitate the processing and recognition of irony later on. Experimental investigations by Schlechtweg & Härtl (2020) have shown that name-mentioning quotes are acoustically observable in the quoted words’ pronunciation. In spoken discourse, there was a difference in pronunciation between the quoted and non-quoted expression with a lengthening effect, i.e. a longer syllable onset, proven for quoted nouns (Schlechtweg & Härtl 2020). In a reading study, participants were asked to read out German name-informing (see 75a) and non-name-informing sentences (see 75b). (75) a. Diese Blütenknospen werden in Essig oder Öl eingelegt. Man nennt sie Kapern/ „Kapern“ in vielen Regionen. ‘These flower buds are pickled in vinegar or oil. One calls them capers/ “capers” in many regions’ - b. Diese Blütenknospen werden in Essig oder Öl eingelegt. Man kennt die Kapern/ „Kapern“ in vielen Regionen. ‘These flower buds are pickled in vinegar or oil. One knows the capers/ “capers” in many regions.’ 84 3 Theoretical background <?page no="85"?> The results further suggest that there is a difference between name-informing and non-name-informing constructions (Schlechtweg & Härtl 2020). Crucially, there were no significant effects detected between constructions with and without quotation marks which can be interpreted as the predicate nennen being enough for an identification of quotation. In addition to this, a large-scale corpus study by (Raue & Cortés Rodríguez 2022) aimed at investigating the individual differences that the occurrence of a determiner induces in name-informing constructions. The corpus-based construction analysis focused on NIQs in English systematically using Sketch‐ Engine and the English Web 2020 (enTenTen20) corpus. Our extraction examines N=2,000 constructions using the naming-informing predicate call following the pattern given down below in (76). (76) a. W calls X “Y”. [-determiner, active] - b. X is called “Y”. [- determiner, passive] - c. X is called a “Y”. [+determiner, active] - d. W calls X a “Y”. [+determiner, passive] Notably, this is the first investigation to look into possible differences between active and passive constructions. All queries used the written marking of quotes for Y. The study aimed to provide a more robust empirical foundation for the individual differences that determiners induces in name-informing constructions. The statistical analysis revealed that there are significantly more valid NIQs when preceded by a determiner. Moreover, there were significantly more valid NIQs in active voice. It was argued that the difference in referentiality of singular generic nouns is reflected in the finding that NIQs with an article occur more frequently in the corpus data. In a follow-up study on NMCs in German corpus data (Raue & Cortés Rodríguez 2025), another examination investigated the German Ten 2020 (de‐ TenTen20) corpus ( Jakubíček et al. 2013) provided by SketchEngine, extracting n=500 hits per query. To ensure a cross-linguistic analysis, all queries followed the same pattern as presented in (76) above and involved the predicate nennen (‘call’). Valid hits were selected manually following the defining criteria of NMCs. Notably, constructions of (76d) in passive and with determiner had the lowest number of hits in the corpus, underscoring the limit of queries that were meant to be extracted (219 hits). After running Generalized Linear Models in R, the results of the German data showed a significant effect for NMCs without determiner, indicating that quoted target words had a preference for not being 3.4 State of the art: Experimental and empirical research 85 <?page no="86"?> preceded by a determiner. In addition, a statistically significant difference was proven to hold between active and passive voice: while passive constructions with determiner had the highest number of hits, the lowest number of valid hits occurred in passive voice when a determiner preceded the quoted nominal. Passive constructions in were less acceptable when the nominal was preceded by a determiner. The high percentage of passive occurences was interpreted as reflecting the definitational nature of NMC constructions, which does not require active voice. Despite this, the findgs need to be set into relation with the previously reported corpus on German and experimental data from German, to explain for cross-linguistic differences between German and English. Kasimir (2008) found evidence that the presence of quotes in reported speech, in names and titles, and in modalizing quotation influences prosodic realization. Contrasting quoted passages with unquoted passages, three different types of subclausal quotation were investigated, using direct reported speech, names and titles, and modal uses of quotation marks in the sense of scare quotes. Prosodic marking seemingly failed to represent a spoken substitute for written quotes as indicated through subjects’ back-translations from the spoken into the written idiom. There was no unique strategy identified for the prosodic marking to differentiate between quoted and unquoted material. As this overview has shown, the exact nature of the quoted material in NMCs, their embedding as well as underlying cognitive processes when reading quotation marks have not been studied empirically. In the following section, the research gap will further be defined to develop research questions for empirical studies. 3.5 Research gap & objectives The debates on quotation, in which NMCs are argued to be located for this investigation, have primarily been addressed from a philosophical angle. Several theories have been and are still being developed, still lacking consensus in the scientific debate. The classification of the various types of quotation and the different kinds of quotation markers have sparked interest mostly from a theoretical approach. Research interests have predominantly been centred on reported speech as well as on irony comprehension. One pivotal area of inquiry revolves around the distinctions among different types of reported speech, with a particular focus on direct speech, indirect speech, and the nuanced category of free indirect speech. Scholars endeavour to unravel the semantic nature of direct speech, probing its syntactic status and shedding light on the intricacies that characterize its representation in language. Moreover, the use-mention distinc‐ 86 3 Theoretical background <?page no="87"?> tion has been addressed by theories of logic, and the philosophy of language. Further research topics include quotative markers, serving as linguistic devices to introduce direct speech. The grammaticalization of new quotative markers in English (for example like or be like which is often used in oral speech), constitutes another fertile ground for investigation (e.g. Golato 2000). The quotative use of like is comparative in nature and represents an approximation of another utterance, illustrating what has previously been uttered without using the exact words (Buchstaller 2001), see the sample sentences by Ranger (2015) taken from COCA: (77) a. We don’t see a lot of her because our schedules clash really badly. I see her for about ten minutes a week. It’s like “Hi… bye” in the door, out of the door. (2) - b. NICE: I know I’m happy because she told me I was happy. I wake up, I’m like -she’s like, How you feeling? I’m like, I’m a little down. She’s like, No you’re not. I’m like, That’s good.’ (5) Be like (see 77b) and like (77a) have only recently entered northern varieties of British English (Beal 2004) and is especially found among younger people. Originating from a conjunction, like stepwise evolved into a grammaticalized quotative construction (Meehan 1991). In addition, empirical studies on reported speech extend their reach to the analysis of spoken language, examining its nuances and pragmatic functions (e.g. Golato 2000; Blackwell & Fox Tree 2012). Experimental research on name-mentioning and pure quotation, as outlined in the previous section, is sparse. Generally speaking, there are no empirical investigations into the linguistic properties of mentioned expressions as well as interactions between linguistic factors and the occurrence of quotes. Further‐ more, cross-linguistic examinations for the languages German and English are missing. As a result of these research gaps, this dissertation aims at providing empirical evidence for the linguistic conditions effective in the composition of NMCs available in German and English. This work attempts at using an experimentally informed approach to study the phenomenon of pure quotation. One of the main goals of this thesis is characterize the phenomenon of name reference in metalinguistic contexts form an using empirical and experimental data. Based on insights for this particular type of quotation, implications can be drawn for the general modelling of quotation as well as its place on the semantics-pragmatics continuum. The following research questions will be addressed empirically: 3.5 Research gap & objectives 87 <?page no="88"?> RQ1: What types of NMCs exist in German and in English? RQ2: What linguistic properties, including word type, lexical frequency, and morphological type, do mentioned expressions have? RQ3: What is the systematicity in the use of quotes? Is their use affected by the linguistic properties of mentioned expressions? RQ4: How does a modalized context influence the processing of lexicalized items as opposed to name-informing contexts? RQ5: How are expressions mentioned in modalizing contexts in contrast to those in name-informing contexts processed? RQ6: How are written quotes processed during online reading? RQ7: Where are quotational constructions located on the semantics-prag‐ matics continuum? The investigation aims at answering the above-mentioned questions with a theoretically-based foundation. Empirical evidence from various studies, using corpus data, rating studies, self-paced reading experiments, and eye-tracking data will subsequently be applied to NMCs and the results discussed with respect to the empirical findings. 88 3 Theoretical background <?page no="89"?> 4 Quotation and reference to names Quotation is a tool by which speakers can denote signs. Among others, Cappelen et al. (2023) have studied the role of quote marks, explaining that they are a metalinguistic tool - a referential technique used to discuss language. This metalinguistic tool highlights the linguistic aspects of the expressions for the recipient. Quotation marking is argued to be employed in name-informing formulations to refer to the name of a lexical concept (see e.g. Härtl 2018; Härtl & Seeliger 2019). For this reason, the names in NMCs are not proper names but the term lexicalized names is instead used referring to terms that are stored in the speaker’s mental lexicon. Within this context, lexicalized names represent a category of terms that have acquired a certain level of use within a speech community and are stored in the mental lexicon of speakers. They have become lexical entries, i.e., well-established concepts. Apart from lexicalized names it is crucial to note that NMCs are used to introduce new terms and that in these cases there are no lexicalized items. In instances where speakers encounter the need to introduce novel terms, the absence of pre-existing lexicalized items necessitates the creation of ad-hoc formations. For example, speakers might use an ad-hoc, i.e., novel, compound to fulfil the naming need in a communicative situation. As a consequence, these names cannot simply be retrieved from the addressee’s mental lexicon but needs to be retrieved using pragmatic reasoning (Allot & Textor 2012). The extension can either be narrowed down or extended by a new ad-hoc concept (Carston 2010). In other words, the concept’s denotation may be more general or specific, constituting either a subset or a superset of the linguistically encoded concept. Within the transition from an ad-hoc formation to becoming a well-established concept stored in the mental lexicon of a specific speech community, name-mentioning constructions play a pivotal role in supporting this process. In doing so, the understanding of quotation marks is one of serving as more than a mere punctuation mark. Instead, they serve as metalinguistic cues that draw attention to the linguistic dimension of expressions. <?page no="90"?> 4.1 Lexical-semantic and pragmatic properties of name-mentioning quotation In this section, the lexical-semantic properties of name-mentioning quotation and how the pragmatic context influences the interpretation of NMCs will be explored. The predicate call has been analysed as a three-place predicate, involving three thematic arguments: an agent x, a theme y, and a relational argument n (Härtl 2020; Cortés Rodríguez et al. 2022). The relational argument introduced the name of the theme argument y. Crucially, the agent is bound generically. The following representation is adapted from Härtl (2020). (78) a. Astronomers call this phenomenon “supernova”. - b. x cally “n” - c. λy λn λx [CALL(x, y, NAME(“n”, y))] - d. GENx [CALL(x, this phenomenon, NAME(“supernova”, this phenom‐ enon))] This NMC introduces the shape “n” of the name of the theme argument y. Furthermore, the sentence adopts a generic meaning in the name-informing construction. This generic binding adds to the flexibility of the construction, allowing it to convey general truths. Genericity allows speakers of a language to “express generalisations and regularities about the world, about individuals, groups of individuals, events, states of affairs and situations” (Seres 2020: 1). Generic sentences are illustrated in (79). (79) a. The dolphin is a mammal. - b. The Earth orbits around the sun. Genericity can be characterized as the “expression of reference to kinds” (Seres 2020: 13) and constitutes a cognitive phenomenon. (79a) above represent instan‐ ces of D-genericity due to the definite singular NP expressing the genericity (Krifka & Gerstner 1987). Sentence-level genericity needs to be distinguished from genericity on a nominal level, whose source is an NP. Kinds have been understood as sortal concepts from a “mental catalogue” (Mueller-Reichau 2011: 35) that categorizes objects. Crucially, following the argumentation presented by Cortés Rodríguez et al. (2022), this analysis excludes sentences with non-ge‐ nerically bound agent arguments. To illustrate, in the sentence The senator called 90 4 Quotation and reference to names <?page no="91"?> the president a “liar” the agent argument is not bound generically and is therefore out of the scope of this analysis. Coming back to the sample sentence exemplified in (78), it can be concluded that generic NMCs are well-suited to express a naming convention. They are located at the interface between lexicon and semantics. While the concept supernova is stored in the speaker’s mental lexicon, the use of the NMC indicates that the word is assumed to be not well-established in the hearer’s mental lexicon. The speaker makes use of the canonical NMC based on the assumption that the recipient is unfamiliar with the quoted nominal. This is strongly linked to the item’s low lexical frequency which renders the sentential construction adequate and acceptable (cf. Härtl 2018). In addition, lexicalized names often carry specialized meanings or connota‐ tions beyond their literal or compositional interpretation. The interpretation of these names may rely heavily on contextual and pragmatic information. Quotation can operate on the shape of a concept’s name. As a consequence, pragmatic factors play a pivotal role in how name-mentioning constructions are interpreted, giving rise to different interpretations. The notion of naming itself consists of two readings. Härtl (2016) outlines that the different interpretations, i.e., name-informing vs. modalizing, relate to the type of binding the agent argument and the event argument of the verbal root (CALL x, y, z, e). The agent (x) and the event variable (e) are tied generically in the name-informing inter‐ pretation of the construction, as argued in (78) above. In contrast, an existential closure of the two arguments is required in a distancing interpretation (Härtl 2016; 2018). Consider the sample sentence presented below. (80) a. Dieses Gebäude nennt man „Hotel“. ‘This building is called “hotel”.’ - b. Die so-genannten „Freunde“ erwiesen sich als nicht vertrauenswürdig. ‘The so-called “friends” turned out to be unreliable.’ In this sentence, a name-informing use of nennen is not acceptable with a literal interpretation. When following a relevance-based implicature to hold in (80b), a distancing interpretation need to be adapted for the highly conventionalized nominal (see Härtl 2018). Therefore, pragmatic reasoning is necessary in order to determine the intended meaning. For an analysis of the lexical-semantic properties of the modifier sogenannt (see Härtl 2018). 4.1 Lexical-semantic and pragmatic properties of name-mentioning quotation 91 <?page no="92"?> 28 The examination centres on NMCs in the languages German and English. This focus might lead to language-specific structures that may affect the generalizability of the findings. While German is highly productive in word formation processes, i.e. the creation of compounds, which impacts the creation of new names, languages with limited compounding might prefer phrases instead of compounds. 4.2 The naming status of nominal mentioned expressions This section delves into the distinction in the naming status of expressions, aiming to explore the mechanisms behind the naming status of nominals. 28 The division between phrases and compounds serves two distinct purposes. The functional divergence between compounds and phrases hinges on the distinc‐ tion between description and naming: phrases are primarily characterized as having a description function, while compounds are said to serve a naming function. This functional dissimilarity was initially underscored by Zimmer (1971) and Downing (1977) and has recently been supported by Olsen (2012) and Bücking (2009). Compounds are characterized by consisting of a combination of at least two free morphemes into a new word. In word-formation studies focused on compounding, there is a prevalent concept emphasizing the naming function of compounds. This viewpoint has been advocated by various scholars (see Dirven & Verspoor 1998; Zimmer 1971; Bücking 2009 & 2010; Hüning 2010, Schlücker & Plag 2011, Härtl 2018 & 2020; Schlücker 2020). According to Schlücker (2012), compounds have a primary role known as the name-function, whereby they serve as naming structures for established concepts. These concepts would in turn be sub-concepts of the head nominals. Zimmer (1971) emphasizes that while descriptions cover a broader range, names are selectively assigned to pertinent or nameworthy categories. The creation of new names is regarded as a fundamental objective within word-formation processes (Fernández-Domínguez 2009), a perspective Bauer (2017) supports by arguing that one of the primary functions of word-formation is to provide names for previously unnamed entities. Speakers create new com‐ pounds to denote relevant categories (Downing 1977). Consequently, speakers create new compounds to denote relevant categories (Downing 1977), driven by the need of a speech community to refer to entities or abstract concepts. New compounds are created in word-formation processes. Their mean‐ ing becomes manifested through lexicalization processes (Schlücker 2012). Compounds have been argued to function either as lexicalized labels or as kind-names. First, according to Štekauer (1998), naming concerns the operation of coining a label for a conceptual entity. On the other, speakers can refer to an 92 4 Quotation and reference to names <?page no="93"?> 29 In contrast to compounds, phrases are ascribed a description function (cf. Schlechtweg 2019). already existing label in naming contexts. The naming act can be regarded as a cognitive phenomenon in which speakers need to identify (semantic) networks of already existing words and the relationship to other lexical entries in order to create a new name (Štekauer 2005). The terms lexical enrichment, labelling, and lexical innovation have been used in the literature to describe the introduction of new names (Fernández-Domínguez 2009). According to Levi (1978), the efficiency of word-formation processes and the semantic specialization - that is, the specialized meaning of compounds - make compounds ideal naming devices. Zimmer (1971) states that whereas descriptions, i.e., phrases, can be used for anything, names are only provided for relevant, nameworthy categories. 29 Compounding would appear to be primarily a device for generating new names across languages; nevertheless, it should be noted that not all nameworthy referents in such languages for which there is no name will inevitably be denoted by a compound. A conventionalized label can be put on a word when a sentence with a naming construction is referring to a lexical item (Härtl 2018). The condition which allows for creating a new label is the need of a speech community for a new expression: they serve to fulfil a naming need. Word-formation processes will only be activated once a designation is needed, i.e., when there is no entry in the mental lexicon (cf. Fernández-Domínguez 2009). Kastovsky (1986) states that the labelling process is naturally linked to the actual or perceived existence of a referent that belongs to a class of referents that need labels. For established words, a permanent label is created (Ryder 1994; Rapp 2015) and they have a new status of a “‘timeless’ name” (Härtl 2016: 15). Generally, the process of labelling a compound consists of three stages: the creation of a category that is a simple noun, the creation of a sub-category by creating a compound with the simple noun as the head constituent combined with a novel non-head constituent, and lastly the labelling of the word. Second, compounds as kind-names are understood as being representative of a broader category or kind. Naming as referring to kinds has initially been introduced by Quine (1960), who regraded kinds as concrete individuals. In addition, Carlson (1977) described kinds as abstract entities that have a link to concrete entities. Kinds are in the following understood as sortal concepts belonging to a general conceptual system (Mueller-Reichau 2011). According to Bücking (2010), nominal compounds in German are underspecified names for kinds; they are kind-denoting. Taking the example of birdhouse the compound 4.2 The naming status of nominal mentioned expressions 93 <?page no="94"?> refers to a type of construction designed for birds. The compound becomes a kind-name by representing a category rather than a specific instance. Composite words would be more prone to adapt a generic interpretation which makes them create kind-denoting terms (Bücking 2010). Nominal compounds have been argued to refer to subordinate categories (Dirven & Verspoor 1998) or sub-kinds (Schlücker 2013). According to Olsen (2012) compounds “name concepts and are closely associated with the denotation of ‘kinds’; they specify a link between a linguistic expression and a complex concept” (2136). Krifka (2003) notes that reference involved in kind reference is connected to an entity being related to specimens. The specimen of a kind is introduced by an existential quantification. Kind-referring terms can be tested for by adding x be extinct, x die out, x exterminate or invent x. This argumentation is based on the idea that only kinds can be invented or wiped out (cf. Krifka & Gerstner-Link 1993; Mueller-Reichau 2011; Härtl 2015; Schlechtweg 2018). (81) a. *A cat is extinct. A clear indicator of kind reference in the language German is when the suffix -art can be added to the NP, e.g., eine Tierart. Additionally, kind-referring nominals are frequently found in metalinguistic definitions used to explain the meaning of concepts. As Seres (2020) puts it: “Definitions represent a mechanism of identification of one kind (concept) with another” (5). In this section, the naming status of compounds was investigated. It has been argued that compounds are well suited to fulfilling a naming function in NMC constructions and should frequently be found in literal sentences. 4.3 The different readings of naming constructions In the following, it will be argued that NMCs give rise to two different readings. Consider the following sample sentence. (82) That’s what I call a friend. Now imagine that the first sentence is uttered in a conversation between a parent and their child, discussing the qualities of a good friend. In this case, the parent might point to someone who has been consistently supportive, understanding, and loyal to the child. In a moment of appreciation, the parent looks at the 94 4 Quotation and reference to names <?page no="95"?> 30 Here in the sense of culturally determined concepts, i.e., purely mental concepts. These concepts are based on our past experiences and are determined by boundary lines that we draw by using words (see e.g., Hoemann et al. 2019). friend and says to their child That's what I call a friend, emphasizing how the person embodies all the qualities they believe make a true friend. For the second scenario, imagine a situation where a person pretends to be friendly but is actually being deceitful or dishonest. Someone could then sarcastically point out this person’s insincerity by uttering the same sentence, implying that the person’s behaviour does not align with genuine friendship qualities. This could occur in a situation where someone acts friendly superficially but does not truly exhibit the loyalty, support, or honesty expected in a real friendship. Notably, the stereotypical concepts 30 of what a “true friend” is are either highlighted or these characteristics need to be rejected. The difference between the two scenarios is that, although using the same exact wording, friend has two different interpretations: the literal, also known as conventionalized meaning, and the non-literal, ironic reading. The verbal root’s event argument and the agent argument’s type of binding are related to the two distinct interpretations of NMCs, a name-informing or modalizing interpretation (see also Chapter 4.1). 4.3.1 Conventionalized readings The terminology conventionalized reading relates to a name-mentioning con‐ struction giving rise to a name-informing, literal interpretation of the quoted ex‐ pression. The function of the sentential construction is to address a specific word metalinguistically and highlight the importance of this specific terminology to the interlocutor. In making use of a conventionalized reading, the addressee is informed about a specific term. The speaker is making use of an agreed-upon interpretation, commonly terminology used by specialists, which allows for establishing a common ground for communication and comprehension among the interlocutors involved in the exchange. In a conventionalized reading, the lexical meaning of the expression in quotes does not deviate from its conventional denotation. The extensional reference does not change in these sentences. For this reason, there can be overlaps between direct and indirect quotes and a conventionalized reading, meaning that there is no deviation from the term’s standard denotation. The terminology was initially introduced by Klockow (1980) and is characterized by referring to the extension of the quoted expression. In his dissertation, Klockow initially proposed to differentiate between three different types of quotation: the P-Zitat 4.3 The different readings of naming constructions 95 <?page no="96"?> (‘pragmatic quote’) and L-Zitat (‘logical quote’), which both belong to the category of conventionalized quotation. The former comprises the types direct quote and mixed quote while the latter can be subsumed as a pure quote. The last type introduced is the modalizing quotation, nowadays also referred to as scare quotes, which would not be redundant in most cases (Stei 2007). This type of quotation will be under discussion in the following sub-chapter. 4.3.2 Modalizing readings In contrast to conventionalized reading, modalized readings refers to the speaker expressing a certain modality. Non-conventionalized readings are characterized by expressing the speakers’ reservation with regard to a certain word or termi‐ nology. Modalizing quotes express, according to Klockow (1980), a reservation of application. He describes modalizing quotes in using examples and also provides the following definition ein bloßes signal für die präsenz einer zusätzlichen sinnebene, eine art aufforderung, den markierten ausdruck nicht in naiver weise zu interpretieren. Wie er zu interpre‐ tieren ist, muß der leser selbst erschließen, wobei nicht nur der kontext, sondern auch die kenntnis sonstiger verwendungszusammenhänge, benutzer etc. des ausdrucks eine rolle spielen kann. [sic! ] (Klockow 1980: 22) a mere signal for the presence of an additional level of meaning, a kind of request not to interpret the marked expression in a naive way. How it is to be interpreted must be discovered by the reader himself, where not only the context, but also the knowledge of other uses, users etc. of the expression can play a role. In this characterization of modalizing quotes, Klockow alludes to the added meaning and also the importance of the context for the interpretation. Name mentioning sentences can comprise a distancing interpretation. This interpre‐ tation is given rise by a modalizing use with which a speaker refers attitudinally to the shape of a certain denotatum’s name (Härtl 2018). This modalizing use is represented in (83) below. (83) a. That’s what I call a “friend”. - b. That’s what I call a “good deal”. In the sentences above, the quoted expressions can adopt a non-convention‐ alized reading when embedded in a (situational) context that gives rise to a non-literal meaning of the quoted item. The speaker expresses a specific 96 4 Quotation and reference to names <?page no="97"?> 31 In this argumentation, I would like to draw a clear distinction between irony and sarcasm. The latter is regarded as having a clear target and transporting some aggressive undertone. 32 Please note that Attardo et al. (2003) do not explicitly mention quotation marks as an indicator of irony but focus on phonological and facial markers. attitude, which negates the stereotypical properties of the expression, e.g. the characteristics of what constitutes a friend. When a word is used modally, the predicate does not fulfil the function of informing the addressee about the name of a concept. Instead, a distancing interpretation is expressed, which is often realized in combination with an ironic or sarcastic under-tone (Härtl 2018). The quotation marking highlights that the semantics depart from the stereo‐ typical interpretation. In accordance with Klockow (1980), a modalizing use of quotes expresses a certain speaker’s reservation concerning the appropriateness of the word. In written language, quotation marks are often used to visualise and highlight ironic content. 31 However, even when lacking irony markers, utterances can be ironical (Attardo 2000). 32 The speakers in (84) express a reservation concerning the word’s application with regards to the word’s referentiality, questioning the referring act of the quoted expression. Gutzmann & Stei (2011) state that quotes mark the expression and “indicate that the standard interpretation of the quotatum is pragmatically blocked” (2662). In some cases, the quotes can be replaced by the word allegedly or so-called. (84) a. The “debate” was in fact a terrible argument. - b. The so-called debate was in fact a terrible argument. - c. The “meeting” turned out to be a complete hoax instead. - d. The alleged meeting turned out to be a complete hoax instead. Schmieder (2013) adds further words that are closely connected with a distanc‐ ing interpretation, namely supposedly, to some extent, and kind of. (85) a. Die „Erde“ auf dem Mond. The “earth” on the moon. - b. Er war ein „Athlet“. He was an “athlete”. - In the above-mentioned sentence, the reference shifts from the stereotypical denotation to a pragmatically enriched meaning. A modalizing use of quotes 4.3 The different readings of naming constructions 97 <?page no="98"?> is also known as expressing a reservation regarding the appropriateness of the word (Schmieder 2013). In addition, ironic sogenannt-constructions entail that the head nominal’s denotatum has been called by the mentioned name before and the results indicate that this meaning component tends to figure as a presupposition. This insight is compatible with an echo approach to verbal irony ( Jorgensen et al. 1984). Note also that the scare quotational use of so-called should be distinguished from its mentioning use in sentences like “Giorgione was so-called because of his size”. The scare quoting meaning is closer to that of alleged and probably belongs to the class of privatives (Stern 2022). 98 4 Quotation and reference to names <?page no="99"?> 5 Characteristics of metalinguistic predicates In German and English, NMCs involve a variety of metalinguistic predicates. The following chapter aims at providing an overview of metalinguistic nam‐ ing-predicates. In addition, the most commonly used predicates, namely nennen and bezeichnen als as well as the English equivalents call and refer to as, will be investigated by looking at their etymology and their linguistic properties. Wilson (2012) describes these predicates as mention verbs. Before doing so, it is crucial to establish a distinction between nomination and naming. (86) a. The person was named the president of the uni‐ versity. [nomination] - b. The king of England was named Edward. [naming] Although both sentences make use of the same verb, there is a sharp contrast between naming verbs and verbs of nomination. The first sentence is an example of an act of nomination: a specific person was given verbally assigned the position of being the president of the university. Individuals receiving a nomination can be assigned a new position or a new status. (87) a. Barack Obama was declared the President of the United States. - b. I hereby pronounce you husband and wife. - c. She was proclaimed/ christened. - d. He was elected president. - e. She was promoted CEO. All the examples illustrate an act of nomination. The predicate refers to a role, position, or a title that has officially been acknowledged as a new status as a consequence of a nomination process (Métairy et al. 2020). Within the taxonomy of speech acts proposed by John Austin (1962), utterances like (87c) change an existing state of affairs by declaring the marriage, clustered into the category of declarations (Plag et al. 2015). Nomination verbs have been argued to take the same syntax as change-of the-state constructions: they both take a small clause complement (see also <?page no="100"?> Stowell 1981; Levin 1993; Matushansky 2008). For this investigation, the focus is exclusively on naming constructions that describe a naming convention. 5.1 The naming predicate call The analysis starts by looking at the most frequently used verb in NMCs, namely the predicate call. In the English Web 2020 (enTenTen20) corpus, call has 21,417,236 hits in total, i.e., 496.63 hits per million tokens, and is herewith the most frequent NMC predicate. Its prevalence in the enTenTen20 corpus attests to its ubiquitous presence in everyday communication. The naming verb call can adopt a variety of uses and has therefore been described as an exceptional verb. Examining call from a cross-linguistic perspective, it stands out as an exceptional verb, as delineated by Matushansky (2008). Call does not exclusively appear in naming constructions. It serves as a vehicle for assigning names but also as a tool for description, communication, instruction, and proclamation. The verb can be used for descriptions in combination with a noun (I wouldn’t call him a great friend), when making a telephone call (We called the firefighters), to order someone to a specific place (They called the prime minister to London), and to make an announcement (We should call on Leona), (see ODE 2023). As a consequence, call is not restricted to naming constructions but can appear in a variety of other syntactical structures. To delve into the etymological roots of call, one can trace it back to Proto-Ger‐ manic *kall. The evolutionary journey of this verb dates back to at least the mid-13 th century, where it was already employed to bestow names upon individuals or objects (ODE 2023). Additionally, its semantic range included not only the act of naming but also encompassed the broader concept of invoking or addressing by a specific designation. This historical dimension illuminates the enduring significance of call as a linguistic device for ascribing identity and facilitating communication. When considering the complement inventory of call, the lexical unit can take either a NP or and AdjP (Herbst et al. 2004). The Cambridge Dictionary lists different readings of the verb. a) to give a name to My name is Alexander, but I’m called Sandy by my friends. b) to regard (something) as I saw you turn that card over - I call that cheating. 100 5 Characteristics of metalinguistic predicates <?page no="101"?> c) to speak loudly (to someone) to attract attention etc She called louder so as to get his attention. d) to summon; to ask (someone) to come (by letter, telephone etc) They called him for an interview for the job. e) to make a visit I shall call at your house this evening f) to telephone I’ll call you at 6 p.m. g) (in card games) to bid As becomes apparent when investigating the sentences, the use of call is not restricted to its function as a naming predicate in NMCs. Given the variety of uses of the predicate call, Anderson (2007) notes that this predicate is used more generally as compared to the predicate name which has a preference for proper names (e.g., I name this child Paul). In addition, call is often used as a phrasal verb in combination with a preposition like in or for. According to Clark & Wong (2002) the “verb be called appeals directly to convention” (187). When considering the naming predicate call in an NMC, the verb functions as a name-selecting predicate which typically involves three arguments in a generic reading, see the representation adapted from Härtl (2020) (see also Chapter 4.1). (88) a. This natural phenomenon is called “moonbow”. - b. x cally “n” - c. λy λn λx [CALL(x, y, NAME(“n”, y))] - d. GENx [CALL(x, this phenomenon, NAME(“moonbow”, this phenom‐ enon))] The construction to be called is described by Matushansky (2008) as an ex‐ ceptional case marking verb that appears in sentences such as She is called Luisa. Because it is the sole name verb to be found with adjectival predicates, call is exceptional (Matushansky 2008). Nonetheless, Matushansky (2008) has demonstrated that name verbs in Germanic languages, such as English and German, exhibit identical syntactical behaviour. Remarkably, call has been perceived as a default verb that possesses certain fundamental characteristics in common with the copula be, see also Chapter 2.2. In addition, call allows for adjectival predicates, here emotionally captivating (Matushansky 2008: 619). 5.1 The naming predicate call 101 <?page no="102"?> (89) She called the performance on the opening night emotionally captivating. In contrast to the German counterpart nennen, call does not have a reflexive use. (90) a. Das Phänomen nennt sich ein Mondregenbogen - b. *This phenomenon is itself called a moonbow. In the German naming construction (90a), a behaves like a co-argument; there‐ fore, sich cannot be used to refer to the subject das Phänomen (see Ruigendijk & Schumacher 2020). 5.2 The naming predicate nennen The German verb nennen, which can be paraphrased by “to mention by name”, and “give a name” has a rich etymological history dating back to the Old High German period. The term originated as nemnen in the 8th century, undergoing assimilation to nemmen around 800 AD, and eventually evolving into nennen in the 9th century (Pfeifer et al. 1993). The Middle High German forms include nemmen and nennen, signifying actions such as “giving a name”, “calling by name”, “establishing” and “determining”. An alternate form of the verb, benuomen in Middle High German, conveys the sense of “making something known by name”. In Middle Dutch, Middle Low German, and Old Frisian, variants like nȫmen, noemen and nōmia are evident, showcasing regional linguistic differences within the Germanic language family. These forms emphasize the act of naming or making something known, reflecting the fundamental role of language in human communication. In the West Germanic linguistic tradition, the verb is reflected in various forms across different languages: ● Old Saxon: nemnian ● Middle Low German: nennen ● Old English: nemnan ● Old Norse: nefna ● Swedish: nämna ● Gothic: namnjan (derived from the Germanic root *namnjan) 102 5 Characteristics of metalinguistic predicates <?page no="103"?> 33 All sample sentences can be found here: www.grammis.ids-mannheim.de/ verbvalenz/ 400802 The term nennen is a derivative of the noun Name, which refers to a name or designation. This linguistic evolution exemplifies the dynamic nature of language, with terms adapting and transforming over time to meet the communicative needs of speakers. The term nennen in German encompasses several nuanced meanings related to giving linguistic designations or mentioning someone or something: 1. giving names or designation 2. mentioning or listing 3. referring to or identifying The spectrum of naming that is in focus for this investigation can be grasped by the meaning of “giving something a name”, “giving the name of something after something”, and “attributing something with such a term”. Seven different readings can be distinguished (Schneider & Lang 2022). 33 a) etwas zur Kentnis geben Der verhaftete Einbrecher nannte der Polizei im Verhör seine Mittäter. ‘The arrested robber named his accomplices to the police during the interroga‐ tion.’ b) etwas einen solchen Namen geben Die Eltern nannten ihren jüngsten Sohn Hans Christian. ‘The parents called her youngest son Hans Christian.’ c) etwas als ein solches charakterisieren Sie nannte ihn einen Dummkopf. ‘She called him an idiot.’ d) jemanden benennen Der Verteidiger des Angeklagten nannte mehrere Zeugen. ‘The defendant of the accused person named several witnesses.’ e) nach etwas den Namen von etwas geben Die Stadt hat die Straßen in diesem Stadtviertel nach Bäumen und Blumen genannt. ‘The city named its streets in this neighbourhood and after trees and flowers.’ f) etwas mit einem solchen Begriff belegen Diese neue konservative philosophische Richtung nennt man Neokonservati‐ vismus. ‘This new conservative philosophical direction is called neo-conservativism.’ g) jemanden als einen solchen aufführen Als Grund für die Verzögerung wird der anhaltende Frost genannt. ‘Persitent frost is stated as the reason.’ 5.2 The naming predicate nennen 103 <?page no="104"?> 34 Please note that the numbers presented here are raw numbers and that the relevant data for solely name-mentioning uses of the verb is expected to be considerably smaller due to the word’s polysemy. 35 Frequency classes are ranging from 1 to 24 with 1 representing the most frequently used word(s). 36 The German News Corpora is based on texts from the year 2021 with 33.323.616 sentences in total. 37 As the Wortschatz Leipzig notes, so genannt as written with a space in between is the new spelling replacing sogenannt. In contrast, Duden (2022) lists so gennant as an alternative way of writing hence preferring sogenannt. In the following, I will stick to the one-word expression for consistency reasons. Not all readings give rise to a naming interpretation. For example, (b) makes use of a proper name, and (c) leads to a non-generic reading of the sentence. However, (f) clearly constitutes a name-informing reading. The lemma nennen has 4,102,809 hits in the German Web 2020 (deTenTen20) corpus and 195.38 hits per million tokens based on texts from 2021 with 33.323.616 sentences. 34 The German News Corpus (Deutsches Nachrichten-Kor‐ pus) reports nennen to belong to the frequency class 10. 35 In contrast to other naming predicates like bezeichnen als, nennen is special as it allows for a reflexive form: sich nennen (‘call.refl’) as in the following sentences. (91) a. Dieses Produkt nennt sich ein Feuerlöscher. ‘This product is called a fire extinguisher.’ - b. Dieses Küchengerät nennt sich ein Mixer. ‘This kitchen aid is called a mixer.’ Additionally, the German language makes use of the participle form sogenannt (‘so-called’). In fact, the abbreviated verb form sog. is found more frequently in written text than the non-shortened verb sogenannt. This is revealed by a corpus investigation with the Wortschatz Leipzig Korpus (Deutsches Nach‐ richten-Korpus based on texts from 2021 with 33.323.616 sentences) 36 from newspaper texts proving a frequency class of 17 for sogenannt but a 14 for the abbreviation sog.  37 Interestingly, the German short form sog. can serve as a fully functioning name-informing construction even when accompanied by brackets. For example, consider the following post on whirlpools: Strudel: An vermeintlich sicheren Badestellen zwischen kleinen Wellen aus Stein (sog. Buhnen) können sich Wirbel und Strudel bilden die Schwimmende nach unten oder zur Flussmitte können. (Hessenschau Instagram Post, July 24, 2022) 104 5 Characteristics of metalinguistic predicates <?page no="105"?> 38 Retrieved May 8, 2022 from www.etymonline.com/ word/ refer 39 Retrieved May 8, 2022 from www.etymonline.com/ search? q=as&ref=searchbar_search hint Strudel: At supposedly safe bathing sites between small waves of stone (so-called bumps) can form whirlpools and strudel swimming down or to the middle of the river. This short text first introduces a phrasal description of the term that is to be explained, followed by the term itself. 5.3 The naming predicate refer In the Oxford English Dictionary, the combination refer to is listed as an intransitive verb and defined as: “Of a word, symbol, etc.: to denote, designate; (of a speaker, writer, etc.) to denote or designate something by means of a word, symbol etc. (frequently with as)”. The verb refer has multiple origins, partly borrowed from French and from Latin. The predicate refer to as is a phrasal verb which etymologically traces back to the late 14 th century word referren and has the following denotation: “to trace back (a quality, etc., to a first cause or origin), attribute, assign” 38 . The Old French term referrer dates back to the 14 th century and was directly borrowed from Latin referre which literally translates to “to carry back”: “from re- ‘back’ (see re-) + ferre ‘to carry, bear’” (OED 2022). The transitive use of “refer to” involves committing something to an authority for consideration, and its sense of directing someone to a source for information dates back to the 17 th century. In naming constructions, we typically find the predicate refer combined with the prepositions to and as. In the Etymology dictionary, as is classified as an adverb, conjunction, or pronoun, and is described as dating back to c. 1200: representing a worn-down form of Old English alswa meaning “quite so”, “wholly so” or “all so”. This term is related to German als, suggesting a linguistic connection between the two languages. It is noteworthy that the German word als traces its roots back to also, emphasizing the commonality in their origins. 39 The by far most commonly used modifier of call is also with 481,821 occurrences in the enTenTen20 corpus as in sentences like It is also called broken heart syndrome. The difference between the predicates call and refer to may be grounded in their etymologies. Understanding the influence of as in refer to as is crucial, considering its role as an adverb, conjunction, or pronoun that has evolved over time. The grammatical considerations of refer to as a phrasal verb involve recognizing it as a transitive verb, just like call. Phrasal verbs consist of a main 5.3 The naming predicate refer 105 <?page no="106"?> 40 Retrieved May 8, 2022 from www.etymonline.com/ search? q=as&ref=searchbar_search hint Retrieved May 8, 2022 from www.dictionary.cambridge.org/ dictionary/ english-germa n/ refer? q=refer+ verb and a preposition adverb (or particle). It is one semantic unit. The phrase refer to as successfully passes the test of object movement for phrasal verbs by Fraser & Bolinger (1971). The particle can be placed before or after the “direct object of transitive phrasal verbs” (Darwin & Gray 1999: 72), as demonstrated by examples such as He referred to this concept as “inclusion” or The concept was referred to as “inclusion”. Beyond its linguistic aspects, the term refer is also linked to other words in English, such as referee originating in the early 17th century, and referendum from the mid-19th century. Both derive from the Latin referre, emphasizing the historical depth and interconnectedness of linguistic evolution. In the Cambridge Dictionary, four different types of reading can be differen‐ tiated. 