Non-Canonical Questions 0192872281, 9780192872289

Citation preview

Non-Canonical Questions

Non-Canonical Questions ANDRE A S TROTZK E

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Andreas Trotzke 2023 The moral rights of the author have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2023941939 ISBN 9780192872289 DOI: 10.1093/oso/9780192872289.001.0001 Printed and bound in the UK by Clays Ltd, Elcograf S.p.A. Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

for Elio

Contents Acknowledgments Abbreviations

ix x

PA RT I I N T R O D U C T I O N T O N O N-C A N O N I C A L Q UE ST I O NS 1. Introduction 1.1 1.2 1.3 1.4

Canonical clause types and speech act syntax Non-canonical clause types and rethinking the role of syntax Canonicity and indirect speech acts Outline of the book

2. Key concepts in the study of questions 2.1 2.2 2.3 2.4

Questions in classical speech act theory and pragmatics Questions in semantic theory Questions in phonology Summary

3. A typology of non-canonical questions 3.1 3.2 3.3 3.4 3.5

Introducing bias, epistemicity, and the speaker’s emotion Non-canonical questions and speaker bias Non-canonical questions and epistemic states Non-canonical questions and the speaker’s emotion Summary

3 5 11 15 19

22 22 25 29 31

32 32 35 38 43 45

PA RT I I T H E S Y N TA X O F N O N-C A N O N I C A L Q U E ST I O NS 4. Non-canonical word order 4.1 4.2 4.3 4.4 4.5

Word order, prosody, and information structure wh-in-situ and echo questions Declarative questions Non-canonical fronting in questions Summary

5. Non-canonical wh-elements 5.1 5.2 5.3 5.4

why-like-what Degree readings of wh-elements in questions How come questions Summary

49 49 54 58 61 64

65 65 68 82 86

viii

CONTENTS

6. Non-canonicity via modal particles 6.1 6.2 6.3 6.4

Modal particles and speech act syntax The syntax of modal particles versus similar elements A cross-linguistic syntax of question particles Summary

88 88 92 96 115

PA RT I I I T H E