40 a) to talk or write (about something); to mention I referred to your theories in my last book. b) To relate to, concern, or apply to My remarks refer to your last letter. c) to send or pass on to someone else for discussion, information, a decision etc The case was referred to a higher law-court d) to look for information (in something) If I’m not sure how to spell a word, I refer to a dictionary In none of the above-mentioned readings, a name-informing one is explicitly stated out. A more precise definition is given in (b) by the Merriam Webster Dictionary. a) to look at or in (something) for information She often refers to her notes when giving a speech. Please refer to our website for more information. b) to have a direct connection or relationship to (something) The word “finch” refers to a kind of bird. The numbers shown in the text refer to footnotes at the bottom of the page. c) to talk about or write about (someone or something) especially briefly: to mention (someone or something) in speech or in writing No one referred to the incident. 106 5 Characteristics of metalinguistic predicates <?page no="107"?> 41 Retrieved May 10, 2022 from www.merriam-webster.com/ dictionary/ refer%20to 42 Retrieved May 10, 2022 from www.duden.de/ rechtschreibung/ bezeichnen 43 Retrieved May 10, 2022 from www.dwds.de/ wb/ etymwb/ bezeichnen 44 Retrieved May 10, 2022 from www.grammis.ids-mannheim.de/ verbs/ view/ 401503/ 1 45 Retrieved May 12, 2022 from www.grammis.ids-mannheim.de/ verbs/ view/ 401503/ 2 The name-informing use of refer to as is reflected in the meaning to call (something or someone) by a specified name. 41 What differentiates refer to as from the other predicates is that it is commonly analysed as a phrasal verb, or as an inherent preposition verb (Dixon 1992). 5.4 The naming predicate bezeichnen When looking at the etymology, the German predicate bezeichnen is a transitive verb that originates from the Middle High German term bezeichenen. In Old High German, bizeihhanen or bizeihhanōn was used in the 9 th -century as a term for a visual representation and also as a word for “meaning”. 42 It was further used as a synonym for “characterise” and “call”. 43 The word as functions as a conjunction to equate two terms or to compare their differences. In Middle High German, alse and als were used for comparison meaning including “because of ” and “as if ”. In New High German, als was reduced to one syllable form of the Middle High German word also. A corpus search in SketchEngine revealed that bezeichnen has a total of 2,368,814 hits and 112.8 hits per million token in the German Web 2020 corpus which has approx. half as many occurrences as the predicate nennen. In other words, nennen is generally used way more frequently in the German language as compared to bezeichnen. However, the verbs do not necessarily have to be used as a name-informing predicate, so the valid numbers of the corpus data might depict a different amount NMC conditions (see Chapter 6). The different readings of bezeichnen als can be clustered into four semantic categories. a) etwas bedeuten Kilokalorien (kcal), im allgemeinen Sprachgebrauch meist Kalorien genannt, bezeichnen die Energiewerte von Nahrungsmitteln. (Mannheimer Morgen, 11.01.2010, S.-13) 44 ‘Kilocalories (kcal), commonly referred to as calories, refer to the energy values of foods.’ b) jemanden/ etwas mit einem Namen versehen Mit dem Fachbegriff Adipositas bezeichnen Ärzte ein krankhaftes Übergewicht. (Mannheimer Morgen, 06.07.2010, S.-17) 45 ‘With the specialist term obesity, doctors refer to a pathological overweight.’ 5.4 The naming predicate bezeichnen 107 <?page no="108"?> 46 Retrieved May 12, 2022 from www.grammis.ids-mannheim.de/ verbs/ view/ 401503/ 3 47 Retrieved May 12, 2022 from www.grammis.ids-mannheim.de/ verbs/ view/ 401503/ 4 48 Sample (a) is a counter-example, at least when considering solely the matrix clause. c) jemanden/ etwas als einen solchen charakterisieren In einem Schreiben zur Aufstellung des nächsten Bundesetats und Finanzplans stimmte er das Kabinett auf einen strikten Sparkurs ein und bezeichnete das Jahr 2011 einen „Wendepunkt“. (Mannheimer Morgen, 14.01.2010) 46 ‘In a letter regarding the nomination of the next federal government and financial plan, he got the cabinet in the mood for a strict austerity course and described the year 2011 as a “turning point”.’ d) etwas markieren Die zu fällenden Bäume sind vom zuständigen Förster bezeichnet worden. (St. Galler Tagblatt, 16.12.1998) 47 ‘The trees that need to be cut down have been designated by the competent cultivator.’ As the sample sentences illustrate, the different readings of bezeichnen als are most commonly of a naming nature. 48 The German News Corpus reports bezeichnen als (‘refer to as’) to belong to the frequency class 11 and is therefore almost as prominently used as the German naming predicate nennen. 5.5 Additional naming predicates and their collocates Apart from frequently occurring predicates call and refer to as, both German and English make use of comparable verbs that function as name-selecting predicates. For English, further predicates include to speak of, describe as, and name. The spectrum of different metalinguistic verbs employed in English NMC is illustrated in (92a-h). (92) a. This area is known as a “magnetic field”. - b. This special kind of moon is nicknamed the “snow moon”. - c. People are speaking of this phenomenon as a “heatwave”. - d. People commonly name the source of a river “headwaters”. - e. Researchers describe this effect as the “cornea reflection”. - f. The expert labelled this issue as a “critical concern”. - g. This phenomenon has been introduced as the “eye-mind assumption” in the literature. 108 5 Characteristics of metalinguistic predicates <?page no="109"?> h. This phenomenon is termed “fast-mapping”. In all of these sample sentences, the verb functions as a name-selecting predicate and provides the name of a lexicalized concept. Similar sentential constructions can be observed in German as well, in which the predicates function as metalinguistic naming predicates. (93) a. Einen Wasserfall unter Wasser betiteln Experten auch als „Sandfall“. ‘A waterfall under water is referred to by experts as a “sandfall”.’ - b. Manche benennen dieses Phänomen als „Sonnenwind“. ‘Some call this phenomenon “solar wind”.’ - c. Bei diesen gesundheitlichen Langzeitfolgen spricht man von Long Covid. ‘These long-term health effects are referred to as long Covid.’ - d. Mein Benutzername lautet Max. ‘My username is Max.’ - e. Die Firma lautet auf den Namen „Apple“. ‘The company is named by the name of “Apple”’. The metalinguistic predicate lauten is frequently found in combination with a proper name. Interestingly, the frequently employed predicate nennen is exceptional as it allows for the wie-question but does not allow for the test with wen oder was (‘who or what’) to ask for the name of the theme. The how-question in (94b) is related to a modality, i.e., targeting e.g. at the way in which the student learned the vocabulary. (94) a. Linguisten nennen diese Konstruktion eine Nominalphrase. ‘Linguists call this construction a noun phrase.’ - b. Wie nennen Linguisten diese Konstruktion? ‘How do linguistic call this construction? ’ The same applies to a majority of the additional metalinguistic predicates presented above. (95) a. This special kind of moon is nicknamed the “snow moon”. How is it nicknamed? - b. People are speaking of this phenomenon as a “heatwave”. How are people speaking of this phenomenon? 5.5 Additional naming predicates and their collocates 109 <?page no="110"?> 49 These categories represent grammatical relations. 50 Please note that most sample sentences presented here focus on nennen in passive voice. c. People commonly name the source of a river “headwaters”. How is it commonly named? - d. Researchers describe this effect as the “cornea reflection”. How is it described? - e. The expert labelled this issue as a “critical concern”. How is it labelled? - f. This phenomenon has been introduced as the “eye-mind assumption” in the literature. How has it been introduced? - g. This phenomenon is termed “fast-mapping”. How is it termed? While both tests are possible in English (see 95a-g), the exceptional behavior of nennen can also be found in other metalinguistic naming-predicates, including betiteln (‘entitle sth. as’), sprechen von (‘speak of ’), and benennen (‘name’). All of the above-mentioned predicates interact with a comparable set of collocations. A search with the word sketch tool in Sketch Engine, a feature that allows for identifying collocations either depending on frequency or their degree of probability, revealed that the verb call is most frequently found in a category 49 named “modifiers of call” that includes “also called” (481,821 occurences), “the so called” (228,818 occurences) as well as “is often called” (117,643 occurences) and “commonly called” (59,519 occurences) in the Ententen20 corpus (see Kilgarriff et al. 2010 for further information on the word sketch function). All of the collocation types can possibly function as a valid NMC. Especially when looking at modifiers, predicates in German and English include e.g. commonly, often, frequently, usually, so, and always. In addition, name-selecting predicates cannot only be found in German and English but also cross-linguistically for instance in French (e.g. appeler), Spanish (conocer; llamarse), Italien (nominare; chiamare), and Latin (nominare). The last sections discussed naming predicates on the languages German and English and also considered additional naming predicates and their collocations to provide a comprehensive overview. 50 What is remarkable here is that met‐ alinguistic predicates studied in this section belong to the central language inventory of the languages under investigation. Despite this, the precise nature of NMCs and their linguistic characteristics in German and English remains unclear. This encompasses various linguistic prop‐ 110 5 Characteristics of metalinguistic predicates <?page no="111"?> erties of the mentioned nominal, including their word type, lexical frequency, and morphological type. Additionally, the presence or absence of quotation marks in NMCs remains a matter of ongoing inquiry, as does the potential interaction between quotation marks and other factors such as word type. In order to further investigate these research questions empirically, the following chapter introduces a corpus analysis aiming at answering the RQs 1-3 (see Chapter 3.5) for NMCs in German and English. 5.5 Additional naming predicates and their collocates 111 <?page no="113"?> 51 NIC is a name-informing construction which does not include quotes as graphemic markers. NIQs, i.e., name-informing quotations, in contrast employ quotes in the written form. 52 The frequency refers to the whole word and not to single constituents 53 Please note that H B and H E do not contradict but complement each other. Multi-word expressions are in the following understood as quoted items that are written with a space in between, e.g. Satire Gipfel (lit. satire conference). These cases are particularly interesting because one of the elements may be enclosed in quotation marks while the other one is not, as reflected in Oberton-“Klangteppich”(lit. overtone-“sound carpet”). 6 Empirical investigation I: Corpus-based construction analysis 6.1 Research questions and hypotheses Name-mentioning predicates are expected to cluster around construction types that can be classified in semantic groups. This means that metalinguistic predicates like call tend to occur in specific grammatical constructions and with a specific set of collocations (see also chapter 5.5) while fulfilling the function of introducing a name. In this study, a representative set of construction types used for name-mentioning will be extracted in a large-scale corpus study and analysed with respect to the factors morphological type of the name, its lexical frequency as well as the occurrence of quotes. In addition, these constructions will be rated concerning their interpretation (name-informing vs. modalizing). The focus is on nominal mentioned expressions and their linguistic properties in the languages German and English. Furthermore, the type of quotation involved in NMCs is under investigation and the following questions will be addressed: What is the systematicity in the use of quotes and is their use affected by the linguistic properties of mentioned expressions? The hypotheses, on which the corpus-based construction analysis is grounded, are formulated below. H A : The valid constructions consist of more NIC than NIQs. 51 H B : More compounds in valid constructions as opposed to simple nouns. H C : Quotation marks appear more often with less frequently used nominals. H D : Valid NMCs prefer a low lexical frequency, i.e., a high frequency class. 52 H E : One-word expressions are preferred over multi-word expressions. 53 <?page no="114"?> 54 Note that written texts, especially when considering newspaper articles, are often edited before being published which may influence the graphemic form of the text and hence the use of quotation marks. 55 https: / / app.sketchengine.eu/ #dashboard? corpname=preloaded%2Fbnc2_tt21 The first hypothesis H A is grounded in the argumentation that valid sentences should show a preference for the nominal to not be accompanied by quotation marks, as supported for example by evidence from Härtl (2020) and Gutzmann & Stei (2011). H B posits a preference for compounds, reinforcing the theoretical framework established in Section 4.2. H C assumes that quotation marks interact with lexical frequently so that quotes are more prevalent with low-frequency nominals while H D hypothesizes that names have a preference for a low lexical frequency, connecting to the argumentation as outlined in Härtl (2018). Last, H E assumes single-word expressions to be favored in valid constructions, which will give further insights into the morpho-syntactic properties of the quoted items. Generally speaking, based on theoretical sketches of NMCs and the assumption that all four predicates function as naming predicates, I assume that the naming predicates trigger a comparable behaviour in both German and English. The findings of the large-scale investigation will be used as a foundation to formulate hypotheses for further empirical investigations. In addition, the aim is to put forward evidence of the linguistic conditions effective in the composition of NMCs available in German and English. 6.2 Methodology The corpus-based construction analysis aims at investigating name-mentioning constructions in German and English. To obtain a representative set of types of NMCs available in German, an investigative search was conducted using the TAGGED-T archive of the DeReKo main archive (IDS Mannheim) and COSMAS II. The archive contains 1.5-billion-word tokens spread over 26 corpora, mostly from newspapers, press agencies, and internet sources. For the analysis of NMCs in English, I used Sketch Engine and the 100-million-word British National Corpus BYU-BNC (Davies 2015: 11). The BNC is a balanced English corpus of contem‐ porary written and spoken language 54 containing 112 million tokens that is processed by TreeTagger pipeline v2.1 and contains a wide range of genres. 55 The large-scale corpus study was centred around the two English name-mentioning predicates call and refer to as well as their German counterparts nennen and bezeichnen. A concordance analysis supported the underlying assumption that call is the most frequently occurring verb by accounting for approx. 0.045% 114 6 Empirical investigation I: Corpus-based construction analysis <?page no="115"?> 56 Please note that these observations of course do not tell anything about the validity of constructions as NICs. 57 The result details for the CQL (Corpus Query language) [lempos_lc="call-v"] that was used in the concordance search in SketchEngine displayed the percentage of the whole corpus with 0.04571% and 0.01215% of the whole corpus for refer CQL [lempos_lc="refer-v"]. The size of the BNC corpus amounts to 112,345,722 tokens and the search for naming verbs can herewith be considered as a mid-frequency phenomenon (see SketchEngine www.lms.sketchengine.eu/ mod/ quiz/ view.php? id=33 0). 58 Please note that the more specified CQL [lempos_lc="call-v"] and [lempos_lc="refer-v"] did not yield a difference in the result details compared to the query reported in Table 4. in the BNC corpus followed by refer (~0.012%) and name (~0.005%). For this reason, the analysis focused only on the first two naming verbs and their German counterparts. However, I would like to acknowledge that a search in the Ententen20 corpus, which is larger in size, revealed that refer and name are almost equally represented in the corpus (~0.012%) whereas call displays the smallest number (~0.009%). 56 Because the investigation is focussing on verbs only, the query for the corpus analysis in English included the part-of-speech role of the word as a verb. 57 Table 4 gives an overview of the queries and the overall hits per corpus. Naming verb Query Concordance size Corpus call CQL [lemma="call" & tag="V.*"] 58 51,351 SketchEngine: Brit‐ ish National Corpus nennen &nennen 208,015 COSMAS II: DeReKo refer to CQL [lemma="refer" & tag= "V.*"] 13,648 SketchEngine: Brit‐ ish National Corpus bezeichnen &bezeichnen 132,503 COSMAS II: DeReKo Tab. 4: Overview of hits per condition Importantly, the queries did not contain any punctuation marks or any addi‐ tional tags that could have restricted the search. The concordance results return the predicate as the KWIC embedded in a left and right context of at least one context sentence. Due to the large output size (see Table 4), the analysis was limited to N=8,000 hits, i.e. 2,000 hits per query. 6.2 Methodology 115 <?page no="116"?> 59 The analysis allowed for all conjugations of the respective verbs. 60 Proper names, abbreviations, foreign words or non-sensical words have been labelled as invalid constructions. 61 The morpho-syntactic type of the name of the name in German sogenannt sentences has been in focus in a corpus study Härtl (2016) and has also been discussed in Härtl (2018) and is not further considered in this analysis. 62 For consistency, variables are written in small caps and their respective levels in italicized small caps. 63 As argued in Chapter 2, quotation marks can sometimes be replaced by capitalization. For this corpus analysis, a syntactic definition of quotation has been adopted. 64 For German: https: / / corpora.uni-leipzig.de/ de? corpusId=deu_news_2020&word=effici ency For English: https: / / corpora.uni-leipzig.de/ de? corpusId=eng_news_2020&word=efficie ncy 6.3 Analysis The constructions extracted from the corpora were manually annotated with regards to fulfilling the criteria as a name-informing construction. Valid NMC constructions were defined as syntactical structures that involved (1) a naming predicate like call, refer, nennen and bezeichnen  59 ; (2) nominal tokens as the argument of the respective verb 60 ; and (3) constructions in which the postcopular nominal refers generically. Regarding the predicated call it is important to mention that the name-mentioning modifier so-called (‘sogenannt’), which represents a participial modifier, was regarded as an invalid NMC as the focus was exclusively on verbs. 61 As a first step, the constructions under investigation were analysed with respect to the nominal being embraced by Q U O T E S or N O Q U O T E S , with the variable labelled as Q U O TATI O N M A R K S62 . All constructions in both corpora dis‐ played solely double quotes both for German and English. Nominals written in bold letters, which were marginal appearances and were only displayed in the English constructions, have not been labelled as proper quotation marks. 63 Please note that other typographical means of quotes such as italics, underlining and bold print which have been discussed in the literature (cf. Davidson 1979, Klockow 1980, Quaßdorf 2016, Cappelen et al. 2023), would not have been represented in the corpus results. The second criteria of the analysis was the N AM E T Y P E : the nominals in focus were divided into the levels S IM P L E N O U N and C O M P O U N D S . Moreover, the W O R D S TAT U S was split into the categories O N E -W O R D and M U L T I -W O R D . In order to investigate the frequency, the L E XI C AL F R E Q U E N C Y for each quoted nominal was calculated using the Wortschatz of the Universität Leipzig. 64 The frequency values are calculated by comparing the value to the most frequent word in German (the article der), displaying how many times 116 6 Empirical investigation I: Corpus-based construction analysis <?page no="117"?> 65 Relative frequency describes the number of occurrences of an item per million tokens (Sketch Engine) whereas the absolute frequency refers to the actual number of occur‐ rences or hits. 66 To illustrate, the compound fire station has 178 hits in the BNC and the formula would for K W is log 2 (5415707 / 178] = 14.89 which was rounded up to a frequency class of 15. 67 For the frequency classes in the Leipzig corpora, Wortschatz Uni Leipzig uses the following explanation for their statistical calculation: “Frequeny class: Words of similar frequency are grouped into classes with the goal that the frequency class of a word does rarely change between different corpora. The frequency of the most frequent word of a corpus is divided by the frequency of the word in question and the logarithm to the basis 2 of the result is rounded up to the next whole number. The most frequent word in a corpus has always frequency class 0; a word in frequency class 1 is around half as often found in the corpus as the most frequent word. In general, a word of frequency class n+1 has half the frequency of a word in frequency class n. Extremely rare words may have a frequency class of 20 or higher in large corpora” (Retrieved July 26, 2023 from www.wortschatz.uni-leipzig.de/ de/ documentation/ faq) 68 To illustrate the enormous difference, consider the following example: 410 occurrences for the word squeeze as a noun as compared to 2,297 hits without any further definition of the word the value is higher. Sketch Engine and the BNC were used to determine the word’s frequency for the analysis in English. As Perkuhn et al. (2012) note, frequency classes are mostly used in German linguistics. For this reason, the statistics used in Sketch Engine are limited to displaying the absolute and relative frequency. 65 To calculate the respective frequency classes based on the word’s relative frequency, the following formula was used: K W = [log 2 (f a (R) / f a (W))]. In this formula, f a (W) stands for the absolute frequency of the word W in the corpus under investigation and f a (R) the absolute frequency of the most frequently used word in the corpus (Perkuhn et al. 2012: 80). For English, the most frequently used word is the with an absolute frequency of 5,415,707 (frequency class=0) and the lemma der in German with the absolute frequency of 11,113,606. 66 Frequency classes have been chosen as they have advantages compared to a token’s absolute frequency: enlarging corpora does not have a huge impact on frequency classes whereas the absolute frequency would vary when adding data to a corpus (Quasthoff 1998); regarding limitations and problems arising with frequency (classes) see Haderlein (2008). 67 As “frequency norms are more informative if they include information about the part-of-speech-roles of the words” (Brysbaert et al. 2018a: 230), and the analysis was only aiming at nomi‐ nals, I used a concordance function that allowed for defining the query type as a ‘word’ and pick the part-of-speech ‘noun’. To illustrate the difference, consider the following sample sentence form the BNC that would have been assigned as a verb regarding its part-of-speech classification: The slit was euphemistically called a squeeze. 68 The variable L E XI C AL F R E Q U E N C Y contains ordinal data ranging 6.3 Analysis 117 <?page no="118"?> 69 Please note that the frequency classes for some corpora might only range between 0-23 as the highest frequency class depends on the size of the corpus. It is debatable where exactly to draw the boundary between a high and a low frequency class. Compare, for example, the SUBTLEX-US database which provides data using the Zipf-value based on calculations using a logarithmic function: “The value ranges from 1 to 7, with low-frequency words covering the range of 1-3, and high-frequency words covering the range of 4-7.” (Brysbaert et al. 2018a: 234). Brysbaert et al. (2018) claim that using the frequency per million words count would be a good standardized measure for word frequency which defined low-frequency words as >5 fpm and high-frequency words with <100fpm. However, please also note that categorizing words as frequent and infrequent using e.g. the frequency classes 0-3 for frequently used words and the classes 4-24 for infrequent words is not applicable to NMC constructions neither in German nor in English as this work focusses solely on nominals, see also a list with the most frequent words for English (www.cls.corpora.uni-leipzig.de/ de/ eng_news_2020/ 3.2.1_ The%20Most%20Frequent%2050%20Words.html) and German (www.cls.corpora.uni-lei pzig.de/ de/ deu_news_2020_1M/ 3.2.1_The%20Most%20Frequent%2050%20Words.html# collapse438). from number 0 to 24. In addition, a further category named F R E Q U E N C Y C LA S S clustered the lexical frequency into the layers H I G H (frequency class from 0-10) and L O W (frequency classes from 11-24). 69 6.4 Results Before presenting the results of the corpus analysis, sample sentences for each of the corpus queries will be illustrated. These stereotypical sentences aim at giving insight into the set of sentences found in the corpus data. All sample sentences listed below have been labelled as valid NMC instances. Properties Concordance Corpus call It does this by demonstrating that once prag‐ matic implications of the sort we shall call im‐ plicature are taken into account, the apparently radical differences between logic and natural language seem to […] SketchEngine: British National Corpus annotation Nominal: Implicature Quotes: No NameType: Simple noun Frequency: Low FrequencyClass: 17 nennen Der Name Dart kommt aus dem Französischen. Die Franzosen setzten bei ihren Schlachten kleine speerähnliche Wurfpfeile ein, die sie COSMAS II: DeReKo 118 6 Empirical investigation I: Corpus-based construction analysis <?page no="119"?> 70 Please note that the frequency classes depend on the corpus size (cf. Perkuhn 2012). In this investigation, the word refer is the predicate with the lowest frequency class ranging to 22. Dart nannten. Die ersten Dart-Wettkämpfe fan‐ den Anfang des 20.-Jahrhunderts in Grossbri‐ tannien statt. annotation Nominal: Dart Quotes: No NameType: Simple noun Frequency: Low FrequencyClass: 16 refer This pernicious combination of rising inflation and rising unemployment, usually referred to as “stagflation”, not only led economists to abandon the simple Phillips curve as a tool for modelling […] SketchEngine: British National Corpus annotation Nominal: Stagflation Quotes: Yes NameType: Compound Frequency: Low FrequencyClass: 17 bezeichnen Ein Illustrator betreibt eine der angewandten Künste. Er zeichnet, malt, aquarelliert und fer‐ tigt Drucke an, um ein Buch, eine Zeitschrift oder andere Schriftstücke zu bebildern. Als Buchillustration bezeichnet man Bilderfolgen, die der Ergänzung und der Erhellung eines Buchtextes dienen. COSMAS II: DeReKo annotation Nominal: Buchillustration Quotes: No NameType: Compound Frequency: low FrequencyClass: 22 Tab. 5: Sample sentences from the corpus data The next question to consider is the following: What are the main properties of the NMC constructions? The corpus-based construction analysis focussed on four different name-mentioning constructions that are defined as factors in the analysis. Table 6 illustrates the total numbers for all valid hits, including their appearance, i.e., with or without quotes, the mean frequency 70 as well as the percentage of valid NMCs in relation to the 2,000 analysed hits. For the 6.4 Results 119 <?page no="120"?> statistical analysis, the data was analysed with the R statistics software (R Core Team 2022). In all analyses, the .05 level of significance was adopted. predicates valid hits % of NMCs 1 N O Q U O T E S Q U O T E S mean frequency call 340 17% 264 76 16 (sd=4.73) refer 103 5.15% 43 59 15 (sd=4.34) nennen 373 18.65% 209 164 19 (sd=4.11) bezeichnen 485 24.25% 310 175 16 (sd=4.47) Tab. 6: Overview of the hits with Q U O T A T I O N M A R K S and L E X I C A L F R E Q U E N C Y | Note: Percentage of valid NMCs in relation to all hits. Fig. 1: Bar plot of valid instances for Q U O T E S and N O Q U O T E S by predicate To investigate the difference in Q U O T E S vs. N O Q U O T E S (H A ) per factor, Chi-squared tests for given probabilities were conducted in R. The critical Chi-squared value for one degree of freedom at a significance level of 0.05 is 3.841. The results prove a statistically significant effect for call (c 2 (1)=103.95, p < .005), nennen (c 2 (1)= 5.429, p = 0.01981) and bezeichnen (c 2 (1)=37.577, p < .005) but did not reach significance for refer (c 2 (1)=2.5098, p = 0.1131). These 120 6 Empirical investigation I: Corpus-based construction analysis <?page no="121"?> findings confirm hypothesis H A for the German constructions and the predicate call, revealing that the valid constructions prefer no quotation marks around the nominal. However, the alternative hypothesis needs to be adopted for constructions involving refer. The large-scale corpus analysis has proven that the graphemic marking in name-mentioning constructions is less frequent. Of further interest was the lexical frequency of the nominals (see hypothesis H C and H D .), which has been divided into low frequency classes ranging from 0 to 10 and high frequency classes from starting at frequency class 11. The exploratory analysis for the frequency classes is presented in Figure 2-5 for each predicate. Fig. 2: Bar plot of the distribution of frequency classes for nennen 6.4 Results 121 <?page no="122"?> Fig. 3: Bar plot of the distribution of frequency classes for call Fig. 4: Bar plot of the distribution of frequency classes for bezeichnen 122 6 Empirical investigation I: Corpus-based construction analysis <?page no="123"?> 71 For a better understanding, I will provide the reader with the number of quoted expres‐ sions belonging to high frequency classes: call=90.76%, nennen=95.71%, refer=85.29%, and bezeichnen 83.38%. 72 See also sample sentences taken from the corpus data presented in Appendix A1. Fig. 5: Bar plot of the distribution of frequency classes for refer The Pearson’s Chi-squared test proved a highly significant effect for frequency classes and the naming predicates [call (c 2 (1)= 171.99, p < .005), nennen (c 2 (1)=224.52, p < .005), refer (c 2 (1)= 56.627, p < .005), bezeichnen (c 2 (1)= 379.31, p < .005]. The statistics for the frequency classes indicate a clear preference for the name z to be grouped in a low frequency class. 71 The statistics herewith support the underlying assumption that a low lexical frequency is a characteristic of mentioned expressions in NMCs. In other words, less frequently used nominals have a predisposition to occur in NMCs. 72 The statistical analysis for N AM E T Y P E proved a highly significant difference for call (c 2 (1)= 20.225, p < .001), nennen (c 2 (1)=31.853, p < .001), and bezeichnen (c 2 (1)=41.037, p < .001). Constructions that involved these predicates proved a preference for compounds, supporting hypothesis H B . In contrast, refer (c 2 (1)= 4.7451, p-<-.05) proved a reversed effect with significantly more simple nouns. 6.4 Results 123 <?page no="124"?> Fig. 6: Bar plot for N A M E T Y P E by predicate As a next step, a more in-depth analysis was conducted to investigate hypothesis H A, which was used to examine individual differences between NICs and NIQs for the W O R D S TAT U S . The statistics with the exact p-values are reported in Table-7. - NMCs without quotes NMCs with Quotes - χ 2 value p-value χ 2 value p-value call 0.003663 0.9517 n.s. 1.3158 0.2513 n.s. nennen 26.914 2.127e -07 *** 7.0488 0.007932* refer 2.814 0.09345 n.s. 2.0508 0.1521 n.s. bezeichnen 35.174 3.015e -09 *** 7.7208 0.005459** Tab. 7: Chi-squared test for given probabilities (Q U O T E S & W O R D S T A T U S ) Most interestingly, there are significant differences in both constructions with and without quotes for the predicates in German, but no effect for English NMCs. As reported above, all constructions with a significant difference prefer a compound instead of simple nouns. 124 6 Empirical investigation I: Corpus-based construction analysis <?page no="125"?> Concerning the W O R D S TAT U S (H E .), the Chi-squared test revealed a signifi‐ cant effect for the predicates nennen (χ 2 (1)= 345.53, p < .001), refer (χ 2 (1)= 6.6275, p < 0.05), and bezeichnen (χ 2 (1)= 346.56, p < .001) with a preference for one-words. Despite this, there was no effect for call (χ 2 (1)=0.82808, p = 0.3628). Fig. 7: Distribution of W O R D S T A T U S by predicate In addition, possible interactions between the different factors have been investigated. First, the question whether there is an interaction between Q U O T E S and F R E Q U E N C Y has been addressed. Linear Regression Models were created with the predicate as dependent variable ad the factors Q U O T E and F R E Q U E N C Y as independent variables. The models did not contain any random effects. Pairwise comparisons were computed using emmeans (Lenth et al. 2018). For the predicate nennen, a statistically significant difference was found for N O Q U O T E S between a high and low lexical frequency (p < .0001). In conditions without quotes, there are significantly more instances with a high lexical frequency as compared to a low lexical frequency. For Q U O T E S , a statistically significant effect (p < .0001) was proven to hold between a high and low lexical frequency as well: there are significantly more instances without quotes in a low lexical frequency. Only significant interactions are reported below. 6.4 Results 125 <?page no="126"?> estimate Std. Error t value Pr(>|t|) (Intercept) 0.06934 0.04044 1.715 0.08663 QuotesYes 0.08023 0.02470 3.248 0.00119 ** Frlow 0.17519 0.04129 4.243 2.36e -05 ** Tab. 8: Pairwise comparison for F R E Q U E N C Y and Quotes contrast estimate SE df t.ratio p.value high noQuotes - low noQuotes -0.1752 0.0413 1333 -4.243 0.0001 high noQuotes - high Quotes -0.0802 0.0247 1333 -3.248 0.0065 high noQuotes - low Quotes -0.2554 0.0483 1333 -5.289 <.0001 low noQuotes - low Quotes -0.0802 0.0247 1333 -3.248 0.0065 high Quotes - low Quotes -0.1752 0.0413 1333 -4.243 0.0001 Tab. 9: Pairwise contrasts for F R E Q U E N C Y by Quotes For both conditions Quotes and noQuotes there were more instances with a low lexical frequency. When considering only a high frequency, more instances occurred in sentences without quotes. The same trend is reflected for a low frequency. Despite this, no effects occurred for the predicates bezeichnen and refer. In the following section, the empirical findings of the corpus study will be summarized and discussed with regards to theoretical approaches to naming predicates. 6.5 Discussion The main aim of this study was to characterize NMC constructions in German and English with respect to the nominal’s graphemic marking, morphological complexity, and frequency. The herewith presented large-scale corpus analysis revealed that the most valid hits were found with the predicate bezeichnen, followed by nennen and call. Contrary to the expectations, there was a wide 126 6 Empirical investigation I: Corpus-based construction analysis <?page no="127"?> 73 Note, however, that Cappelen & Lepore (1997) would strongly oppose this view and argue that written quotations require graphemic marking. range in the percentage of valid constructions, showing that refer to as displayed the lowest number of valid NMCs. From a cross-linguistic perspective, this may indicate that the German equivalent bezeichnen is a more prominent sentential construction in discourse to inform a speaker about a name. When taking into consideration corpus data from the Enhanced Cues Corpus which investigated metalanguage, Wilson (2012) revealed that “80% of appearances of the verb call preceded a candidate instance that was labeled as mentioned language” (6). Therefore, it is no surprise that call is more frequently used in valid NMCs as opposed the refer. Regarding the frequency of the occurrence of NMCs in the corpus data, NMC constructions can be considered as a well-established and frequently occurring linguistic phenomenon in both languages. Regarding hypothesis H A and H B , the statistical analysis confirmed the assumption for all of the predicates except for refer. One aim of the herewith presented study was to answer if there is a systematicity in the use of quotes. An investigation into the valid name-mentioning constructions has shown that the graphemic marking in name-mentioning constructions is less frequent. This finding is in line with the first corpus study reported in Härtl (2020) for the verbs nennen and bezeichnen, in which the total amount of NIQs was lower than the occurrences of NICs. Another corpus study by Gutzmann & Stei (2011) has looked into newspaper texts and reported that the percentage of mentioned expressions marked by quotation marks is higher than those without quotation marks but displayed only 59.02% for Begriff ‘expression’ and 62.73% for constructions with Wort ‘word’. As Gutzmann & Stei (2011) state, the use of quotation marks is not compulsory in metalinguistic contexts. 73 In following the approach to quotation by Washington (1992) it can be argued that quotes are not an essential part of the construction but that they are optional. Due to the contextual embedding, the quotational context becomes obvious (cf. Cappelen & Lepore 1999), as the context gives rise to a mentioning reading of the respective word, leading speakers to omit quotes. The question of the necessity of quotation marks for quotation has also been discussed by Saka (1998), who claims: Quote marks are often omitted in writing as well (contra Reimer 1996): it is downright normal, outside of scholarly writing, to exclude quote marks, especially in construc‐ tions like ‘The word cats is a noun’; and even in logic publications, where one might expect the greatest exactitude, it is common for quote marks to be omitted. (118) 6.5 Discussion 127 <?page no="128"?> 74 Bold print as used in the original. Washington (1992) further adds that quotes are oftentimes omitted in oral communication by referring to the formulation “quotes-unquote”, intonational cues, and also gestural air quotes. Cappelen & Lepore (2007) devote a whole chapter in their work “Language turned on itself ” to the omission of quotation marks. They ask two central questions: 1) Are sentences without quotation marks grammatical? 2) Is the semantic value the same with and without quotation marks? In their argumentation, the answer to both questions is ‘no’. Reimer (1996) claims that any theory of quotation claiming quotes to be a defining feature of quotation to be mistaken: quotation does not require quotation marks. Quotes are, according to De Brabanter (2023), relying on contextual factors to make sure the recipient recognizes the unmarked quotation. These contextual factors might be linguistic as metalinguistic predicates (as in Boston is disyllabic). De Brabanter (2023) introduces the following examples involving the predicative complement of call to bolster his argumentation that neither the absence nor the presence of quotation marks would make a semantic difference: (96) a. She called him “Donald” from the start. He asked if he could call her “Secretary Clinton”. - b. I instinctively started calling him Uncle, as is the custom for Asian elders. He started calling me “beta”, or son, (…). - - (De Brabanter 2023: 7) 74 He further notes that the author of the second example must have consciously decided to mark only one of the quotational occurrences. In fact, philosophical approaches to quotation might need to be updated - meaning that some philosophical assumptions base their line of argumentation on a non-empirically based approach. For example, Bazzoni (2016: 123) claims that there would be no way of distinguishing between the two following sentences in spoken English: (97) a. The Evening Star has three words. - b. “The Evening Star” has three words. To take this argumentation one step further, this would support a pragmatically informed theory for NMCs. 128 6 Empirical investigation I: Corpus-based construction analysis <?page no="129"?> Depending on the source, there might be style guides that use quotation marks as convention. For example, in the Wikipedia style manual, mentioned language should be either presented in italics, bold letters or quote marks and style guides for other journals might also require similar writing practices. Yet, especially with online resources, manuals might not be followed strictly by the writers or simply not read before writing a text (see also Wilson 2012). This explains the wide variety in the use of quotation marks in the corpus data as this search was not restricted to a specific source in order to have a wider spectrum of natural language in the data. Despite this, it needs to be noted that a low percentage of quotes in the corpus data does not necessarily indicate that quotation marks are redundant or optional. In fact, the results indeed reveal liberated writing practices - probably even more than for other punctuation marks like commas. The conclusion to be drawn here is that quotes in NMCs are often left out (see also Härtl 2018). Concerning the type of the name, the vast amount of compound nouns in the corpus data supports the underlying assumption that compounds are inclined to fulfil a naming function, an idea that is widespread in the literature on the word-formation phenomenon compounding (see section 4.2; e.g. Dirven & Verspoor 1998; Zimmer 1971; Bücking 2009 & 2010; Hüning 2010; Härtl 2018 & 2020; Schlücker 2020). Bauer states (2017) that one of the main functions of word-formation is to name things that were not given a name yet as can be seen in dictionaries of neologisms. In contrast to word-formation products, whose basic function is to name, i.e. refer to entities, it is commonly argued that syntactic constructions would “assert or describe” (Bauer 2017: 63; see also Zimmer 1971; Downing 1977; Warren 1978; Ryder 1994). The compound-phrase division is functionally characterized by the functional difference between description and naming. As the alternative hypothesis need to be accepted for refer for the hypothesis H A and H B , however, I argue that these results need to be taken with caution for the following reason: the data frame with refer is the only one with a low number of valid hits, which might have resulted in the statistics pointing into another direction and not being as reliable as the other three predicates. Alternatively, refer might have to be considered distinct from the predicate call, with a different set of properties being preferred in NMCs. Concerning H E , which the word status of the quoted item, another charac‐ teristic feature of the item is that it is predominantly a one-word expression. This finding indicates that for example adjective-noun compounds as well as adjective-noun phrases only represent marginal cases in the corpus data. This 6.5 Discussion 129 <?page no="130"?> is in line with the argumentation presented in Härtl (2015) on the acceptability of A-N phrases and compounds in German nennen-sentences. The analysis further revealed that low-frequent nominals have a predisposi‐ tion for NMCs. From a broader perspective it can be concluded that a high frequency class of the nominal in focus has a predisposition to occur in name-mentioning constructions, which also becomes evident when considering the mean frequency classes as reported in Table 6. The corpus-based analysis clearly proves that name-selecting predicates are sensitive to the expression’s degree of conventionalization: highly conventionalized expressions are less compatible with naming verbs. This allows us to confirm hypothesis H C and H D for all four predicates. In addition, a high number of quoted nominals in frequency classes 23 and 24 indicates that these words are ad-hoc formations or novel compounds, supporting again the naming function of NMC constructions. In order to interpret these findings, it is of interest to have a closer look at a number of neurobiological studies, investigating the lexical frequency and the influence of the sentence context on brain’s response. The so-called word frequency effect describes the effect that high-frequency words are processed more efficiently, i.e. faster, than low-frequency words (Brysbaert et al. 2018b). According to the study using MEG technology, lexical frequency affects brain response: high frequently used words are more easily processed in our brain and take less effort as opposed to less frequent words (Huizeling 2022). Lexical frequency influences not only word processing but also the lexical access itself: The effect of lexical frequency progressively moved anteriorly through temporal and frontal cortex throughout word processing, supporting that lexical frequency influences multiple stages of word processing, such as lexical access and integration with the sentential context. (Huizeling 2022: 170) It thus suggests that lexical retrieval timing is influenced by sentential embed‐ ding of a lexical item through prediction processes. Taking these findings into consideration, the NMC structure may facilitate the processing of convention‐ alized names in our brain. The reoccurring sentential pattern of NMCs can substitute for the low lexical frequency and help detecting possible new words. Notably unlike initial predictions, there were almost no adjective-noun compounds found in the data. In a corpus study reported in Härtl (2016), adjective-noun compound names like in (98a) occurred less frequently in a construction with sogenannt (‘so-called’) than analogous adjective-noun phrasal names (98b). 130 6 Empirical investigation I: Corpus-based construction analysis <?page no="131"?> (98) a. Das sogenannte Schwarzlicht ist eine Form von ultraviolettem Licht, das für das menschliche Auge nicht sichtbar ist. ‘The so-called blacklight a form of ultraviolet light that is not visible to the human eye.’ - b. Berufsfahrer haben heute die sogenannte grüne Welle ausgenutzt. ‘Drivers took advantage of the so-called synchronized traffic today.’ Another finding of the corpus study showed that phrases occurred more frequently in quotes than adjective-noun compounds which was interpreted as reflecting a more pronounced name status of compounds as compared to their phrasal counterparts. Due to the limitations of the corpus data, it is difficult to make predictions about the behaviour of adjective-noun compounds in NMCs. Despite this, the low percentage in the corpus data reveals that they are not preferably used in these constructions and are therefore not that frequent. 6.6 Intermediate summary The large-scale corpus investigation presented in this chapter has revealed a first insight into the nature of NMCs in the languages German and English and revealed the complexity of NMCs with regards to quoted nominal’s lexical frequency, word type, and morphological type. Further, empirical evidence on the use of quotes in different sentential constructions has been presented. As argued in section 4.3, name-mentioning constructions can give rise to different readings and name-mentioning predicates like nennen (‘call’), involve three thematic arguments: an agent x, a theme y, and the name z of the theme y. On the one hand, name-mentioning constructions give rise to a name-inform‐ ing interpretation when the agent (x) and the event variable (e) are bound generically. On the other hand, a modalizing interpretation an existential closure of the two arguments is entailed (Härtl 2016, 2018). In contrast to a name-informing interpretation, speakers express a specific modality, i.e., by expressing a negative or evaluative component of the lexicalized expression. The modalizing use of an expression leads to a distancing interpretation: a speaker refers attitudinally to the shape of a certain denotatum’s name. In following a pragmatic approach, quotes serve the function as a pragmatic marker. They indicate the non-stereotypical use of an expression (cf. Klockow 1978; Gutzmann & Stei 2011; Härtl 2018). Additionally, name-informing quota‐ tion as represented below in (99a) need to be distinguished from a modalizing use of quotation marks. 6.6 Intermediate summary 131 <?page no="132"?> (99) a. Die Heimfahrt führte die Familie über eine sogenannte „Schnell‐ straße“. (The way back home leads the family onto a so-called “expressway”.) - b. Die Heimfahrt führte die Familie über eine sogenannte „Schnell‐ straße“. (The unsurfaced path was, according to the route map, a so-called “street”.) In the name-mentioning construction (99a), reference to the name Schnellstraße is established by the quotation marking and by sogenannt. In the example (99b), the nominal head, which is a highly familiar expression, occurs in quotation marks with which the speaker expresses a certain modality, i.e., signals that the quoted expression is not appropriate in this context and therefore deviates from the stereotypical semantics (cf. Klockow 1978; Gutzmann & Stei 2011). In this “modalizing” use (Klockow 1978), the construction signals the non-standard use of an expression, which is often accompanied by what is known in the literature as scare quotes (Predelli 2003; Recanati 2001). The speaker distances themself from the name for a concept which often leads to an ironic or sarcastic interpretation of the quoted nominal (Härtl 2018). Therefore, the following empirical studies presented in section 7.1 and 7.2 focus on investigating name-mentioning constructions in modalizing contexts. After manually labelling a representative sample of NMCs, the following question will be in focus of the next analysis: Does the corpus data show a preference for a literal, i.e. name-informing, reading or rather a distanced interpretation? 132 6 Empirical investigation I: Corpus-based construction analysis <?page no="133"?> 7 Experimental investigation I: The interpretation of NMCs 7.1 NI vs. MOD: A rating study This study aims at investigating the interpretation of NMCs in German, drawing on the corpus data and using a forced-choice rating design. The forced-choice task allows for examining interpretative preferences through a scalar rating system, reflecting a spectrum of possible interpretations for single sentences. The design enables the isolated analysis and comparison of participant rat‐ ings, allowing for an investigation of interpretive variations and any notable deviations across contexts. The interpretive judgments are expected to reveal which interpretations of NMCs are more natural or preferable within specific contextual embeddings. 7.1.1 Research question and hypothesis Building on the data extracted in the corpus study, this analysis focusses on the interpretation of NMCs in German with ironic and non-ironic interpretation in relation to lexical frequency. The forced-choice study implemented here build on the investigation reported in Schrader (2017), which proved an interaction be‐ tween the lexical frequency of a mentioned nominal involved in sogenannt-con‐ struction and its interpretation as either literal or modalizing, such that highly frequent nominals are more likely to link to ironic uses of the construction. In accordance with these findings, the underlying assumption was that compounds with a low lexical frequency would exhibit a preference for name-informing constructions. High frequently used nominals should then show a preference for modalizing uses. Moreover, an ordinal interaction between the WordType and the lexical frequency was expected so that highly frequent simple nouns are more inclined toward modalizing utterances. 7.1.2 Method In order to investigate the interpretation (name-informing vs. modalizing) in NMCs, a forced-choice rating study was implemented. The data set consisted of 536 constructions for the predicate bezeichnen and a total of 372 constructions for the German predicate nennen taken from COSMAS II DEReKo corpus. The analysed data is lower than the reported data set from the corpus study (see <?page no="134"?> 75 Please note that Schrader (2017) introduced to term konventionelle Anführung (lit. ‘conventional quotation’) to substitute for conventionalized reading. 76 Rating a two on the scale: to show a tendency for a name-informing interpretation; rating a four on the scale: to show a tendency for a modalizing interpretation 77 Note that this sample sentence was a sample sentence taken from Schrader (2017). It was decided to use this sample sentence in order to ensure that all items of the corpus data could be labelled. Chapter 6). This is grounded in the fact that although the corpus data contained enough context to be labelled as a valid NMC, some constructions were lacking enough context to be properly labelled with regards to their interpretation. These data points have therefore been excluded from the analysis. In total, 2 people rated the corpus data using a Likert-scale ranging from 1 for ‘conventionalized reading’ to a ‘modalizing reading’ 75 with a 5 on the scale. Prior to the rating, the participants were given an instruction sheet which included a definition of the (a) name-informing constructions [to be labelled as a 1], (b) modalizing interpretation [to be labelled as a 5], (c) a borderline case [to be labelled as a 3, 2, or 4 76 ]. All three categories were illustrated by sample sentences in German. The conventionalized, i.e., literal reading, was defined as any sentential construction in which an expression with an extensional reference is presented. In conventionalized readings, the quoted expression is commonly connected to a definition, description or explanation of the nominal, as becomes obvious in the following sample sentence that was given to the participants: „Mit Inkunabeln bezeichne man Bücher aus der Wiege der Buchdruckkunst seit 1450 bis 1500” 77 (‘Books from the cradle of the printing art from 1450 to 1500 are referred to as incunables’). Further, the participants were told that in contrast to a name-informing reading, modalizing sentences represent a non-stereotypical or a distancing use of a word. The stereotypical interpretation receives an additional interpretation that is influenced by the context: „Den Umgang mit der Doppelbelastung Beruf und politisches Amt bezeichnet er als „Spagat”. Einerseits wolle, wer Ja zu einem Amt sage, dieses auch bestmöglich erfüllen.” (‘He describes the dealing with the double burden of occupation and political office as “balancing act”. On the one hand, those who say yes to a position want to do the best they can.’). For the investigation, the scale was used in order to enable participants to choose tendencies. Appendix B1 contains the instructions provided to the participants and an overview of the rating schema and scale. Appendix B2 exemplarily illustrates how the participants rated the sample sentnecs using the mean rating. For a list of the total counts for the predicates see Appendix B3. 134 7 Experimental investigation I: The interpretation of NMCs <?page no="135"?> 7.1.3 Analysis & Results As a first step, the ratings were investigated for possible differences between the participants using the ordinal data ratings from 1 to 5. All instances with missing values were removed prior to the analysis. The descriptive statistics already reveal a tendency for a name-informing interpretation with 45.31% for nennen and 47.76% for bezeichnen. We are first examining the statistics for bezeichnen followed by nennen. The statistical analysis was conducted in R (2023). The mean rating ranged at 2.43 (sd = 1.08). The distribution of the rating is presented below. Fig. 8: Distribution of mean ratings for nennen For the scale ratings 1 - a name-informing construction - and a 5 on the scale - a non-literal interpretation, a Chi-squared test was conducted in R (2023). When considering solely the clearly identifiable sentences with these two ratings, the results proved a statistically significant effect (χ 2 (1)=34.571, p < .001). The corpus data contained significantly more name-informing sentences than a modalizing reading. A similar result is observed when taking together the ratings for a name-informing reading with the ratings displaying the tendency for this rating, i.e. 1 to 2.5 and 3.5 to 5. The difference between a name-informing and non-name-informing reading is statistically significant as well (χ 2 (1)=68.167, p <-.001). As a next step, the word type for the scale ratings was considered. The Fisher’s Exact Test results indicates a p-value of 1, which suggests that there is no association between W O R D T Y P E and rating. 7.1 NI vs. MOD: A rating study 135 <?page no="136"?> Fig. 9: Amount of ratings for the Interpretation by W O R D T Y P E for nennen Regarding the frequency, there was no single instance with a high lexical frequency associated with a scale rating of 1 or 5. For a low frequency class, there was no significant effect between a name-informing reading and a modalizing reading using the Fisher’s Exact Test for count data (p = 1). The interaction between frequency and rating was not investigated further as the ratings 1 and 5 consisted exclusively of a low lexical frequency. We will now move on the ratings for the predicate bezeichnen. The mean value was 2.86 (sd = 1.21). When considering the ratings 1 and 5 on the scale, a significant effect is observed with a preference for a name-informing reading (χ 2 (1) = 13.954, p < .0001). The same effect shows up when taking the trends also into account (χ 2 (1) = 4.7918, p < .05). 136 7 Experimental investigation I: The interpretation of NMCs <?page no="137"?> Fig. 10: Distribution of mean ratings for bezeichnen Based on the ANOVA results, there is no significant difference in the means of the groups defined by the W O R D T Y P E variable (Pr(>F) = .239). Regarding the frequency class, i.e., a difference between low and high, for the interpretation, no significant difference was detected (χ 2 (3) = 4, p = .2615). Fig. 11: Amount of ratings for the Interpretation by W O R D T Y P E for bezeichnen 7.1 NI vs. MOD: A rating study 137 <?page no="138"?> For the predicate bezeichnen, compounds had the highest amount in both readings, indicating that the word type is not affected by the interpretation of the NMC. The next sub-chapter presents a discussion of the findings. 7.1.4 Discussion The study presented here aimed at investigating the properties of NMC con‐ struction of the previously conducted corpus study with regards to their interpretation. The rating study challenges the hypothesis and results outlined by Schrader (2017), as it reveals that highly frequent nominals do not exhibit a higher likelihood of being associated with modalizing uses of the construction. Notably, the analysis focuses exclusively on the verb bezeichnen with no occurrences observed in nennen for compounds with a low lexical frequency. Summarizing the findings of the rating study, both predicates revealed a statistically significant effect for a name-informing interpretation instead of a modalizing reading. This means that a high percentage of the corpus data consist of literal NMCs. In addition, both predicates have a strong effect for compound words instead of simple nouns. As argued in Chapter 4.2, name-informing constructions were expected to exhibit a preference for compounds. In contrast to Schrader’s (2017) findings, the statistical analysis did not reveal significant effects concerning the dependence of lexical frequency on the sentence’s interpretation as either modalizing or conventional. Furthermore, no interaction was observed between a) the frequency class multiplied by the mean rating and b) the frequency class and Name-Type for bezeichnen. It is worth noting that the extensive number of valid constructions - espe‐ cially for the predicate bezeichnen - in combination with the extended duration required for reading and rating the corpus data may introduce a potential fatigue effect among participants. It should be acknowledged that this aspect could not have been controlled for. Another limitation of this investigation pertains to the restricted number of participants and the potential constraint of the corpus data context, which may hinder a definitive determination of the intended readings. Consequently, the results derived from the judgment of the valid NMCs may not necessarily be generalizable and only reflect the interpretation of NMCs for the language German. Further research is needed to explore the potential impact of NMCs in English. To address these limitations, a more controlled experimental setting with a larger participant pool and a specific experimental design is deemed necessary to thoroughly explore potential differences between modalizing and name-informing uses of NMC constructions. Despite this, the investigation 138 7 Experimental investigation I: The interpretation of NMCs <?page no="139"?> 78 I would like to note that the self-paced reading study presented in this section is based on a paper previously published in the Proceedings of the 45th Annual Conference of the Cognitive Science Society (CogSci 2023) under a Creative Commons Attribution 4.0 International License (CC BY 4.0). The original publication Raue (2023) shares content with this thesis. The version presented here includes additional statistical analyses and a revised modeling approach (Cumulative Link Mixed Model instead of ANOVA Single Factor analyses), resulting in minor differences in reported outcomes. 79 Situational irony is characterized by lacking intentionality and being a condition of events in opposition to the expected event (cf. Lucariello 2007), while dramatic irony is related to fictional storytelling. 80 Other types of irony include socratic irony (pretending to ignore a topic for pedagogical purposes), and irony of fate which also corresponds to situational irony. revealed the difficulties in rating different readings of NMCs and provided first material for further experimental studies. 7.2 NI vs. MOD: A self-paced reading study Building again on the findings of the corpus analysis, the studies to be conducted in this phase investigates the characteristics of name-mentioning in modalizing contexts. This study aims at testing whether the interpretation of literal and non-literal NMCs is influenced by two factors, namely lexical frequency and word type of the nominal. In order to empirically study possible interactions between these two factors, a self-paced reading experiment will be presented in this chapter. The experimental design is well-suited for a precise measurement of how participants process different text segments, which allows for insights into the real-time processing of NMCs. By allowing participants to control the pace of reading, this approach can reveal individual-specific differences in reading times (cf. Frinsel & Christiansen 2024) and reveal how and if readers adjust reading times when encountering NMCs in literal and/ or non-literal contexts. The next section will briefly introduce theoretical considerations that form the basis of the argumentation on NMCs in modalizing contexts. 78 7.2.1 Accounts of verbal irony In the following investigation, the focus is on what is known as verbal irony in the literature. In contrast to, for example, situational and dramatic irony 79 , verbal irony can be defined as expressing something other than the literal meaning, i.e., most commonly the opposite of the literal meaning (Attardo 2000). 80 Verbal irony evaluatively refers back to a preceding event used by the speaker to express attitudinal content. This attitude can either be in a positive or negative 7.2 NI vs. MOD: A self-paced reading study 139 <?page no="140"?> way: as ironic criticism or ironic praise. Verbal irony occurs on a variety of linguistic levels from one-word phrases to rhetoric questions, exclamations, and interjections. Generally, irony can be described as expressing the opposite of the literal meaning (cf. Attardo 2000; Garmendia 2018). Irony usually creates an incon‐ gruity between the context and the utterance itself (Barbe 1995). There’s consensus in the literature that ironic utterances are contextually inappropriate but given that they are uttered intentionally and expected to be recognized by the addressee, the utterance is still relevant (Attardo 2000). Traditionally, irony has been described as a trope, i.e., a figure of speech which goes back to the distinction between language and metalanguage (cf. Kroeger 2018; Attardo 2000). In the following section, three approaches on figurative language comprehension will be reviewed in more detail: the standard pragmatic model, the graded salience hypothesis, and the direct access view. In the Gricean tradition, the standard pragmatic view has been an influential theoretical approach to figurative language comprehension. Ironic utterances flout Grice’s Cooperative Principle (see Grice 1989), more specifically the Maxim of Quality, as the speaker asserts something that they believe to be false (Meibauer 2007) which results in irony functioning as a conversational implicature. Irony stereotypically implies a negative, critical attitude towards some object. This basic idea can also be observed in Grice’s work as he states that irony is connected to the “expression of a feeling, attitude, or evaluation” (Grice 1978: 53-54). According to the Gricean standard pragmatic view, the discrepancy between the two interpretations is crucial for irony processing. Under this account, the literal meaning has been argued to be processed first followed by applying an alternative interpretation due to the discrepancy. Consequently, irony processing should be more costly and take more time than the literal interpretation. Post-Gricean approaches to irony treat verbal irony as a type of echoic allusion to an attributed utterance or thought (Sperber & Wilson 1981). The speaker echoes a thought or an utterance in order to express a critical or mocking attitude. According to this approach, the aim of the speaker is to express a dissociative attitude to a tacitly attributed utterance or thought, based on some perceived discrepancy between the way it represents the world and the way things actually are. The ironic utterance indicates dissent from the utterance (Sperber & Wilson 1981). Another well-known approach to irony describes verbal irony as a type of pretence, where a speaker pretends to assert some content while expecting the addressee to see through the pretence and recognise the critical attitude behind 140 7 Experimental investigation I: The interpretation of NMCs <?page no="141"?> it. This so-called joint pretence account “invites one’s interlocutors to consider a situation in which the utterance would be true and notice how absurd this situation is” (Kumon-Nakamura et al. 1995). Both post-Gricean accounts reject the claim that the hallmark of irony is to communicate the opposite of the literal meaning. The importance of context for disambiguating the ironic meaning has further been emphasized in the direct access view which states that the ironic meaning can sometimes be accessed directly (Gibbs 1994). Gibbs argues that both literal and figurative meaning make use of the same mechanisms and that the understanding of irony “does not necessarily require special cognitive processes beyond those used to comprehend literal speech” (Gibbs 1994: 437). In other words, the literal meaning was argued to be not necessarily processed first if the context supports the access of the literal content. From a semantic point of view, verbal irony involves a form of indirect negation (Giora 1995). The theory of irony as negation proposes that irony is a form of negation, in which the speaker negates the literal meaning of an utterance and asserts the opposite meaning. Consider this viewpoint by looking at a classic example. (100) a. Context: It is raining cats and dogs outside. A: What a nice weather. - b. nice(weather) According to the graded salience hypothesis, more salient meanings have a priority in the interpretation. Salient meaning is defined as meaning that can directly be retrieved from the mental lexicon (Giora & Fein 1999). In two experiments, Giora & Fein (1999) found evidence to support the role of salient meaning in irony processing, attributing the direct access to salience than an effect of context. Based on this idea, Attardo states that the processing of irony involves processing of meaning of the text, rejection of this interpretation on pragmatic grounds, subsequent reinterpretation of the text (Attardo 2000). Crucially, Katz & Pexman (1997) claim that graded salience and contextual factors interact in the comprehension of irony. When a speaker utilizes irony, they are making use of a well-established philosophical concept that distinguishes between actively using and mentioning expressions (Colston & Gibbs 2007). The use-mention distinction offers possi‐ bility to make “reference to some state of affairs that was predicted, expected or desired, either because of some explicit prediction or based upon a mutually shared domain of knowledge” (Colston & Gibbs 2007: 5). 7.2 NI vs. MOD: A self-paced reading study 141 <?page no="142"?> As discussed in the previous section, name-mentioning constructions can give rise to two distinct types of reading: a name-informing and a modalizing reading. Consider the following examples. (101) a. The politicians reject actions that could have helped to fight against climate change. Something like this is called climate protection. [non-literal] - b. The moderator talks about government action that aim at reducing global warming. Something like this is called climate protection. [literal] The nominal climate protection in (101a) adopts a non-literal meaning, i.e. an ironic interpretation, instead of a name-informing one (101b). When considering the examples in (101), the context sentence serves as the common ground shared by the speaker and addressee. The addressee can only interpret the ironic content as such in a communicative situation as presented in the sample sentences. Hence, the context plays a crucial role in disambiguating the intended meaning. Furthermore, the speaker of the utterance in (101a) expresses a negative evaluation which is directed at the addressee (Sperber & Wilson 1986). In other words, the speaker criticizes the actions to fulfil the function of climate protection. In addition, NMC can be used to attitudinally allude to an official title, expressing that the established title does not fit to the person. The non-literal interpretation of titles is emphasized when accompanied by a reflexive verb, as illustrated below. (102) a. Der Chef überträgt alle wichtigen Entscheidungen an seine Mitarbei‐ tenden. ‘The CEO leaves important decisions to the employees.’ - b. So jemand bezeichnet sich als Geschäftsführer, kommentiert die Mitar‐ beiterin. 'Someone like that calls himself CEO’, comments an employee.’ In the construction, a person is introduced and has been given the name CEO. Stereotypically, a CEO is defined as someone who makes important decisions. In (102b), this defining characteristic of a CEO, i.e. the competence of decision-making, is denied which gives rise to the ironic interpretation. Therefore, the descriptive content of the utterance expresses that the person X is not really or should not be called a CEO. The expressive content of this sample relates to the speaker of the direct quotation in expressing a negative attitude towards the competence or behaviour of the CEO. The modalizing construction 142 7 Experimental investigation I: The interpretation of NMCs <?page no="143"?> presents and alternate meaning to the literal one: it entails that the nominal’s denotatum, i.e. CEO, has been given that name before and that the speaker echoes some previously uttered words. Importantly, the ironic content can only be understood as such when both the speaker and addressee have a shared common ground based on their experiences, more specifically that the person is the company’s CEO. Interestingly, written communication on social media platforms such as Twitter has recently introduced a way of disambiguating short texts. To facilitate the understanding of text-picture relations for example for people with autism spectrum, the text is followed by a tone indicator like “/ li” for literally, “/ ij” for inside joke, and “/ S” for sarcastic (cf. Christantil et al. 2022). This trend illustrates not only the importance of context for a proper understanding of ironic content but also that ironic content must be learned and interpreted by the addressee. Before presenting the self-paced reading study, the state of the art regarding the processing of irony is discussed. 7.2.2 The processing of irony: Empirical evidence This chapter addresses the state of the art regarding empirical evidence on the processing of irony. Experimental and empirical investigations in the field of irony have mainly focused on the processing of irony (Dews & Winter 1999), cognitive mechanisms involved in figurative language comprehension, and realisations of irony including facial, lexical, and vocal cues (Attardo et al. 2003). The psycholinguistic literature on irony comprehension aimed at investigating how cognitive impairments like aphasia affect irony comprehension (McDonald 2007). In recent studies, ironic utterances have been contrasted with literal meanings, basing their understanding of irony on the echoic-mention approach ( Jorgensen et al. 1984). Dews and Winner (1999) investigated the processing of irony, presenting empirical evidence of two studies that showed first that the literal meaning of irony is processed and, second, that the non-literal meaning of ironic criticism is obligatorily processed. They argue that their experiments support a multiple meaning model of processing irony which underlines that verbal irony is not only a pragmatic phenomenon but combines both fields of semantics and pragmatics. One basic assumption is that cognitive process in irony comprehension take longer as involve more complex mental operations (cf. Jiang 2011; Kowatch et al. 2013). These longer processing times are consequently reflected in a longer reaction time (RT). In a moving window reading experiment, Pexman et al. (2000) proved that the target statement is processed slower in the ironic 7.2 NI vs. MOD: A self-paced reading study 143 <?page no="144"?> 81 For evidence on metaphors see e.g. Arzouan et al. (2007), Coulson & Van Petten (2007), and Kazmerski et al. (2003). Studies on idioms are reported in Laurent et al. (2006) and Ferretti et al. (2007). as compared to conventional metaphors and salient-based literal conditions. Further empirical evidence comes from tracking eye movements, in which sarcastic utterances were proven to be more difficult to understand, leading to longer reading times, as opposed to literal conditions (Olkoniemi et al. 2016). Moreover, these findings are bolstered by another eye-tracking experiment which confirmed longer reading times for ironic conditions as opposed to literal utterances (Filik & Moxey 2010). It is commonly assumed that context-based inferences are required when inter‐ preting an utterance as ironic (cf. Paunov et al. 2022: 416). The comprehension of ironic and non-ironic content involves partially distinct neurocognitive processes which is reflected in the difference in sensitivity of the left and right hemisphere (cf. Regel 2009). In contrast to core linguistic processes, non-literal cognitive processes like irony have been argued to be located in the right hemisphere. According to Fedorenko & Thompson-Schill (2014), language regions of the right hemisphere are more flexible and less specialized. In order to investigate the timing of language processing, event-related brain potentials (ERPs) have been used due to their high temporal resolution in millisecond rage (cf. Regel 2009). ERP analyses have mainly focussed on the N400 and P600 components as they have been proven to be sensitive to “pragmatic inference making” (Spotorno et al. 2013). Different EEG studies present evidence for a “P600/ late positivity components” (Spotorno et al. 2013: 3) for ironic sentences compared to literal sentences. Neurocognitive processes on irony comprehension have investigated figurative language, looking especially at metaphors, idioms, and irony. 81 Evidence from ERP studies report greater amplitudes of the N400 component during the comprehension of ironic statements in contrast to literal utterances (Cornejo et al. 2007; Katz et al. 2004; Regel 2009). This effect indicates that processing semantic information is costlier and that difficulties occur in the early integration of lexical semantic information. Empirical research on metaphors, proverbs, and idioms shows that when figurative language is used, processing semantic information becomes more complex and expensive when figurative language is involved. This effect is manifested in an increased N400 component (cf. Regel 2009). A recent ERP study on verbal irony (Regel 2009) contrasted literal conditions with prosodic information, reporting a difference in the pattern of brain activity for the target word: irony showed “early and late ERP effects” and a “sustained left anterior negativity starting around 200 ms (referred to as sustained LAN) followed by an additional late posterior positivity in the latency range of 500-900 144 7 Experimental investigation I: The interpretation of NMCs <?page no="145"?> 82 Theory of mind is understood as the ability of a person to reason about mental states of other people. ms” (Regel 2009: 60). In two subsequent ERP experiments, Regel (2009) included quotation marks around the critical word as an additional factor when presenting the target sentence segmented word-by-word. The findings indicate that not only quotation marks facilitate the processing of irony, but also pragmatic/ contextual information had an influence on the initial processing of irony. Further, contextual cues as well as sensitivity for irony can facilitate the comprehension and processing (Olkoniemi & Kaakinen 2021). Katz et al. (2004) report several studies on sarcasm which requires social-cultural factors to be understood and a sarcastic interpretation early on in the processing. Other psycholinguistic studies have shown how the interpretation of ironic utterances depends on an incongruity between the context and the utterance (Colston & O’Brien 2000; Ivanko & Pexman 2003). Despite the importance of context-related factors for the comprehension of irony, recent research has emphasized the importance of reader-related factors, including a persons’ working memory capacity (WMC) and Theory of Mind (ToM) (cf. Bosco et al. 2018; Hagoort & Levinson 2014) 82 . In addition to the language-selective network, which has been shown to support literal comprehen‐ sions (Fedorenko et al. 2020), language understanding recruits the ToM network which allows for social inferences (Saxe & Kanwisher 2003). The working memory capacity has been argued to positively influence the efficiency of processing irony (Giora 1999). The ToM network plays a crucial role in the comprehension of non-literal language especially metaphor and irony (cf. Blasko & Kazmerski 2006). The importance of the ToM network is supported by evidence from a meta-analysis of the functional neuroimaging technique fMRI targeting at non-literal language processing: non-literal processing was supported by “i) mechanisms that process literal linguistic meaning, and ii) mechanisms that support general social inference” (Hauptman et al. 2023: 96). They thus undermine a strong divide between literal and non-literal aspects of language and challenge the claim that non-literal processing requires additional executive resources (Hauptman et al. 2023). In addition, an fMRI study on the processing of irony revealed a “functional connectivity between language and the ToM networks during irony processing” (Spotorno et al. 2013: 1). They argue that this interaction is related to integrative processing when comprehending an ironic utterance. Moreover, individual factors involved in the processing and comprehension of irony has been proven to be influenced by an emotional component. The processing of irony is facilitated with subjects who have a good ability to process 7.2 NI vs. MOD: A self-paced reading study 145 <?page no="146"?> 83 Verb-noun compounds and verbal compounds, which also appear frequently in NMCs, have not been the focus of in this SPR. emotional information (Olkoniemi 2018). Generally speaking, reader-related factors are therefore a crucial component in the research on irony comprehension. For the following investigation, it is important to make a note on the type of quotation effective in this study. The quotational constructions under investigation in this self-paced reading experiment are direct quotations, which are used to convey an “utterance verbatim” (Brendel et al. 2011: 96). In the previous section, it has been argued, that NMCs can be analysed as representing an instance of pure quotation. The following investigation focuses on sentences like And someone like this calls himself “expert” where the nominal expert expresses a “specific speaker modality implying a certain reservation with respect to the semantic appropriateness of the expression in quotes” (Härtl 2020: 771). As argued previously, quotations in distancing constructions have been classified as instances of scare quotes (Recanati 2001; Predelli 2003; Meibauer 2007; Härtl 2018). 7.2.3 Methodology Thirty-eight participants (mean age=24.21; sd=3.13; age range 19-31) took part in the experiment. They were all undergraduate students from the University of Kassel and received course credit for participating in the experiment. Prior to the experiment, the participants completed a questionnaire regarding personal information. All participants were self-reported native speakers of German and had corrected or corrected-to-normal vision. The participants were naïve with respect to the research question and hypotheses to ensure that the manipulations were not obvious to the participants. 7.2.3.1 Material and design The empirical focus of this study is on German. The experimental design was a between-items design with matched sets, meaning that all subjects were opposed to half of the items from each condition. The test material is comprised of 32 paired scenarios, i.e., a total of 64 experimental contexts (see Table 10 for an example). The paired scenarios consist of, on the one hand, a literal condition in which the speaker introduces a name-mentioning construction whereas the second condition gives rise to a modalized reading, i.e., ironic interpretation of the sentence. I carefully selected 32 nominals that are either noun compounds 83 146 7 Experimental investigation I: The interpretation of NMCs <?page no="147"?> or simple nouns. All target nominals are syntactically integrated into a sentence and are related to either a literal or a modalizing reading of the target sentence. The self-paced reading study exhibits a 2x2x2 design: I N T E R P R E TATI O N (NI vs. Mod.), W O R D C LA S S (compound vs. simple noun) and L E XI C AL F R E Q U E N C Y (low vs. high), resulting in the following conditions. Word class Frequency class Modalizing (MOD) Name-informing (NI) compound high Condition 1 Condition 5 compound low Condition 2 Condition 6 simple noun high Condition 3 Condition 7 simple noun low Condition 4 Condition 8 Tab. 10: Overview of experimental conditions Each condition was made up of 8 sentences out of which 4 contained the naming predicate nennen (‘call’) and 4 the predicate bezeichnen als (‘refer to as’). Prior to the experiment, two student assistants were asked to carefully read all experimental items and decide whether the setting of the context sentences was clearly giving rise to a modalizing reading as opposed to a non-modalizing reading of the respective name-informing counterparts. The pre-test was further used to ensure contextual plausibility of the experimental items. It is important to note that it was ensured that the conditions were balanced with regard to presenting learning-new-words contexts to avoid systematic confounds. These learning contexts are preferred in NMC constructions, e.g. when a new word is introduced in a teaching situation between teacher and students or parents and their children. For this reason, learning contexts were also used in the modalizing conditions. In addition, all experimental items were checked for sentence plausibility prior to the experiment. The experimental items followed the pattern of context sentences followed by the target sentence. The material is comprised of 64 different context sentences that aim at setting the situational context and triggering either a literal or modalizing interpretation. The overall sentence structure, as well as the sentence length of the introductory sentences, was comparable. The target sentence consists of a name, a verb, and the target noun. For a greater variation in naming predicates, sentences with nennen (‘call’) have been divided into equal parts containing nennt man (‘call’) and sich nennen (‘call.refl’). Crucially, the target noun was placed sentence-medially and was always either a 7.2 NI vs. MOD: A self-paced reading study 147 <?page no="148"?> 84 I would like to thank the team of the Natural language processing group of the University of Leipzig for sharing a word list of the German News Corpora (2020) including the respective frequency with me. 85 https: / / www.wortschatz.uni-leipzig.de/ de 86 Please note that there is a difference in punctuation between German and English. While the latter has a comma before the quotation marks, the comma is placed after the direct quote in German. compound or a simple noun. The frequency of the nominal 84 was controlled with the corpus of the German language provided by the Leipzig Corpora Collection (LCC). 85 The frequency of the target items ranged between the frequency classes 1-9 for a high lexical frequency and 16-24 for a low lexical frequency (see Appendix C3 for an overview of all target words and their lexical frequency). The target sentence was presented in italics to indicate the direct quotation in a spoken discourse situation. All paired scenarios contained the same target sentence so that 32 sentences could be presented to the participants without having to present any target noun twice. For all sentences, the reaction times (RT) were measured in milliseconds (ms), recording the time from the onset of the stimulus to the participant’s response. Response times are measures as they allow for drawing inferences on cognitive processes involved in different tasks for language processing ( Jiang 2011). The direct quotation used a pattern in which the reporting clause comes second, following the syntax/ sentential structure of a so-called reported clause as commonly used in Standard German. In this type of direct speech, the subject [S] and the reporting verb [V] are inverted: (103) So etwas nennt sich Pressekonferenz , sagt Lisa. - [direct quote] [V] [S] As a result, the direct quotation was always followed by a comma 86 and a non-italicized reported clause. In order to set the target phrase graphemically apart from the reported clause, the direct quotation was graphemically set apart by using only italics but no quotation marks in the written mode. The spill-over region contained between two to three words with the clause being introduced with one of the following verbs: sagen (‘say’), anmerken (‘remark’), kommentieren (‘comment on’), erklären (‘explain’), beschreiben (‘describe’). The reporting clauses varied across the experimental items, resulting in a total of 64 different clauses. The last part of the experimental item was a comprehension question which participants had to answer either with ‘yes’ or ‘no’. The semantic questions 148 7 Experimental investigation I: The interpretation of NMCs <?page no="149"?> addressed three different parts of the experimental items: the context sentence, the direct quotation, and the reported clause. The experimental items were balanced across the two groups so that half of the semantic questions were correctly answered with a yes and the other half with a no. All participants were asked to do self-paced reading of paragraphs as illustrated in the schematized test design below. Context sentences: NI MOD Der Besitzer der Weinkellerei erklärt die Besonderheiten von verschiedenen Wei‐ nen. ‘The owner of the wine cellar explains the difference between his wines.’ Kim berichtet in einer Mail von einem Treffen mit einem Kollegen, der ver‐ korkten Wein servierte. ‘Kim reports in an email about a meeting with a colleague who served wine that was corked’ MOD Target sentence: So jemand nennt sich Weinkenner, merkt Kim an. Someone like that calls [call.REFL] himself/ herself wine connois‐ seur. - Target 1 Target 2 Target 3 Spillover region Comprehension question: Hat Kim ihre Mutter getroffen? ‘Did Kim meet her mother? ’ NI Target sentence: So jemand nennt sich Weinkenner, sagt Maria. Someone like that calls [call.REFL] himself/ herself wine connois‐ seur, says Maria. Target 1 Target 2 Target 3 Spillover region Comprehension question: Besitzt der Mann den Weinkeller? ‘Does the man own the wine cellar? ’ Ja / Nein ‘Yes’ / ‘No’ Tab. 11: Schematized test design for a minimal pair 7.2 NI vs. MOD: A self-paced reading study 149 <?page no="150"?> 87 I would like to thank Álvaro Cortés Rodríguez for pointing out that the reading regions, i.e. the part of the sentence measured with RT, are different across items and that the referent in the antecedent is in different positions. Note that in order to create an ironic context sentence, the target nominal had to be paraphrased in the antecedent, e.g. i.e. the paraphrase somebody who knows a lot about a wine to describe the term wine connoisseur, which made it impossible to have the same distance between the target word and description across items. 88 The sample Moving Window experiment is available as an .es3 data file here: https: / / s upport.pstnet.com/ hc/ en-us/ articles/ 115001914408-Moving-Window-30019- (retrieved October 10, 2022) 89 I would like to thank Todd Shafer for his input on adapting the code in E-Studio. All of the context sentences contain a defining description of the target word. 87 That way, the target nominals were syntactically integrated in the target sentence. Regarding the so-called SlideObjects in E-Studio, each trial consisted of 4 slides: 1. a fixation asterisk 2. a context sentence followed by the target sentence and spillover region 3. a question mark 4. a comprehension question. In the following section, the underlying hypotheses will be sketched for each of the target reading regions. 7.2.3.2 Procedure The experiment was conducted in a controlled setting in the language laboratory on the campus of the University of Kassel. All participants were invited to sit in front of a computer screen in a silent setting. The experiment lasted for approximately 20 minutes. The experiment was conducted using the E-Prime 3.0 software for psycho‐ linguistic experiments (runtime version 3.0.3.80; Psychology Software Tools, Pittsburgh, PA 2016). To create the experiment, the code and experimental design were adapted using a sample Moving Window experiment as provided on the PST website which takes sentences (or phrases) and breaks them into several windows. 88 Importantly, the context sentence was visually presented in one line as a whole while the target sentence and spillover were presented in the line underneath. 89 In E-Studio, 2 sub-experiments called Group A and Group B were created. Each experiment contained 3 practice sentences and 23 target sentences. The experimental design was a within-subject design as all subjects were exposed 150 7 Experimental investigation I: The interpretation of NMCs <?page no="151"?> 90 See also https: / / support.pstnet.com/ hc/ en-us/ articles/ 115000901928-TIMING-What-is -Millisecond-Accuracy-22846) to all conditions. To avoid effects of presentation order, the list of items was presented in different random orders to the participants (Gibson & Fedorenko 2013). The experiment started with two slides that presented the instructions to the participants (see Appendix C1 for the instructions). The participants were encouraged to read the sentences at their own speed and were offered the opportunity to ask questions after reading the instructions and after completing a short trial. The practice round consisted of 3 sentences with two items giving rise to a modalized interpretation and one literal reading. The input device for the self-paced reading experiment was restricted to the keyboard with the following keys: the SPACE button, and the button with the number ‘1’ and ‘9’. The keys were labelled with a ‘WEITER’ (‘NEXT’) on the space button, ‘JA’ (‘YES’) on the key with the number ‘1’, and a ‘NEIN’ (‘NO’) on the number ‘9’. Between each of the trials, a fixation asterisk was presented for 2000ms in the centre of the screen. In general, the text was displayed in black on a white screen. Thus, the stimuli were aligned on each slide. Crucially, the slide with the target stimulus contained multiple SlideObjects in order to be able to present the target sentence, i.e. the direct quotation, in italic. This way, a smaller TextDisplay could be used which has the advantage of improving the timing (Spapé et al. 2019). The comprehension accuracy was measured for each sentence. As the target sentence was split up into different individual attributes, the reaction times (RTs) were recorded between the key presses. All reaction times are measured in milliseconds and the E-Prime software is known for having a standard deviation of less than one millisecond/ sub-millisecond accuracy (PST Support E-Prime 3). 90 7.2.4 Hypotheses The aim of the experimental investigation was to test if the type of the expression (complex vs. simple nouns) is also a factor in modalizing contexts triggering ironic interpretations of the mentioned name. The following hypoth‐ eses concern the target sentences by looking at three main reading regions, i.e., the three self-paced reading regions as presented in the following example: 7.2 NI vs. MOD: A self-paced reading study 151 <?page no="152"?> 91 The results for Reading Time 1 and Reading Time 2 are discussed in Raue (2023). Reading time 1: Someone/ Something like this Reading time 2: is called/ referred to as/ call themselves Reading time 3: target nominal Reading time 4: Spillover region As the context sentences functioned as either ironically or literally biasing context, the context sentences were different across the conditions. Therefore, the reaction time for the context sentence was not statistically explored and no significant effects were expected for reading time 1 and 2. 91 For Reading time 3, the following hypotheses will be investigated. The first hypothesis posits that the type of interpretation as either name-informing or modalizing influences reading time. Second, a low lexical frequency is expected to be easier processed in modalizing contexts, which is manifested in a shorter RT in modalizing constructions in contrast to name-informing ones. The opposite is expected to be reflected in the shorter RTs in name-informing conditions with a high lexical frequency. Third, the W O R D T Y P E is expected to be manifested in the interpretation as well. Compounds are hypothesized to be preferred in name-informing conditions whereas simple noun will be preferred in modalizing sentences. H A : I NT E R P R E TATI O N (NI vs. MOD) • significantly longer reading times are expected for MOD condition as opposed to NI readings H B : L E XI C AL F R E Q U E N C Y (high vs. low) • a low lexical frequency is expected to give rise to faster RTs in modalizing constructions • a high lexical frequency in NI reading times is expected to be longer than reading times for a low lexical frequency • in both NI and MOD, a low lexical frequency should have faster RTs for low frequency items while H C : W O R D T Y P E (compound vs. simple noun) significantly longer reading times for MOD condition for compounds in NI, simple nouns should be read faster than compounds 152 7 Experimental investigation I: The interpretation of NMCs <?page no="153"?> The following predictive plots illustrate the central hypotheses for the target reading region. Fig. 12: Predictive plot for W O R D T Y P E and I N T E R P R E T A T I O N (target word) Fig. 13: Predictive plot for F R E Q U E N C Y and I N T E R P R E T A T I O N (target word) For the spillover region, here called Reading Time 4, modalizing conditions were expected to be processed longer as compared to name-informing ones as the modalizing interpretation can “spill over” to the next reading region, leading 7.2 NI vs. MOD: A self-paced reading study 153 <?page no="154"?> 92 Please note that six out of the eleven data points ranged between 1800 and 1900ms and were therefore close to reaching the threshold. 93 The z-score ranged between 0.22 and 1.15; the mean RT was 967.4ms for the target nominal. 94 The z-score for item types ranged between 0.239 and 1.294 to longer processing times. To empirically test these hypotheses, the following study was designed. 7.2.5 Analysis and Results Prior to the statistical analyses, the raw data was prepared using E-Merge to combine the data files and E-DataAid3 to filter and export the data. In total, the data consists of 1216 individual cases (38 subjects x 32 cases per subject). As a first step, I examined the responses to the comprehension question. The threshold of accuracy was set to 70 % per subject, and the analysis revealed that all subjects met the requirement. In total, there were 132 incorrect responses across the conditions which makes up 10.86% of incorrect responses in relation to all responses. As all subjects reached the accuracy threshold for the comprehension question, the data points were not excluded from the analysis of Reading Time 1 and Reading Time 2. Furthermore, an understanding and proper reading of the context sentence was crucial for investigating the difference between an NI and MOD reading. Therefore, all data points with a reaction time lower than 2000 ms were removed. As a result of this restriction, a total of 11 data points were excluded for all further statistical analyses. 92 After this outlier treatment, the lowest RT was found in the condition labelled as MODlowS_nm with a mean of 4355.16 ms and the longest RT in the condition NIhighS_bez with a mean RT of 14048.1 ms. In addition, the target condition was checked for outliers using the z-score. The threshold was set to the recommended standard deviation of plus/ minus 2 from the overall mean (cf. Loewen & Plonsky 2016) These data points were regarded as extreme values and removed prior to the analysis. Only one subject had significantly longer RT with a mean of 22461 ms (SD = 19450) across all conditions (z-score = 2.14). 93 In addition, all individual items were checked for using the z-score as well. None of the items had to be removed as all data points had a standard deviation of less than plus/ minus 2 from the overall mean (723.71 ms). 94 As there were no outliers detected in the carefully selected item types, they were treated as equally functioning conditions and were not further investigated in the following analyses. 154 7 Experimental investigation I: The interpretation of NMCs <?page no="155"?> 95 In addition, the readxl package (Wickham et al. 2019), ggplot2 (Wickham 2016), and emmeans (Lenth et al. 2018) were used. The statistical analysis of the spillover region, for the mean RT ranged between 709.1 ms and 1829 ms, will not be reported here in detail. All statistical analyses were analysed in R (R Core Team 2021), using linear mixed effects models with the lme4 package (Bates et al. 2015). 95 Reading Time for the target word This section presents the results of the analysis of Reading Time 3, i.e., for the target nominal. Additional statistical analyses for the target word not reported in detail in this section can be found in Appendix C6. The first analysis investigated a possible interaction between the two main condition, i.e., the modalized and name-informing reading. - Estimate Std. Error t value (Intercept) 767.245 45.541 16.847 NI -10.026 33.359 -0.301 WordLengthmiddle -88.532 40.304 -2.197 WordLengthshort -145.898 41.561 -3.510 NI: WordLengthmiddle -8.771 43.088 -0.204 NI: WordLengthshort 46.237 43.305 1.068 Formula: lmer(TargetWindow ~ 1 + MainCondition * WordLength + (1|Subject) + (1|MinimalPair), REML = FALSE) Tab. 12: Fixed effects statistics for I N T E R P R E T A T I O N 7.2 NI vs. MOD: A self-paced reading study 155 <?page no="156"?> Fig. 14: Mean RTs for the target word in the main conditions. Error bars represent confidence intervals +/ - 95% The modalizing targets were read faster (677.9542 ms; SD = 312.8661) than the literal conditions (688.6946 ms; SD = 355.8751) but the effect was statistically not significant, revealed by emmeans pairwise comparison (p = 0.8834). Second, the difference between the W O R D T Y P E was investigated. - Estimate Std. Error t value (Intercept) 782.10 42.69 18.322 WordTypeMODS -125.79 42.38 -2.968 WordTypeNIC -38.33 23.28 -1.646 WordTypeNIS -85.12 42.81 -1.988 WordLengthmiddle -59.30 34.98 -1.695 WordLengthshort -37.77 51.11 -0.739 Formula: lmer(TargetWindow ~ 1 + WordType + WordLength + (1|MinimalPair) + (1|Subject), REML="FALSE") Tab. 13: Fixed effects statistics for WordType and Interpretation The emmeans pairwise comparison for the linear model reveals a statistically significant effect for a modalizing interpretation between compounds and simple nouns. 156 7 Experimental investigation I: The interpretation of NMCs <?page no="157"?> Contrasts Estimate SE df t.ratio p.value MODC - MODS 125.8 44.6 45.4 2.823 0.0343 MODC - NIC 38.3 23.3 948.0 1.644 0.3547 MODC - NIS 85.1 45.0 47.2 1.893 0.2448 MODS - NIC -87.5 44.6 45.9 -1.959 0.2186 MODS - NIS -40.7 22.9 945.4 -1.773 0.2870 NIC - NIS 46.8 45.1 47.7 1.038 0.7281 - Tab. 14: Pairwise contrasts for W O R D T Y P E and I N T E R P R E T A T I O N Fig. 15: Mean RTs for the target word in the main conditions by W O R D T Y P E . Error bars represent confidence intervals +/ - 95% The difference between compounds and simple nouns was statistically non-sig‐ nificant. Despite this, compounds took generally longer to be processed than simple nouns (see Figure 15). As a third step, it was investigated whether the data showed effects between a low and a high lexical frequency. The fixed effects statistics is reported in Table-15. 7.2 NI vs. MOD: A self-paced reading study 157 <?page no="158"?> Fig. 16: Mean RTs for the spillover regions for F R E Q U E N C Y . Error bars represent confidence intervals +/ - 95% - Estimate Std. Error t value (Intercept) 810.257 43.980 18.423 MODlow -83.471 28.378 -2.941 NIhigh -8.504 23.352 -0.364 NIlow -70.775 28.427 -2.490 WordLengthmiddle -102.932 29.520 -3.487 WordLengthshort -133.354 30.606 -4.357 Formula: lmer(TargetWindow ~ 1 + Frequency + WordLength + (1|MinimalPair) + (1|Subject), REML="FALSE") Tab. 15: Fixed effects statistics for I N T E R P R E T A T I O N and F R E Q U E N C Y The linear model revealed a non-significant effect for the lexical frequency. Despite this, the emmeans contrast looking into pairwise comparisons revealed a statistically significant effect for modalizing sentences between a high and low lexical frequency. 158 7 Experimental investigation I: The interpretation of NMCs <?page no="159"?> Contrasts Estimate SE df t.ratio p.value MODhigh - MODlow 83.5 29.5 75.4 2.830 0.0297 MODhigh - NIhigh 8.5 23.4 953.6 0.363 0.9836 MODhigh - NIlow 70.8 29.5 76.1 2.396 0.0865 MODlow - NIhigh -75.0 29.5 75.7 -2.539 0.0620 MODlow - NIlow -12.7 23.0 943.0 -0.552 0.9459 NIhigh - NIlow 62.3 29.6 76.5 2.105 0.1608 Formula: lmer(TargetWindow ~ 1 + Frequency * Quotes + WordLength + (1|Minimal‐ Pair) + (1|Subject), REML="FALSE") Tab. 16: Pairwise contrasts for I N T E R P R E T A T I O N and F R E Q U E N C Y Spillover region For the spillover region, i.e., the reporting clause, it was of interest to analyse whether the interpretation of the construction was also reflected in word following the target nominal. Statistical analyses that are not reported in detail, e.g. pairwise comparisons, here can be found in Appendix C7. The modalizing targets took longer to read (855.7729 ms; SD=581.6553) than the literal conditions (813.8703 ms; SD=537.4041). A linear model was fitted for the main conditions to investigate the difference empirically. The conditions under investigation relate to the linguistic properties of the target nominal to investigate whether the WordType and Frequency have an impact on the succeeding words. Fig. 17: Mean RTs for the spillover region in the main conditions. Error bars represent confidence intervals +/ - 95% 7.2 NI vs. MOD: A self-paced reading study 159 <?page no="160"?> Estimate Std. Error t value (Intercept) 854.52 50.43 16.945 MainConditionNI -45.83 30.75 -1.491 Formula: lmer(SpilloverRegion ~ 1 + MainCondition + (1|Subject) + (1|MinimalPair), REML = FALSE) Tab. 17: Fixed effects statistics for I N T E R P R E T A T I O N The results for the interpretation are statistically not significant. The pairwise comparison is reported in Appendix C7. Fig. 18: Mean RTs for the spillover region for W O R D T Y P E . Error bars represent confidence intervals +/ - 95% - Estimate Std. Error t value (Intercept) 875.27 58.05 15.072 MODS -41.32 57.40 -0.720 NIC -69.31 43.31 -1.600 NIS -63.52 57.55 -1.104 lmer(SpilloverRegion ~ 1 + WordType + (1|MinimalPair) + (1|Subject), REML="FALSE") Tab. 18: Fixed effects for W O R D T Y P E and I N T E R P R E T A T I O N 160 7 Experimental investigation I: The interpretation of NMCs <?page no="161"?> As indicated in the statistics, there are no significant effects for W O R D T Y P E and a modalizing or name-informing interpretation. Next, the lexical frequency in the spillover region was investigated. Fig. 19: Mean RTs for the spillover regions for F R E Q U E N C Y . Error bars represent confidence intervals +/ - 95% The linear model does not indicate any significant interactions due to the small differences in reading times. The fixed effects statistics are reported in Table 19. - Estimate Std. Error t value (Intercept) 844.44 58.32 14.478 MODlow 19.34 57.42 0.337 NIhigh -10.26 43.91 -0.234 NIlow -60.77 57.51 -1.057 lmer(SpilloverRegion ~ 1 + Frequency + (1|MinimalPair) + (1|Subject), REML="FALSE") Tab. 19: Fixed effects for I N T E R P R E T A T I O N and F R E Q U E N C Y In a name-informing reading, a high lexical frequency was read faster than a low frequency word while a reverse effect is reflected in the mean reading times for a modalizing interpretation. 7.2 NI vs. MOD: A self-paced reading study 161 <?page no="162"?> 7.2.6 Discussion The herewith presented self-paced reading experiment investigated whether lexical frequency and the complexity of the mentioned nominal affects the processing of name-informing and modalizing constructions in German senten‐ ces. The reading time was investigated at two different positions in the target name-mentioning construction. The basic assumption was that response times for modalizing conditions should take longer than for literal meanings. This idea is bolstered by empirical evidence from various experiments in which comprehending irony required additional processing. The main analysis for the target nominal, i.e., Reading Time 3, revealed that there was statistically no significant difference between the literal and non-literal conditions, indicating that no strong conclusion can be drawn. The mean reaction time revealed that literal statements have a tendency to take longer to be read than literal utterances when considering the target word. The effect was however not significant and there was only a small difference in the mean reading time for the two main conditions. Given the possibility of additional factors impacting the comprehension of modalizing sentences, significance testing may not always capture subtle effects. Similar findings are reported in Gibbs (1986) who provided empirical evidence for a non-significant but yet a shorter RT for sarcastic utterances. Ronderos et al. (2023) argue that irony comprehension can be as fast or even faster than the literal interpretation, which is particularly influenced by the context. In another self-paced reading experiment (see Schwoebel et al. 2000), the non-significant difference for the reading times for either ironic or literal target words was confirmed. Although reporting that ironic conditions took longer to be processed, the authors also reported a spillover effect which was shown in the reading times for the word following the target word, manifested in the reporting clause that followed the target word. Counter-evidence is e.g. presented by Giora et al. (2007) in which ironic targets took longer to read than their literal counterparts. The authors found that although reading times did not vary in the critical condition for ironic and non-ironic interpretations, a spillover effect was found for the words following the critical reading region. This means that ironic spillovers were read slower than non-ironic ones. The authors claim that this effect reflects the initial lexical access of the literal meaning with subsequent processing of the ironic content being reflected in later stages of the processing. Further empirical evidence was presented by Schwoebel et al. (2000) who found a significantly longer latencies for irony-biasing contexts as opposed to literal-biasing contexts. As this discussion is ongoing and is impacted by various 162 7 Experimental investigation I: The interpretation of NMCs <?page no="163"?> factors that are subject-specific, such as the WMC and exposure to irony in everyday communication, no clear conclusion can be drawn here. The statistical analysis for W O R D T Y P E revealed that there was no interaction for the reaction times between compounds and simple nouns. Despite this, compounds took longer to be processed as indicated by the mean reading times. The lexical frequency of the target nominal revealed a non-significant interaction between a high and low lexical frequency. Overall, a high frequency class, i.e. words with a low lexical frequency, were processed slower than words with a low frequency class. This finding is consistent with the idea that has been introduced by Monsell et al. (1989) and claims that high-frequency words are processed faster because of their familiarity and their location in the mental lexicon. In lexical decision tasks, participants responded faster to high-frequency words than to low-frequency words (cf. Gardner et al. 1987). Eye-tracking experiments have proven that the target word receives increased fixation times for low-frequency words in comparison to high-frequency words (cf. Kretschmar et al. 2015). Additionally, the word length had no significant influence in any of the analyses. Moreover, the spillover region was investigated. In contrast to the target word the mean reading time was longer for a modalizing reading. This might indicate that the modalizing interpretation was further processed in the reading region following the target word. Yet, the difference is only present in the mean reading times but not in the linear model. For the W O R D T Y P E condition, a similar finding is found: the statistic reveal no significant effects. In the last analysis, the frequency of the preceding nominal did not have an effect on the reading time in the spillover region. A recent study by Schlechtweg & Härtl (2023) has shown that the processing of irony in name-mentioning constructions is facilitated if the nominal is enclosed in quotation marks. This finding indicates that quotation marks can support ironic content and that ironic sentences might have been processed faster and probably would have resulted in stronger empirical effects when being enclosed by quotation marks. Previous studies that have investigated the processing of irony in written language have also been interested in reader-related factors influencing the processing arguing that the RTs and understanding of irony can be highly individual. One of the factors that impacts the processing of irony has been referred to as the WMC, comparing the processing of irony for participants with a high and a low WMC. People with a high WMC have been shown to “start processing the intended meaning early, during the first-pass reading of the 7.2 NI vs. MOD: A self-paced reading study 163 <?page no="164"?> phrase” (Olkoniemi et al. 2019) while people with a low WMC use compensating strategies such as look-backs to disambiguate the ironic content. Although the current study has not been controlled for reader-related factors, empirical evidence has been put forward which indicate that embedding irony in short contexts such as three to four sentences does not bring this effect up (Olkoniemi et al. 2016). Despite the fact that context is crucial in arriving at the speaker-intended interpretation, evidence from reading studies have confirmed that context had no effect on the early stages of comprehension (cf. Giora & Fein 1999; Giora 1998; Katz et al. 2004). To conclude, the reaction times indicate that (i) there are differences in the processing of literal and ironic language involving name-mentioning constructions and (ii) that the lexical frequency as well as the word type play a role in disambiguating the intended meaning. Crucially, highly frequent simple nouns facilitate the processing of irony in NMCs whereas compounds are inclined to fulfil a name-informing function in literal sentences. Given the fact that the self-paced reading study mostly reflects trends in the mean rating but only few significant effects, the methodological approach to this research question might be too sensitive for a self-paced reading design. This is additionally reflected in the results for word length, which is not reflected in the present data although it is a well-pronounced effect. It could be argued that a self-paced reading design is not suitable for this type of investigation due to the wide variety of factors and that a more fine-grained analysis is required. We can summarize the findings of the herewith presented self-paced reading study as follows: while name-informing sentences were read longer than modalizing ones, a reverse effect was detected in the spillover region in which modalizing contexts were read longer than non-ironic spillovers. 164 7 Experimental investigation I: The interpretation of NMCs <?page no="165"?> 96 Parts of this chapter were written in collaboration with Holden Härtl and Álvaro Cortés Rodríguez, and has in parts been published in Raue et al. (2025). This Chapter includes a more detailed presentation on the characteristics of eye movements during reading and provides additional statistical analyses for the Second Fixation Duration (see Section 8.4). 97 Until recently, the vast majority of empirical research on reading processes stem from native speakers of English. Studies investigating reading from a cross-linguistic perspective are still scarce (cf. Rau et al. 2015; Siegelman et al. 2022). 8 Experimental investigation II: Evidence from eye-tracking 8.1 The processing of quotation marks in German NMCs The following study aims at investigating the processing of quotation marks in German NMCs in the written mode. More specifically, the study focusses on the research question whether and how quotation marks affect the eye movements in NMCs during reading. This study builds on findings presented in the corpus study presented in Chapter 6. 96 . 8.2 Characteristics of eye movements in reading This section reviews central findings in eye-tracking research on reading and introduces the basic inventory of eye movements that are crucial for the purpose of this study. Since the first observations on eye-movements during reading in 1879 by Hering and by Lamare (Wade 2010), eye-tracking has successfully been implemented in reading research and developed into a powerful tool to study underlying cognitive processes. Eye-tracking is a methodology that allows for insights into the moment-to-moment (online) cognitive processes involved in the processing of language, i.e., during reading, and the temporal order of these processes. Basic eye movements that are measured by the eye-tracker include saccades, i.e., ballistic eye movements from one viewing position to another, and fixations, i.e., the time period or pauses between saccades in which the eyes are relatively stable on a visual input. Fixations concern the physical motion of the eyes. The basic characteristics of eye-movements during silent reading for a mean fixation duration are 225 ms with a mean saccadic size of 2, which are about 8 letters (Rayner 1998). Rayner (1978) initially reported the average typical fixation duration for skilled readers of English with 200-250 ms. 97 More recent data <?page no="166"?> 98 The mean fixation durations for oral reading, scene perception, and music reading are longer. 99 Longer regression most likely reflect problems regarding the text comprehension. 100 Parafoveal preview means that the reader is looking ahead while still fixating a specific word (see Pynte & Kennedy 2007). presented in Rayner (2009) reports fixation durations of 225-250 ms. Various studies have proven that there is a great variability in fixation duration ranging from 50 ms to 500 ms (Rayner & Pollatsek 2006), and that the average fixation duration depends on the stimuli and across tasks (Holmqvist et al. 2011). 98 Without fixating a certain word, it is typically not possible to extract the word’s meaning and information about the morphological composition of the word (Staub & Rayner 2007). A commonly applied distinction for fixations in reading research are progressive fixations, that is looking ahead or forward movements, and regressive fixations, i.e. backward eye movements. Typically, in languages with a reading direction from left to right, readers move their foveal vision area into the direction aligned with the text. For German and English there exist mostly forward saccades with an average saccade length of 7-9 letters for skilled readers (Rayner 1978). Saccades that go back to a previously fixated word or region of text are called regressions. Most frequently the eyes perform a regression from word n to n-1. 99 The length of regressions is on average relatively short, meaning that regressions commonly appear between two words (Raner & Pollatsek 1994). Out of all saccades, regressions make up approximately 10 % (Vitu & McConkie 2000). The reading process is, however, not restricted to direct fixations, i.e., the foveal vision. In the parafoveal view, surrounding words can also be processed without receiving a fixation, especially words right to the gaze position. In contrast to regressions, pre-processing happens when fixating word n while word is n+1 already pre-processed, known as the parafoveal preview effect (cf. Pynte & Kennedy 2007). Preview information on a word is obtained in reading without n+1 receiving a fixation. Crucially, the parafoveal view already gathers information regarding word length, orthography, and phonological information (Attardo & Pickering 2023). If readers have a valid preview, the following word in the reading direction have a shorter fixation time, with the preview benefit usually being 30-50 ms (Rayner 2009). Another possibility is that items that have been decoded parafoveally, i.e., outside of the foveal area, appear as skipped words in the reading data, which simply means that the word or AOI has not received a fixation. This is often the case with high frequent words such as articles, particles, and function words. 100 Skipping words only affect progressive reading. Experimental data revealed a higher skipping 166 8 Experimental investigation II: Evidence from eye-tracking <?page no="167"?> 101 McConkie et al. (1992) argue that words with a fixation of less than 140 ms are affected by lexical properties. probability for short words and words with a high lexical frequency (Engbert & Kliegl 2011). When looking at reading at the word-level, three different effects are well established in the literature. These factors are lexical influences that have been proven to have an impact on eye movements in reading. 101 The characteristics are related to processing in the foveal vision. First, the word length effect (Rayner et al. 2011) describes the finding for gaze duration and total time that longer words are processed longer than shorter words. This effect is grounded in the increased probability of refixations for longer words and also longer fixation durations of the word (cf. Rayner 2009). Word length has an influence on the word skipping probability with longer words being less likely to be skipped as opposed to shorter words: in 67 % of the time, words with three letters are skipped. In contrast, only in 20 % of the time longer words with 7-8 letters are skipped (cf. Rayner et al. 2011; Rayner & McConkie 1976). In turn, this means that words with 8 letters are almost always fixated. Second, the predictability effect describes the phenomenon that readers process words faster if these are more predictable within the context of the sentence are processed faster (Ehrlich & Rayner 1981). The fixation duration is therefore shorter if the context makes a word highly predictable. Third, the frequency effect is known as the effect that more frequently occurring words are faster processed during reading (Inhoff & Rayner 1986; Rayner & Duffy 1986). Fixation times are thus influenced by the word frequency effect (cf. Brysbaert et al. 2018b) and predictability effect. Research has shown that frequency is strongly correlated with word length of highly frequent words being shorter and vice versa (Attardo & Pickering 2023). Even when matched for word length, readers fixate for shorter durations on words with a high frequency compared to low-frequency words and for shorter durations on highly predictable words compared to low-predictability words (Rayner et al. 2011). Frequency nonetheless has a strong effect on fixation times. In addition, Rayner et al. (1996) report that the lexical frequency has an influence on word skipping: when comparing words of equal length, words with a high lexical frequency are skipped more frequently as opposed to low frequency words. Furthermore, unpredictable words or words with a low frequency generally receive multiple fixations. The individual fixations durations are however not extended (Engbert & Kliegl 2011). 8.2 Characteristics of eye movements in reading 167 <?page no="168"?> 102 Word familiarity is, in contrast to word frequency, not determined by corpus counts but from rating norms. As a consequence, two words can have the same frequency value but regarded as having a different degree of familiarity (Clifton et al. 2016). 103 Holmqvist et al. (2011) argue that attention and fixation should not be interpreted as synchronous events in referring to a study that proved attention being up to 250 ms ahead of the eye. Further influences on eye-movements include factors such as the reader’s familiarity with the word, 102 the acquisition age, the orthographic transparency, semantic and morphological properties of the word(s), as well as syntactic structures and contextual aspects (see e.g. Clifton et al. 2016, Attardo & Pickering 2023 for an overview). Within the theory of reading, the so-called eye-mind hypothesis proposed by Just and Carpenter (1982) has been highly influential. The underlying assumption of the theory defends that for the duration of the processing of a word, the eye remains fixated on this word. In simple words, the hypothesis states that fixating an item means that it is processed without delay for the time the fixation is recorded. However, research has shown that the eye-mind theory is only true to some extent. Essentially, assuming that gaze is synonymous with attention poses a challenge. 103 It is crucial to differentiate between overt atten‐ tion and covert attention. Covert attention comes into play in eye movements for planning the next saccade. While gaze correlates with attention other factors have an impact too, including the type of the stimuli, the individuals’ cognitive load, and the attention being overt or covert. Taking into consideration that visual and linguistic variables have an impact on eye movements, a weaker form of the eye-mind hypothesis should be assumed (see e.g. Reichle et al. 2009). Although there is a lag between the eyes and the mind due to preview and spillover effects, Staub and Rayner (2007) conclude that the duration readers look at a certain word can still be regarded as a good reflection of the processing time needed for this word. An updated formulation of the eye-mind theory has been proposed by Attardo and Pickering (2023) who state that “the mind tends to process what the eyes are gazing at, although there may be temporary decouplings between gaze and attention” (60). Different types of fixations have been argued to be associated with different types of cognitive processing. The first fixation on a word is considered to serve the function of lexical activation. In contrast, fixations that appear temporarily at a later point are considered to fulfil discourse integrative processes (Holmqvist et al. 2011). In addition, findings have indicated that there are “functional links between what is fixated and cognitive processing of that item - the longer the fixation the ‘deeper’ the processing” (Holmqvist et al. 2011: 382). For the purpose 168 8 Experimental investigation II: Evidence from eye-tracking <?page no="169"?> of this investigation, we follow a cognitive-linguistic position which assumes that the duration of fixating a word is influenced by contextual factors and properties of the word, meaning that the time spent processing reflects the time needed for accessing the meaning and incorporating it into the meaning of the text or sentence (see Rayner & Liversedge 2012). Opponents of the cognitive view argue for the oculomotor view which assumes a relatively independent relationship between the processing of fixation times and moment-to-moment cognitive processing. According to the oculomotor view, eye movements are a very restricted tool for the study of reading processes (Rayner & Liversedge 2012; Vitu 2012). 8.3 Empirical evidence on punctuation marks This section discusses the state of the art on the processing of punctuation marks and more specifically on quotation marks. In general, literature on the influence of punctuation marks on the reading process using eye-tracking methodology is sparse. An exception is the eye-tracking study presented in Hill & Murray (2000), who investigated the effect of commas and spaces on natural reading processes. In their study, they investigated three different sentential constructions, namely early and late closure, reduced and unreduced relatives, and NPs or VPs attached to prepositional phrases, see examples from Hill & Murray (2000: 9) repeated below. (104) a. Once the dog stopped scratching(,) the nice vet laughed out loud before sitting down. Early Closure - b. Once the dog stopped scratching the nice vet(,) he laughed out loud before sitting down. Late Closure - c. The critic(,) played the music(,) listened very atten‐ tively before saying no. Reduced Rel‐ ative - d. The critic(,) who was played the music(,) listened very attentively before saying no. Unreduced Relative - e. The vet injected the cat(,) with the collar(,) before leaving for a rather late lunch. Noun Phrase - f. The vet injected the cat(,) with the needle(,) before leaving for a rather late lunch. Verb Phrase The sentences appeared with or without a comma, grounded in the assumption that commas should facilitate processing difficulties. Based on an analysis of 8.3 Empirical evidence on punctuation marks 169 <?page no="170"?> the gaze duration on the target items, the authors conclude that commas and also spaces have an effect on text segmentation. While commas did not have an impact on the landing position and first fixation duration of a word followed by a comma, an effect was found for extended reading time and a higher proportion of fixation counts. Further empirical evidence on punctuation marks comes from Hirotani et al. (2006), revealing that commas in clause-final position extend first pass reading times on the end of the clause but facilitating the duration spend on reading the whole sentence. Punctuation marks had clear effects on three different eye-tracking measures in both investigations outlined above, namely a longer mean gaze duration and mean fixation count for punctuated words, as well as the mean landing position being shifted to the right of the word preceding a punctuation mark. In addition to eye-tracking data on commas, Yao & Scheepers (2011) present evidence from two studies on the reading of written stories in direct and indirect quotations, in which the speaking rate was modulated using several sentences to create different contexts with appropriate adverbs giving rise either to a fast or slow-speaking context (e.g. She said slowly/ quickly). In both oral reading and silent reading experiments, the context affected the reading rates. In contrast to indirect speech, direct speech was read significantly faster in fast-speaking quoted contexts in both oral and silent reading settings, which was interpreted as reflecting the vivid representation, i.e., the speed simulation, of a person’s inner speech. These findings are supported by Stites et al. (2013), who modulated the speaking rate using a single adverb in direct and indirect quotations. In their research, they observed that when readers encountered direct quotes characterized as being uttered “quickly” as opposed to “slowly”, they tended to spend less time reading them. Interestingly, these effects were not evident in a nearly identical section of text when presented as an indirect quote. Previous investigations into the nature of NMCs have used corpus data (cf. Härtl 2020, Raue 2022, Raue & Cortés Rodríguez 2022), acceptability judge‐ ments (cf. Cortés Rodríguez et al. 2022), and reading studies implementing the self-paced reading paradigm (cf. Schlechtweg & Härtl 2023; Raue 2023). Empirical evidence from the large-scale corpus study (see Chapter 6) found that name-mentioning constructions in German involving the predicate nennen (‘call’) have a significant preference for not being accompanied by quotation marks. In sentential constructions with the same predicate, a significant effect has thus been proven for a low lexical frequency of the mentioned nominal as opposed to expressions with a high lexical frequency. In addition, a recent study using the self-paced reading paradigm revealed that the processing of irony in NMCs is facilitated if the nominal is enclosed in quotation marks (Schlechtweg 170 8 Experimental investigation II: Evidence from eye-tracking <?page no="171"?> & Härtl 2023). Quotation marks have been argued to have an effect on language processing, creating a processing burden first but facilitating reading times at the end of quotation-marked target sentences, helping the reader to recognize ironic content. Despite this, the nature of the reading process involved in the processing of metalinguistic sentences and of the special type of pure quotation, has not been investigated using eye-tracking methodology. We regard this approach to NMCs as particularly beneficial to gain insights into the moment-to-moment processing with millisecond accuracy. 8.3.1 Research question and hypotheses The present study aims at investigating the processing of quotation marks in the written mode. More specifically, the study aims at answering the research question whether and how quotation marks affect the eye movements in NMCs during reading. To investigate the effect of quotes in NMCs, minimal pairs were created in German following the sentential pattern illustrated in (105). We expect a difference to arise in eye movements and fixation durations between the conditions Quotes (see 106a) and noQuotes (see 105b). (105) a. Ella erfährt, dass man dieses Gefäß „Vase“ nennt, und erzählt Niko davon. ‘Ella realizes that this jar is called “vase” and tells Niko about it.’ [Quotes] - b. Ella erfährt, dass man dieses Gefäß Vase nennt, und erzählt Niko davon. ‘Ella realizes that this jar is called vase and tells Niko about it.’ [noQuotes] Appropriate eye-measures for analysing individual words include, among others, First Fixation Duration (FFD), Gaze Duration/ Dwell Time and Single Fixation Duration (Conklin et al. 2019). For the purpose of this study, we are particularly interested in fixations and the duration of fixations as these measures are more responsive to linguistic factors (Staub & Rayner 2007; Conklin et al. 2019) as opposed to saccades. The reading measures employed in this study are commonly used reading measures related to language processing. As demonstrated by Holmqvist et al. (2011), there is a great variety in the range of terminology used by researchers for the different kinds of measures. To avoid terminological confusion, all relevant measures are defined here. For the target reading region, the following reading time measures are considered: 8.3 Empirical evidence on punctuation marks 171 <?page no="172"?> (i) the first fixation duration, (ii) the second fixation duration, (iii) the third fixation duration, (iv) interest area dwell time, (v) total number of fixations (total fixation count), and (vi) regression count. Both fixations and dwells are position measures. Fixations describe a period in which the eye is relatively static. In this time period, the foveal vision fixates a small area, so that the position of the fovea is measured by the eye tracker. Thus, fixations are the time spent between saccades. The dwell time, in contrast, is defined as the time spent within a certain AOI which includes not only the first fixation on a word but also all subsequent fixations including returns to that AOI. Consequently, the position measure for dwells is the respective AOI. While the first run dwell time sums up the duration spent in one IA before entering either to the right or left, the (total) dwell time takes into consideration the whole duration spent in one IA, including re-fixations. Different position measures for fixations, i.e. first fixation and later fixations, have been chosen for this study because the different types of fixations can reflect different cognitive processes. It is expected that a combination of investigating both early and late measures is necessary to gain a comprehensive understanding of how quotes influence reading. Generally speaking, measure‐ ments of fixations are classified into either early or late measures, reflecting different stages of reading processing. As the name indicates, early measures focus on the initial stages of processing, and “are seen primarily as a reflection of highly automatic word recognition and lexical access processes while later measures tend to reflect more conscious, controlled, strategic processes” (Con‐ klin et al. 2019: 66). While first fixation and first run dwell time clearly reflect early measures, regressions can be regarded as a late measure (Rayner 2009). For the purpose of this study, early measures can provide insight into how quotes capture the readers initial attention while late measures give insight into high-level cognitive processes. More specifically, the early measurement called First Fixation Duration, a measure that describes only the first fixation on a visual target, activates the lexical process. The dwell time measurement, in turn, has been employed as a late measure (Carter & Luke 2020; Siyanova-Chanturia et al. 2023), and has been discussed with respect to a higher informativeness and difficulties regarding the extraction of word information (Holmqvist et al. 2011). The hierarchical structure of reading can be clustered into the following four processes: (i) the perceptual level in which the raw perception of letters takes place, (ii) the stage of the lexical level, in which the reader extracts the meaning of words, followed by (iii) the propositional level where clauses and propositions are formed, and finally (iv) the situational model in which the addressee forms a representation of the mental model by taking the context 172 8 Experimental investigation II: Evidence from eye-tracking <?page no="173"?> into account (Attardo & Pickering 2023). Reading research using eye-tracking methodology has especially contributed to understanding the second and third stages of the hierarchical model. Based on empirical evidence, different models of eye control have been proposed, most notably the E-Z reading model (Reichle et al. 1998), and the SWIFT model (Engbert et al. 2002). Both models differ in their assumptions on the causes of regressions. While in the E-Z model reader’s regressions are considered to occur because of post-lexical processing, the SWIFT model attributes regressions to reflect incomplete lexical processing and no regressions due to post-lexical processing (Reichle et al. 2009). As we have explained above, the processing system is sensitive to punctuation marks such as commas. Consistent with these findings, we hypothesize (H A ) the processing system to be sensitive to the processing of quotes, reflected in extended fixation durations on the word embedded in quotes as well as a higher number of fixations on the quoted item). Based on the pragmatic approach to quotes as outlined in Schlechtweg & Härtl (2023), a difference is expected to manifest between early and late eye measurements. We hypothesize that the first fixation duration should not be sensitive to this effect as the first fixation is commonly associated with lexical access and quotes should not have an impact in this early measurement. However, we expect a difference to be manifested in later measures like the dwell time of the target IA (H B ). The last hypothesis is derived from regressive saccades to be argued to involve higher order linguistic information and may be a result of a disruptive processing (Rayner & Liversedge 2012). We therefore hypothesize a higher amount of regressions to quoted items (H C ). The hypotheses are stated out below. H A : More fixations and a longer fixation duration in the target IA H B : A significant effect in the dwell time with longer duration on quoted item H C : Higher amount of saccadic regressions from n+1 to target word n To test for these hypotheses, we conducted the eye-tracking experiment pre‐ sented below. 8.3.2 Participants Twenty-four native speakers of German participated in the eye-tracking experi‐ ment. Subjects were paid 8€ for their participation. All participants reported to have normal or corrected-to normal vision. The mean age per participant was 25 years (SD = 4.93) and a total of 24 participants took part in the study (20 8.3 Empirical evidence on punctuation marks 173 <?page no="174"?> 104 Please note that closing the eyes independently from each other caused difficulties for some participants. In this case, the right eye was tracked as the vast majority of people have ocular dominance on the right eye, see e.g. Roth et al. (2002) who report 65-70% of right eye dominance. 105 Retrieved August 12, 2023 from www.sr-research.com/ support/ thread-214.html female, 4 male). 13 people had eye dominance on the right eye and 10 on the left eye. An ethics and integrity statement was signed by all participants prior to the study. The experiment was approved by the Research Ethics Committee of the University of Kassel. 8.3.3 Apparatus All eye-movements were collected using the SR-Research EyeLink 1000 Plus (SR Research Ltd., Toronto, Ontario, Canada). The eye-tracker was running at 500 Hz sampling rate. Viewing for the participants was binocular but only the dominant eye was tracked. For the eye-tracking setup, we used a desktop mount combined with a chinrest and forehead reast. A 35 mm lens was used and the camera and illuminator were placed at a distance of approx. 57 cm between the eyes and the monitor as recommended by SR Research (User Manual 2009: 52). The monocular tracking mode was used based on the participants dominant eye, also referred to as “ocular dominance”. As argued in Hooge et al. (2019), tracking one eye should in some case be preferred over binocular tracking. Prior to the experiment, the Porta Test (also known as the Porta Dot Test) was implemented to test for eye dominance (cf. Bossi et al. 2018). For this test, the participants were instructed to extend their arms to form a rectangular shape with their hands through which they were able to see a dot glued to the wall the lab. When opening and closing the eyes in succession, the dominant eye remains aligned with the dot. 104 SR Research report a high accuracy of approx. 0.5 degrees of visual angle for the EyeLink 1000 tracker which relates to the recorded gaze position and the actual position of the eye. 105 A high accuracy of eye-tracker is especially important for reading studies because of the small stimuli, i.e., the letters, presented to the participants. Regarding the precision, EyeLink 1000 reports the value of about 0.01 degrees of visual angle which describes the consistency of the measurement. This precision has been argued the be reliable at detecting small changes in the participants gaze position. The experimental items were presented in black letters 20pt Times New Roman font and the line spacing for the target items was single. The text 174 8 Experimental investigation II: Evidence from eye-tracking <?page no="175"?> 106 The re-calibration was used as the participants were encouraged to move around in the lab and get some refreshments to avoid fatigue effects. Further, re-calibrations have the advantage of avoiding “drifts”, i.e. a deteriorating accuracy of the eye-tracker (see Attardo & Pickering 2023). 107 Filtered for during participant recruitment. resource was horizontally aligned to the left and vertically aligned to the top. The sampling rate for each trial was set to 500 Hz. The EyeLink 1000 PlusHost computer was used, which has a vertical refresh rate of at least 100 Hz with a microssacade resolution of 0.05º for a head supported mode. 8.3.4 Materials and experimental design In this section, the experimental design and material used for this study will be presented. The software used for creating and presenting the stimuli was Experiment Builder (SR Research 2023). The experimental items consisted of 80 trials out of which 60 were fillers. After randomly presenting 40 items, there was an approx. 5-minute break followed by a re-calibration of the eye-tracker. 106 The stimuli consisted of a sentence, either a target sentence of filler item, followed by a comprehension question. The comprehension question was implemented to ensure that the participants are paying attention throughout the experiment. All experimental items were divided into four groups and the design was a between-subjects design. Crucially, the position of the target stimuli was not varied across the target sentences. The order of the items was randomized for each participant. The experimental items were created in German, which was also the native language of the participants. 107 The target and filler items were created for the purpose of the study. The target items consisted of 20 minimal pairs which were manipulated by varying the presence and absence of quotation marks around the nominal in NMC constructions. All participants were exposed to a total of 10 critical items. As a consequence, participants read five target sentences that contained no quotation marks, i.e., condition noQuotes, and five constructions with quotation marks, condition Quotes. Each participant was further exposed to a total of 60 filler items, which consisted of sentences with multiple sluicing and gendered terms. Crucially, none of the filler items contained quotation marks. The comprehension of the sentence was controlled for by asking a yes-no comprehension question after each sentence which required a button press for confirmation. All target nominals were controlled for lexical frequency such that the frequency classes ranged between 11 and 22, here considered as a low lexical frequency (see Appendix 1 for an overview of the target nominals with their 8.3 Empirical evidence on punctuation marks 175 <?page no="176"?> 108 Please note that instead of using syllables, other researchers might use (average) word length based on letters of the respective word(s), cf. Eskenazi 2023. This calculation is based on the word frequency effect reported in Rayner et al. (2011), in which target words were clustered into 3 different types of length (short: 4-5 letters, medium: 7-9 letters, and long: 10-12 letters. Word length has been reported as influencing the fixation duration and word skipping probability (Rayner et al. 2011). 109 The average word length is commonly calculated by dividing the number of letters by the number of words. 110 The categories are used following Rayner et al. (2011) with the only exception that long words ranged between 10-12 letters in their study. word length and lexical frequency). Regarding the word length, the amount of syllables ranged between 2-4 syllables per word. 108 The average word length was 8.35 letters (SD=2.39). 109 The mean lexical frequency ranged at 15.55 (SD=2.85), and the mean number of syllables was 2.6 (SD=0.58). Target nominals were grouped into the following categories: short (4-6 letters), medium (7-9 letters), and long (10-13 letters). 110 The location of the target nominal was consistent across all critical items and placed sentence-medially as illustrated in the following minimal pair. (106) a. Kim weiß, dass man dieses Verfahren „Neuwahl“ nennt, und belehrt Anna darüber. ‘Kim knows that this procedure is called “re-election” and informs Anna about this.’ [Quotes] - b. Kim weiß, dass man dieses Verfahren Neuwahl nennt, und belehrt Anna darüber. ‘Kim knows that this procedure is called re-election and informs Anna about this.’ [noQuotes] In contrast to the eye-tracking experiment on punctuated and unpunctuated sentences with comma presented in Hill & Murray (2000), we decided not to add any additional spacing for the conditions noQuotes because this would have resulted in a non-naturalistic representation of the sentence on the screen. The eye-tracking during reading paradigm was implemented for this study, a method that has been proven to be useful for investigating both sentence-level and word-level phenomena. 176 8 Experimental investigation II: Evidence from eye-tracking <?page no="177"?> 111 Please note that this time is highly individual and depends on (i) the participant setup in front of the eye-tracker prior to the experiment, (ii) the duration of the calibration and validation process, (iii) the length of the mid-experiment break. 112 Retrieved August 14, 2023 from www.sr-research.com/ support/ showthread.php? tid=2 44 8.3.5 Procedure and analysis The experiment lasted between 30 to 45 minutes. 111 Prior to the experiment, participants were asked to sign the declaration of confirmation. All participants wearing glasses cleaned their glasses in advance to avoid tracking issues. The lids for the windows were closed to ensure that no sunlight would interfere with the infra-red light of the eye-tracker and the lights were turned off at least one minute before the experiment started. Then, it was ensured that the subjects were comfortably seated in front of a computer. In order to minimize head movements, the experiment was conducted using solely the so-called head-fixed mode, which is characterized by using a chinrest and forehead rest. Prior to the experiment, it was ensured that there were no distractions in the peripheral field. All phones were silenced beforehand, and it was ensured that the participants were feeling comfortable and not distracted before starting the experiment. Participants were encouraged to read the sentences at their normal speed, and they were told to read all sentences silently. The experiment started with a welcome slide with information on the experiment duration and a brief description of the study’s aim. After pressing the enter key, participants read a slide with instructions concerning the experiment in which they were informed that the study aims at investigating the under‐ standing of sentences (see Appendix A1). The experiment started with questions regarding their gender, handedness, and dominant eye to let them settle in and feel relaxed during the experiment. Afterwards, a 9-point-calibration, which is recommended for a head-fixed mode, was performed followed by a 9-point-validation. In case that the calibration was asymmetrical, i.e., a poor calibration, we either used a manual calibration instead of an automatic or the calibration was re-done to ensure data quality. The validation was only accepted once it was marked as a “good” validation, which means that the worst point error was < 1.5° with an average error of < 1.0°. 112 Another calibration and validation procedure was used after participants read half of the items as multiple calibrations are necessary to ensure data quality. In the next step, participants continued with a practice trial which consisted of 4 sentences including feedback on selecting the answer to the comprehension question. The first four sentences were presented to familiarize the participants 8.3 Empirical evidence on punctuation marks 177 <?page no="178"?> with the procedure and were not related to the target items. Participants were allowed to ask questions, if necessary, and received feedback on correct or incorrect answers to the comprehension question. After presenting another slide announcing the start of the experiment and reminding them that no questions could be answered in between, the experiment started with a key press. A trial always started with a fixation cross on the screen which was aligned to the beginning of the sentence and presented for 2 seconds. Implementing a fixation cross as a noise display is especially important for short trials in reading experiments as these may have an impact on the average first fixation duration and the last fixation of the previously presented trial can spillover to the first fixation in the following trial (cf. Holmqvist et al. 2011). In order to analyze the data, Interest Areas (IA), also known as Areas of Interest (AOI) or Regions of Interest (ROI), were created around each word of the target sentence, as displayed down below. Tom erfährt, dass man diese Krankheit Tom learns that one this disease “Sepsis” nennt, und erzählt Anna davon. “sepsis” call and tells Anna about it. To create IAs, the default parser was used in Experiment Builder. A total of 12 IAs were created for each target sentence. The IA templates with the segmentation information was constant for each condition. Importantly, the variable of interest, i.e., the target nominal, was placed sentence-medially in order to be able to take delayed effects of the processing of quotes into consideration (see also Schlechtweg & Härtl 2023), and to avoid the wrap-up effects (see Hirotani et al. 2006; Conklin et al. 2019). 8.4 Data cleaning and results Before moving on to reporting the results of the eye-tracking experiment, the data processing will be described, using data cleaning methods and details regarding outlier removal. Data Viewer (SR Research 2023) was used for visual inspection and pre-processing of the data as a preparation for data analysis. For the analysis of the sentences of interest, i.e. the critical conditions Quotes and noQuotes, all trials with a wrong answer to the comprehension question were 178 8 Experimental investigation II: Evidence from eye-tracking <?page no="179"?> 113 When the inspector window in Data Viewer did not show any fixations or only 2 fixations per sentence, known as track loss, the data was removed. 114 Manually correcting fixations in reading experiments is a common practice (cf. Cohen 2013) but should never been done horizontally. 115 A total of 26 individual trials showed misaligned fixations which would have resulted in about 6.42% of data loss. 116 Translation: ‘Luisa learns that this process is called nuclear fusion and informs Luis about it.’ 117 Fixations are indicated in blue colour and saccades in green. As indicated by the arrows, the reading direction is from left-to-right sequence. removed. Filtering for correct answers only resulted in a total of 18 incorrect answers out of the 480 critical items, meaning that 3.75% of the data were removal. In the next step of the analysis, the 462 items were visually inspected trial-by-trial and data points with no or only few fixations and saccades were removed. 113 A total of 27 critical items were removed, out of which 19 belonged to one recording session. As a consequence, a total of 435 critical items are left for the analysis. After applying templates across all conditions, fixation adjustments were made using the drift correction in Data Viewer if possible. 114 Vertically mis‐ aligned fixations are in principle correctable for single-line texts with minor adjustments on the y-axis allowing for taking into consideration more trials in the analysis. 115 A total of 26 items showed vertically misaligned fixations out of which 20 were corrected and 6 discarded for further analysis because of severe vertical drift that could not be corrected in Data Viewer. Fig. 20: Sample 116 of vertically misaligned data 117 | Translation: ‘Luisa learns that this process is called nuclear fusion, and she telly Luis about it.’ Data cleaning methods for fixations in reading research usually include fixation merging methods, temporal cut-offs, and outlier removal. The data files were analysed following the most commonly used data cleaning method, which has been described as the baseline approach by Eskenazi (2023). This approach is 8.4 Data cleaning and results 179 <?page no="180"?> 118 As the analysis of three different cleaning method approaches by Eskenazi (2023) for the measure first fixation duration did not indicate a major impact, the baseline approach was chosen for this analysis. 119 Participant idiosyncrasy is reflected in the variety of average fixation duration across participants (Holmqvist et al. 2011). characterized by using no fixation merging techniques, temporal cut offs or outlier treatments. 118 The following overlay image illustrate how the final data set is visualized containing fixations. Fig. 21: Overlay images from EyeLink Data Viewer | Translation: ‘Leonie knows that this procedure is called blood transfusion, and she teaches Manuel about it.’ The fixations on this passage are illustrated as blue circles with larger circles indicating longer fixations on the word. In order to reduce the impact of idiosyncratic factors and explain the participant’s 119 and item’s effect and contribution to the variance, a multilevel modelling was required as the statistical approach. For each eye-tracking measure reported below, separate models were built. The eye-tracking measures constituted the dependent variable and the condition with the two levels Quotes and noQuotes as well as the word length were the independent variables. Participant and items were entered as random intercepts to the model. Overall, the mean count of fixations for the target IA was 2.35 for noQM and 2.78 for Quotes, revealing a slightly higher count for conditions with quotation marks. The distribution of the amount of fixations for Quotes and noQuotes by fixations are presented below. 180 8 Experimental investigation II: Evidence from eye-tracking <?page no="181"?> Fig. 22: Total amount of fixation durations by condition on the target item In addition, regressions from a higher IA into the target IA, i.e. regressions from right to left, were analysed. For the target sentences these regressions were performed from word “nennt” back to the target nominal. The regression-in count revealed that the highest number of regressions were 3 into the target IA. Notably, all words receiving a third regression were enclosed in quotes. In 152 cases, the target IA received a regression while 241 received no regression. When considering solely the instances with regressions, there were 88 regressions for Quotes and 64 regressions for noQuotes. The analysis showed a statistically significant effect between the two conditions (χ²(1) = 8.6879, p = .003) with more regression to nominals enclosed in quotes. This indicates that conditions with quotes lead to a higher amount of regressions. For the statistical analysis of the fixation durations, i.e., the dependent variables First Fixation Duration, Second Fixation Duration, First Run Dwell Time, and IA dwell time, linear mixed effects models (LMEM) were implemented using the lmer function of the lme4 package (Bates et al. 2015) in the statistics software R (R Core Team 2024). For each of the continuous variables, we used nested models to obtain the best fitting. The variables Q U O T E S and W O R D L E N G TH were entered in the model as fixed effects. For subjects and items, random effects and random slopes were also entered into the model. Models were selected based on the lowest AIC value (Akaike 1974), which indicates a good balance between model fit and complexity (see Venables & Ripley 2002), the best fitting model was determined following Matuschek et al.’s (2017) approach, whereby optimal random effect structure by including them only when they enhanced the model's goodness of fit. If the model 8.4 Data cleaning and results 181 <?page no="182"?> did not converge, we followed Barr et al. (2013) by systematically eliminating random slopes by participant or item that accounted for the least. Only the models with the best fit were selected and are reported here. In the following, the eye movement measures are reported in the order of the time of processing during reading, starting with the descriptive statistics and the model results. First Fixation Duration For the early measure of the FFD, there was a skipping rate of 2.76% for the target item (see column ‘Missing Value’ in Table 20): six skips for short words, three skips for words with medium length, and three skips for long words. Crucially, all skips happened exclusively in the noQuotes condition. Condition N Mean Variance Standard Deviation Missing Values Quotes 213 252.8592 7679.829 87.63463 0 noQuotes 222 265.5429 8303.379 91.12288 12 Tab. 20: Descriptive statistics for first fixation duration 15All missing values were excluded prior to the statistical analysis. By definition, skipped items have a first fixation of 0 ms which would be problematic if included in further analyses (see Conklin et al. 2019). The mean duration of the FFD measure obtained for each corresponding word length is illustrated in Figure 23. Fig. 23: Mean first fixation durations and standard error per condition 182 8 Experimental investigation II: Evidence from eye-tracking <?page no="183"?> The inferential statistics with the best fitting LMEM and its corresponding formula are given in Table-21. - Estimate Std. Error df t value Pr(>|t|) (Intercept) 243.344 12.346 37.580 19.710 <2e -16 *** Q U O T E S Quotes -10.977 8.009 385.015 -1.371 0.1713 W O R D L E N G T H me‐ dium 29.709 13.668 20.007 2.174 0.0419 * W O R D L E N G T H long 32.600 13.743 20.416 2.372 0.0276 * a The significance levels reported here are the following: p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***. Formula: lmer(FFD ~ 1+ Q U O T E S + W O R D L E N G T H + (1 | Subject) + (1 | Item), REML= “FALSE”) Tab. 21: Summary of statistical analysis for FFD a Even though descriptively shorter first fixation durations were obtained for target items with quotes, the result of the models show a non-significant difference between the conditions Quotes and noQuotes. The significant effects in W O R D L E N G TH in Table 21 reveal that first fixation durations were significantly longer for medium and long words compared to short words. In order to obtain the difference between the levels medium and long, we relevel this factor and set the level long as the baseline. This model shows a non-significant difference between W O R D L E N G TH medium and long (t = -0.22; p = .828). No interaction was obtained between the experimental factors. Second Fixation Duration The descriptive statistics for the conditions of the second fixation duration in the target IA are given in Table 22 and the mean values for SFD are plotted in Figure 24. Condition N Mean Variance Standard Deviation Missing Values Quotes 213 223.5939 13027.88 114.1397 48 noQuotes 222 233.9533 14546.72 120.6098 72 Tab. 22: Descriptive statistics for second fixation duration 8.4 Data cleaning and results 183 <?page no="184"?> Fig. 24: Mean second fixation durations and standard error per condition The results of the LMEM with the best fit for the variable SFD are presented in Table-23. 184 8 Experimental investigation II: Evidence from eye-tracking <?page no="185"?> Estimate Std. Error df tvalue Pr(>|t|) (Intercept) 227.72 15.21 28.04 14.97 < .001 *** Q U O T E S NoQuotes 4.02 6.46 287.85 0.62 0.534 W O R D L E N G T H me‐ dium 2.12 18.43 21.03 0.12 0.909 W O R D L E N G T H long 0.83 18.23 19.87 0.05 0.964 a The significance levels reported here are the following: p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***. Formula: lmer(SFD ~ 1+ Q U O T E S + W O R D L E N G T H + (1 | Subject) + (1 | Item), REML= “FALSE”) Tab. 23: Summary of statistical analysis for FFD a The W O R D L E N G TH factor was releveled to set short as the baseline. The results reveal no significant effect for W O R D L E N G TH in the SFD measurement (t = -0.22, p = .828). The QM condition receives shorter mean fixation durations for the W O R D L E N G TH short and medium but a longer mean fixation duration in W O R D L E N G TH long, see also Figure 24. Despite this, the LMEM did not reveal any significant differences in the duration of the SFD in the target IA. First Run Dwell Time The descriptive statistics for the conditions of the first pass dwell time in the target IA are given in Table 24 and the mean values per condition for FRD are plotted in Figure 25. Condition N Mean Variance Standard Deviation Missing Values Quotes 213 377.6761 40399.84 200.9971 0 noQuotes 222 393.5524 77405.22 278.2179 12 Tab. 24: Descriptive statistics for first run dwell time 8.4 Data cleaning and results 185 <?page no="186"?> Fig. 25: Mean first run dwell times and standard error per condition The results of the LMEM with the best fit for the variable FRD are presented in Table-25. - Estimate Std. Error df t value Pr(>|t|) (Intercept) 274.82 51.29 26.45 5.358 1.24e-05 *** Q U O T E S Quotes -5.28 17.92 382.65 -0.295 0.76848 W O R D L E N G T H me‐ dium 118.75 64.52 19.75 1.840 0.08079 W O R D L E N G T H long 204.20 64.61 19.85 3.161 0.00495 ** a The significance levels reported here are the following: p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***. Formula: lmer(FRD ~ 1 + Q U O T E S + W O R D L E N G T H + (1 | Subject) + (1 + | Item), REML="FALSE") Tab. 25: Summary of statistical analysis for FRD a These results reveal a significant effect for W O R D L E N G TH between long and short words and a marginal effect between medium and short words. To compare the levels medium and long, we relevel the factor, setting long as the baseline. The model results indicate no significant differ-ence between medium and long words (t = -1.377; p = 0.184). The noQuotes condition receives longer mean 186 8 Experimental investigation II: Evidence from eye-tracking <?page no="187"?> fixation times as can be seen in Figure 25, however, for the factor Q U O T E S the linear mixed effects model did not reveal any significant differences in the duration of the first run dwell time in the target IA. Lastly, no interaction was obtained between the experimental factors. Interest Area Dwell Time The descriptive statistics for the IA dwell time are given in Table 26 and the mean dwell times are presented in Figure 26. Condition N Mean Variance StandardDeviation Quotes 213 660.2676 154144.5 392.6124 noQuotes 222 598.3468 184644.5 429.7028 Tab. 26: Descriptive statistics for dwell time Fig. 26: Mean dwell times and standard error per condition In contrast to the measures explored beforehand, sentences in the Quotes condition have a longer mean duration in the dwell time measure. The model with the best fit for the dwell time in the Interest Area is given in Table-27. - Estimate Std. Error df t value Pr(>|t|) (Intercept) 402.75 82.72 33.76 4.869 2.58 e-05 *** Q U O T E S Quotes 70.90 28.15 393.39 2.518 0.01220 * 8.4 Data cleaning and results 187 <?page no="188"?> W O R D L E N G T H me‐ dium 210.79 94.12 19.20 2.239 0.03715 * W O R D L E N G T H long 340.00 94.26 19.32 3.607 0.00184 ** a The significance levels reported here are the following: p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***. Formula: Dwell_Time ~ 1 + Q U O T E S + W O R D L E N G T H + (1 + | Item) + (1 + Subject) Tab. 27: Summary of statistical analysis for dwell time a The statistical model for the dwell time reveals a significant effect for Q U O T E S , whereby the Quotes condition received significantly longer dwell time than the noQuotes condition (see Figure 26). Similar to the results reported in previous models, the main effects obtained for W O R D L E N G TH showed a significant difference between long and short words as well as between medium and short words, whereby items with less lengthy words received shorter dwell times. To examine the difference between medium and long levels, we releveled the factor by setting long as the baseline. The model results show no significant difference between medium and long words (t = -1.426; p = .170). The model showed no interaction between the factors. The results for the early measurements (i.e., first fixation duration, second fixation duration, and first run dwell time) will be discussed together with the results obtained for the late measurement (i.e., total dwell time at the IA) in the following section. 8.5 Discussion The aim of this investigation was to provide empirical evidence for reading dif‐ ferences between sentences with and without quotes. In the present eye-track‐ ing study on the reading of NMCs, we have observed that in the first fixation only 12 out of 435 target IAs are not fixated and that crucially all the skipped items were of the noQuotes condition. The skipping rate is usually influenced by visual factors and linguistic information like length, frequency, and lexical status (Conklin et al. 2019). Although words with only 4 letters were included as experimental items, which are frequently skipped in reading - more specifically not fixated but only included in the parafoveal view - they were never skipped when surrounded by quotes. We therefore accept H A in claiming that the target IA receives more fixations. This indicates that quotes make a substantial contribution to the understanding of the sentence. 188 8 Experimental investigation II: Evidence from eye-tracking <?page no="189"?> Moreover, a significant effect materialized in the dwell time measurement. In contrast to the FFD and FRD, where the noQuotes condition had longer, though not statistically significant, duration, the dwell time measure reveals that quoted nominals are fixated on significantly longer than non-quoted nominals. This finding is supported by a reading-time experiment reported in Schlechtweg (2022), who found longer reading times for quoted nominals as opposed to non-quoted, yet the effect obtained in the present experiment is not statistically significant. The significantly longer fixation duration for the Quotes condition in the dwell time leads us to accept H B . Quotes activate processing in later stages, which becomes evident from the contrast between early and late measures. We interpret this effect as reflecting a top-down process typically associated with higher cognitive discourse-based functions (e.g., Rayner & Pollatsek 1994; Orquin & Mueller Loose 2013; Orquin & Holmqvist 2018). Top-down processes mean that the interpretation of the sentence relies on higher cognitive functions, such as discourse processing or contextual understanding. This interpretation is compatible with a pragmatic account of pure quotation (Schlechtweg & Härtl 2023). In addition, the significant difference for the lexical fac-tor word length found in all three measurements with varying significant effects is reflected in the word length effect, which implies that longer words generally take longer to process than shorter words. While this effect is commonly attributed to several cognitive mechanisms, including visual and lexical processing demands, we argue that the familiarity of the word and its semantic complexity have further impacted the variability across measures in the data. We consider the word length effect to be independent of NMC constructions but to be present in other types of quotation as well. The higher percentage of re-fixations on the target nominal corresponds with the increased number of regressions. This leads us to accept hypothesis H C due to the higher number of saccadic regressions from n+1 to the target word n in the Quotes condition as opposed to noQuotes. We argue that this finding is consistent with proposals within the EZ-Reader framework, based on the assumption that post-lexical integration failures can trigger regressions, given that between-word regressions, i.e., short, regressions are considered to be a consequence of higher-level language processing (Reichle et al. 2009). The obtained results suggest that the presence of quotation marks introduces addi‐ tional cognitive demands during reading, disrupting the post-lexical integration process and giving rise to saccadic regressions. In summary, the data obtained from our eye-tracking study confirm the effects associated with quotes in fixation count, dwell time, and regressions. In comparison to FFDs, regressions and dwell time are reflective of later, i.e., 8.5 Discussion 189 <?page no="190"?> post-lexical, stages of the reading process. The variance observed in dwell time implies that quotation marks introduce processing effects in the later stages of reading, indicative of higher cognitive processes. During earlier stages, no effects attributable to the presence of quotes were discerned. We take the lack of early-stage effects as evidence that participants tended to overlook the influence of quotes during their initial processing. We believe that our findings lend support for the view that quotes are represented and processed as part of the pragmatic domain of language and not as part of the compositional representa‐ tion. The present investigation therefore aligns with previous evidence from Schlechtweg & Härtl (2020, 2023) which also supports a pragmatic approach to the study of quotation. We conclude that the quotational meanings are construed pragmatically and are processed in the discourse-integrative stages during the reading process. To the best of our knowledge, this study represents the first of its kind in this field, providing insights into the cognitive processing of pure quotations. Future research should address the processing of a variety of quotation types, contrasting pure quotes with scare and mixed quotes, and investigate the processing of quotation (marks) from a cross-linguistic perspective. 190 8 Experimental investigation II: Evidence from eye-tracking <?page no="191"?> 9 Conclusion This dissertation has focused on a special type of quotational construction that involves naming predicates such as call and refer to as. The central aim of this investigation was to investigate the metalinguistic uses of names from a theoretical and empirical angle, focussing on names used for lexicalized concepts. This section provides a summary of the main findings, discusses practical applications, and points out open research questions. Investigating the nature of mentioning constructions, a variety of different methodological approaches has been applied. In the following, the research questions spelled out in chapter 2.5 will be addressed in connecting theoretical approaches to the empirical data obtained from the different studies. A summarizing answer will be given to the research questions. The research questions (RQs) are repeated below. RQ1: What types of NMCs exist in German and in English? RQ2: What linguistic properties, including word type, lexical frequency, and morphological type, do mentioned expressions have? RQ3: What is the systematicity in the use of quotes? Is their use affected by the linguistic properties of mentioned expressions? From an empirical perspective, the first three questions can be answered by looking at the large-scale corpus study on German and English NMCs, which explored the lexical inventory used to refer to the name of a concept. The experimental results show a wide variety of sentential constructions for the naming predicates and their different semantic uses leading to distinct readings of the verbs. The syntactic variety of NMCs has been addressed in Chapter 5 and becomes evident when considering the different sample sentences. Etymo‐ logically, metalinguistic predicates have a long tradition within the language system where one of their arguments is a name that refers to a lexicalized concept. To illustrate, the predicate call has been documented to date back to at least the mid-13 th century. Predicates that occur in metalinguistic sentences are not restricted to the predicates nennen, call, refer, and bezeichnen but might for instance be replaced with speak of or name. Their cross-linguistic appearance, highly frequent use in everyday communication, and rich collocational patterns confirm their importance within the language system. Metalinguistic predicates have been analysed following a copula based approach (cf. Härtl 2020), meaning <?page no="192"?> that NMCs involve either an identificational copula or an equative copula. The corpus study revealed that NMCs are not restricted to a specific register as reflected in the sources of the corpus data, although a preference for encyclo‐ paedia entries and educational resources can be detected due to their inherent function of explaining a wide variety of different concepts and defining them. In the corpus data, the predicate with the highest amount of valid instances was bezeichnen with approx. 24 % followed by nennen with 18 % and call with about 17 % while refer to as had the lowest percentage of valid instances with only 5 %. While refer to as surprisingly stands out with the lowest number of hits, the German counterpart bezeichnen was the most frequently used metalinguistic predicate to introduce a concept. An analysis of the word type for the mentioned nominal in NMCs revealed a clear preference for compound words in combination with nennen, bezeichnen als, and the English predicate call. The lexical frequency for all predicates had a tendency for a low lexical frequency with several ad-hoc formations present in NMCs. When clustering the frequency classes into the categories low and high, a statistically significant difference appears between the two categories. A low lexical frequency is significantly more common than a high lexical frequency. Due to the deviating pattern found for refer, which might be related to its status as a phrasal verb, it is of interest to further investigate possible differences between refer and call from the perspective of cognitive processing and the relation to further metalinguistic predicates as listed in Chapter 5.5. Notably, the corpus study revealed that the target nominal was rarely an adjective-noun compound but preferentially a noun-noun compound. This finding contradicted previous assumptions as formulated, e.g., in Härtl (2016), Kotowski et al. (2014), and Schlechtweg & Härtl (2016). Adjective-noun compounds were expected to be present in the corpus data and the low numbers should be taken into account when considering the naming status of compound words as adjective-noun compounds have been argued to function as naming units (cf. Schlechtweg 2018). When investigating the occurrence and absence of quotation marks around the nominal nennen, the English equivalent call and bezeichnen, a statistically significant effect for no quotation marks was detected. Mentioned nominals were most commonly not highlighted by quotes. Quotation marks are often used in NMCs constructions, and they interact with the item’s lexical frequency: specifically, the analysis showed significant differences between instances with quotation marks between a high and low lexical frequency, meaning that there are significantly more instances with a low lexical frequency. For refer, however, there was a reversed effect proven with a significant effect for quotes. 192 9 Conclusion <?page no="193"?> To summarize, the empirically-based understanding of the mechanisms that govern the metalinguistic reference to names of concepts obtained from this investigation revealed a preference for compound words with a one-word status and predominantly a low lexical frequency. Having addressed RQs 1-3, the next two RQs will be approached. RQ4: How are expressions mentioned in modalizing contexts in contrast to those in name-informing contexts processed? RQ5: How does a modalized context influence the processing of lexicalized items as opposed to name-informing contexts? RQs 4-5 have been addressed in a rating study as well as in a self-paced reading study to examine how expressions mentioned in modalizing contexts contrast with those in name-informing contexts. The argumentation is based on the observation that metalinguistic predicates in NMCs can give rise either to a name-informing, i.e., literal, interpretation of the quoted nominal or a distancing, iconic reading. The rating study (see Chapter 7.1) build upon the valid experimental items extracted from the corpus study on German data, involving nennen and be‐ zeichnen. Both constructions predominantly elicited a preference for a literal, name-informing reading over a modalizing one. This preference aligns with the primary function of metalinguistic predicates, which is to introduce a new term. Modalizing uses do not occur as frequently as a literal interpretation, reflecting the ability of the predicate to give rise to an ironic meaning but without a frequent implementation of this use in the corpus data. This discrepancy might also be attributed to the register used in the corpus data, which was frequently a formal register as it can be found in newspaper articles and encyclopaedic online articles. Another explanation for this deviation is a possible difference in the use of NMC in oral and written speech. Modalizing uses might be found more frequently in oral, informal communications instead of in written online formats. While the corpus data consists predominantly of written text of formal register, this effect might be explained for. Moreover, a strong effect for compound words instead of simple nouns was proven for both predicates, aligning with previous assumption. This finding confirms the difference in acceptability of functioning as a naming device between the two word types: compounds inherently function well to refer to names. Despite these findings, interactions between word type and lexical frequency were not present. 9 Conclusion 193 <?page no="194"?> Due to the high number of name-informing occurrences in the corpus data, it can be concluded that readers would expect metalinguistic predicates to trigger a literal reading in NMCs, which should be processed faster than a modalizing reading. To test this hypothesis further, the self-paced reading study (see Chapter 7.2) investigated the interplay between word type and lexical frequency by interpretation. For the target word, although there was no significant effect found for lexical frequency, a low frequency had generally shorter reading times as opposed to a high frequency. A statistically significant effect was detected for a modalizing reading between simple nouns and compounds with the latter receiving significantly slower reading times. In a name-informing reading, the effect was too small to reach significance but still simple nouns were read faster than compounds. The results for word type are consistent with the predicted reading times. Concerning the spillover region following the target word, mean reading times were faster for a name-informing interpretation as compared to a modalizing reading. No significant effects were observed regarding the frequency and word type. The data obtained from the two studies does not provide a clear indication of the impact of lexical frequency on the interpretation of NMCs, indicating the need for further research with meticulously designed materials to delve deeper into this issue. The limitations of the studies can be related to the number of test participants or confounding variables like the experimental items. Investigations in future research could include a discussion of potential differences between oral and written language usage, and the impact of register on the use of NMCs. Additionally, expanding the research to include data from English speakers would enrich our understanding of the distinct interpretations of NMCs and expand the understanding of the multifaceted interplay of factors involved in their interpretation. Having gained insight into the occurrence of quotes from the corpus study, a considerable amount of sentential constructions was shown to be accompanied by quotes despite the wide variety of delineating mentioned language. While no systematicity that could be attributed to linguistic factors of the construction was detected, the impact of quotes on cognitive mechanisms and reading processes remained an open question. To get insights into the online processes, the following RQ was formulated. RQ6: How are written quotes processed during online reading? 194 9 Conclusion <?page no="195"?> The last RQ has been addressed using eye-tracking methodology to gain insights into temporal processes. The experimental design used name-informing con‐ structions in which the nominal was either written with or without quotation marks. The statistical exploration revealed a difference between early and late eye-tracking measures. Early measures included the First Fixation Duration, Second Fixation Duration, and First-Past Reading Time while the Dwell Time for the target Interest Area was considered to reflect later stages in the reading process. Only in the late eye-tracking measure Dwell Time an effect was found with a significantly longer fixation duration in the condition with quotes. In other words, the eyes of the participants were fixated for a longer period of time on words with quotation marks as compared to words without quota‐ tion marks in NMCs. This finding was interpreted as reflecting higher-order cognitive processes due to the late effects (Raue et al. 2025). This argumentation is based on the assumption that the cognitive processing of semantic content happens earlier than pragmatic processing. Counter evidence for this claim comes for example from an ERP study which argues that pragmatic processing can be detected at the same time or even before semantic information is processed (Egorova et al. 2013). In addition, regressions from the following word were considered in the analysis. Significantly more regressions back to the target word were detected when this word was accompanied by quotes. Word enclosed in quotation marks were furthermore never skipped during reading but word skipping occurred in the non-quoted sentences. The last RQ can be answered combining different studies in connection with theoretical approaches to the theory of quotation. RQ7: Where are quotational constructions located on the semantics-prag‐ matics continuum? The question outlined in 7 is closely related to two further questions: (i) how exactly can we define semantics and pragmatics and (ii) do we adapt a distinction or interface view for the two linguistic phenomena? Answering the first question is beyond the scope of this investigation. As argued in Chapter 2.4, theoretical approaches to quotation have proposed semantic and pragmatic accounts to quotation. Central questions have been how quotations refer and about what they refer (Cappelen et al. 2023). On the one side, semantic analyses argue that quotes have truth-conditional effects, meaning that the quotes content is truth-conditionally relevant (Predelli 2003). On the other side, pragmatic theories argue that the manifestation of quotes are not part 9 Conclusion 195 <?page no="196"?> of the quotational constructions compositional semantic representation (Härtl & Schlechtweg 2023). The lack of consensus on theories of quotation is not restricted to the individual weaknesses of the theories but also related to the inability of any theory to account for all the different types of quotation at once. Despite this, an interface view, which assumes semantics and pragmatics to overlap partly, has been applied by Härtl & Schlechtweg (2023) and was also applied in this thesis for the analysis of NMCs. The interface view is advantageous as it acknowledges an interaction between the two domains, which allows for incorporating both semantic and non-semantic content in quotational constructions. In an attempt to enrich theoretical considerations on quotation with empiri‐ cally based data, this thesis has put forward empirical evidence on NMCs from eye-tracking data. The findings indicate that there are temporal differences in the processing of constructions with quotation marks around the target nominal. To briefly repeat the main findings, the first fixation duration and first-past dwell time, which are oftentimes argued to constitute early measure‐ ments, did not show any effects. In later processing windows like the dwell time, a significant effect was found with a faster processing of nominals without quotation marks. These findings were interpreted as reflecting higher cognitive processes which are based on integrative processes, i.e., pragmatic instead of semantic processing. This argumentation is based on the assumption that semantic processing precedes pragmatic processing. The empirical evidence from eye-tracking therefore suggests that quotation marks lead to processing differences, and that they serve a specific function and should not be considered redundant in NMCs. Instead, it can be argued that quotation marks are optional in written language but not redundant, as they impact cognitive processing (see also RQ4 & 5). Commonly, the optionality of quotes has served as an argument in favour of a pragmatic approach. This is based on the idea that quotes are only a punctuation device that are not an essential part of the utterance (Washington 1992). For instance, contextual clues might be sufficient to construe the quotational meaning (De Brabanter 2023). Within the context of NMCs, contextual cues can for example be a context that requires the speaker to explain the name of a concept or a person asking for a clarification of a specific term. In these contextual settings the use of metalanguage is expected, which might facilitate a quotational interpretation of the utterance. However, it is still an open question whether and how metalinguistic predicates facilitate the interpretation of mentioned language, and how the word order (SVO or SOV order in German sentences) impacts the reading process. 196 9 Conclusion <?page no="197"?> 120 It is still an open research question whether further empirical evidence on NMCs, most notably manipulating contextual cues, graphemic markers, and alternated verb position support the argumentation that quotational readings are pragmatically construed instead of semantically encoded. The interaction between semantic and pragmatic information is a pivotal aspect of the comprehension process, particularly in the context of quotations. This is especially relevant given that the framework utilized in the eye-tracking study aligns with a pragmatic approach to quotation as advocated for by De Brabanter. 120 While the primary focus of this investigation was not on formal semantic-pragmatic aspects, it is essential to discuss aspects that contribute for a more comprehensive perspective on how these findings inform our understanding of the compositional processes underlying quotational mean‐ ing. The phenomenon of quotation presents challenges to the principle of compositionality, a cornerstone of semantic theory. For instance, Bazzoni (2016) argues that standard compositional semantics struggles to accommodate pure quotations due to their distinctive semantic and syntactic properties, suggesting to treat quoted items as semantic entities rather than syntactic ones. Pagin and Westerståhl (2010) propose a notion of “general compositionality” that considers linguistic context, yet they acknowledge a limitation: this framework applies primarily to written language, where quotations function as syntactic operators signalling a shift in context. In contrast, Rabern (2023) introduces the concept of a QUOTE operator, which adheres to principles of constituency, compositionality, and synonymy. The empirical evidence obtained from the corpus study revealed that the status of a mentioning use is often indicated through naming predicates, register, and the contextual embedding of the utterance. In written language quotation marks serve as an optional tool rather than a mandatory one while the absence of quotation marks does not lead to ungrammaticality in NMCs, supporting a pragmatic view. Moreover, sentences with quotation marks are associated with discourse based cognitive functions during reading, which adds empirical evidence to the argumentation that quotations should not be regarded as a (purely) semantic phenomenon. While the meaning of the quotational construction is composed from an interaction between lexico-semantic and pragmatic content, the exact nature of this interaction is still an open research question. A comprehensive model of comprehension processes for quotations should (at least) take into consideration that (i) the initial lexico-semantic composition of a quotational NMC gets refined due to the presence of quotation marks in later processing stages, acknowledging a pragmatic effect on quotational constructions (ii) these effects by quotation marks are not present in non-quotation marked NMCs (iii) naming 9 Conclusion 197 <?page no="198"?> predicates give rise to quotational readings which might impact entangling the meaning of quoted items. Another possible differentiation might need to be made between literal and non-literal readings of the sentence. In order to advance our understanding of metalinguistic predicates and their role in other research directions, several promising avenues for further research emerge across various disciplines. Experimental and empirical research on NMCs holds practical applications in language acquisition and pedagogy. The corpus study presented in this work has confirmed that NMCs are frequently used in everyday language and that they are necessary to acquire new concepts. NMCs play a key role in language acquisition of young children, specifically in how children use definitional constructions to acquire new vocabulary (cf. Krifka 2012; Seres 2020). While it is already documented that a 3-year-old made use of metalinguistic sentences with call and name (Perlis et al. 1998), whether the same effect can be observed with a larger amount of participants and also applies to German or further languages has not yet been explored. Further expanding research in language acquisition, particularly inspired by the work of Perlis et al. (1998), is essential. Addressing research questions such as the age at which children begin uttering NMCs can offer valuable insights into the development of metalinguistic awareness and linguistic competence in young learners. Interdisciplinary approaches from a language-learning and teaching perspec‐ tive hold promise for enhancing teaching of vocabulary. Employing naming predicates is expected to facilitate vocabulary learning and the acquisition of new terms, which should consistently be used by teachers to explain new (abstract) concepts. Integrating NMCs in lessons across disciplines - such as chemistry, biology, or physics - can aid in teaching technical terminology. (107) a. This chemical element is called “copper”. - b. The process of water turning into vapor is called “evaporation”. Expanding the focus in teaching to include less commonly discussed types of quotations, such as pure and scare quotes, is also beneficial. While direct and indirect types of quotation are standardly taught in German school classes in both German and English lessons, other types of quotation are typically not considered. By incorporating pure quotations into language instruction, educators can raise students’ awareness of the use-mention distinction, which is particularly helpful in foreign language classes where pure quotations can 198 9 Conclusion <?page no="199"?> explain or clarify terms in another language (108a-b). In a bilingual classroom, teachers could reinforce terms by using NMCs in both languages (108c). (108) a. In English, we call this piece of equipment a “saw”. - b. The term for “Brücke” in English is bridge. - c. In English, this is called a ‘tree,’ but in German, it’s called a ‘Baum.’ Scare quotes are also valuable in explaining the complexity of meaning, i.e., a differentiation between literal and non-literal utterances. In order to grasp the modalizing reading of NMCs, students need to consider the contextual setting, i.e., pragmatic content of an utterance, along with prosody, and multimodal elements like air quotes and facial expressions. Given the frequent use of scare quotes in political discourse, teaching students to critically evaluate the complexity of meaning and how scare quotes impact meaning is crucial beyond the classroom. On top of that, another desideratum in this research context is that quotation marks are regarded solely as a punctuation device and that learning the different systematics, i.e., language-specific notations with upper and lower quotes as well as the combination with further punctuation marks in the different types of sentential structures (see Section 2.1.1) often leads to confusion for young writers. In moving beyond the simplistic view of seeing quotation marks as a mere punctuation device - an idea that is often reflected in teaching material like school books - empirically-informed didactics should guide learn‐ ers in understanding the complexities of quotational constructions. Exploring pedagogical strategies that leverage metalinguistic awareness could enhance language instruction in educational settings. Another field of application is the area of natural-language processing. In the realm of AI language technologies, there is a need to develop advanced methods for detecting metalanguage and understanding the mechanisms of metalanguage use, particularly focusing on the use-mention distinction. Build‐ ing on this, it is crucial to create algorithms capable of recognizing the nuanced functions of quotational constructions to accurately interpret the intended meaning. Additionally, it is desirable to develop a more fine-grained metalan‐ guage corpus, building upon existing frameworks such as the one proposed by Wilson (2012), taking quotational constructions also into consideration. Such resources are invaluable for training and evaluating AI models and advancing natural language processing capabilities. Detailed characteristics of NMCs and indicators of mentioned language are necessary to detect instances 9 Conclusion 199 <?page no="200"?> of metalanguage and “learn” the distinction between use and mention. The empirical insights obtained from the corpus study can be used for detecting mentioned language and improve language models. In addition, an understanding of the nature of scare quotes and the multi‐ modal factors involved in highlighting the scare-quotational content such as prosody, contextual markers, or visual cues could support advancements in multimodal AI models. Building on the empirical findings obtained for the interpretation of NMCs could be expanded using eye-tracking technology to include facial expression, gestures, and also acoustic analyses to differentiate between literal and modalizing sentences. This would be particularly valuable in conversational AI and virtual assistants, allowing for more natural interactions. In sum, incorporating these insights into NLP systems could foster more sophisticated language technologies that recognize and respond to instances of metalanguage in human communication. The last paragraphs are devoted to discussing directions for future research. Further experimental evidence should take into consideration cross-linguistic evidence like for example from English-speaking participants. English data is particularly interesting with a focus on understanding the differences between refer and call. These investigations should elucidate any deviations of refer as observed in the corpus data. Second, a deeper understanding of quotation at the semantics-pragmatics interface is warranted. Expanding experimental work in this area could involve designing studies that deliberately violate the semantic and pragmatic content of sentences, for example, by incorporating scalar implicatures. This research would contribute to unravelling the complex interplay between semantics and pragmatics and allow for a theory of quotation for NMCs. Moreover, psycholinguistic investigations could explore reader-related fac‐ tors that impact the interpretation of NMCs. For instance, examining the influence of working memory capacity on the processing of NMCs could provide valuable insights into cognitive mechanisms underlying language comprehen‐ sion and the interpretation of NMCs. In summary, NMCs are firmly embedded in the language repertoire and hold a central position therein, solely due to their frequency of use - also of relevance cross-linguistically. Mentioned language has a long tradition in language, which is confirmed from a historical perspective. Another noticeable fact about this special type of quotational construction is that NMCs have emerged into a necessary tool to communicate about language and specific terminology and introduce words in a definition-like structure. The diverse directions for further research outlined in the previous section underscores the interdisciplinary 200 9 Conclusion <?page no="201"?> nature of investigating metalinguistic predicates, encompassing areas such as experimental psychology, linguistics, language acquisition, pedagogy, and com‐ putational linguistics. This reaffirms the versatile and prevalent use of NMCs in everyday communication, highlighting the importance of exploring the cognitive processes underlying metalanguage and the linguistic characteristics involved in comprehending metalinguistic predicates and mentioned nominals. 9 Conclusion 201 <?page no="203"?> References Abbott, Barbara. 2005. Some notes on quotation. Belgian Journal of Linguistics, 17(1), 13-26. DOI: 10.1075/ bjl.17.02abb Abbott, Barbara. 2010. Reference. Oxford: Oxford University Press. Ágel, Vilmos. 2017. Grammatische Textanalyse. Textglieder, Satzglieder, Wortgruppen‐ glieder. Berlin: De Gruyter. DOI: 10.1515/ zfs-2018-0004 Aijmer, Karin. 2015. Quotative markers in ‘A corpus of English Dialogues 1560-1760’. In: Jenny Arendholz, Wolfram Bublitz & Monika Kirner-Ludwig (eds.) The Pragmatics of Quoting Now and Then (Topics in English Linguistics 89). Berlin: De Gruyter, 231-254. DOI: 10.1515/ 9783110427561-011 Akaike, Hirotugu. 1974. A new look at the statistical model identification. IEEE Transac‐ tions on Automatic Control 19(6), 716-723. DOI: 10.1109/ TAC.1974.1100705 Allan, Keith & Katarzyna Jaszczolt. 2012. The Cambridge handbook of pragmatics. Cambridge handbooks in language and linguistics. Cambridge: Cambridge University Press. DOI: 10.1017/ CBO9781139022453. Allott, Nicholas & Mark Textor. 2012. Lexical pragmatic adjustment and ad hoc concepts. International Review of Pragmatics 4(2), 185-208. DOI: 10.1163/ 18773109-00040204 Anderson, John M. 2007. The grammar of names. Oxford: Oxford Academic. DOI: 10.109 3/ acprof: oso/ 9780199297412.001.0001 Anderson, Luvell & Ernie Lepore. 2013. What did you call me? Slurs as prohibited words setting things up. Analytic Philosophy 54(3), 350-363. DOI: 10.1111/ phib.12023 Anderson, Michael, Bryant Lee, Jon Go, Shuda Li, Ben Sutandio & LuoYan Zhou. 2006. On the Types, Frequency, Uses and Characteristics of Meta-language in Conversation. Proceedings of the Annual Meeting of the Cognitive Science Society 28, 973-978. Arche, María J., Antonio Fábregas & Rafael Marín. 2019. Main questions in the study of copulas: Categories, structures, and operations. In: María J. Arche, Antonio Fábregas & Rafael Marín (eds.) The Grammar of Copulas Across Languages (Oxford Studies in Theoretical Linguistics). Oxford: Oxford University Press, 1-30. DOI: 10.1093/ oso/ 97 80198829850.003.0001 Aristodemo, Valentina, Beatrice Giustolisi, Giorgia Zorzi, Doriane Gras, Charlotte Hauser, Rita Sala, Jordina Sánchez-Amat, Caterina Donati & Carlo Cecchetto. 2022. Insights from a comprehension study in different populations of LIS, LSC and LSF signers. Natural Language & Linguistic Theory 41, 459-500. DOI: 10.1007/ s11049-022-09539-0 Arzouan, Yossi, Abraham Goldstein & Miriam Faust. 2007. Brainwaves are stethoscopes: ERP correlates of novel metaphor comprehension. Brain Research 1160, 69-81. DOI: 10.1016/ j.brainres.2007.05.034 <?page no="204"?> Attardo, Salvatore & Lucy Pickering. 2023. Eye tracking in linguistics. London: Blooms‐ bury Academic. DOI: 10.1093/ llc/ fqad081 Attardo, Salvatore, Jodi Eisterhold, Jennifer Hay & Isabella Poggi. 2003. Multimodal markers of irony and sarcasm. Humor 16(2), 243-260. DOI: 10.1515/ humr.2003.012 Attardo, Salvatore. 2000. Irony as relevant inappropriateness. Journal of pragmatics 32(6), 793-826. DOI: 10.1016/ S0378-2166(99)00070-3 Austin, John L. 1962. Speech acts. In: J. Hillis Miller (ed.) Speech Acts in Literature. Stanford: Staford University Press. DOI: 10.1515/ 9780804764322-003 Bach, Kent. 2007. Regressions in Pragmatics (and Semantics). In: N. Burton-Roberts (ed.) Pragmatics, 24-44. Basingstoke: Palgrave Macmillan. Barbe, Katharina. 1995. Irony in Context. Pragmatics & Beyond (New Series 34). Amster‐ dam: John Benjamins. Barr, Dale J., Roger Levy, Christoph Scheepers & Harry J. Tily. 2013. Random effects structure for confirmatory hypothesis testing: Keep it maximal. Journal of Memory and Language 68(3), 255-278. DOI: 10.1016/ j.jml.2012.11.001 Bates, Douglas, Martin Machler, Ben Bolker & Steve Walker. 2015. Fitting linear mixed-ef‐ fects models using lme4. Journal of Statistical Software 67(1), 1-48. DOI: 10.18637/ jss. v067.i01 Bauer, Laurie. 2017. Compounds and compounding. Cambridge: Cambridge University Press. Bazzoni, André. 2016. Pure quotation, metalanguage and metasemantics. Linguistics and Philosophy 39(2), 119-149. DOI: 10.1007/ s10988-015-9186-8 Beal, Joan. 2004. English dialects in the North of England: morphology and syntax. In: Bernd Kortmann, Edgar Schneider, K. Burridge, R. Mesthrie, & C. Upton (eds.) A Handbook of Varieties of English. Vol. 2: Morphology and Syntax. Berlin: De Gruyter, 114-141. DOI: 10.1515/ 9783110208399.2.373 Benelli, Beatrice, Carmen Belacchi, Gianluca Gini & Daniela Lucangeli. 2006. ‘To define means to say what you know about things’: the development of definitional skills as metalinguistic acquisition. Journal of Child Language 33, 71-79. DOI: 10.1017/ S0305000905007312 Beran Michael J., Johannes Brandl, Josef Perner & Joëlle Proust. 2013. Founda‐ tions of metacognition. Oxford: Oxford University Press. DOI: 10.1093/ acprof: oso/ 9780199646739.001.0001 Bezeichnen. 2022. Duden. Retrieved March 14, 2022 from https: / / www.duden.de/ rechtsc hreibung/ bezeichnen. Bezeichnen. 2022. Etymologisches Wörterbuch des Deutschen (1993). In: Wolfgang Pfeifer et al. (eds.) Digitales Wörterbuch der deutschen Sprache. Retrieved December 12, 2022 from https: / / www.dwds.de/ wb/ etymwb/ bezeichnen. 204 References <?page no="205"?> Blackwell, Natalia & Jean E. Fox Tree. 2012. Social factors affect quotative choice. Journal of Pragmatics 44, 1150-1162. DOI: 10.1016/ j.pragma.2012.05.001 Blasko, Dawn G. & Victoria. A. Kazmerski. 2006. ERP correlates of individual differences in the comprehension of nonliteral language. Metaphor and Symbol 21, 267-284. DOI: 10.1207/ s15327868ms2104_4 Borghi, Anna M., Ferdinand Binkofski, Cristiano Castelfranchi, Felice Cimatti, Claudia Scorolli & Luca Tummolini. 2017. The challenge of abstract concepts. Psychological Bulletin 143(3), 263-292. DOI: 10.1037/ bul0000089 Bosco, Francesca M., Maurizio Tirassa & Ilaria Gabbatore. 2018. Why Pragmatics and Theory of Mind Do Not (Completely) Overlap. Frontiers in Psychology 9, 1-7. DOI: 10 .3389/ fpsyg.2018.01453 Bossi, Manuela, Lisa M. Hamm, Annegret Dahlmann-Noor & Steven C. Dakin. 2018. A comparison of tests for quantifying sensory eye dominance. Vision Research 153, 60-69. DOI: 10.1016/ j.visres.2018.09.006 Brendel, Elke, Jörg Meibauer & Markus Steinbach. 2011. Exploring the Meaning of Quotation. In: Elke Brendel (ed.) Understanding quotation. Berlin: De Gruyter, 1-34. DOI: 10.1515/ 9783110240085.1 Brendel, Ursula. 2017. Das Interpunktionssystem des Deutschen. In: Angelika Linke & Helmuth Feilke (eds.) Oberfläche und Performanz. Untersuchungen zur Sprache als dynamische Gestalt. Tübingen: Max Niemeyer, 25-41. DOI: -10.1080/ 00393274.2018.15 31249 Brysbaert, Marc, Paweł Mandera & Emmanuel Keuleers. 2018a. Corpus Linguistics. In: Anette M. B. de Groot & Peter Hagoort (eds.) Research Methods in Psycholinguistics and the Neurobiology of Language - A Practical Guide. London: Wiley, 230-246. DOI: 10.1002/ 9781394259762.ch12 Brysbaert, Marc, Paweł Mandera & Emmanuel Keuleers. 2018b. The Word Frequency Effect in Word Processing: An Updated Review. Current Directions in Psychological Science 27(1), 45-50. DOI: 10.1177/ 0963721417727521 Buchstaller, Isabelle. 2001. An alternative view of like: its grammaticalisation in conversa‐ tional American English and beyond. Edinburgh Working Papers in Applied Linguistics 11, 21-41. Bücking, Sebastian. 2009. How Do Phrasal and Lexical Modification Differ? - Contrasting Adjective-Noun Combinations in German. Word Structure 2(2), 184-204. DOI: 10.336 6/ e1750124509000403 Bücking, Sebastian. 2010. German Nominal Compounds as Underspecified Names for Kinds. In: Susan Olsen (ed.) New Impulses in Word-Formation. Hamburg: Buske, 253- 281. References 205 <?page no="206"?> Buekens, Filip. 2011. Faultless Disagreement, Assertions and the Affective-Expres‐ sive Dimension of Judgments of Taste. Philosophia 39, 637-655. DOI: 10.1007/ s11406-011-9318-5 Burton-Roberts, Noel. 1986. Implications of the pragmatics of non-descriptive definition. Journal of Linguistics 22, 311-329. DOI: 10.1017/ S0022226700010811 Caduff, Corina. 2009. Das einfache Anführungszeichen. Zeichen auf Distanz. In: Christine Abbt & Tim Kammasch (eds.) Punkt, Punkt, Komma, Strich? : Geste, Gestalt und Bedeutung philosophischer Zeichensetzung. Bielefeld: Transcript Verlag, 153-162. DOI: 10.1515/ 9783839409886-013 Call. 2022. In: Cambridge Dictionary. 2022. Retrieved March 14, 2022 from dictionary.ca mbridge.org/ dictionary/ english-german/ call. Call. 2023. In: Online Dictionary Etymology - Etymonline. Retrieved April 25, 2023 from www.etymonline.com/ search? q=call. Cann, Ronnie, Ruth Kempson & Eleni Gregoromichelaki. 2009. Semantics: An Introduction to Meaning in Language. Cambridge: Cambridge University Press. Cappelen, Herman & Ernest Lepore. 1997. Varieties of Quotation. Mind 106, 429-450. DOI: 10.1093/ acprof: oso/ 9780199641338.003.0012 Cappelen, Herman & Ernest Lepore. 1999. Using, mentioning and quoting: a reply to Saka. Mind 108(432), 741-750. DOI: 10.1093/ mind/ 108.432.741 Cappelen, Herman & Ernest Lepore. 2007. Language Turned on Itself: The Semantics and Pragmatics of Metalinguistic Discourse. Oxford: Oxford University Press. DOI: 10.109 3/ acprof: oso/ 9780199231195.001.0001 Cappelen, Herman, Ernest Lepore & Matthew McKeever. 2023. Quotation. In: Edward N. Zalta & Uri Nodelman-(eds.) The Stanford Encyclopedia of Philosophy (Summer 2023 Edition). plato.stanford.edu/ archives/ sum2023/ entries/ quotation/ Carlson, Gregory Norman. 1977. Reference to kinds in English. New York: Garland. Carston, Robyn. 2010. Explicit communication and ‘free’ pragmatic enrichment. In: B. Soria & E. Romero (eds.) Explicit Communication: Robyn Carston’s Pragmatics. Basingstoke: Palgrave Macmillan, 217-285. DOI: 10.1057/ 9780230292352_14 Carter, Benjamin T. & Steven G. Luke. 2020. Best practices in eye tracking research. Int ernational Journal of Psychophysiology 155, 49-62. DOI: 10.1016/ j.ijpsycho.2020.05.0 10 Christantil, Maria Febiana, Puri Bestari Mardani & Khansa Ayu Fadhila. 2022. Analysing the Meaning of Tone Indicators by Neurodivergent Community in Twitter. Interna‐ tional Journal of Social Science Research and Review 5(1), 5-15. DOI: -10.47814/ ijssrr.v5 i1.118 Cirillo, Letizia. 2019. The pragmatics of air quotes in English academic presentations. Journal of Pragmatics 142, 1-15. DOI: 10.1016/ j.pragma.2018.12.022 206 References <?page no="207"?> Clark, Eve & Andrew D.-W. Wong. 2002. Pragmatic directions about language use: Offers of words and relations. Language in Society 31(2), 181-212. DOI: -10.1017/ s004740450 1020152 Clark, Herbert H. & Richard Gerrig. 1990. Quotations as demonstrations. Language 66, 764-805. Clifton, Charles, Fernanda Ferreira, John M. Henderson, Albrecht W. Inhoff, Simon P. Liversedge, Erik D. Reichle & Elizabeth R. Schotter. 2016. Eye movements in reading and information processing: Keith Rayner’s 40 year legacy. Journal of Memory and Language 86, 1-19. DOI: 10.1016/ j.jml.2015.07.004 Cohen, Andrew L. 2013. Software for the automatic correction of recorded eye fixa‐ tion locations in reading experiments. Behavior Research Methods 45(3), 679-683. DOI: -10.3758/ s13428-012-0280-3 Colston, Herbert L. & Jennifer O’Brien. 2000. Contrast and pragmatics in figurative language: Anything understatement can do, irony can do better. Journal of Pragmatics 32, 1557-1583. DOI: 10.1016/ s0378-2166(99)00110-1 Colston, Herbert L. & Raymond W. Gibbs. 2007. A brief history of irony. In: Raymond W. Gibbs & Herbert L. Colston (ed.) Irony in language and thought: A cognitive science reader. New York: Lawrence Erlbaum Associates, 467-498. Conklin, Kathy, Ana Pellicer-Sánchez & Gareth Carrol. 2019. Eye-Tracking: A Guide for Applied Linguistics-Research. Cambridge: Cambridge University Press. DOI: -10.1017 / 9781108233279 Copula. 2023. Glossary of linguistic terms. Retrieved April 25, 2023 from glos‐ sary.sil.org/ term/ copula. Copula. 2023. Online Etymology Dictionary (OED). Retrieved May 8, 2023 from etymon line.com/ word/ copula. Cornejo, Carlos, Franco Simonetti, Nerea Aldunate, Agustín Ibáñez, Vladimir Lopez & Lucía Melloni. 2007. Electrophysiological evidence of different interpretative strategies in irony comprehension. Journal of Psycholinguistic Research 36, 411-430. DOI: -10.1007/ s10936-007-9052-0 Cortés Rodríguez, Álvaro, Holden Härtl, Natascha Raue & Kristina Weissbecker. 2022. Referring nouns in name-informing quotation: An empirical investigation. Linguistic Research 39(2), 327-354. DOI: 10.17250/ khisli.39.2.202206.004 COSMAS I/ II (Corpus Search, Management and Analysis System). 1991-2021. Leibniz-In‐ stitut für Deutsche Sprache, Mannheim. Coulmas, Florian. 1986. Reported speech: some general issues. In: Florian Coulmas (ed.) Direct and Indirect Speech. Berlin: De Gruyter, 1-28. DOI: -10.1515/ 9783110871968.1 Coulson, Seana & Cyma Van Petten. 2007. A special role for the right hemisphere in metaphor comprehension? : ERP evidence from hemifield presentation. Brain Research 1146, 128-145. DOI: 10.1016/ j.brainres.2007.03.008 References 207 <?page no="208"?> Cruse, David Alan. 2004. Meaning in Language. An Introduction to Semantics and Pragmatics (Oxford Textbooks in Linguistics, 2nd ed.). Oxford: Oxford University Press. Curnow, Timothy Jowan. 1999. Towards a Cross-linguistic Typology of Copula Construc‐ tions. Proceedings of the 1999 Conference of the Australian Linguistic Society, 1-9. D’Arcy, Alexandra. 2015. Quotation and advances in understanding syntactic systems. The Annual Review of Linguistics 1, 43-61. DOI: 10.1146/ annurev-lin‐ guist-030514-125220 Darwin, Clayton M. & Loretta S. Gray. 1999. Going after the Phrasal Verb: An Alternative Approach to Classification. TESOL Quarterly 33, 65-83. DOI: -10.2307/ 3588191 Davidson, Donald. 1977. Reality without reference. Dialectica 31, 247-258. DOI: -10.1111 / j.1746-8361.1977.tb01287.x Davidson, Donald. 1979. Quotation. Theory and Decision 11, 27-40. DOI: -10.1007/ bf00126690 Davidson, Donald. 1984. Inquiries Into Truth And Interpretation. Oxford: Oxford Univer‐ sity Press. Davies, Mark. 2015. Corpora: An introduction. In: Douglas Biber & Randi Reppen (eds.) The Cambridge Handbook of English Corpus Linguistics (Cambridge Handbooks in Language and Linguistics), 11-31. DOI: 10.1017/ CBO9781139764377.002 De Brabanter, Philippe. 2003. Making Sense of Mention, Quotation, and Autonymy. A Semantic and Pragmatic Survey of Metalinguistic Discourse. Phd thesis, Université Paris-Sorbonne---Paris IV. http: / / tel.archives-ouvertes.fr/ ijn_00000420 De Brabanter, Philippe. 2010. The Semantics and Pragmatics of Hybrid Quotations. Lan guage and Linguistics Compass 4(2), 107-120. DOI: 10.1111/ j.1749-818X.2009.00185.x De Brabanter, Phillipe. 2023. Quotation does not need marks of quotation. Linguistics 61(2), 285-316. DOI: 10.1515/ ling-2021-0087 De Brabanter, Phillippe. 2017. Why quotation is not a semantic phenomenon, and why it calls for a pragmatic theory. In: Ilse Depraetere & Raphael Salkie (eds.) Semantics and pragmatics: Drawing a Line (Logic, Argumentation & Reasoning 11). Berlin: Springer, 227-254. DOI: 10.1007/ 978-3-319-32247-6_14 De Roeck, Marijke. 1994. A functional typology of speech reports. In: Elisabeth Eng‐ berg‐Peterson, Lisbeth Falstr‐Jakobson & Lone Schak Rasmussen (eds.) Function and Expression in Universal Grammar. Berlin: De Gruyter, 331-351. Declerck, Renaat. 1988. Studies on Copular Sentences, Clefts and Pseudo-Clefts. Berlin: De Gruyter. DOI: 10.1515/ 9783110869330 Dews, Shelly & Ellen Winner. 1999. Obligatory processing of literal and nonliteral meanings in verbal irony. Journal of Pragmatics 31(12), 1579-1599. DOI: 10.1016/ s037 8-2166(99)00005-3 208 References <?page no="209"?> Dirven, René & Verspoor, Marjolijn. 1998. Cognitive exploration of language and linguis‐ tics. Amsterdam: Benjamins. DOI: 10.1075/ clip.1(1999) Dixon, Robert Malcolm Ward. 1992. A new approach to English grammar, on semantic principles. Oxford: Clarendon Press. DOI: 10.1093/ oso/ 9780198242727.001.0001 Dobrovie-Sorin, Carmen & Claire Beyssade. 2012. Redefining Indefinites. Dordrecht: Springer. DOI: 10.1007/ 978-94-007-3002-1 Docherty, Peter & Tomáš Mach. 2023. Direct Quotation, Rhetorical function and appli‐ cations for teaching. In: Rosemary Wette (ed.) Teaching and Learning Source-Based Writing. Current Perspectives and Future Directions. New York: Routledge, 241-256. DOI: -10.4324/ 9781003283485 Downing, Pamela. 1977. On the creation and use of English compound nouns. Language 53(4), 810-842. DOI: 10.2307/ 412913 Egorova, Natalia, Yury Shtyrov & Friedemann Pulvermüller. 2013. Early and parallel processing of pragmatic and semantic information in speech acts: neurophysiological evidence. Frontiers in human neuroscience 7(86), 1-13. DOI: 10.3389/ fnhum.2013.00086 Ehrlich, Susan F. & Keith Rayner. 1981. Contextual effects on word perception and eye movements during reading. Journal of Verbal Learning and Verbal Behavior 20(6), 641-655. DOI: 10.1016/ S0022-5371(81)90220-6 Engbert, Ralf & Reinhold Kliegl. 2011. Parallel graded attention models of reading. In Simon P. Liversedge, Iain Gilchrist & Stefan Everling (eds.) The Oxford Handbook of Eye Movements. Oxford: Oxford University Press, 787-800. DOI: 10.1093/ oxfordhb/ 97 80199539789.001.0001 Engbert, Ralf, André Longtin & Reinhold Kliegl. 2002. A dynamic model of saccade generation n reading based on spatially distributed lexical processing. Vision Research 42(5), 621-636. DOI: 10.1016/ s0042-6989(01)00301-7 Eskenazi, Michael A. 2023. Best practices for cleaning eye movement data in read‐ ing research. Behavior Research Methods. Online ahead of print. DOI: 10.3758/ s13428-023-02137-x Evans, Nicholas. 2012. Some problems in the typology of quotation: a canonical approach. In: Dunstan Brown, Marina Chumakina & Greville G. Corbett (eds.) Canonical Morphology and Syntax. Oxford: Oxford Academic, 66-98. DOI: 10.1093/ acprof: oso/ 9 780199604326.001.0001 Fara, Delia Graff. 2015. Names are predicates. Philosophical Review 124(1), 59-117. DOI: 10.1215/ 00318108-2812660 Fedorenko, Evelina & Shraon L. Thompson-Schill. 2014. Reworking the language net‐ work. Trends in Cognitive Sciences 18(3), 120-126. DOI: 10.1016/ j.tics.2013.12.006 Fedorenko, Evelina, Idan Asher Blank, Matthew Siegelmann & Zachary Mineroff. 2020. Lack of selectivity for syntax relative to word meanings throughout the language network. Cognition 203, 1-24. DOI: 10.1016/ j.cognition.2020.104348. References 209 <?page no="210"?> Fernández-Domínguez, Jesús. 2009. Productivity in English Word-formation: An approach to N+N compounding. Bern: Lang. Ferretti, Todd R., Christopher A. Schwint & Albert N. Katz. 2007. Electrophysiological and behavioral measures of the influence of literal and figurative contextual constraints on proverb comprehension. Brain and Language 101, 38-49. DOI: -10.1016/ j.bandl.200 6.07.002 Filik, Ruth & Linda M. Moxey. 2010. The on-line processing of written irony. Cognition 116(3), 421-436. DOI: 10.1016/ j.cognition.2010.06.005 Finkbeiner, Rita. 2015. „Ich kenne da so einen Jungen … kennen ist gut, wir waren halt mal zusammen weg.“: On the Pragmatics and Metapragmatics of „X ist gut“ in German. In: Jenny Arendholz, Wolfram Bublitz & Monika Kirner-Ludwig (eds.) The Pragmatics of Quoting Now and Then (Topics in English Linguistics 89). Berlin: De Gruyter, 147-176. DOI: 10.1515/ 9783110427561-008 Fontaine, Lise, Katy Jones & David Schönthal. 2023. Referring in language: an integrated approach. Cambridge: Cambridge University Press. DOI: 10.1017/ 9781316534625 Fraser, Bruce & Dwight Bolinger. 1971. The Phrasal Verb in English. Cambridge: Harvard University Press. DOI: 10.2307/ 412224 Frege, Gottlob. 1892. On sense and reference. In: Peter Geach & Max Black (eds.) Translations from the Philosophical Writings of Gottlob Frege 3, 56-78. Oxford: Basil Blackwell. Frinsel, Felicity F. & Morten H. Christiansen. 2024. Capturing individual differences in sentence processing: How reliable is the self-paced reading task? Behavior research methods 56(6), 6248-6257. DOI: 10.3758/ s13428-024-02355-x García-Carpintero, Manuel. 2003. Double-Duty Quotation: The Deferred Ostension Account. Belgian Journal of Linguistics 17(1), 89-108. DOI: -10.1075/ bjl.17.06gar García-Carpintero, Manuel. 2012. Minimalism on quotation? Critical Review of Cappelen and Lepore’s Language Turned On Itself. Philosophical Studies 161(2), 207-225. DOI: 10.1007/ s11098-011-9730-5 García-Carpintero, Manuel.-2005. Double-Duty Quotation: The Deferred Ostension Acc ount. Belgian Journal of Linguistics 17, 89-106. DOI: 10.1075/ bjl.17.06gar Gardner, Michael K., Ernst Z. Rothkopf, Richard Lapan & Toby Lafferty. 1987. The word frequency effect in lexical decision: Finding a frequency-based component. Memory & Cognition 15(1), 24-28. DOI: 10.3758/ BF03197709 Garmendia, Joana. 2018. Irony. Key Topics in Semantics and Pragmatics. Cambridge: Cambridge University Press. DOI: 10.1017/ 9781316136218 Geach, Peter. 1957. Mental Acts. Their content and their objects. London: Routledge. DOI: 10: 0710038313- Geist, Ljudmila. 2006. Die Kopula und ihre Komplemente. Zur Kompositionalität in Kopulasätzen. Tübingen: Max Niemeyer Verlag. DOI: 10.37307/ j.2198-2430.2008.01.13 210 References <?page no="211"?> Giannakidou, Anastasia & Alda Mari. 2019. Veridicality, subjectivity and grammatical mood. The view from Greek and Italian. Talk at Sinn und Bedeutung 24, Universität Osnabrück. Gibbs, Raymond W. 1986. On the psycholinguistics of sarcasm. Journal of Experimental Psychology: General 115, 3-15. DOI: 10.1037/ / 0096-3445.115.1.3 Gibbs, Raymond W. 1994. The poetics of mind: Figurative thought, language, and understanding. Journal of Pragmatics 31, 919-929. DOI: 10.1590/ s0102-445019970002 00009 Gibson, Edward & Evelina Fedorenko. 2013. The need for quantitative methods in syntax and semantics research. Language and Cognitive Processes 28(1-2), 88-124. DOI: 10.10 80/ 01690965.2010.515080 Giora, Rachel & Ofer Fein. 1999. Irony: Context and Salience. Metaphor and Symbol 14(4), 241-257. DOI: 10.1207/ S15327868MS1404_1 Giora, Rachel, Ofer Fein, Dafna Laadan, Joe Wolfson, Michal Zeituny, Ran Kidron, Ronie Kaufman & Ronit Shaham. 2007. Expecting irony: Context versus salience-based effects. Metaphor and Symbol 22(2), 119-146. DOI: -10.1080/ 10926480701235346 Giora, Rachel. 1995. On Irony and Negation. Discourse Processes 19(2), 239-264. DOI: -10. 1080/ 01638539509544916 Giora, Rachel. 1998. When is relevance? On the role of salience in utterance interpreta‐ tion. Revista Alicantina de Estudios Ingleses 11, 85-94. DOI: -10.14198/ raei.1998.11.08 Giora, Rachel. 1999. On the priority of salient meanings: Studies of literal and figurative language. Journal of pragmatics 31(7), 919-929. DOI: 10.1016/ s0378-2166(98)00100-3 Golato, Andrea. 2000. An innovative German quotative for reporting on embodied actions: Und ich so/ und er so ‘and I’m like/ and he’s like’. Journal of Pragmatics 32(1), 29-54. DOI: 10.1016/ s0378-2166(99)00030-2 Gómez-Torrente, Mario. 2003. Remarks on impure quotation. Belgian Journal of Linguis‐ tics 12(1), 129-151. DOI: 10.1075/ bjl.17.08gom Gómez-Torrente, Mario. 2017. Semantics vs. Pragmatics in Impure Quotation. In: Paul Saka & Michael Johnson (eds.) Perspectives in Pragmatics, Philosophy & Psychology. Cham: Springer, 135-167. DOI: 10.1007/ 978-3-319-68747-6_6 Grice, Herbert Paul. 1978. Further Notes on Logic and Conversation. In: Peter Cole & Jerry L. Morgan (eds.) Syntax and Semantics: Pragmatics 9. New York: Academic Press, 183-97. Grice, Herbert Paul. 1989. Studies in the way of words. Cambridge: Harvard University Press. Güldemann, Tom. 2008. Quotative indexes in African languages: A synchronic and dia‐ chronic survey. Berlin: De Gruyter. DOI: 10.1515/ 9783110211450 Gundel, Jeanette K. & Barbara Abbott. 2019. The Oxford Handbook of Reference. Oxford: Oxford University Press. References 211 <?page no="212"?> Gupta, Anil & Stephen Mackereth. 2023. Definitions. In: Edward N. Zalta & Uri Nodel‐ man-(eds.) The Stanford Encyclopedia of Philosophy (Fall 2023 Edition). https: / / plato.s tanford.edu/ archives/ fall2023/ entries/ definitions/ Gupta, Anil. 1980. The logic of common nouns. An investigation in quantified modal logic. New Haven: Yale University Press. Gutzmann, Daniel & Erik Stei. 2011. How quotation marks what people do with words. Journal of Pragmatics 43(10), 265-2663. DOI: 10.1016/ j.pragma.2011.03.010 Gutzmann, Daniel. 2007. Zitate und die Semantik/ Pragmatik-Schnittstelle. In: Elke Brendel, Jörg Meibauer & Markus Steinbach (eds.) Zitat und Bedeutung, 111-133. Hamburg: Buske. Gutzmann, Daniel. 2021. Semantics vs. pragmatics. In: Daniel Gutzmann, Lisa Matthew‐ son, Cécile Meier, Hotze Rullmann & Thomas Ede Zimmermann (eds.) The Companion to Semantics. Oxford: Wiley, 1-31. DOI: -10.1002/ 9781118788516.sem049 Haderlein, Veronika. 2008. Das Konzept zentraler Wortschätze. Bestandsaufnahme, theor‐ etisch-methodische Weiterführung und praktische Untersuchung. Doctoral Dissertation, LMU München. DOI: 10.5282/ edoc.8066 Hagoort, Peter & Stephen C. Levinson. 2014. Neuropragmatics. In: Michael S. Gazzaniga, & George R. Mangun. The cognitive neurosciences, 667-674. Cambridge: MIT Press. DOI: 10.7551/ mitpress/ 9504.003.0073 Härtl, Holden & Marcel Schlechtweg. 2023. Quotation as an interface phenomenon. Linguistics 61(2), 275-284. DOI: 10.1515/ ling-2023-0029 Härtl, Holden & Heiko Seeliger. 2019. Is a so-called “beach” a beach? An empirically based analysis of secondary content induced by ironic name use. In: Daniel Gutzmann & Katharina Turgay (eds.) Secondary Content - The Semantics and Pragmatics of Side Issues (Current Research in the Semantics/ Pragmatics Interface 37). Leiden: Brill. DOI: / 10.1163/ 9789004393127_009 Härtl, Holden. 2015. Semantic non-transparency in the mental lexicon: On the relation between word-formation and naming. In: Claudia Brinker-von der Heyde, Nina Kalwa, Nina-Maria Klug & Paul Reszke (eds.) Eigentlichkeit: Zum Verhältnis von Sprache, Sprechern und Welt. Berlin: De Gruyter, 395-416. DOI: -10.1515/ 97831103354 77.395 Härtl, Holden. 2016. Normality at the boundary between word-formation and syntax. In: Franz J. d'Avis, & Horst Lohnstein (eds.) Linguistische Berichte, Sonderheft 22, 65-92. Hamburg: Helmut Buske. Härtl, Holden. 2018. Name-informing and distancing ‚sogenannt’ (‘so-called’): Name mentioning and the lexicon-pragmatics interface. Zeitschrift für Sprachwissenschaft 37(2), 139-169. DOI: 10.1515/ zfs-2018-0008 Härtl, Holden. 2020. Referring nouns in name-informing quotation: A copula-based approach. In: Michael Franke, Nikola Kompa, Liu Mingya, Jutta Mueller & Juliane 212 References <?page no="213"?> Schwab (eds.) Proceedings of Sinn und Bedeutung 24, 291-304. Osnabrück: University Osnabrück. Hauptman, Miriam, Idan Blank & Evelina Fedorenko. 2023. Non-literal language pro‐ cessing is jointly supported by the language and Theory of Mind networks: Evidence from a novel meta-analytic fMRI approach. Cortex 162, 96-114. DOI: -10.1101/ 2022.03 .08.481056 Heller, Daphna & Lynsey Wolter. 2008. That is Rosa. Identificational sentences as intension-al predication. In: Atle Grønn (ed.) Proceedings of Sinn und Bedeutung 12, 226-240. Oslo: University of Oslo. Herbst, Thomas, David Heath, Ian F. Roe & Dieter Götz. 2004. A Valency Dictionary of English: A Corpus-Based Analysis of the Complementation Patterns of English Verbs, Nouns and Adjectives. Berlin: De Gruyter. DOI: -10.1515/ 9783110892581 Higgins, Roger. 1979. The Pseudo-Cleft Construction in English. New York: Garland Publishing. Hill, Robin L. & W. Murray. 2000. Commas and spaces: Effects of punctuation on eye movements and sentence parsing. In: Alan Kennedy, Ralph Radach, Dieter Heller & Joël Pynte (eds.) Reading as a Perceptual Process. Oxford: Elsevier, 565-589. DOI: -10.1 016/ b978-008043642-5/ 50027-9 Hirotani, Masako, Lyn Frazier & Keith Rayner. 2006. Punctuation and intonation effects on clause and sentence wrap-up: Evidence from eye movements. Journal of Memory and Language 54(3), 425-443. DOI: 10.1016/ j.jml.2005.12.001 Hoekstra, Jarich. 2023. From naming verb to copula: The case of Wangerooge Frisian Heit. Journal of Germanic Linguistics 35(2), 97-147. DOI: -10.1017/ S1470542722000113 Hoemann, Katie, Fei Xu & Lisa Feldman Barrett. 2019. Emotion words, emotion concepts, and emotional development in children: A constructionist hypothesis. Developmental psychology 55(9), 1830-1849. DOI: 10.1037/ dev0000686 Hofmann, Peter & Natascha Raue. 2024. Quotation Unbound - Bridging the Divide Across Quotation Types. Workshop Abstract for the DGfS 2025 on “Variation and Change”, 47.-Jahrestagung der Deutschen Gesellschaft für Sprachwissenschaft. Holmqvist, Kenneth, Marcus Nyström, Richard Andersson, Richard Dewhurst, Jarodzka Halszka, Halszka Jarodzka, & Joost van de Weijer. 2011. Eye Tracking: A comprehensive guide to methods and measures. Oxford: Oxford University Press. Hooge, Ignace T. C., Gijs A. Holleman, Nina C. Haukes & Roy S. Hessels. 2019. Gaze tracking accuracy in humans: One eye is sometimes better than two. Behavior Research Methods 51, 2712-2721. DOI: 10.3758/ s13428-018-1135-3 Horton, Bruce. 1996. What are copula verbs? Cognitive Linguistics in the Redwoods, 319-346. DOI: 10.1515/ 9783110811421.319 References 213 <?page no="214"?> Huizeling, Eleanor, Sophie Arana, Peter Hagoort & Jan-Mathijs Schoffelen. 2022. Lexical Frequency and Sentence Context Influence the Brain’s Response to Single Words. Neurobiology of Language 3(1), 149-179. DOI: 10.1162/ nol_a_00054 Hüning, Matthias. 2010. Adjective + Noun constructions between syntax and word formation in Dutch and German. In: Sascha Michel & Michael Onysko (eds.) Cognitive approaches to word formation. Berlin: De Gruyter, 195-216. DOI: 10.1515/ 9783110223606.195 Inhoff, Albrecht W. & Keith Rayner. 1986. Parafoveal word processing during eye fixations in reading: Effects of word frequency. Perception & Psychophysics 40(6), 431-439. DOI-: 10.3758/ BF03208203 Issel-Dombert, Sandra. 2018. Eine Frage des Stils - Zur Rolle doppelter Anführungszei‐ chen bei der extrême droite als Strategie der Entdiabolisierung. In: Kristin George, Miriam Langlotz, Urania Milevski & Katharina Siedschlag (eds.) Interpunktion im Spannungsverhältnis zwischen Norm und stilistischer Freiheit. Sprachwissenschaftliche, sprachdidaktische und literaturwissenschaftliche Perspektiven. Frankfurt: Peter Lang, 47-62. DOI: 10.3726/ b11682 Ivanko, Stacey L. & Penny M. Pexman. 2003. Context incongruity and irony processing. Discourse Processes 35(3), 241-279. DOI: 10.1207/ s15326950dp3503_2 Jakubíček, Miloš, Vojtěch Kovář, Pavel Rychlý & Vít Suchomel. 2013. The TenTen corpus family. 7th International Corpus Linguistics Conference CL, 125-127. Jiang, Nan. 2011. Conducting Reaction Time Research in Second Language Studies. New York: Routledge. DOI: 10.4324/ 9780203146255 Johnson, Michael & Ernie Lepore. 2011. Misrepresenting misrepresentation. In: Elke Brendel, Jörg Meibauer & Markus Steinbach (eds.) Understanding Quotation. Berlin: De Gruyter, 231-248. DOI: 10.1515/ 9783110240085.231 Johnson, Michael. 2011. The Punctuation Theory of Quotation. In: Elke Brendel, Jörg Meibauer & Markus Steinbach (eds.) Understanding Quotation. Berlin: De Gruyter, 209-230. DOI: -10.1515/ 9783110240085.209 Jones, Bob Morris. 2018. Copular Clauses in Welsh. Aberystwyth University. https: / / user s.aber.ac.uk/ bmj/ Ymchwil/ Copula_clauses_in_Welsh.pdf Jones, Jan-Erik. 2023. “Locke on Real Essence”. In: Edward N. Zalta & Uri Nodelman (eds.) The Stanford Encyclopedia of Philosophy (Winter 2022 Edition). https: / / plato.stanford. edu/ entries/ real-essence/ Jorgensen, Julia, George A. Miller & Dan Sperber. 1984. Test of the mention theory of irony. Journal of Experimental Psychology: General 113(1), 112-120. DOI: 10.1037/ / 009 6-3445.113.1.112 Jouravlev, Olessia, David Zheng, Zuzanna Balewski, Alvince Le Arnz Pongos, Zena Levan, Susan Goldin-Meadow & Evelina Fedorenko. 2019. Speech-accompanying 214 References <?page no="215"?> gestures are not processed by the language-processing mechanisms. Neuropsychologia 132, 107-132. DOI: 10.1016/ j.neuropsychologia.2019.107132 Just, Marcel Adam & Patricia A. Carpenter. 1982. Paradigms and processes in reading comprehension. Journal of Experimental Psychology: General 111, 228-238. DOI: -10.1 037/ / 0096-3445.111.2.228 Kasimir, Elke. 2008. Prosodic Correlates of Subclausal Quotes. ZAS Papers in Linguistics 49, 67-77. DOI: / 10.21248/ zaspil.49.2008.364 Kastovsky, Dieter. 1986. The problem of productivity in word-formation. Linguistics 24, 585-600. DOI: 10.1515/ ling.1986.24.3.585 Katz, Albert N. & Penny M. Pexman. 1997. Interpreting figurative statements: Speaker occupation can change metaphor to irony. Metaphor and Symbol 12(1), 19-41. Katz, Albert N., Dawn G. Blasko & Victoria A. Kazmerski. 2004. Saying what you don’t mean: Social influences on sarcastic language processing. Current Directions in Psychological Science 13, 186-189. DOI: 10.1111/ j.0963-7214.2004.00304.x Kazmerski, Victoria A. & Dawn G. Blasko & Banchiamlack G. Dessalegn. 2003. ERP and behavioral evidence of individual differences in metaphor comprehension. Memory and Cognition 31, 673-689. DOI: 10.3758/ bf03196107 Kerstens, Johan, Eddy Ruys & Joost Zwarts. 2023. Lexicon of Linguistics. Utrecht institute of Linguistics. Utrecht: Utrecht University OTS. Kilgarriff, Adam, Vojtěch Kovář, Simon Krek, Irena Srdanović & Carole Tiberius. 2010. A quantitative evaluation of word sketches. Proceedings of the 14th EURALEX Interna‐ tional Congress, 372-379. Kiss, Katalin É. 2006. Focussing as predication. In: Valéria Molnár & Susanne Winkler (eds.) The Architecture of Focus. Berlin: De Gruyter, 169-193. DOI: -10.1515/ 978311092 2011.169 Klewitz, Gabriele & Elizabeth Couper-Kuhlen. 1999. Quote-unquote? The role of prosody in the contextualization of reported speech sequences. Pragmatics 9(4), 459-485. DOI: 10.1075/ prag.9.4.03kle Klockow, Reinhard. 1978. „Kein kalter Kaffee“: Zur modalisierenden Verwendung von Anführungszeichen. Linguistische Berichte 55, 14-24. Klockow, Reinhard. 1980. Linguistik der Gänsefüsschen: Untersuchungen zum Gebrauch d. Anführungszeichen im gegenwärtigen Deutsch. Frankfurt: Haag und Herchen. Kotowski, Sven, Katja Böer & Holden Härtl. 2014. Compounds vs. Phrases: The Cognitive Status of Morphological Products. In: Franz Rainer, Wolfgang U. Dressler, Francesco Gardani & Hans-Christian Luschützky (eds.)-Morphology and Meaning. Selected Pa pers from the 15th International Morphology Meeting (Current Issues in Linguistic Theory). Amsterdam: Benjamins, 191-204. References 215 <?page no="216"?> Kowatch, Kristy, Juanita W. Whalen & Penny M. Pexman. 2013. Irony comprehension in action: A new test of processing for verbal irony. Discourse Processes 50(5), 301-315. DOI: 10.1080/ 0163853X.2013.799934 Kretschmar, Franziska & Matthias Schlesewsky & Adrian Staub. 2015. Dissociating Word Frequency and Predictability Effects in Reading: Evidence From Coregistration of Eye Movements and EEG. Journal of Experimental Psychology: Learning, Memory, and Cognition 41(6), 1648-1662. DOI: 10.1037/ xlm0000128 Krifka, Manfred & Claudia Gerstner. 1987. An Outline in Genericity. SNS-Bericht. Seminar für natürlich-sprachliche Systeme der Universität Tübingen. Krifka, Manfred & Claudia Gerstner-Link. 1993. Genericity. In: Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld & Theo Vennemann (eds.) Syntax - An internatio nal handbook of contemporary research, 966-978. Berlin: De Gruyter. Krifka, Manfred. 2003. Bare NPs: Kind-referring, Indefinites, Both, or Neither? In: Robert B. Young & Yuping Zhou (eds.) SALT XIII. New York: Cornell University, 180-203. DOI: -10.3765/ salt.v13i0.2880 Krifka, Manfred. 2012. Definitional Generics. In: Alda Mari, Claire Beyssade & Fabio Del Prete (eds.) Genericity. Oxford Studies in Theoretical Linguistics, 372-389. Oxford: Oxford University Press. DOI: 10.1093/ acprof: oso/ 9780199691807.003.0015 Kroeger, Paul R. 2018. Analyzing meaning: An introduction to semantics and pragmatics. Textbooks in Language Sciences 5. Berlin: Language Science Press. DOI: -10.1080/ 072 68602.2019.1680090 Kronmüller, Edmundo, Ira Noveck, Natalia Rivera, Francisco Jaume-Guazzini & Dale Barr. 2017. The positive side of a negative reference: the delay between linguistic processing and common ground. Royal Society Open Science 4, 1-14. DOI: -10.1098/ rsos.160827 Kumon-Nakamura, Sachi, Sam Glucksberg & Mary Brown. 1995. How about another piece of pie: The allusional pretense theory of discourse irony. Journal of Experimental Psychology: General 124(1), 3-21. DOI: 10.1037/ / 0096-3445.124.1.3 Laurent, Jean-Paul, Guy Denhieres, Christine Passerieux, Galina Iakimova & Marie-Christine Hardy-Bayle. 2006. On understanding idiomatic language: The sali‐ ence hypothesis assessed by ERPs. Brain Research 1068(1), 151-160. DOI: -10.1016/ j.br ainres.2005.10.076 Lenth, Russell V. 2016. Least-squares means: The R package lsmeans. Version 2.30.0. Journal of Statistical Software 69(1), 1-33. DOI: 10.18637/ jss.v069.i01 Levi, Judith N. 1978. The syntax and semantics of complex nominals. New York: Academic Press. Levin, Beth. 1993. English verb classes and alternations: a preliminary investigation. Chicago: The University of Chicago Press. 216 References <?page no="217"?> Levinson, Stephen C. 2000. Presumptive Meanings. The Theory of Generalized Conversa‐ tional Implicature. Cambridge: MIT Press. DOI: 10.7551/ mitpress/ 5526.001.0001 Locke, John. 1689. An Essay concerning Human Understanding, edited by P. H. Nidditch (1975). Oxford: Oxford University Press. Loewen, Shawn & Luke Plonsky. 2016. An A - Z of applied linguistics research methods. London: Palgrave. DOI: 10.1007/ 978-1-137-40322-3 Lucariello, Joan. 2007. Situational irony: a concept of events gone away. In: Gibbs Raymond & Colston Herbert (eds.) Irony in Language and Thought. London: Taylor & Francis Group, 467-498. DOI: 10.1037/ / 0096-3445.123.2.129 Maienborn, Claudia. 2003. Against a Davidsonian analysis of copula sentences. In: Makoto Kadowaki & Shigeto Kawahara (eds.) NELS Proceedings 33, 2-20. Maier, Emar, 2014. Mixed quotation: the grammar of apparently transparent opacity. Semantics and Pragmatics 7(7), 1-67. Maier, Emar. 2007. Mixed quotation: between use and mention. Proceedings of LENLS 240, 1-15. Maier, Emar. 2020. Mixed Quotation. In: Daniel Gutzmann, Lisa Matthewson, Cécile Meier, Hotze Rullman & Ede Zimmerman (eds.) The Companion to Semantics. New Jersey: Wiley Blackwell. DOI: 10.1002/ 9781118788516 Matthews, Peter.-2014.-The Concise Oxford Dictionary of Linguistics. Oxford: Oxford University Press. Matuschek, Hannes, Reinhold Kliegl, Shravan Vasishth, Harald Baayen & Douglas Bates. 2017. Balancing Type I error and power in linear mixed models. Journal of Memory and Language 94, 305-315. DOI: 10.1016/ j.jml.2017.01.001 Matushansky, Ora. 2008. On the Linguistic Complexity of Proper Names. Linguistics and Philosophy 31(5), 573-627. DOI: 10.1007/ s10988-008-9050-1 McConkie, George W., Michael D. Reddix & David Zola. 1992. Perception and Cognition in Reading: Where is the Meeting Point? In: Keith Rayner (ed.) Eye Movements and Visual Cognition (Springer Series in Neuropsychology). Springer: New York, 293-303. DOI: 10.1007/ 978-1-4612-2852-3_17 McDonald, Skye. 2007. Neuropsychological Studies of Sarcasm. In: Raymond W. Gibbs & Herbert L. Colston. Irony in language and thought, 217-230. DOI: 10.4324/ 9781410 616685 McGhee-Bidlack, Betty. 1991. The development of noun definitions: a metalinguistic analysis. Journal of Child Language 18, 417-34. DOI: 10.1017/ s0305000900011132 McKeever, Matthew. 2023. A new theory of quotation. Unpublished paper (draft). Retrieved December 10, 2023 from mipmckeever.weebly.com/ writings-etc.html McMurtrie, Douglas C. 1934. Concerning Quotation Marks. New York: Privately printed. References 217 <?page no="218"?> Meehan, Teresa. 1991. It’s like, ‘What’s happening in the evolution of like? ’ A theory of grammaticalization. Kansas Working Papers in Linguistics 16, 37-51. DOI: -10.17161/ k wpl.1808.423 Meibauer, Jörg. 2007. Syngrapheme als pragmatische Indikatoren: Anführung und Auslassung. In: Sandra Döring & Jochen Geilfuß-Wolfgang (eds.) Von der Pragmatik zur Grammatik. Leipzig: Universitätsverlag, 21-37. DOI: 10.37307/ j.2198-2430.2009.0 1.10 Meibauer, Jörg. 2014. Lying at the Semantics-Pragmatics Interface. Mouton Series in Pragmatics 14. Boston: De Gruyter. DOI: 10.1515/ 9781614510840 Meibauer, Jörg. 2015. Only “Nur”. Scare quoted (exclusive) focus particles at the semantics/ pragmatics interface. In: Jenny Arendholz, Wolfram Bublitz & Monika Kirner-Ludwig (eds.) The Pragmatics of Quoting Now and Then (Topics in English Linguistics 89). Berlin: De Gruyter, 177-208. DOI: 10.1515/ 9783110427561-009 Meta. 2023. Dictionary.com. Retrieved May 17, 2023. dictionary.com/ browse/ meta- Métairy, Justine, Peter Lauwers, Renata Enghels, Miriam Taverniers & Marleen Van Peteghem. 2020. A micro-typological perspective on resultative secondary predicates: the case of nomination verb constructions. Language Sciences 78, 1-18. DOI: -10.1016/ j.langsci.2019.101253 Metalanguage. 2023. Collins Dictionary. Retrieved May 17, 2023 from collinsdictionary.c om/ dictionary/ english/ metalanguage. Mikkelsen, Line. 2011. Copular clauses. Specification, Predication and Equation. In: Clau‐ dia Maienborn, Klaus von Heusinger & Paul Portner (eds.) Semantics. An International Handbook of Natural Language Meaning 1. Berlin: De Gruyter, 1805-1829. DOI: -10.15 15/ 9783110255072.1805 Mikulová, Jana. 2016. Marking the end of direct speech in Late Latin. Graeco-Latina Brunensia 21(2), 169-182. Monsell, Simon, Mary C. Doyle & Patrick N. Haggard. 1989. Effects of frequency on visual word recognition tasks: Where are they? Journal of Experimental Psychology: General 118(1), 43-71. Moore, Andrew W. 2019. How Significant is the Use/ Mention Distinction? Language, World, and Limits: Essays in the Philosophy of Language and Metaphysics, 11-26. DOI: 10.1093/ oso/ 9780198823643.003.0001 Morzycki, Marcin. 2011. Metalinguistic comparison in an alternative semantics for imprecision. Natural Language Semantics 19(1), 39-86. DOI: 10.1007/ s11050-010-9063 -5 Mueller-Reichau, Olav. 2011. Sorting the World: On the Relevance of the Kind/ Object-Dis‐ tinction to Referential Semantics. Heusenstamm: Ontos. DOI: -10.1515/ 9783110323580 Müller, Marcus. 2022. “Ich will das hier nicht ausführlich erläutern; denn das ist viel zu kompliziert“. Terminologiearbeit und terminologische Arbeitsverweigerung in Ple‐ 218 References <?page no="219"?> nardebatten des deutschen Bundestages. Zeitschrift Korpora Deutsch als Fremdsprache 2(1), 95-122. Nennen. 2022. Etymologisches Wörterbuch des Deutschen (1993). In: Wolfgang Pfeifer et al. (eds.) Digitales Wörterbuch der deutschen Sprache. Retrieved March 14, 2022. dwds.de/ wb/ etymwb/ nennen Olkoniemi, Henri & Johanna K. Kaakinen. 2021. Processing of irony in text: A systematic review of eye-tracking studies. Canadian Journal of Experimental Psychology/ Revue canadienne de psychologie expérimentale 75(2), 99-106. DOI: -10.1037/ cep0000216 Olkoniemi, Henri, Eerika Johander & Johanna K. Kaakinen. 2019. The role of look-backs in the processing of written sarcasm. Memory & Cognition 47(1), 87-105. DOI: -10.37 58/ s13421-018-0852-2 Olkoniemi, Henri, Henri Ranta & Johanna K. Kaakinen. 2016. Individual differences in the processing of written sarcasm and metaphor: Evidence from eye movements. Journal of Experimental Psychology: Learning, Memory, and Cognition 42, 433-450. DOI: 10.10 37/ xlm0000176 Olkoniemi, Henri. 2018. Individual differences in processing written irony. Doctoral dissertation, University of Turku. Olsen, Susan. 2012. Semantics of Compounds. In: Claudia Maienborn, Klaus von Heu‐ singer & Paul Portner (eds.) Semantics: An International Handbook of Natural Language Meaning 3. Berlin: De Gruyter, 2120-2150. DOI: 10.1353/ lan.2013.0026 Orquin, Jacob L. & Kenneth Holmqvist. 2018. Threats to the validity of eye-movement research in psychology. Behavior Research Methods 50(4), 1645-1656. DOI: 10.3758/ s13428-017-0998-z Orquin, Jacob L. & Simone Mueller Loose. 2013. Attention and choice: A review on eye movements in decision making. Acta Psychologica 144(1), 190-206. DOI: 10.1016/ j.actpsy.2013.06.003 Pagin, Peter & Dag Westerståhl. 2010. Pure quotation and general compositionality. Linguistics and Philosophy 33(5). 381-415. DOI: 10.1007/ s10988-011-9083-8 Paunov, Anna M., Itamar A. Blank, Oleg Jouravlev, Zoe Mineroff, Jules Gallée & Evelina Fedorenko. 2022. Differential tracking of linguistic vs. mental state content in natu‐ ralistic stimuli by language and theory of mind (ToM) brain networks. Neurobiology of Language 3(3), 413-440. DOI: 10.1162/ nol_a_00071 Perkuhn, Rainer, Holger Keibel, Marc Kupietz. 2012. Korpuslinguistik. Paderborn: Fink. Retrieved June 8, 2022 from homepage.ruhr-uni-bochum.de/ stephen.berman/ Korpus linguistik/ H%C3%A4ufigkeitsma%C3%9Fe.html Perlis, Donald, Khemdut Purang & Carl Andersen. 1998. Conversational adequacy: mistakes are the essence. International Journal of Human-Computer Studies 48 (5), 553-575. DOI: 10.1006/ ijhc.1997.0181 References 219 <?page no="220"?> Pettita, Giulia, Mark Halley & Brenda Nicodemus. 2018. “What’s the sign for nitty gritty? ” Managing metalinguistic references in ASL-English dialogue interpreting. Translation and Interpreting Studies 13(1), 49-70. DOI: 10.1075/ tis.00004.pet Pexman, Penny M., Todd R. Ferretti & Albert N. Katz. 2000. Disourse factors that influence online reading of metaphor and irony. Discourse Processes 29(3), 201-222. DOI: 10.120 7/ S15326950dp2903_2 Plag, Ingo, Sabine Arndt-Lappe, Maria Braun & Mareile Schramm. 2015. The meaning of words and sentences. In: Ingo Plag (ed.) Introduction to English linguistics. Berlin: De Gruyter, 145-180. DOI: 10.1515/ 9783110378382-010 Polenz, Peter von. 1999. Deutsche Sprachgeschichte vom Spätmittelalter bis zur Gegenwart. 19. und 20.-Jahrhundert 3. Berlin: De Gruyter. DOI: -10.1515/ 9783110805918 Potts, Christopher. 2007. The Dimensions of Quotation. In Chris Barker & Pauline Jacobsen (eds.) Direct Compositionality. Oxford: Oxford University Press, 405-431. DOI: 10.1093/ oso/ 9780199204373.003.0012 Predelli, Stefano. 2003. Scare Quotes and their Relation to Other Semantic Issues. Linguistics and Philosophy 26(1), 1-28. DOI: 10.1023/ a: 1022278209949 Psychology Software Tools, Pittsburgh, PA, Inc. 2016. E-Prime 3.0. https: / / support.pstne t.com Pustet, Regina. 2003. Copulas. Universals in the Categorization of the Lexicon. DOI: 10.1093/ acprof: oso/ 9780199258505.001.0001 Pynte, Joël & Alan Kennedy. 2007. The influence of punctuation and word class on dis‐ tributed processing in normal reading. Vision research 47(9), 1215-1227. DOI: 10.1016/ j.visres.2006.12.006. Quaßdorf, Sixta. 2016. “A little more than kin” - Quotations as a Linguistic Phenomenon: A Study Based on Quotations from Shakespeare’s Hamlet. Doctoral dissertation, University of Freiburg. Quasthoff, Uwe. 1998. Tools for automatic lexicon maintenance: acquisition, error correction, and the generation of missing values. LREC, 853-856. Quer, Josep. 2011. Reporting and quoting in signed discourse. Understanding quotation 7, 277-302. DOI: 10.1515/ 9783110240085.277 Quine, Willard Van Orman. 1940. Mathematical Logic, Boston, MA: Harvard University Press. Quine, Willard Van Orman. 1960. Carnap and logical truth. Synthese 12, 350-374. DOI: 1 0.1007/ bf00485423 Quotation. 2023. Macmillan Dictionary. Retrieved May 10, 2023. Quotation. 2023. Merriam-Webster.com. Retrieved August 10, 2023 from merriam-webst er.com/ dictionary/ quotation. 220 References <?page no="221"?> Quotation. 2023. Oxford Advanced American Dictionary (OAAD). Retrieved August 10, 2023 from oxfordlearnersdictionaries.com/ definition/ american_english/ quotation? q= quotation. R Core Team. 2023. R: A language and environment for statistical computing. R founda‐ tion for statistical computing. Vienna, Austria. https: / / www.R-project.org (Version 4.1) [Computer software] Rabern, Brian. 2023. Pure quotation in linguistic context.-Journal of Philosophical Logic 52 (2), 393-413. DOI: 10.1007/ s10992-022-09675-3 Ranger, Graham. 2015. Quotative LIKE in contemporary non standard English. Variation, ajustement, interprétation, 1-13. DOI: 10.4000/ books.pufc.8887 Rapp, Irene. 2015. On the Temporal Interpretation of Present Participles in German. Journal of Semantics 3(32), 477-523. DOI: 10.1093/ jos/ ffu005 Rau, Anne K., Kristina Moll, Margaret J. Snowling & Karin Landerl. 2015. Effects of or‐ thographic consistency on eye movement behavior: German and English children and adults process the same words differently. Journal of Experimental Child Psychology 130, 92-105. DOI: 10.1016/ j.jecp.2014.09.012 Raue, Natascha & Álvaro Cortés Rodríguez. 2022. The interplay between quotation and referentiality: An empirical investigation into name-mentioning constructions. Talk at the 11th meeting of Linguistic Evidence,-Université de Paris - CNRS. Raue, Natascha & Álvaro Cortés Rodríguez. 2024. The interplay between quotation and referentiality: An empirical investigation into name-mentioning constructions. In preperation. Raue, Natascha, Álvaro Cortés Rodríguez & Holden Härtl. 2025. The processing of quotation marks in German: Evidence from eye tracking. Zeitschrift für Sprachwissen‐ schaft-44(1).-DOI: 10.18148/ zs/ 2025-2006 Raue, Natascha. 2020. Reference to Names: A Theoretical and Empirical Exploration of Name Mentioning. State examination thesis, University of Kassel. Raue, Natascha. 2023. Scare quotes in name-informing constructions: A self-paced reading study on the processing of modalizing quotational constructions. In: Micah Goldwater, Florencia Anggoro, Brett Hayes & Desmond Ong (eds.) Proceedings of the Annual Meeting of the Cognitive Science Society 45, 1702-1708. Rayner, Keith & Alexander Pollatsek. 1994. The psychology of reading. New York: Routledge. DOI: -10.4324/ 9780203357798 Rayner, Keith & Alexander Pollatsek. 2006. Eye-movement control in reading. In: Mat‐ thew J. Traxler (ed.) Handbook of psycholinguistics, 613-657. Amsterdam: Academic Press. DOI: 10.1016/ b978-012369374-7/ 50017-1 Rayner, Keith & G. W. McConkie. 1976. What guides a reader’s eye movements? Vision Research 16(8), 829-837. DOI: 10.1016/ 0042-6989(76)90143-7 References 221 <?page no="222"?> Rayner, Keith & Simon P. Liversedge. 2012. Linguistic and cognitive influences on eye movements during reading. In: Simon P. Liversedge, Iain Gilchrist & Stefan Everling (eds.) The Oxford Handbook of Eye Movements, 752-766. DOI: 10.1093/ ox‐ fordhb/ 9780199539789.013.0041 Rayner, Keith & Susan A. Duffy. 1986. Lexical complexity and fixation times in reading: Effects of word frequency, verb complexity, and lexical ambiguity. Memory & Cogni‐ tion, 14(3), 191-201. DOI: 10.3758/ BF03197692 Rayner, Keith, Denis Drieghe & Simon P. Liversedge. 2011. Eye movements and word skipping during reading: Effects of word length and predictability. Journal of Experi‐ mental Psychology: Human Perception and Performance 37(2), 514-528. DOI: -10.1037/ a0020990 Rayner, Keith, Sara C. Sereno & Gary E. Raney. 1996. Eye movement control in reading: A comparison of two types of models. Journal of Experimental Psychology: Human Perception and Performance 22(5), 1188-1200. DOI: 10.1037/ 0096-1523.22.5.1188 Rayner, Keith, Timpthy J. Slattery, Denis Drieghe & Simon P. Liversedge. 2011. Eye movements and word skipping during reading: effects of word length and predicta‐ bility. Journal of experimental psychology. Human perception and performance 37(2), 514-528. DOI: 10.1037/ a0020990 Rayner, Keith. 1978. Eye movement latencies for parafoveally presented words. Bulletin of the Psychonomic Society 11(1), 13-16. DOI: 10.3758/ bf03336753 Rayner, Keith. 1998. Eye movements in reading and information processing: 20 years of research. Psychological Bulletin 124(3), 372-422. DOI: 10.1037/ 0033-2909.124.3.372 Rayner, Keith. 2009. Eye movements and attention in reading scene perception and visual search. Quarterly Journal of Experimental Psychology 62(8), 1457-1506. DOI: -10.1080/ 17470210902816461 Recanati, Francois. 2001. Open Quotation. Mind 110(439), 637-687. DOI: -10.1093/ mind/ 110.439.637 Recanati, François. 2008. Open quotation revisited. Philosophical Perspectives 22(1), 443-471. DOI: 10.1093/ acprof: oso/ 9780199226993.003.0009 Refer. 2022. Cambridge Dictionary. Retrieved March 14, 2022 from dictionary.cambridge .org/ dictionary/ english-german/ refer? q=refer+ Refer. 2022. Merriam Webster Dictionary. Retrieved March 14, 2022 from https: / / www.m erriam-webster.com/ dictionary/ refer%20to Refer. 2022. Online Dictionary Etymology - Etymonline. Retrieved March 14, 2022 from etymonline.com/ search? q=as&ref=searchbar_searchhint Regel, Stefanie. 2009. The Comprehension of Figurative Language: Electrophysiological Evidence on the Processing of Irony. Doctoral dissertation, University of Potsdam. Leipzig: MPI. Reichenbach, Hans. 1947. Elements of Symbolic Logic. New York: Free Press. 222 References <?page no="223"?> Reichle, Erik D., Alexander Pollatsek, Donald L. Fisher & Keith Rayner. 1998. Toward a model of eye movement control in reading. Psychological Review 105(1), 125-157. DOI: 10.1037/ 0033-295x.105.1.125 Reichle, Erik D., Tessa Warren & Kerry McConnell. 2009. Using E-Z Reader to model the effects of higher level language processing on eye movements during reading. Psychonomic bulletin & review 16(1), 1-21. DOI: 10.3758/ PBR.16.1.1 Reimer, Marga. 1996. Quotation Marks: Demonstratives or Demonstrations? Analysis 56(1), 131-42. DOI: 10.1093/ analys/ 56.3.131 Riegelnik, Štefan. 2011. Contextualism and the use-mention distinction. Lodz Papers in Pragmatics 7(2), 281-290. DOI: -10.2478/ v10016-011-0015-4 Ronderos, Camilo R., John M. Tomlinson Jr. & Ira Noveck. 2023. When irony is faster than its literal control: The role of mindreading during irony comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 50(3), 509-522. DOI: -10 .1037/ xlm0001219 Roth, Heide, Andrea N. Lora & Kenneth M. Heilman. 2002. Best practices for cleaning eye movement data in reading research. Behavior Research Methods 125, 2023-2035. DOI: 10.3758/ s13428-023-02137-x Rothstein, Susan. 2004. Copular construction in English. In: Susan Rothstein (ed.) Predicates and Their Subjects (Studies in Linguistics and Philosophy 74), 235-270. Dordrecht: Springer. DOI: -10.1007/ 978-94-010-0690-3_9 Ruigendijk, Esther & Petra B. Schumacher. 2020. Variation in reference assignment processes: psycholinguistic evidence from Germanic languages. The Journal of Com‐ parative Germanic Linguistics 23(1), 39-76. DOI: 10.1007/ s10828-019-09112-x Ryder, Mary E. 1994. Ordered chaos: the interpretation of English noun-noun compounds. Berkeley: University of California Press. Saka, Paul. 1998. Quotation and the use-mention distinction. Mind 107(425), 113-135. DOI: 10.1093/ mind/ 107.425.113 Saka, Paul. 2003. Quotational constructions. In: Philippe De Brabanter (ed.) Hybrid Quotations (Belgian Journal of Linguistics 17). Amsterdam: John Benjamins, 187-212. DOI: 10.1075/ bjl.17.11sak Saxe, Rebecca & Nancy Kanwisher. 2003. People thinking about thinking people: the role of the temporo-parietal junction in “theory of mind”. Neuroimage 19(4), 1835-1842. DOI: 10.1016/ s1053-8119(03)00230-1 Schlechtweg, Marcel & Holden Härtl. 2016. Memorization and the morphology-syntax d ivide: A cross-linguistic investigation.-SKASE Journal of Theoretical Linguistics-13(1), 46-68. Schlechtweg, Marcel & Holden Härtl. 2020. Do we pronounce quotation? An analysis of name-informing and non-name-informing contexts. Language and Speech 63(4), 769-798. DOI: 10.1177/ 0023830919893393 References 223 <?page no="224"?> Schlechtweg, Marcel & Holden Härtl. 2023. Quotation marks and the processing of iron y in English: Evidence from a reading time study. Linguistics 61(2), 355-390. DOI: -10 .1515/ ling-2021-0079 Schlechtweg, Marcel. 2018. The naming potential of compounds and phrases: An empir‐ ical study on German adjective-noun constructions. Word Structure 11(3), 359-384. Schlechtweg, Marcel. 2019. The compound-phrase divide and the lexicon: Insights from non-lexicalized adjective-noun combinations in German. Lingua 227(1), 1-24. DOI: -1 0.1016/ j.lingua.2019.05.009 Schlechtweg, Marcel. 2022. Quotation marks in advanced language acquisition: A reading time experiment on English pure quotation. Linguistik online 118(6), 113-128. DOI: 10.13092/ lo.118.9047 Schlücker, Barbara & Ingo Plag. 2011. Compound or phrase? Analogy in naming. Lingua 121(9), 1539-1551. Schlücker, Barbara. 2012. Die deutsche Kompositionsfreudigkeit. Übersicht und Einfüh‐ rung. In: Livio Gaeta & Barbara Schlücker (eds.) Das Deutsche als kompositionsfreudige Sprache. Strukturelle Eigenschaften und systembezogene Aspekte (Linguistik---Impulse & Tendenzen 46). Berlin: De Gruyter, 1-25. DOI: -10.1515/ 9783110278439.1 Schlücker, Barbara. 2013. Non-classifying compounds in German. Folia Linguistica 47(2), 449-480. Schlücker, Barbara. 2020. Between word-formation and syntax: Cross-linguistic perspec‐ tives on an ongoing debate. Zeitschrift für Wortbildung 4(1), 26-74. DOI: -10.3726/ zwj w.2020.01.02 Schmieder, Falko. 2013. Anführungszeichen als Symptom. Zum historischen Bedeutungs- und Funktionswandel einer Zeichenform. In: Ernst Müller (ed.) Forum interdisziplinäre Begriffsgeschichte (FIB) 2(1), 61-66. DOI: 10.1002/ bewi.201001462 Schneider, Roman & Christian Lang. 2022. Das grammatische Informationssystem grammis - Inhalte, Anwendungen und Perspektiven. Zeitschrift für germanistische Linguistik 50(2), 407-427. Berlin: De Gruyter. DOI: 10.1515/ zgl-2022-2060 Schrader, Sabine. 2017. Konventionelle Anführung vs. modalisierende Anführung: Das Zusammenspiel sprachlicher und pragmatischer Faktoren bei der Erzeugung von un‐ terschiedlichen Anführungslesarten in Benennungskontexten. State examination thesis, University of Kassel. Schwoebel, John, Shelly Dews, Ellen Winner & Kavitha Srinivas. 2000. Obligatory processing of the literal meaning of ironic utterances: Further evidence. Metaphor and Symbol 15, 47-61. DOI: 10.1207/ s15327868ms151&2_4 Scott, Kate. 2019. Referring Expressions, Pragmatics, and Style. London: Kingston Univer‐ sity. DOI: 10.1017/ 9781316822845 Searle, John. 1969. Speech acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. DOI: 10.1017/ cbo9781139173438 224 References <?page no="225"?> Searle, John. 1990. Foreword. In: A. Kronfeld (ed.) Reference and Computation: An Essay in Applied Philosophy of Language (Studies in Natural Language Processing). Cambridge: Cambridge University Press, xiii-xviii. Seres, Daria & M. Teresa Espinal. 2019. Russian definitional generic sentences. Glossa: a journal of general linguistics 4(1) 59, 1-30. DOI: 10.5334/ gjgl.760 Seres, Daria. 2020. The Expression of Genericity in Languages with and without Articles Russian as Compared to English and Romance. Doctoral Dissertation, Universitat Autònoma de Barcelona. Shan, Chung-Chieh. 2010. The character of quotation. Linguistics and Philosophy 33, 417-443. DOI: 10.1007/ s10988-011-9085-6 Shea, Nicholas. 2018. Metacognition and abstract concepts. Philosophical Transactions of the Royal Society B: Biological Sciences 373(1752), 3-7. DOI: -10.1098/ rstb.2017.0133 Siegelman, Noam, Sascha Schroeder, Cengiz Acartürk, Hee-Don Ahn, Svetlana Alexeeva, Simona Amenta, Raymond Bertram, Rolando Bonandrini, Marc Brysbaert, Daria Chernova, Sara Maria Da Fonseca, Nicolas Dirix, Wouter Duyck, Argyro Fella, Ram Frost, Carolina A. Gattei, Areti Kalaitzi, Nayoung Kwon, Kaidi Lõo, Marco Marelli, Timothy C. Papadopoulos, Athanassios Protopapas, Satu Savo, Diego E. Shalom, Natalia Slioussar, Roni Stein, Longjiao Sui, Analí Taboh, Veronica Tønnesen, Kerem Alp Usal & Victor Kuperman. 2022. Expanding horizons of cross-linguistic research on reading: The Multilingual Eye-movement Corpus (MECO). Behavior Research Methods 54 (6), 2843-2863. DOI: 10.3758/ s13428-021-01772-6 Simchen, Ori. 1999. Quotational mixing of use and mention. Philosophical Quarterly 49(196), 325-336. DOI: 10.1111/ 1467-9213.00145 Siyanova-Chanturia,-Anna & Irina Elgort. 2023. Anlysing reading with eye tracking. In: Sandrine Zufferey & Pascal Gygax (eds.) The Routledge Handbook of Experimental Linguistics, 250-267. DOI: 10.4324/ 9781003392972 Snow, Catherine E. 1990. The development of definitional skill. Journal of Child Language 17, 697-710. DOI: 10.1017/ S0305000900010953 Sogenannt. 2022. Duden. Retrieved March 14, 2022 from https: / / www.duden.de/ rechtsch reibung/ sogenannt. Spapé, Michiel & Rinus Verdonschot & Henk van Steenbergen. 2019. The E-Primer: An Introduction to Creating Psychological Experiments in E-Prime. Amsterdam: Leiden University Press. DOI: 10.1017/ 9789400603363 Sperber, Dan & Deirdre Wilson. 1981. Irony and the Use-Mention Distinction. In: Peter Cole (ed.) Radical Pragmatics. New York: Academic Press, 295-318. Sperber, Dan & Deirdre Wilson. 1986. Relevance: Communication and Cognition. Oxford: Blackwell. References 225 <?page no="226"?> Spotorno, Nicola, Anne Cheylus, Jean-Baptiste Van Der Henst & Ira A. Noveck. 2013. What’s behind a P600? Integration Operations during Irony Processing. PLOS ONE 8(6), 1-10. DOI: 10.1371/ journal.pone.0066839 SR Research Ltd. 2013. EyeLink 1000 Plus.-[Apparatus and software]. https: / / www.sr-re search.com/ eyelink-1000-plus/ SR Research Ltd. 2024.-Data Viewer-(Version 4.4.1) [Computer software]. SR Research Ltd. https: / / www.sr-research.com/ data-viewer/ SR Research Ltd. 2024.-Experiment Builder-(Version 2.5.90) [Computer software]. SR Research Ltd. https: / / www.sr-research.com/ experiment-builder/ SR Reserach Ltd. 2009. EyeLink 1000 User Manual (Version 1.5.0). https: / / fchetail.ulb.ac. be/ wp-content/ uploads/ EyeLink-1000-User-Manual-1.5.0.pdf Staub, Adrian & Keith Rayner. 2007. Eye movements and on-line comprehension pro‐ cesses. In: Shirley-Ann Rueschemeyer & M. Gareth Gaskell (eds.) The Oxford handbook of psycholinguistics. Oxford: Oxford University Press, 327-342. DOI: 10.1093/ oxfordhb/ 9780198568971.013.0019 Stei, Erik. 2007. Über Gebrauch und „Bedeutung“ von Anführungszeichen. In: Elke Brendel, Jörg Meibauer & Markus Steinbach (eds.) Zitat und Bedeutung. Hamburg: Buske, 215-235. DOI: 10.46771/ 978-3-87548-955-2 Štekauer, Pavol. 1998. An onomasiological theory of English word-formation. Studies in Functional and Structural Linguistics. Amsterdam: John Benjamins. DOI: 10.1075/ sfsl.46 Štekauer, Pavol. 2005. Onomasiological approach to word-formation. In: Pavol Štekauer & Rochelle Lieber (eds.) Handbook of word-formation. Studies in Natural Language and Linguistic Theory 64. Dordrecht: Springer, 207-232. DOI: 10.1007/ 1-4020-3596-9_9 Stern, Josef. 2022. Quotations as Pictures. Cambridge: The MIT Press. DOI: -10.7551/ mitp ress/ 11210.001.0001 Sternberg, Robert J. & Talia Ben-Zeev. 2001. Complex cognition. The psychology of human thought. New York: Oxford University Press. Stites, Mallory C., Steven G. Luke & Kiel Christianson. 2013. The psychologist said quickly, “Dialogue descriptions modulate reading speed! ”. Memory & Cognition 41, 137-151. DOI: 10.3758/ s13421-012-0248-7 Stowell, Timothy Angus. 1981. Origins of phrase structure. Doctoral Dissertation, Massa‐ chusetts Institute of Technology. Svanlund, Jan. 2018. Metalinguistic comments and signals: What can they tell us about the conventionalization of neologies? Pragmatics & Cognition 25(1), 122-141. DOI: -1 0.1075/ pc.18005.sva Tarski, Alfred. 1933. The concept of truth in formalized languages. In: Alfred Tarski (ed.) Logic, Semantics, Metamathematics 2. Indianapolis: Hackett, 152-278. Tarski, Alfred. 1956. Logic, Semantics, Metamathematics. Oxford: Oxford University Press. 226 References <?page no="227"?> Venables, William N. & Brian D. Ripley. 2002. Modern Applied Statistics with S. New York: Springer. DOI: 10.1007/ 978-0-387-21706-2. Vitu, Françoise & George W. McConkie. 2012. Regressive saccades and word perception in adult reading. In: Alan Kennedy, Ralph Radach, Dieter Heller & Joël Pynte (eds.) Reading as a Perceptual Process. North-Holland: Elsevier Science Publishers, 301-326. DOI: 10.1016/ B978-008043642-5/ 50015-2 Vitu, Françoise. 2012. On the role of visual and oculomotor processes in reading. In: Simon P. Liversedge, Iain Gilchrist & Stefan Everling (eds.) The Oxford Handbook of Eye Movements. Oxford: Oxford Library of Psychology, 732-749. DOI: -10.1093/ oxfordhb/ 9780199539789.001.0001 Völkel, Svenja & Franziska Kretzschmar. 2021. Introducing linguistic research. Cambridge: Cambridge University Press. DOI: 10.1017/ 9781316884485 Wade, Nicholas J. 2010. Pioneers of eye movement research. I-perception 1(2), 33-68. DOI: 10.1068/ i0389 Wang, Xiaofei. 2018. Quotation and truth-conditional pragmatics. London: Routledge. Warren, Beatrice. 1978. Semantic Patterns of Noun-Noun Compounds. Doctoral Disserta‐ tion, Gothenburg Studies in English. Washington, Corey. 1992. The identity theory of quotation. Journal of Philosophy 89, 582-605. DOI: 10.2307/ 2941057 Watson, Rita. 1985. Towards a theory of definition. Journal of Child Language 24, 535-65. DOI: 10.1017/ S0305000900006309 Waugh, Linda R. 1980. The Poetic Function in the Theory of Roman Jakobson. Poetics Today, 2(1), 57-82. DOI: 10.2307/ 1772352 Wickham, Hadley, Jennifer Bryan, Marcin Kalicinski, Komarov Valery, Christophe Leitienne, Bob Colbert, David Hoerl, Evan Miller & Jennifer Bryan. 2019. R Package ‘readxl’. Version 1.3. Wickham, Hadley. 2016. ggplot2: Elegant Graphics for Data Analysis. New York: Springer. https: / / ggplot2.tidyverse.org Wildt Eugenia, Katharina J. Rohlfing & Ingrid Scharlau. 2019. The Role of Sali‐ ency in Learning First Words. Frontiers in Psychology 10(1150), 1-8. DOI: -10.3389/ fpsyg.2019.01150 Williamson, Timothy. 2000. Knowledge and its Limits. Oxford: Oxford University Press. Wilson, Deirdre & Dan Sperber. 1992. On Verbal Irony. Lingua 87(1), 53-76. DOI: -10.101 6/ 0024-3841(92)90025-e Wilson, Shomir. 2011. A computational theory of the use-mention distinction in natural language. Doctoral Dissertation, University of Maryland. Wilson, Shomir. 2012. The creation of a corpus of English metalanguage. Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics, 638-646. Wiślicki, Jan. 2014. Semantics of quotation. Proceedings of PhML 2014, 238-249. References 227 <?page no="228"?> Wiślicki, Jan. 2021. Quotation as a modality. Unpublished paper (draft) available here: h ttps: / / ling.auf.net/ lingbuzz/ 005324 Yao, Bo & Christoph Scheepers. 2011. Contextual modulation of reading rate for direct versus indirect speech quotations. Cognition 121(3), 447-453. DOI: 10.1016/ j.cognitio n.2011.08.007 Zimmer, Karl E. 1971. Some general observations about nominal compounds. Working Papers on Language Universals 5. 228 References <?page no="229"?> 121 The factors are listed as follows: 1) Verb, 2) Nominal, 3) Quotes, 4) NameType, 5) WordStatus, 6) Frequency, 7) FrequencyClass Appendix Appendix A: Corpus study Excepts from DeReKo and BNC Concordance Factors 121 Weil sich die Bevölkerungszahl Amriswils per 31. Dezember 1988 auf 9710 erhöht hatte, kam die Diskussion auf, ob sich das „Welt‐ dorf “ bald „Weltstadt“ zu nennen hätte. Dies, zumal in Amriswil durchaus bereits städtische Verhältnisse herrschten, wie der Zeitungskolumnist mit Hinweis auf mehrere Vorfälle von Einbrü‐ chen und Diebstählen feststellte. Zur Stadt schlug sich Amriswil erst etliche Jahre nach Überschreitung der 10000er-Grenze, näm‐ lich 2005. ‘Since the population of Amriswils had increased to 9,710 as of 31 December 1988, the debate arose as to whether the “town” should soon be called “city”. This is especially because urban conditions already prevailed in Amriswil, as the newspaper columnist noted with reference to several incidents of burglary and theft. Amris‐ wil joined the city only several years after crossing the 10000s limit, namely in 2005.’ 1) Nennen 2) “Weltstadt” 3) Yes 4) Compound 5) One-word 6) 16 7) Low Am 19. Januar hatte ein nordkoreanischer Militärsprecher mit Kampfhandlungen gedroht, nachdem die südkoreanische Küstenwacht illegal eingedrungene Fischerboote nach Nord‐ korea zurückgedrängt hatte. Pjöngjang nannte dies „Provokatio‐ nen“ der südkoreanischen Marine, die bereits eine „Gefahrenlinie überschritten“ hätten. ‘On January 19th, a North Korean military spokesman threatened to fight after the South Korean Coast Guard illegally pushed back fishing boats into North Korea. Pyongyang called this “provocations” by the South Korean Navy, which had already “crossed a dangerous line”.’ 1) Nennen 2) “Provokatio‐ nen” 3) Yes 4) Simple noun 5) One-word 6) 13 7) Low Pargätzi wohnte 40 Jahre an der Unteren Mariabergstrasse. 2007, nach dem Tod seiner Gattin Margrit, zog er in den Teil für betreutes Wohnen des regionalen Pflegeheims. In seinem Zimmer ist er umgeben von Filmrollen und Fotoalben, die ihn nicht nur an seine Familie, sondern auch jederzeit an einmaligen Erlebnisse in den Bergen und seine Leistungen als Erfinder und Fotograf erinnern. Er ist rüstig, nimmt Anteil am öffentlichen Leben und hofft, gesund seinen 90. Geburtstag am Weihnachtstag 2009 erleben zu dürfen. Die heutigen Extremkletterer bezeichnet er als Artisten. 1) Bezeichnen 2) Artisten 3) No 4) Simple noun 5) One-word 6) 15 7) Low <?page no="230"?> ‘Pargätzi lived for 40 years on the lower Mariaberg street. In 2007, after the death of his wife Margrit, he moved to the part of the regional nursing home. In his room he is surrounded by film roles and photo albums that remind him not only of his family, but also of unique experiences in the mountains and his achievements as an inventor and photographer. He is busy, participating in public life and hoping to be able to celebrate his 90th birthday on Christmas Day 2009. He calls today extreme climber’s artists.’ Sein „Gambrinus“-Konzert im 2006 wurde vom spärlichen Publi‐ kum als „the best thing ever“ und im St. Galler Tagblatt schlicht als „Geniestreich“ bezeichnet. Der vietnamesische Trompeter Cuong Vu verfügt über eine unglaubliche Musikalität, gepaart mit Virtuosität und einer überbordenden Kreativität. Im Trio kreiert er mit Bass, Drums und elektronischen Effekten phantas‐ tische Sounds. ‘His “Gambrinus”-concert in 2006 was labeled "the best thing ever” by the sparse ALS audience and simply “stroke of genius” in the St. Galler news outlet. The Vietnamese trumpet player Cuong Vu has an incredible musicality, coupled with virtuosity and overwhelming creativity. In the trio, he creates fantastic sounds with bass, drums, and electronic effects.’ 1) Bezeichnen 2) Geniestreich 3) No 4) Simple noun 5) One-word 6) 16 7) Low yes I know what he means now. </ s><s> A meteor is is something rock floating through space. </ s><s> When it comes into our atmosphere it's called a meteorite and most of them burn off as they come through the atmosphere. </ s><s> Some of them and yet you get huge damage. 1) Call 2) Meteorite 3) No 4) Simple noun 5) One-word 6) 16 7) Low Well, it needs to be Q P9 then doesn’t it? </ s><s> Where we were just now with the research then talk about questionnaires called research approval forms. No there is a separate one for qualitative and it is different. 1) Call 2) Research appro‐ val forms 3) No 4) Compound 5) Multi-word 6) 19 7) Low Subordinate entities are connected to the root and further subor‐ dinates connected to them. </ s><s> These are often referred to as parent/ child relationships. </ s><s> Fig. 4.13 Hierarchy with siblings </ s><s> Figure 4.13 shows the basic structure. 1) Refer 2) parent/ child re‐ lationships 3) No 4) Compound 5) Multi-word 6) 20 7) Low 230 Appendix <?page no="231"?> And she was absolutely exhausted. </ s><s> She wondered if she was getting sick. </ s><s> If this was what Mandy happily referred to as romance, she decided that being an intern at a busy inner-city hospital was marginally less stressful. 1) Refer 2) Romance 3) No 4) Simple noun 5) One-word 6) 12 7) Low Appendix A: Corpus study 231 <?page no="232"?> Appendix B: Rating Study Appendix B1: Instructions for rating the interpretation Die Einordnung der Lesart soll wie in der Tabelle auf einer Skala von 1 bis 5 bewertet werden: ‘The classification of the interpretation needs to be classified as illustrated in the Table-on a scale from 1 to 5: ’ 232 Appendix <?page no="233"?> Namensinformierende Lesart ‘name-informing interpretation’ Grenzfälle ‘borderline cases’ Modalisierende Lesart ‘modalizing interpretation’ Label für die Analyse ‘labels for the analysis’ 1 2 3 4 5 Beispielsätze ‘sample sentences’ Neben dem schottischen Instru‐ ment hat sich in der Bordun‐ musik Mitteleuropas eine Fami‐ lie eng verwandter Dudelsäcke mittlerer Lautstärke durchge‐ setzt, die Sackpfeifen genannt werden. ‘In addition to the Scottish in‐ strument, a family of closely related medium-volume pipes, known as bagpipes, has estab‐ lished itself in the Bordun music of Central Europe.’ - - - - Der junge Karlsruher Kunsthis‐ toriker Wolfgang Ullrich nennt solche inhaltsleere Bilder Joker: Sie sind überall einsetzbar, weil man auf ihnen weder die Insel Reichenau noch das Dorf Berlin‐ gen erkennt, sie haben weltweite Gültigkeit. ‘Wolfgang Ullrich, the young art historian from Karlsruhe, calls such content-free images jokers: they can be used everywhere as one cannot recognize the is‐ land of Reichenau or the village of Berlingen; they are valid all around the world.’ Tab. 28: Instructions for rating the interpretation Appendix B: Rating Study 233 <?page no="234"?> Detaillierte Instruktionen zur Einordung der Lesart (‘detailed instructions for labelling’): Definitionen (‘definitions’) Namensinformierende Lesart (konventionelle Anführung) [Label: 1] ‘Name-informing interpretation (conventionalized reading) [label 1]’ In der konventionellen Anführung wird ein Ausdruck als Name präsentiert, der einen extensionalen Bezug hat. Das Namensverben nennen oder bezeichnen als beziehen sich auf ein Nomen, welches eingeführt wird. Dieses Nomen informiert über den konventionellen Namen eines lexikalisierten Begriffes: Diese Krankheit nennt man ‚Sepsis‘. Häufig sind diese namensinformierenden Konstruktionen mit einer Erklärung, Beschreibung oder Definition des Wortes verbunden. ‘In a conventionalized interpretation, an expression is presented as a name that has an extensional reference. The naming predicate nennen (‘call’) or bezeichnen als (‘refer to as’) is related to the nominal being introduced. The nominal informs about the conventionalized name of a lexicalized term: This disease is called “sepsis”. These name-informing constructions are oftentimes associated with an explanation, description or definition of the word.’ Beispielsätze aus den Daten: ‘sample sentences from the corpus data: ’ mit nennen: „Diese Tatsache wird auch heute gerne von der Medizin genutzt - man nennt das den Placeboeffekt.“ ‘involving nennen (‚call‘): “This fact is still used in medicine today - it’s called the placebo effect.” mit bezeichnen als: „Mit Inkunabeln bezeichne man Bücher aus der Wiege der Buchdruckkunst seit 1450 bis 1500.“ ‘involving bezeichnen als (‘refer to as’): “Books from the cradle of the printing art from 1450 to 1500 are referred to as incunables” Modalisierende Lesart (modalisierende Anführung) [Label: 5] Eine modalisierende Lesart weist auf eine nicht-standardmäßige Verwendung (bei Ironie) oder auf eine unangebrachte Verwendung (bei Distanzierung = ein Sprecher distanziert sich gegenüber einer Äußerung; negative oder wertende Sprechereinstellung) eines Begriffs hin. Im Gegensatz zu der namensinformier‐ enden Lesart erhält die modalisierende Lesart neben der Standardinterpretation eine weitere, vom Kontext beeinflusste Interpretation. Modalisierende Anfüh‐ rung benutzt einen lexikalisierten Ausdruck, der im Fall von Ironie nicht mit seiner konventionellen Extension benutzt wird oder im Fall von Distanzierung dessen Extensionszugehörigkeit ablehnt. 234 Appendix <?page no="235"?> ‘A modalizing reading refers to a non-standard use (in the case of irony) or to an inappropriate use (a distancing use = a speaker distances themself from an expression; negative or judging speaker attitude) of a term. In con‐ trast to the name-informed interpretation, the modalizing reading receives an interpretation influenced by context in addition to the standard interpretation. Modalizing readings make use of a lexicalized expression that is not used with a conventional extension. In distancing utterances, the extension’s affiliation is rejected.’ Ironischer Beispielsatz: Kontext: es regnet ‘context: it’s raining’ A: „Tolles Wetter heute.“ ‘A: “Nice wheather.”’ Bedeutung: Das Wetter ist nicht toll. ‘Meaning: The weather is not nice.’ Der Satz transportiert eine Sprechereinstellung, die nur aus dem Kontext ersichtlich wird. ‘The sentence has an underlying meaning and transports a speaker’s attitude that only becomes evident when taking the context into consideration.’ Beispielsätze mit sogenannt: Und gerade die Unfähigkeit der politischen Entscheidungsträger anstehende Probleme wirklich lösen zu wollen, zu kön‐ nen, beweist diese sogenannte „Gesundheitsreform“/ „Das Verhalten der beiden sogenannten Volksparteien zeigt deutlich, dass sie ihre demokratischen Verp‐ flichtungen vor lauter Machtstreben nicht mehr kennen.“ ‘Examples with so-called: The inability of political decision makers to really want to solve the problems ahead, is proven by the so-called “health reform”/ “The behaviour of the two so-called people’s parties clearly shows that they no longer know their democratic obligations because of their desire for power.”’ Beispielsatz mit bezeichnen als: Den Umgang mit der Doppelbelastung Beruf und politisches Amt bezeichnet er als „Spagat“. ‘Examples with refer to as: He describes dealing with the double burden of occupation and political office as “balancing act”. On the one hand, those who say yes to a position want to do the best they can.’ Grenzfälle [Label: 3] ‘borderline cases [label 3]’ Beispielsatz mit sogenannt: Zum ersten Mal zur Prämierung zugelassen waren sogenannte Exoten, also Brände aus ursprünglich nicht regionalen Grundstof‐ fen, und Liköre. Appendix B: Rating Study 235 <?page no="236"?> 122 All sample sentences provided here are taken from the DeReKo corpus (COSMAS II). ‘Examples with so-called: For the first time, the so-called exotics, i.e. brandies from originally non-regional raw materials, and liqueurs, were allowed to be awarded.’ Dieser Satz ist ein Grenzfall: Der Kontext des Satzes verweist darauf, dass der Verfasser eine gewisse Einstellung bzw. seinen Unmut äußert, welches durch die Verwendung des Wortes ‚Exoten‘ deutlich wird. Andererseits enthält der Satz eine Definition, die häufig bei namensinformierenden Lesarten vorkommen. ‘This sentence is a boundary case: The context of the sentence indicates that the author expresses a certain attitude or disapproval, which is evident by the use of the word ‘exotics’. On the other hand, the sentence contains a definition that is often found in name-informing readings.’ Label 2: Sätze mit Interpretation, die eher zur namensinformierenden Lesart tendieren ‘label 2: sentences with a tendency to a name-informing reading’ Label 4: Sätze mit Interpretation, die eher zur modalisierenden Lesart tendieren ‘label 4: sentences with a tendency to a modalizing reading’ Appendix B2: Sample sentences with mean rating of participants 122 Ratings for bezeichnen Scale rating 1 Mit Anlagen von rund 2,43 Milliarden Franken in den Büchern schloss sie mit einem Deckungsgrad von 115 Prozent. Mit dem Deckungsgrad wird die Fähigkeit der Kasse bezeichnet, alle versprochenen Leistungen zu erfüllen. ‘She closed with a coverage rate of 115 percent with investments of around 2.43 billion francs. The degree of coverage refers to the ability of the cashier to fulfil all promised services.’ Scale rating 2 Mit den Dscheladas haben Affen Einsitz im neuen Gehege genommen, die schon seit 50 Jahren im Zürcher Zoo wohnen. Als eine Rasselbande bezeichnet der Zoo-Kurator Robert Zingg die 25köpfige Affengruppe. ‘With the Jeladas, monkeys have taken their place in the new cage, which have been living in the Zurich Zoo for 50 years. The Zoo curator Robert Zingg describes the 25-member group as a racing band.’ 236 Appendix <?page no="237"?> Scale rating 3 Die Diskussion darüber ist entbrannt, ob es in Ordnung ist oder nicht, wenn UBS-Manager schon wieder Boni einbzw. abkassieren. Die Frage wäre einfach zu beantworten, doch flugs haben die Boni-Befürworter die Diskussion auf eine andere Ebene gehoben, nämlich zur Frage nach der Definition: Ist das, was die Manager bekommen sollen, überhaupt als Bonus zu bezeichnen oder nicht. ‘There has been a debate about whether it is okay or not for UBS managers to reclaim bonuses again. The question would be simple to answer, but soon the bonus advocates have taken the discussion to a different level, namely to the question of the definition: is what managers get to call a bonus at all or not.’ Scale rating 4 In den Korb der schlechten Nachrichten gehörten das Polittheater in Bern und die Finanzkrise mit den negativen Auswirkungen auf die Wirtschaft. Als Wort oder Unwort des Jahres bezeichnete Walter Luginbühl das „Rettungspaket“. ‘The basket of bad news included the political theatre in Bern and the financial crisis with its negative impact on the economy. Walter Luginbühl named the “rescue package” the word of the year.’ Scale rating 5 Gerade dreissig Tage sind vergangen, seit die SVP ihr Positionspapier zum Bankgeheimnis präsentierte. Darin heisst es: «Eine Aufweichung kommt für die SVP NICHT in Frage! » Alles Schnee von gestern. Nun schwenkt die SVP auf die Linie des Bundesrates und der anderen Parteien um, die sie in Inseraten vom Wochenende noch als Landesverräter bezeichnet hatte. Glaubwürdige Politik sieht anders aus. ‘Only thirty days have passed since the SVP presented its position paper on banking secrecy. It says: “A mitigation is NOT in question for the SVP! ” All yesterday's snow. Now the SVP turns to the line of the Bundesrat and the other parties, which it had described in ads from the weekend as “traitor”. Credible politics looks different.’ Ratings for nennen Scale rating 1 Neben dem schottischen Instrument hat sich in der Bordunmusik Mitteleur‐ opas eine Familie eng verwandter Dudelsäcke mittlerer Lautstärke durchgesetzt, die Sackpfeifen genannt werden. Dazu gehören die französische Cornemuse, die deutsche Schäferpfeife und der flämische Dudelsack. ‘In addition to the Scottish instrument, a family of closely related me‐ dium-volume dudelsacks, known as sackpipes, has established itself in Central Appendix B: Rating Study 237 <?page no="238"?> Europe’s bordun music. These include the French Cornemuse, the German Shepherd pipe and the Flemish bagpipe.’ Scale rating 2 Wird es draussen kalt, dann bedeutet das für unseren Körper: aufgepasst! Viele Thermorezeptoren in der Haut melden dem Hypothalamus im Gehirn, dass da draußen etwas los ist. Jetzt bildet sich, was wir eine Gänsehaut nennen. Die Haut zieht sich zusammen zu unzähligen kleinen Hügelchen. ‘If it gets cold outside, it means for our bodies: be careful! Many thermo-re‐ ceptors in the skin tell the hypothalamus in the brain that something is going on out there. Now what we call a goose bump is forming. The skin draws together into countless small hills.’ Scale rating 3 Die Anstrengung aller verantwortlichen Lehrkräfte, auch des Schulratspräsi‐ denten Daniel Kürsteiner, trug bekanntlich Früchte. Bedauerlicherweise wurde dann unter der Ägide gewisser Leute eine Schulpolitik aufgegleist, die ich nicht mittragen konnte. An dieser Stelle muss deshalb auch einmal das „Hauptübel“ in Lichtensteig genannt werden: Vetternwirtschaft und Kurzsichtigkeit. ‘The efforts of all responsible teachers, including school council chairman Daniel Kürsteiner, were well-known. Unfortunately, under the auspices of some people, a school policy was set up which I was unable to support. At this point, therefore, the “main problem” in Lichtensteig must be called: nepotism and short-sightedness.’ Scale rating 4 Am 19. Januar hatte ein nordkoreanischer Militärsprecher mit Kampfhand‐ lungen gedroht, nachdem die südkoreanische Küstenwacht illegal eingedrun‐ gene Fischerboote nach Nordkorea zurückgedrängt hatte. Pjöngjang nannte dies „Provokationen“ der südkoreanischen Marine, die bereits eine „Gefahrenlinie überschritten“ hätten. ‘On 19 January, a North Korean military spokesman threatened fighting after the South Korean Coast Guard pushed illegal fishing boats back into North Korea. Pyongyang called this a “provocation” by the South Korean Navy, which had already “crossed a line of danger”.’ Scale rating 5 Es dauert allerdings, bis das klar wird, und in der Zwischenzeit werden wir höchst unterhaltsam und durchaus realistisch in den Alltag einer Klinik eingeführt. Wir lernen den unangenehmen Chefarzt kennen, von seinen armen 238 Appendix <?page no="239"?> Untergebenen „Kugelblitz“ genannt, und auch die charmante Patientin, deren Schenkelhals operiert werden soll. ‘However, it takes time to make this clear, and in the meantime, we are introduced to the everyday life of a clinic in a very reckless and quite realistic way. We get to know the unpleasant chief physician, called “ball lightning” by his poor subjects, and also the charming patient whose thigh neck is to be operated.’ Appendix B3: Total counts for the predicates (mean values by participants) Rating on the scale Counts for nennen Counts for bezeichnen 1 50 74 1.5 75 41 2 69 65 2.5 38 76 3 55 71 3.5 25 58 4 33 66 4.5 21 48 5 6 35 Total count 372 534 Tab. 29: Total counts for the predicates Appendix B: Rating Study 239 <?page no="240"?> Appendix C: Self-paced reading study Appendix C1: Instructions for the SPR-study Instructions presented on Slide 1 (originally in German, translated under‐ neath) Willkommen in unserem Sprachlabor! In unserem linguistischen Experiment untersuchen wir das Satzverstehen. Die Ergebnisse sind anonym und werden ausschließlich für wissenschaftliche Zwecke verwendet. Die Studie dauert ca. 15 Minuten. Gleich werden Sie immer jeweils einen Satz auf dem Bildschirm sehen. Zuerst sehen Sie nur Striche, die anzeigen, wie lang der Satz jeweils sein wird. Um den ersten Teil des Satzes zu sehen, drücken Sie bitte die LEERTASTE. Sobald Sie wieder die LEERTASTE drücken, erscheint der nächste Teil und der vorhergehende verschwindet wieder und so weiter. WEITER mit LEERTASTE Welcome to our language lab! In our linguistic experiment, we examine the understanding of sentences. The results are anonymous and are used exclusively for scientific purposes. The study takes approximately 15 minutes. Immediately you will always see one sentence on the screen. First, you will see only lines that indicate how long the sentence will be. To view the first part of the sentence, please press the empty key. Once you press the empty button again, the next part appears and the previous part disappears again and so on. Press the SPACE BAR to continue Instructions presented on Slide 2 Bitte lesen Sie die Sätze zügig, aber in Ihrem eigenen, nor‐ malen Lesetempo, sodass Sie den Inhalt verstehen. Lassen Sie wäh‐ rend des Lesens immer Ihren linken oder rechten Zeigefinger auf der LEERTASTE. Betätigen Sie stets die LEERTASTE, nachdem Sie einen Satzabschnitt gelesen haben, um den nächsten zu sehen. Im Anschluss wird Ihnen jeweils eine Verständnisfrage gestellt, die sich ausschließlich auf den zuvor gelesenen Satz bezieht. Bitte beantworten Sie diese Frage entweder mit "1" = JA oder "9" = NEIN, wie auf der 240 Appendix <?page no="241"?> Tastatur angegeben. Eine einmal gegebene Antwort können Sie nicht mehr revidieren. Machen Sie einfach mit dem nächsten Satz weiter. Bevor das Experiment startet, gibt es einen Probedurch‐ gang mit drei Beispielsätzen. Sollten Sie noch Fragen ha‐ ben, wenden Sie sich bitte jetzt an die Versuchsleiterin. LEERTASTE startet den ÜBUNGSBLOCK Please read the sentences quickly, but at your own normal reading rate, so that you understand the content. Always keep your left or right index finger on the empty button while reading. Always press the empty key after reading a sentence to see the next one. You will then be asked a comprehension question, which relates exclu‐ sively to the previously read sentence. Please answer this question with either "1" = YES or "9" = NO, as indicated on the keyboard. You can‐ not revise an answer once given. Just proceed with the next sentence. Before the experiment starts, there is a sample pass with three sample sentences. If you have any further questions, please contact the pilot now. Press the SPACE BAR to start the EXERCISE BLOCK Appendix C2: Training Items of the SPR-Study The following three practice items are listed in the order they have been presented to the participants in the practice trial. The dashes are used here to indicate where the sentence is broken up into segments. Practice Item 1 (MODC_ns): Die Pizza des italienischen Kochs ist überhaupt nicht durchgebacken. / So jemand / nennt sich / Pizzabäcker, / kommentiert Mia. / Wollte Mia eine Pizza essen? ‘The Italian chef ’s pizza is not well cooked at all. / Someone like that / calls himself / pizza maker, / commented Mia. Did Mia want to eat a pizza? ’ (Correct answer: Yes) Practice Item 2 (NIhighC_ns): Der Biologe hat eine neue Pflanzenart bestimmt. / So jemanden / bezeichnet man als / Forscher, / erklärt Julian. Bezeichnet man so etwas als Politik? ‘The biologist has identified a new plant species. / Appendix C: Self-paced reading study 241 <?page no="242"?> Someone like that / is referred to as / researcher, / explains Julian. Is that what you call politics? ’ (Correct Answer: No) Practice Item 3 (MODC_nm): Der Leiter des Sinfonieorchesters hat seinen Dirigentenstab zu Hause verges‐ sen./ So jemanden / nennt man / Dirigenten, / kommentiert die Pianistin. / Erklärt Susanne etwas? ‘The conductor of the symphony orchestra has forgotten his conductor’s staff at home./ Someone like that/ is called / conductor, / commented the pianist. Does Susanne explain anything? ’ (Correct Answer: No) Appendix C3: Target words and their respective lexical frequency word type & frequency target word lexical frequency mean value SD compound, high frequency class 16.875 0.599 - Weinkenner wine connoisseur 18 - - - Rabattaktion discount promotion 16 - - - Honigbiene honey bee 17 - - - Beamtenstatus civil servant status 17 - - - Eisbrecher icebreaker 16 - - - Kundenbetreuer customer service officer 17 - - - Literaturverfilmung literary adaption 17 - - - Pflanzenwachstum plant growth 17 - - 242 Appendix <?page no="243"?> compounds, low frequency class 8.75 0.433 - Pressekonferenz press conference 9 - - - Bürgermeister mayor 8 - - - Klimaschutz climate protection 9 - - - Staatsanwaltschaft public prosecutor‘s office 8 - - - Landesregierung state government 9 - - - Gesundheitsminister health minister 9 - - - Geschäftsführer general manager 9 - - - Gemeinderat district council 9 - - Simple noun, high frequency class 17.125 1.536 - Artistik artistry 16 - - - Schorle spritzer 19 - - - Fabrikant industrialist 18 - - - Konditor confectioner 16 - - - Pomologe pomologist 20 - - - Film film 16 - - - Kosmetiker cosmetician 16 - - - Veranda veranda - 16 - - - Simple noun, low frequency class 8.125 1.165 - Freund friend 9 - - - Schauspieler actor 9 - - - Unternehmen company 6 - - - Ermittler investigator 9 - - - Haus house 7 - - Appendix C: Self-paced reading study 243 <?page no="244"?> Straße street 7 - Künstler artist 9 - - - Richter judge 9 - - Tab. 30: SPR target words and Frequency | Note: SD has been rounded to the third decimal place Appendix C4: Overview of all experimental items Context Sentence Target Sentence Spillover 1,1, MODhighC_ns Kim berichtet in einer Mail von einem Treffen mit einem Kollegen, der verkorkten Wein ser‐ vierte. So jemand nennt sich Weinkenner, merkt Kim an. ‘Kim writes in an email about a meeting with a colleague, who served bottled wine.’ ‘Someone like that calls themselves wine connoisseur,’ ‘remarks Kim’ 1, 2, NIhighC_ns Der Besitzer der Weinkellerei erklärt die Beson‐ derheiten von verschiedenen Weinen. So jemand nennt sich Weinkenner, sagt Maria. ‘The owner of the winery explains the peculiar‐ ities of different wines.’ ‘Someone like that calls themselves wine connoisseur,’ ‘says Maria.’ -2, 1. MODhigh_ns- Die Kundin stellt entsetzt fest, dass das Angebot der Woche im Supermarkt genau so viel kostet wie sonst. So etwas nennt sich Rabattaktion, kommentiert ein Mann. ‘The customer is horrified to find that the weekly offer at the supermarket costs just as much as it normally does.’ ‘They call that a dis‐ count promotion,’ ‘comments a man.’ 2, 2, NIhighC_ns Die Politiker diskutieren angesichts gestiegener Energiepreise über eine finanzielle Erleichter‐ ung für alle Bürger. So etwas nennt sich Rabattaktion, Kommentiert ein Journalist. ‘Politicians discuss financial relief for all citi‐ zens in the face of rising energy prices.’ ‘Something like that is called a dis‐ count promotion,’ ‘comments a journalist.’ 244 Appendix <?page no="245"?> 3, 1, MODhighC_nm Die Kindergartenkinder versuchen erfolglos, eine wählerische Biene mit Honig zu füttern. So etwas nennt man Honigbiene, sagt Luis. ‘Kindergarten children unsuccessfully try to feed a picky bee with honey.’ ‘Something like that is called a honey bee,’ ‘says Luis.’ 3, 2, NIhighC_nm- Die Kinder beobachten ein kleines Fluginsekt, das an einer Blüte Nektar trinkt. So etwas nennt man Honigbiene, erklärt Susan-ne. ‘The children watch a small flying insect drink‐ ing nectar from a flower.’ ‘Something like that is called a honey bee,’ ‘explains Sus‐ anne.’ 4, 1, MODhighC_nm Die verbeamtete Lehrkraft bemängelt, dass sie seit Wochen kein Geld erhalten hat. So etwas nennt man Beamtenstatus, kommentiert eine Freun‐ din. ‘The civil servant teacher complains that she has not received any money for weeks.’ ‘That is called civil servant status,’ ‘comments a friend.’ 4, 2, NIhighC_nm Die Angestellte erklärt ihrem Freund, dass sie eine Beschäftigung im öffentlichen Dienst ohne Streikrecht ausübt. So etwas nennt man Beamtenstatus, erklärt Sam. ‘The employee tells her friend that she is em‐ ployed in the public service without the right to go on strike.’ ‘Something like that is called public servant status,’ ‘explains Sam. ’ 5, 1, MODhighC_bez Das speziell für vereiste Routen konstruierte Schiff ist im dicken Packeis stecken geblieben. So etwas bezeichnet man als Eisbrecher, sagt Luise. ‘The ship that is specifically designed for icy routes got stuck in the thick pack ice.’ ‘Something like that is called an ice‐ breaker,’ ‘says Luise.’ 5, 2, NIhighC_bez Auf der Führung durch das Marinemuseum deu‐ tet der Guide auf das schwerste Schiff der Aus‐ stellung, welches Packeis durchqueren kann. So etwas bezeichnet man als Eisbrecher, erklärt der Guide. ‘On a guided tour at the maritime museum, the guide points to the heaviest ship of the exhibition, which can cross pack ice’ ‘Something like this is called an ice‐ breaker,’ ‘explains the guide.’ Appendix C: Self-paced reading study 245 <?page no="246"?> 6, 1, MODhighC_bez Elisa wartet schon seit Tagen auf eine Antwort ihres Ansprechpartners für eine Rücksendung. So jemanden be‐ zeichnet man als Kundenbetreuer, kommentiert Leon. ‘Elisa has been waiting for days for a response from her contact regarding a return shipment.’ ‘Someone like this is called a customer service officer,’ ‘comments Leon.’ 6, 2, NIhighC_bez Der Abteilungsleiter erklärt Leonie, was ihre Arbeit im Kundenservice ist. So jemanden be‐ zeichnet man als Kundenbetreuer, sagt eine Mi‐ tarbeiterin. ‘The head of the department explains to Leonie her work in customer service.’ ‘Someone like this is called a customer service officer,’ ‘says a cow‐ orker.’ 7, 1, MODhighC_bez- Die Bibliothekare diskutieren hitzig über einen Film, dessen Handlung maßgeblich von der Buchvorlage abweicht. So etwas bezeichnet man als Literatur‐ verfilmung, kommentiert Tom. ‘The librarians are heatedly discussing a film whose plot differs significantly from the book template.’ ‘Something like this is called a literary adaptation’ ‘comments Tom.’ 7, 2, NIhighC_bez Das Open-Air Kino präsentiert einen Filmklas‐ siker, der auf einem literarischen Werk basiert. So etwas bezeichnet man als Literatur‐ verfilmung, sagt der Re‐ porter. ‘The open-air cinema presents a movie based on a literary piece of work.’ ‘Something like this is called literary adaptation,’ ‘says the re‐ porter.’ 8, 1, MODhighC_bez Die gesäte Kürbispflanze ist den ganzen Som‐ mer über kaum gewachsen. So etwas bezeichnet man als Pflanzen‐ wachstum, sagt Antonia. The squash plant that was planted barely grew throughout the summer. ‘Something like this is called plant growth,’ ‘says Anto‐ nia.’ 8, 2, NIhighC_bez Die Botaniker dokumentieren über einen Zei‐ traum, wie sich die Pflanze entwickelt. So etwas bezeichnet man als Pflanzen‐ wachstum, beschreibt der Experte. 246 Appendix <?page no="247"?> ‘The botanists document the plants develop‐ ment over a period of time.’ ‘Something like this is called plant growth,’ ‘The expert describes.’ 9, 3, MODlowC_ns Auf der organisierten Veranstaltung haben die Pressesprecherinnen keine Fragen der Journal‐ isten beantwortet. So etwas nennt sich Pressekonferenz, kommentiert Ben. ‘At the organized event, the press speakers did not answer any questions from journalists.’ ‘Something like this is called press con‐ ference,’ ‘comments‐ Ben.’ 9, 4, NIlowC_ns Die Regierungssprecherin hat zu einer Veran‐ staltung geladen, auf der sie für die Medien Fragen beantwortet. So etwas nennt sich Pressekonferenz, sagt Lisa. ‘The government spokesperson has invited to an event where she answers questions for the media.’ ‘Something like this is called press con‐ ference,’ ‘says Lisa.’ 10, 3, MODlowC_ns Der Leiter der Stadtverwaltung drückt sich vor der Umsetzung von Beschlüssen. So jemand nennt sich Bürgermeister, kommentiert eine Frau. ‘The head of the city administration avoids the implementation of decisions.’ ‘Someone like this calls themselves mayor,’ ‘comments a woman.’ 10, 4, NIlowC_ns Auf einer öffentlichen Veranstaltung hält ein Mann eine Rede. So jemand nennt sich Bürgermeister, erklärt die Mutter. ‘A man is giving a speech at a public event.’ ‘Someone like this calls themselves mayor,’ ‘explains the mother.’ 11, 3, MODlowC_nm Die Politiker verwerfen eine Maßnahme, die dem Klima zugutegekommen wäre. So etwas nennt man Klimaschutz, kommentiert eine Bür‐ gerin. ‘Politicians reject a measure that would have benefited the climate.’ ‘Something like this is called climate protection,’ ‘comments a citizen.’ 11, 4, NIlowC_nm Appendix C: Self-paced reading study 247 <?page no="248"?> Der Moderator berichtet über nachhaltige Maßnahmen, die die Erderwärmung reduzieren sollen. So etwas nennt man Klimaschutz, erklärt Katrin ‘The moderator reports on sustainable measures to reduce global warming.’ ‘Something like this is called climate protection,’ ‘explains Ka‐ trin.’ 12, 3, MODlowC_nm Die staatliche Behörde hat unabsichtlich sämtli‐ che Dokumente zu einem Ermittlungsverfahren gelöscht. So etwas nennt man Staatsanwaltschaft, merkt Alex an. ‘The state authority has unintentionally deleted all documents related to an investigation proce‐ dure.’ ‘Something like this is called public prosecutor’s office,’ ‘notes Alex.’ 12, 4, NIlowC_nm Die vom Gericht abhängige Justizbehörde hat die Aufgabe der Leitung von Ermittlungen. So etwas nennt man Staatsanwaltschaft, sagt Oliver. ‘The judicial authority under the jurisdiction of the court is responsible for conducting investi‐ gations.’ ‘Something like this is called public prosecutor’s office,’ ‘says Oliver.’ 13, 3, MODlowC_bez Die neue Regierung kann sich seit Monaten nicht auf einen Koalitionsvertrag einigen. So etwas bezeichnet man als Landesre‐ gierung, kommentiert Anna. ‘The new government has been unable to agree on a coalition contract for months.’ ‘Something like this is called state gov‐ ernment,’ ‘comments Anna.’ 13, 4, NIlowC_bez Der gemeinnützige Verein klärt auf der Straße über die Regierung des Bundeslandes auf. So etwas bezeichnet man als Landesre‐ gierung, erklärt Emma. ‘The non-profit organization educates people on the government of the federal state on the street.’ ‘Something like this is called state gov‐ ernment,’ ‘explains Emma.’ 14, 3, MODlowC_bez Der Experte erläutert den Kindern, dass Süßig‐ keiten gesund sind. So jemanden be‐ zeichnet man als Gesundheits‐ minister, kommentiert die Erzie‐ herin. 248 Appendix <?page no="249"?> ‘The expert tells the children that sweets are healthy.’ ‘Someone like this is called health min‐ ister,’ ‘comments the educator.’ 14, 4, NIlowC_bez Im Politikunterricht sprechen die Schüler über den Leiter des Gesundheitsministeriums. So jemanden be‐ zeichnet man als Gesundheits‐ minister, sagt eine Schülerin. ‘Students speak about the head of the ministry of health in their politics class.’ ‘Someone like this is called health min‐ ister,’ ‘says a stu‐ dent says.’ 15, 3, MODlowC_bez Der Chef überträgt alle wichtigen Entscheidun‐ gen an seine Mitarbeitenden. So jemanden be‐ zeichnet man als Geschäftsführer, sagt die Mi‐ tarbeiterin. ‘The boss leaves all important decisions to his co-workers.’ ‘Someone like that is called CEO,’ ‘says the coworker.’ 15, 4, NIlowC_bez - Im Rahmen eines Berufsinformationstages sprechen die Schülerinnen über die höchste Positionen innerhalb eines Betriebes. So jemanden be‐ zeichnet man als Geschäftsführer, sagt Sophie. ‘As part of a career information day, students talk about the highest position in a company.’ ‘Someone like this is called general manager,’ ‘says Sophie.’ 16, 3, MODlowC_bez Die Ratsmitglieder haben seit Wochen keine Entscheidung über das Müllproblem getroffen. So etwas bezeich‐ net man als Ge‐ meinderat, sagt David. ‘The council members have not made a decision concerning the trash problem for weeks.’ ‘Something like this is called a district council,’ ‘says David.’ 16, 4, NIlowC_bez In dem bayrischen Dorf steht die Wahl der polit‐ ischen Vertretung der Bürgerinnen und Bürger der Gemeinde an. So etwas bezeich‐ net man als Ge‐ meinderat, kommentiert Lina. ‘In the Bavarian village it is time for the election of political representatives of the people of the community.’ ‘Something like this is called a district council,’ ‘comments Lina.’ Appendix C: Self-paced reading study 249 <?page no="250"?> 17, 5, MODhighS_ns Der Balanceakt im weltberühmten Circus endet für die Artisten in einer Blamage vor dem gan‐ zen Publikum. So etwas nennt sich Artistik, kommentiert Hannah. ‘The balancing act in the world-famous circus ends in an embarrassment for the artists in front of the entire audience.’ ‘Something like this is called artistry,’ ‘comments Hannah.’ 17, 6, NIhighS_ns Die Zuschauer verfolgen gespannt die Darbie‐ tung der Darsteller im Varieté. So etwas nennt sich Artistik, erklärt Emil. ‘The spectators are eagerly watching the artists’ performance in the variety show.’ ‘Something like this is called artistry,’ ‘explains Emil.’ 18, 5, MODhighS_ns Das Wasser im Restaurant enthält nur einen winzigen Spritzer Apfelsaft. So etwas nennt man Schorle, kommentiert Emilio. ‘The water at the restaurant only contains a small amount of apple juice.’ ‘Something like this is called a spritzer,’ ‘comments Emilio.’ 18, 6, NIhighS_ns Die ausländischen Gäste gehen in Deutschland in ein Restaurant und bekommen einen Wein mit Mineralwasser serviert. So etwas nennt man Schorle, sagt Markus. ‘The foreign guests go to a restaurant in Ger‐ many and are served a wine with mineral water.’ ‘Something like this is called a spritzer,’ ‘says Mar‐ kus.’ 19, 5, MODhighS_nm Der Produzent teilt der Kundin mit, dass die Firma wegen Lieferengpässen keine Ware lie‐ fern kann. So jemand nennt sich Fabrikant, merkt Luca an. ‘The manufacturer informs the customer that the company is unable to deliver goods due to delivery shortages.’ ‘Someone like that calls themselves an industrialist,’ ‘notes Luca.’ 19, 6, NIhighS_nm Die Firma arbeitet eng mit einem Mann zusam‐ men, der eine bestimmte Ware herstellt. So jemand nennt sich Fabrikant, erklärt Paul. ‘The company closely works with a man pro‐ ducing a specific good.’ ‘Someone like that calls themselves an industrialist,’ ‘explains Paul.’ 20, 5, MODhighS_nm 250 Appendix <?page no="251"?> Die Zeitung berichtet, dass der ortsansässige Feinbäcker nur Tiefkühltorten verkauft. So jemand nennt sich Konditor, kommentiert Tim. ‘The newspaper reports that the local pastry cook only sells refrigerated cakes.’ ‘Someone like that calls themselves confectioner,’ ‘comments Tim.’ 20, 6, NIhighS_nm Ein Experte erklärt, wie aus einem Teig ein luftiger Obstboden entsteht. So jemand nennt sich Konditor, sagt Johanna. ‘An expert explains how a light fruit crust is made from a dough.’ ‘Someone like that calls themselves confectioner,’ ‘says Jo‐ hanna.’ 21, 5, MODhighS_bez Der Fachmann hat auf der Vorstellung den Jon‐ agold Apfel mit der Boskop Sorte verwechselt. So jemand bezeich‐ net sich als Pomo‐ loge, sagt Jana. ‘The expert confused the Jonagold apple with the Boskop type at the presentation.’ ‘Someone like that calls themselves a pomologist,’ ‘says Jana.’ 21, 6, NIhighS_bez Ein Obstbaukundler erklärt die Lehre der Arten und Sorten von Obst. So jemand bezeich‐ net sich als Pomo‐ loge, erklärt Man‐ uela. ‘A fruit planting expert explains the study of the kinds and types of fruit.’ ‘Someone like that calls themselves a pomologist,’ ‘explains Manuela.’ 22, 5, MODhighS_bez Die Freundinnen kommen enttäuscht aus dem Kinosaal raus. So etwas bezeichnet man als Film, kommentiert Johanna. ‘The friends disappointedly leave the movie theatre.’ ‘Something like that is called a film,’ ‘comments Johanna.’ 22, 6, NIhighS_bez Die Schulklasse redet bei einem Museumsbe‐ such über eine mit der Filmkamera aufgenom‐ mene Abfolge von bewegten Bildern. So etwas bezeichnet man als Film, erklärt Ella. ‘During a museum visit, the school class talks about a series of moving pictures recorded with a film camera.’ ‘Something like this is called a film,’ ‘explains Ella.’ 23, 5, MODhighS_bez Appendix C: Self-paced reading study 251 <?page no="252"?> Der Experte empfiehlt, keinen Sonnenschutz in den Sommermonaten zu tragen. So jemanden be‐ zeichnet man als Kosmetiker, sagt die Kriti‐ kerin. ‘The expert recommends not to put on sun protection during the months of summer.’ ‘Something like this is called a cosmeti‐ cian,’ ‘says the critic.’ 23, 6, NIhighS_bez Die Teenager schauen ein Video über einen Laboranten in der Kosmetikindustrie. So jemanden be‐ zeichnet man als Kosmetiker, erklärt der Kommenta‐ tor. ‘The teenagers watch a video about a laboratory assistant in the cosmetic industry.’ ‘Someone like that is called a cosmeti‐ cian,’ ‘explains the commenta‐ tor.’ 24, 5, MODhighS_bez Die Verlobten sitzen auf einem Vorbau, dessen Überdachung nur aus morschen Holzbrettern besteht. So etwas bezeichnet man als Veranda, sagt die Ver‐ lobte. ‘The fiancés are sitting on a porch with a roof that consists of rotten wooden boards.’ ‘Something like that is called a ve‐ randa,’ ‘says the fiancé.’ 24, 6, NIhighS_bez Die Immobilienmakler bieten ein Wohnhaus an, das einen verglasten Vorbau besitzt. So etwas bezeichnet man als Veranda, sagt der Ma‐ kler. ‘The estate agent offers a house for sale which has a porch made of glass.’ ‘Something like that is called a ve‐ randa,’ ‘says the es‐ tate agent.’ 25, 7, MODlowS_ns John hat schon zum dritten Mal in Folge den Geburtstag seines Kumpels Kai verpasst. So jemand nennt sich Freund, sagt Kai. ‘John missed his buddy’s birthday three times in a row.’ ‘Someone like that calls themselves a friend,’ ‘says Kai.’ 25, 8, NIlowS_ns Die Kindergärtnerin liest ein Buch über zwei Kinder vor, die sich gegenseitig unterstützen. So jemand nennt sich Freund, sagt die Erzie‐ herin. ‘The kindergarten teacher reads out loud a book about two children supporting each other.’ ‘Someone like that calls themselve a friend,’ ‘says the edu‐ cator.’ 252 Appendix <?page no="253"?> 26, 7, MODlowS_ns Das Publikum beobachtet, wie ein Akteur im Theater seinen Sprechtext vergessen hat. So jemand nennt sich Schauspieler, sagt eine Zu‐ schauerin. ‘The audience watches a player forgetting his lines in the theatre.’ ‘Someone like that calls themselves an actor,’ ‘says a mem‐ ber of the au‐ dience.’ 26, 8, NIlowS_ns Die Jugendlichen haben einen Workshop in einem Theater gebucht und beobachten einen Akteur auf der Bühne. So jemand nennt sich Schauspieler, erklärt der Dramaturg. ‘The teenagers booked a workshop at a theatre and are watching a player on stage.’ ‘Something like that calls them‐ selves an actor, ‘explains the dramaturge.’ 27, 7, MODlowS_nm Der Betrieb besteht lediglich aus einem kleinen Büro und zwei Mitarbeitern. So etwas nennt man Unternehmen, kommentiert Finn. ‘The business consists only of a small office and two workers.’ ‘Something like this is called a company,’ ‘comments Finn.’ 27, 8, NIlowS_nm Die Seniorengruppe besucht einen Betrieb, der aus mehreren Filialen besteht. So etwas nennt man Unternehmen, merkt Noah an. ‘The senior group visits a business consisting of multiple branches.’ ‘Something like that is called a com‐ pany,’ ‘notes Noah.’ 28, 7, MODlowS_nm Der Detektiv hat die Spuren am Tatort verse‐ hentlich unbrauchbar gemacht. So jemanden nennt man Ermittler, kommentiert die Lokalzei‐ tung. ‘The detective accidentally rendered the evi‐ dence at the crime scene useless.’ ‘Something like that is called an in‐ vestigator,’ ‘comments the local newspaper.’ 28, 8, NIlowS_nm Die Kinder sehen zwei Polizisten, die einen Unfall genauer untersuchen. So jemanden nennt man Ermittler, erklärt der Betreuer. ‘The children see two police officers who are closely investigating an accident.’ ‘Someomne like that is called an in‐ vestigator,’ ‘explains the carer.’ Appendix C: Self-paced reading study 253 <?page no="254"?> 29, 7, MODlowS_bez Das Navi führt das Ehepaar zu einer morschen Hütte am Waldesrand. So etwas bezeichnet man als Haus, kommentiert der Ehemann. ‘The GPS guides the couple to a rotten cabin at the edge of the woods.’ ‘Something like that is called a house,’ ‘comments the husband.’ 29, 8, NIlowS_bez - Die Hausbewohner begrüßen ihren Besuch aus Argentinien vor ihrem Gebäude. So etwas bezeichnet man als Haus, erklärt die Gastgeberin. ‘The home owners greet their visitors from Argentina in front of their building.’ ‘Something like that is called a house, ‘explains the hostess.’ 30, 7, MODlowS_bez Auf dem Weg in das nächstgelegene Dorf muss die Freundesgruppe über einen geflickten und holprigen Weg fahren. So etwas bezeichnet man als Straße, kommentiert Lars. ‘On their way to the village close by the group of friends had to drive across a patched and bumpy path.’ ‘Something like that is called a street,’ ‘comments Lara.’ 30, 8, NIlowS_bez Die Familie ist mit den Fahrrädern unterwegs und erklärt ihren Kindern den Straßenverkehr. So etwas bezeichnet man als Straße, sagt der ältere Bruder. ‘The family is on a biking trip and explains traffic control to their their children.’ ‘Something like that is called a street,’ ‘says the older brother.’ 31, 7, MODlowS_bez Die Freundinnen hören den Vortrag über einen Mann, der seine Kunstwerke plagiiert hat. So jemanden be‐ zeichnet man als Künstler, sagt Felix. ‘The friends are listening to a lecture about a man who plagiarized works of art.’ ‘Someone like that is called an artist,’ ‘says Felix.’ 31, 8, NIlowS_bez Auf einer Ausstellung präsentiert ein Maler stolz seine Ausstellungsstücke. So jemanden be‐ zeichnet man als Künstler, sagt Emma‐ nuel. ‘A painter is proudly presenting his pieces at an exhebition.’ ‘Someone like this is called an artist,’ ‘says Emma‐ nuel.’ 254 Appendix <?page no="255"?> 32, 7, MODlowS_bez Der Mann konnte die Sachlage nicht objektiv betrachten und hat trotz dessen ein Urteil ge‐ fällt. So jemanden be‐ zeichnet man als Richter, kommentiert eine Zeitung‐ sleserin. ‘The man could not look at the situation objec‐ tively but still passed a sentence.’ ‘Someone like that is called a judge,’ ‘comments a reader of the newspaper.’ 32, 8, NIlowS_bez Die Familie sieht einen Film über eine Person, die die Rechtsprechung ausübt. So jemanden be‐ zeichnet man als Richter, sagt der Va‐ ter. ‘The family watches a movie about a person administering justice.’ ‘Someone like that is called a judge,’ ‘says the fa‐ ther.’ Tab. 31: Overview of all SPR experimental items Appendix C5: Word length of the target words Target word Word length Schorle spritzer short Freund friend short Schauspieler actor middle Landesregierung state government long Pflanzenwachstum plant growth long Kosmetiker cosmetician middle Weinkenner wine connoisseur middle Kundenbetreuer custom service middle Bürgermeister mayor middle Unternehmen company middle Künstler artist short Haus house short Artistik artist short Eisbrecher ice breaker middle Appendix C: Self-paced reading study 255 <?page no="256"?> Literaturverfilmung literary adaption long Film film short Gesundheitsminister health minister long Pomologe pomologist short Geschäftsführer manager long Konditor conditor short Richter judge short Staatsanwaltschaft public prosecutor’s office long Veranda veranda short Klimaschutz climate protection middle Fabrikant manufacturer middle Rabattaktion discount action middle Honigbiene honeybee middle Ermittler investigator middle Pressekonferenz press conference long Beamtenstatus civil servant status long Gemeinderat city council middle Straße street short Tab. 32: Word length for the target words Appendix C6: Reading Time for the target word MainCondition emmean SE df lower.CL upper.CL MOD 706 52.4 71.4 602 811 NI 756 52.3 70.7 651 860 Tab. 33: Pairwise comparison for the main condition 256 Appendix <?page no="257"?> MainCondi‐ tion Wor‐ dLength emmean SE df lower.CL upper.CL MOD long 804 77.0 60.1 650 958 NI long 853 76.9 59.9 700 1007 MOD middle 718 55.3 73.1 607 828 NI middle 767 55.3 73.5 656 877 MOD short 598 72.8 62.8 452 743 NI short 647 72.5 61.9 502 792 Tab. 34: Pairwise comparison for the main condition by WordLength contrast estimate SE df t.ratio p.value MODC - MODS 79.3 84.8 62.3 0.935 0.7863 MODC - NIC 27.2 56.8 981.4 0.480 0.9636 MODC - NIS -47.3 85.5 64.4 -0.554 0.9452 MODS - NIC -52.0 84.3 61.0 -0.617 0.9262 MODS - NIS -126.6 57.2 980.5 -2.213 0.1205 NIC - NIS -74.6 85.0 63.1 -0.877 0.8166 Tab. 35: Pairwise comparison for WordType contrast estimate SE df t.ratio p.value MODhigh - MODlow 28.0 70.5 86.1 0.397 0.9787 MODhigh - NIhigh -83.3 57.7 986.0 -1.443 0.4726 MODhigh - NIlow 12.2 70.5 86.1 0.173 0.9981 MODlow - NIhigh -111.2 69.9 83.2 -1.591 0.3895 MODlow - NIlow -15.8 56.5 976.9 -0.279 0.9924 NIhigh - NIlow 95.5 69.9 83.6 1.365 0.5246 Tab. 36: Pairwise comparison for WordType and Frequency Word‐ Type Wor‐ dLength emmean SE df lower.CL upper.CL Appendix C: Self-paced reading study 257 <?page no="258"?> MODC long 843 80.8 70.6 681 1004 MODS long 763 109.8 51.7 543 984 NIC long 815 80.7 70.2 655 976 NIS long 890 110.3 52.8 669 1111 MODC middle 758 69.2 86.0 621 896 MODS middle 679 71.2 81.2 537 820 NIC middle 731 68.6 83.4 594 867 NIS middle 805 72.0 85.2 662 948 MODC short 637 107.2 53.3 422 852 MODS short 558 76.8 75.6 405 711 NIC short 610 106.8 52.6 396 824 NIS short 685 76.4 73.7 533 837 Tab. 37: Pairwise comparison for WordType and Interpretation Appendix C7: Reading Time for the spillover region contrast estimate SE df t.ratio p.value MOD - NI 45.8 30.8 977 1.489 0.1367 Tab. 38: Pairwise comparison for the main condition contrast estimate SE df t.ratio p.value MODC - MODS 41.32 58.3 63.7 0.709 0.8934 MODC - NIC 69.31 43.4 978.1 1.598 0.3802 MODC - NIS 63.52 58.5 64.4 1.086 0.6990 MODS - NIC 27.99 58.0 62.5 0.483 0.9627 MODS - NIS 22.20 43.7 977.4 0.508 0.9572 NIC - NIS -5.79 58.1 63.2 -0.100 0.9996 Tab. 39: Pairwise comparison for WordType 258 Appendix <?page no="259"?> contrast estimate SE df t.ratio p.value MODhigh---MODlow -19.3 58.3 63.6 -0.332 0.9873 MODhigh - NIhigh 10.3 44.0 981.6 0.233 0.9955 MODhigh - NIlow 60.8 58.4 64.0 1.040 0.7265 MODlow - NIhigh 29.6 58.1 62.8 0.509 0.9566 MODlow - NIlow 80.1 43.1 974.4 1.859 0.2465 NIhigh---NIlow 50.5 58.2 63.3 0.867 0.8216 Tab. 40: Pairwise comparison for Frequency and Interpretation Appendix C: Self-paced reading study 259 <?page no="260"?> Appendix D: Eye-Tracking study Appendix D1: Instructions for the experiment Instruktionen Folie 1 (Slide 1) Willkommen in unserem Sprachlabor! In unserem linguistischen Experiment untersuchen wir das Satzverstehen. Die Ergebnisse sind anonym und werden ausschließlich für wissenschaftliche Zwecke verwendet. Die Studie dauert ca. 30-45 Minuten (‘Slide 2: Welcome to our language lab! In our linguistic experiment, we examine the understanding of sentences. The results are anonymous and are used exclusively for scientific purposes. The study takes approximately 30-45 minutes.’) Folie 2 Gleich werden Sie einen Satz auf dem Bildschirm sehen. Bitte lesen Sie die Sätze zügig, aber in Ihrem eigenen, normalen Lesetempo, sodass Sie den Inhalt verstehen. Betätigen Sie anschließend die ENTER-TASTE. Im Anschluss wird Ihnen jeweils eine Verständnisfrage gestellt, die sich ausschließlich auf den zuvor gelesenen Satz bezieht. Bitte beantworten Sie diese Frage entweder mit “1” = JA oder “9” = NEIN, wie auf der Tastatur angegeben. Eine einmal gegebene Antwort können Sie nicht mehr revidieren. Machen Sie einfach mit dem nächsten Satz weiter. Bevor das Experiment startet, gibt es einen Probedurchgang mit vier Beispielsätzen. Sollten Sie noch Fragen haben, wenden Sie sich bitte jetzt an die Versuchsleiter*in. Drücken Sie eine beliebige Taste, um den Probedurchlauf zu beginnen. (‘Slide 2: You will first see a sentence on the screen. Please read the sentences quickly, but at your own normal reading rate, so that you understand the content. Then press ENTER. You will then be asked a comprehension question, which relates exclusively to the previously read sentence. Please answer this question with either “1” = YES or “9” = NO, as indicated on the keyboard. You cannot revise an answer once given. Just proceed with the next sentence. Before the experiment starts, there is a sample trial with four sentences. If you have any further questions, please contact the researcher now. Press any button to start the sample process.’) 260 Appendix <?page no="261"?> Appendix D2: Overview of the stimuli ID Target Nominal 1 Letter count Lexical Frequency 2 1 Neuwahl re-election 7 13 2 Sepsis sepsis 6 15 3 Kernfusion nuclear fusion 10 14 4 Supermond supermoon 9 14 5 Bürgergeld citizens’ funds 10 11 6 Ultraschall ultrasound 11 15 7 Radio radio 5 11 8 Stockrose hollyhock 9 20 9 Bahnradsport track cycling 13 18 10 Verfilzen felted material 9 22 11 Demenz dementia 6 12 12 Blutspende blood donation 10 14 13 Harfe harp 5 15 14 Vase vase 4 15 15 Polarlichter Northern lights 12 16 16 Bleichen bleach 8 17 17 Wandregal shelve 9 19 18 Platzregen cloudburst 10 16 19 Sitzsack beanbag 8 19 20 Klippe cliff 6 15 Tab. 41: Target stimuli of the eye-tracking study Note English translation given in italics. 2 Lexical frequencies calculated based on the German News Corpora 2022 (Deutsches Nachrichten-Korpus 2022) Wortschatz Uni Leipzig. Appendix D: Eye-Tracking study 261 <?page no="262"?> Appendix D2: Experimental items Condition Experimental Items Quotes Kim weiß, dass man dieses Verfahren „Neuwahl“ nennt, und belehrt Anna darüber. ‘Kim knows that this procedure is called “re-election”, and she teaches Anna about it.’ noQuotes Kim weiß, dass man dieses Verfahren Neuwahl nennt, und belehrt Anna darüber. ‘Kim knows that this procedure is called re-election, and she teaches Anna about it.’ Quotes Tom erfährt, dass man diese Krankheit „Sepsis“ nennt, und erzählt Anna davon. ‘Tom learns that this disease is called “sepsis”, and tells Anna about it.’ noQuotes Tom erfährt, dass man diese Krankheit Sepsis nennt, und erzählt Anna davon. ‘Tom learns that this disease is called sepsis, and tells Anna about it.’ Quotes Luisa erfährt, dass man diesen Prozess „Kernfusion“ nennt, und be‐ richtet Luis davon. ‘Luisa learns that this process is called “nuclear fusion” and reports Luis about it.’ noQuotes Luisa erfährt, dass man diesen Prozess Kernfusion nennt, und berichtet Luis davon. ‘Luisa learns that this process is called nuclear fusion and reports Luis about it.’ Quotes Martin erfährt, dass man dieses Phänomen „Supermond“ nennt, und informiert Lisa darüber. ‘Martin learns that this phenomenon is called “super moon”, and she in-forms Lisa about it.’ noQuotes Martin erfährt, dass man dieses Phänomen Supermond nennt, und informiert Lisa darüber. ‘Martin learns that this phenomenon is called super moon, and she in-forms Lisa about it.’ Quotes Angelina lernt, dass man diese Zahlung „Bürgergeld“ nennt, und berichtet Hanno davon. ‘Angelina learns that this payment is called “citizen’s benefit”, and reports this to Hanno.’ noQuotes Angelina lernt, dass man diese Zahlung Bürgergeld nennt, und berich‐ tet Hanno davon. ‘Angelina learns that this payment is called citizen’s benefit, and reports this to Hanno.’ Quotes Nele erfährt, dass man dieses Verfahren „Ultraschall“ nennt, und informiert Holger darüber. 262 Appendix <?page no="263"?> ‘Nele learns that this procedure is called “ultrasound”, and she informs Holger about it.’ noQuotes Nele erfährt, dass man dieses Verfahren Ultraschall nennt, und infor‐ miert Holger darüber. ‘Nele learns that this procedure is called ultrasound, and she informs Holger about it.’ Quotes Charlotta erfährt, dass man dieses Gerät „Radio“ nennt, und informiert Paul darüber. ‘Charlotta learns that this device is called “radio”, and she informs Paul about it.’ noQuotes Charlotta erfährt, dass man dieses Gerät Radio nennt, und informiert Paul darüber. ‘Charlotta learns that this device is called radio, and she informs Paul about it.’ Quotes Maria weiß, dass man diese Pflanze „Stockrose“ nennt, und erzählt Paul davon ‘Mary knows that this plant is called a “hollyhock”, and she tells Paul about it.’. noQuotes Maria weiß, dass man diese Pflanze Stockrose nennt, und erzählt Paul davon. ‘Mary knows that this plant is called a hollyhock, and she tells Paul about it.’ Quotes Christoph weiß, dass man diese Sportart „Bahnradsport“ nennt, und berichtet Tina darüber. ‘Christoph knows that this sport is called “rail bicycle”, and she tells Tina about it.’ noQuotes Christoph weiß, dass man diese Sportart Bahnradsport nennt, und berichtet Tina darüber. ‘Christoph knows that this sport is called rail bicycle, and she tells Tina about it.’ Quotes Annika weiß, dass man diesen Prozess „Verfilzen“ nennt, und berichtet Toni davon. ‘Annika knows that this process is called “felting”, and she tells Toni about it.’ noQuotes Annika weiß, dass man diesen Prozess Verfilzen nennt, und berichtet Toni davon. ‘Annika knows that this process is called felting, and she tells Toni about it.’ Quotes Emanuel lernt, dass man diese Krankheit „Demenz“ nennt, und berich‐ tet Helena davon. ‘Emanuel learns that this disease is called “dementia”, and she tells Helena about it.’ noQuotes Emanuel lernt, dass man diese Krankheit Demenz nennt, und berichtet Helena davon. Appendix D: Eye-Tracking study 263 <?page no="264"?> ‘Emanuel learns that this disease is called dementia, and she tells Helena about it.’ Quotes Leonie weiß, dass man diese Verfahren „Blutspende“ nennt, und be‐ lehrt Manuel darüber. ‘Leonie knows that this procedure is called “blood donation”, and she teaches Manuel about it.’ noQuotes Leonie weiß, dass man diese Verfahren Blutspende nennt, und belehrt Manuel darüber. ‘Leonie knows that this procedure is called blood donation, and she teaches Manuel about it.’ Quotes Marcel erfährt, dass man dieses Instrument „Harfe“ nennt, und berich‐ tet Carla davon. ‘Marcel learns that this instrument is called a “harp”, and he tells Carla about it.’ noQuotes Marcel erfährt, dass man dieses Instrument Harfe nennt, und berichtet Carla davon. ‘Marcel learns that this instrument is called a harp, and he tells Carla about it.’ Quotes Ella erfährt, dass man dieses Gefäß „Vase“ nennt, und erzählt Niko davon. ‘Ella learns that this piece is called a “vase”, and she tells Niko about it.’ noQuotes Ella erfährt, dass man dieses Gefäß Vase nennt und erzählt Niko davon. ‘Ella learns that this piece is called a vase, and she tells Niko about it.’ Quotes Maria weiß, dass man diese Erscheinung „Polarlichter“ nennt und erzählt Markus davon. ‘Mary knows that this phenomenon is called “northern lights” and tells Mark about it.’ noQuotes Maria weiß, dass man diese Erscheinung Polarlichter nennt, und erzählt Markus davon. ‘Mary knows that this phenomenon is called northern lights and tells Mark about it.’ Quotes Oliver erfährt, dass man dieses Verfahren „Bleichen“ nennt, und infor‐ miert Beate darüber. ‘Oliver learns that this procedure is called “bleaching”, and she informs Beate about it.’ noQuotes Oliver erfährt, dass man dieses Verfahren Bleichen nennt, und informiert Beate darüber. ‘Oliver learns that this procedure is called bleaching, and she informs Beate about it.’ Quotes Helena weiß, dass man diese Konstruktion „Wandregal“ nennt, und erzählt Kim davon. ‘Helena knows that this construction is called a “shelf ”, and she tells Kim about it.’ 264 Appendix <?page no="265"?> noQuotes Helena weiß, dass man diese Konstruktion Wandregal nennt, und erzählt Kim davon. ‘Helena knows that this construction is called a shelf, and she tells Kim about it.’ Quotes Anton weiß, dass man dieses Phänomen einen „Platzregen“ nennt, und erzählt Miriam davon. ‘Anton knows that this phenomenon is called a “sudden downpour”, and he tells Miriam about it.’ noQuotes Anton weiß, dass man dieses Phänomen einen Platzregen nennt, und erzählt Miriam davon. ‘Anton knows that this phenomenon is called a sudden downpour, and he tells Miriam about it.’ Quotes Felicia erfährt, dass man diese Sitzgelegenheit „Sitzsack“ nennt, und erzählt Johannes davon. ‘Felicia learns that this seat is called a “bean bag”, and she tells Johannes about it.’ noQuotes Felicia erfährt, dass man diese Sitzgelegenheit Sitzsack nennt, und erzählt Johannes davon. ‘Felicia learns that this seat is called a bean bag, and she tells Johannes about it.’ Quotes Andreas weiß, dass man diese Felsformation „Klippe“ nennt, und berich‐ tet Daniela davon. ‘Andreas knows that this rock formation is called a “cliff ”, and tells Daniela about it.’ noQuotes Andreas weiß, dass man diese Felsformation Klippe nennt, und berichtet Daniela davon. ‘Andreas knows that this rock formation is called a cliff, and tells Daniela about it.’ Tab. 42: Overview of Target sentences Appendix D3: Participant-specific details Session label Eye tracked gender handedness age Quotes_1A Left female right_handed 23 Quotes_2B Left female right_handed 21 Quotes_3C Right female right_handed 24 Quotes_5A Left female right_handed 30 Quotes_6B Right female right_handed 20 Quotes_7C Right female right_handed 38 Appendix D: Eye-Tracking study 265 <?page no="266"?> Quotes_8D Left male right_handed 23 Quotes_9A Right female right_handed 26 Quotes_10B Left female right_handed 21 Quotes_11C Right female right_handed 20 Quotes_12D Right male right_handed 28 Quotes_13A Right female right_handed 23 Quotes_14B Right female right_handed 23 Quotes_15C Right male right_handed 20 Quotes_16D Right female right_handed 37 Quotes_17A Left female right_handed 28 Quotes_18B Left male right_handed 28 Quotes_19C Left female right_handed 28 Quotes_20D Left female right_handed 22 Quotes_21A Left female right_handed 23 Quotes_22B Right female right_handed 30 Quotes_23C Right female right_handed 25 Quotes_24D Right female right_handed 20 Tab. 43: Participant-specific details of the eye-tracking study 266 Appendix <?page no="267"?> Appendix D4: Heat maps for target items Quotes and noQuotes Fig. 27: Heat map for noQuotes target sentence Appendix D: Eye-Tracking study 267 <?page no="268"?> Fig. 28: Heat map for Quotes target sentence 268 Appendix <?page no="269"?> List of Figures Fig. 1: Bar plot of valid instances for Q U O T E S and N O Q U O T E S by predicate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Fig. 2: Bar plot of the distribution of frequency classes for nennen 121 Fig. 3: Bar plot of the distribution of frequency classes for call . . . 122 Fig. 4: Bar plot of the distribution of frequency classes for bezeichnen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Fig. 5: Bar plot of the distribution of frequency classes for refer . . 123 Fig. 6: Bar plot for N AM E T Y P E by predicate . . . . . . . . . . . . . . . . . . . . 124 Fig. 7: Distribution of W O R D S TAT U S by predicate . . . . . . . . . . . . . . . 125 Fig. 8: Distribution of mean ratings for nennen . . . . . . . . . . . . . . . . 135 Fig. 9: Amount of ratings for the Interpretation by W O R D T Y P E for nennen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Fig. 10: Distribution of mean ratings for bezeichnen . . . . . . . . . . . . . 137 Fig. 11: Amount of ratings for the Interpretation by W O R D T Y P E for bezeichnen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Fig. 12: Predictive plot for W O R D T Y P E and I NT E R P R E TATI O N (target word) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Fig. 13: Predictive plot for F R E Q U E N C Y and I N T E R P R E TATI O N (target word) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Fig. 14: Mean RTs for the target word in the main conditions. Error bars represent confidence intervals +/ - 95% . . . . . . . . . . . . . 156 Fig. 15: Mean RTs for the target word in the main conditions by W O R D T Y P E . Error bars represent confidence intervals +/ - 95% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 Fig. 16: Mean RTs for the spillover regions for F R E Q U E N C Y . Error bars represent confidence intervals +/ - 95% . . . . . . . . . . . . . . . . . 158 Fig. 17: Mean RTs for the spillover region in the main conditions. Error bars represent confidence intervals +/ - 95% . . . . . . . . 159 Fig. 18: Mean RTs for the spillover region for W O R D T Y P E . Error bars represent confidence intervals +/ - 95% . . . . . . . . . . . . . . . . . 160 Fig. 19: Mean RTs for the spillover regions for F R E Q U E N C Y . Error bars represent confidence intervals +/ - 95% . . . . . . . . . . . . . . . . . 161 Fig. 20: Sample of vertically misaligned data | Translation: ‘Luisa learns that this process is called nuclear fusion, and she telly Luis about it.’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 <?page no="270"?> Fig. 21: Overlay images from EyeLink Data Viewer | Translation: ‘Leonie knows that this procedure is called blood transfusion, and she teaches Manuel about it.’ . . . . . . . . . . . 180 Fig. 22: Total amount of fixation durations by condition on the target item . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Fig. 23: Mean first fixation durations and standard error per condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 Fig. 24: Mean second fixation durations and standard error per condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 Fig. 25: Mean first run dwell times and standard error per condition 186 Fig. 26: Mean dwell times and standard error per condition . . . . . . 187 Fig. 27: Heat map for noQuotes target sentence . . . . . . . . . . . . . . . . . 267 Fig. 28: Heat map for Quotes target sentence . . . . . . . . . . . . . . . . . . . 268 270 List of Figures <?page no="271"?> List of Tables Tab. 1: Overview of quotation marks in the German News Corpora 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Tab. 2: Overview of quotation marks in the German News Corpora 2023 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Tab. 3: Overview of quotation marks in the English News Corpora 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Tab. 4: Overview of hits per condition . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Tab. 5: Sample sentences from the corpus data . . . . . . . . . . . . . . . . . . . 118 Tab. 6: Overview of the hits with Q U O TATI O N M A R K S and L E XI C AL F R E Q U E N C Y | Note: Percentage of valid NMCs in relation to all hits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Tab. 7: Chi-squared test for given probabilities (Q U O T E S & W O R D S TAT U S ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Tab. 8: Pairwise comparison for F R E Q U E N C Y and Quotes . . . . . . . . . . . . . 126 Tab. 9: Pairwise contrasts for F R E Q U E N C Y by Quotes . . . . . . . . . . . . . . . . 126 Tab. 10: Overview of experimental conditions . . . . . . . . . . . . . . . . . . . . . 147 Tab. 11: Schematized test design for a minimal pair . . . . . . . . . . . . . . . . 149 Tab. 12: Fixed effects statistics for I NT E R P R E TATI O N . . . . . . . . . . . . . . . . . . 155 Tab. 13: Fixed effects statistics for WordType and Interpretation . . . . . . 156 Tab. 14: Pairwise contrasts for W O R D T Y P E and I NT E R P R E TATI O N . . . . . . . 157 Tab. 15: Fixed effects statistics for I NT E R P R E TATI O N and F R E Q U E N C Y . . . . . 158 Tab. 16: Pairwise contrasts for I NT E R P R E TATI O N and F R E Q U E N C Y . . . . . . . . 159 Tab. 17: Fixed effects statistics for I NT E R P R E TATI O N . . . . . . . . . . . . . . . . . . 160 Tab. 18: Fixed effects for W O R D T Y P E and I N T E R P R E TATI O N . . . . . . . . . . . . 160 Tab. 19: Fixed effects for I N T E R P R E TATI O N and F R E Q U E N C Y . . . . . . . . . . . . . 161 Tab. 20: Descriptive statistics for first fixation duration . . . . . . . . . . . . . 182 Tab. 21: Summary of statistical analysis for FFD a . . . . . . . . . . . . . . . . . . . 183 Tab. 22: Descriptive statistics for second fixation duration . . . . . . . . . . 183 Tab. 23: Summary of statistical analysis for FFD a . . . . . . . . . . . . . . . . . . . 185 Tab. 24: Descriptive statistics for first run dwell time . . . . . . . . . . . . . . . 185 Tab. 25: Summary of statistical analysis for FRD a . . . . . . . . . . . . . . . . . . 186 Tab. 26: Descriptive statistics for dwell time . . . . . . . . . . . . . . . . . . . . . . 187 Tab. 27: Summary of statistical analysis for dwell time a . . . . . . . . . . . . . 187 Tab. 28: Instructions for rating the interpretation . . . . . . . . . . . . . . . . . 233 Tab. 29: Total counts for the predicates . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 <?page no="272"?> Tab. 30: SPR target words and Frequency | Note: SD has been rounded to the third decimal place . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 Tab. 31: Overview of all SPR experimental items . . . . . . . . . . . . . . . . . . . 244 Tab. 32: Word length for the target words . . . . . . . . . . . . . . . . . . . . . . . . 255 Tab. 33: Pairwise comparison for the main condition . . . . . . . . . . . . . . . 256 Tab. 34: Pairwise comparison for the main condition by WordLength . 257 Tab. 35: Pairwise comparison for WordType . . . . . . . . . . . . . . . . . . . . . . . 257 Tab. 36: Pairwise comparison for WordType and Frequency . . . . . . . . . . 257 Tab. 37: Pairwise comparison for WordType and Interpretation . . . . . . . 257 Tab. 38: Pairwise comparison for the main condition . . . . . . . . . . . . . . . 258 Tab. 39: Pairwise comparison for WordType . . . . . . . . . . . . . . . . . . . . . . . 258 Tab. 40: Pairwise comparison for Frequency and Interpretation . . . . . . . 259 Tab. 41: Target stimuli of the eye-tracking study . . . . . . . . . . . . . . . . . . . 261 Tab. 42: Overview of Target sentences . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Tab. 43: Participant-specific details of the eye-tracking study . . . . . . . . 265 272 List of Tables <?page no="273"?> TÜBINGER BEITRÄGE ZUR LINGUISTIK (TBL) Bisher sind erschienen: Frühere Bände finden Sie unter: https: / / www.narr.de/ linguistik-kat/ linguistikreihen-kat? ___store=narr_starter_de 562 Martina Zimmermann Distinktion durch Sprache? Eine kritisch soziolinguistische Ethnographie der studentischen Mobilität im marktwirtschaftlichen Hochschulsystem der mehrsprachigen Schweiz 2017, 304 Seiten €[D] 68,- ISBN 978-3-8233-8144-0 563 Philip Hausenblas Spannung und Textverstehen Die kognitionslinguistische Perspektive auf ein textsemantisches Phänomen 2018, 256 Seiten €[D] 88,- ISBN 978-3-8233-8155-6 564 Barbara Schäfer-Prieß, Roger Schöntag (Hrsg.) Seitenblicke auf die französische Sprachgeschichte Akten der Tagung Französische Sprachgeschichte an der Ludwig-Maximilians- Universität München (13. - 16. Oktober 2016) 2018, 558 Seiten €[D] 128,- ISBN 978-3-8233-8118-1 565 Vincent Balnat L’appellativisation du prénom Étude contrastive allemand-français 2018, 298 Seiten €[D] 78,- ISBN 978-3-8233-8185-3 566 Silvia Natale Informationsorganisation und makrostrukturelle Planung in Erzählungen Italienisch und Französisch im Vergleich unter Berücksichtigung bilingualer SprecherInnen 2018, 212 Seiten €[D] 68,- ISBN 978-3-8233-8209-6 567 Ilona Schulze Bilder - Schilder - Sprache Empirische Studien zur Text-Bild-Semiotik im öffentlichen Raum 2019, 227 Seiten €[D] 59,- ISBN 978-3-8233-8298-0 568 Julia Moira Radtke Sich einen Namen machen Onymische Formen im Szenegraffiti 2020, 407 Seiten €[D] 88,- ISBN 978-3-8233-8330-7 571 Melanie Kunkel Kundenbeschwerden im Web 2.0 Eine korpusbasierte Untersuchung zur Pragmatik von Beschwerden im Deutschen und Italienischen 2020, 304 Seiten €[D] 78,- ISBN 978-3-8233-8364-2 573 Mario Franco Barros Neue Medien und Text: Privatbrief und private E-Mail im Vergleich 2020, ca. 750 Seiten €[D] 119,90 ISBN 978-3-8233-8377-2 Reihenübersicht.indd 1 Reihenübersicht.indd 1 21.11.2025 10: 39: 32 21.11.2025 10: 39: 32 <?page no="274"?> 574 Sofiana Lindemann Special Indefinites in Sentence and Discourse 2020, 250 Seiten €[D] 68,- ISBN 978-3-8233-8381-9 575 Junjie Meng Aufgaben in Übersetzungslehrbüchern Eine qualitative und quantitative Untersuchung ausgewählter deutschchinesischer Übersetzungslehrbücher 2020, 206 Seiten €[D] 48,- ISBN 978-3-8233-8382-6 576 Anne-Laure Daux-Combaudon, Anne Larrory- Wunder (Hrsg.) Kurze Formen in der Sprache / Formes brèves de la langue Syntaktische, semantische und textuelle Aspekte / aspects syntaxiques, sémantiques et textuels 2020, 392 Seiten €[D] 78,- ISBN 978-3-8233-8386-4 577 Bettina Eiber Wikipedia und der Wandel der Enzyklopädiesprache Ein französisch-italienischer Vergleich 2020, 473 Seiten €[D] 88,- ISBN 978-3-8233-8407-6 578 Lidia Becker, Julia Kuhn. Christina Ossenkop, Anja Overbeck, Claudia Polzin-Haumann, Elton Prifti (Hrsg.) Fachbewusstsein der Romanistik Romanistisches Kolloquium XXXII 2020, 327 Seiten €[D] 98,- ISBN 978-3-8233-8418-2 579 Lidia Becker, Julia Kuhn. Christina Ossenkop, Anja Overbeck, Claudia Polzin-Haumann, Elton Prifti (Hrsg.) Romanistik und Wirtschaft Romanistisches Kolloquium XXXIII 2020, 274 Seiten €[D] 98,- ISBN 978-3-8233-8420-5 580 Claudia Schweitzer Die Musik der Sprache Französische Prosodie im Spiegel der musikalischen Entwicklungen vom 16. bis 21. Jahrhundert 2021, 201 Seiten €[D] 58,- ISBN 978-3-8233-8493-9 581 Roger Schöntag Das Verständnis von Vulgärlatein in der Frühen Neuzeit vor dem Hintergrund der questione della lingua Eine Untersuchung zur Begriffsgeschichte im Rahmen einer sozio- und varietätenlinguistischen Verortung: Die sprachtheoretische Debatte zur Antike von Leonardo Bruni und Flavio Biondo bis Celso Cittadini (1435-1601) 2022, 763 Seiten €[D] 138,- ISBN 978-3-8233-8540-0 582 Lea Schäfer Onymische Flexion Strukturen und Entwicklungen kontinentalwestgermanischer Dialekte 2021, 486 Seiten €[D] 98,- ISBN 978-3-8233-8521-9 583 Sarah Brommer, Kersten Sven Roth, Jürgen Spitzmüller (Hrsg.) Brückenschläge Linguistik an den Schnittstellen 2022, 324 Seiten €[D] 78,- ISBN 978-3-8233-8518-9 Reihenübersicht.indd 2 Reihenübersicht.indd 2 21.11.2025 10: 39: 32 21.11.2025 10: 39: 32 <?page no="275"?> 584 Mohcine Ait Ramdan Konzeptualisierung von Konkreta und Abstrakta Eine kulturorientierte, kognitionslinguistische Vergleichsstudie zwischen dem Deutschen, dem Arabischen und dem Französischen 2022, 245 Seiten €[D] 68,- ISBN 978-3-8233-8556-1 585 Steffen Hessler Autorschaftserkennung und Verstellungsstrategien Textanalysen und -vergleiche im Spektrum forensischer Linguistik, Informationssicherheit und Machine-Learning 2023, 426 Seiten €[D] 88,- ISBN 978-3-8233-8561-5 586 Jakob Wüest Une histoire des connecteurs logiques Causalité, argumentation, conséquence, finalité et concession 2023, 319 Seiten €[D] 58,- ISBN 978-3-8233-8615-5 587 Bin Zhang Metapherntheorie und Konstruktionsgrammatik Ein vierdimensionaler Ansatz zur Analyse von Metaphern und metaphorischen Konstruktionen 2023, 410 Seiten €[D] 78,- ISBN 978-3-8233-8614-8 588 Maria Schädler Die Proposition mit Kopula Urteilscharakter, logisch-semantische Valenz und formalisierte Sprache 2023, 872 Seiten €[D] 118,- ISBN 978-3-381-10781-0 589 Eva Lavric, Gerhard Pisek (eds.) Language and Football 2024, 384 Seiten €[D] 88,- ISBN 978-3-8233-8624-7 590 Laura Guse Bewegungskonstruktionen des Deutschen Korpusstudien zur Versprachlichung von Bewegungsereignissen aus konstruktionsgrammatischer Perspektive 2024, 358 Seiten €[D] 88,- ISBN 978-3-381-11031-5 591 Norbert Dittmar, Eva Neuland (Hrsg.) Wirkungsgeschichte der Soziolinguistik des Deutschen Dokumentarische und autobiographische Rekonstruktionen 2025, 350 Seiten €[D] 88,- ISBN 978-3-381-13271-3 592 Kirsten Schindler, Alexandra Lavinia Zepter (Hrsg.) Sprachliche Kreativität in Bildungs- und Fachsprache Textproduktion und Diskurspraktiken 2025, 315 Seiten €[D] 88,- ISBN 978-3-381-10941-8 593 Natascha Raue Reference to Names An empirically based study on metalinguistic uses of names of lexicalized concepts 2025, 273 Seiten €[D] 78,- ISBN 978-3-381-14641-3 Reihenübersicht.indd 3 Reihenübersicht.indd 3 21.11.2025 10: 39: 32 21.11.2025 10: 39: 32 <?page no="276"?> www.narr.de TBL Tübinger Beiträge zur Linguistik How do people talk about words? is work explores naming constructions such as is formation is called a “coral reef ” that are used to introduce lexicalized concepts like coral reef through metalinguistic reference. Drawing on the theoretical foundations in philosophy of language as well as recent work in semantics and pragmatics, experimental evidence from corpus analyses and psycholinguistic experiments—including self-paced reading and eye-tracking data—uncover the linguistic and cognitive processes behind these naming constructions. Special attention is given to the use and processing of quotational constructions during reading. Bridging (psycho)linguistics and philosophy of language, this interdisciplinary work o ers new insights into how people use quotational constructions to talk about language, and how readers interpret and process naming constructions with and without quotation marks. 593 Raue Reference to Names Reference to Names An empirically based study on metalinguistic uses of names of lexicalized concepts Natascha Raue ISBN 978-3-381-14641-3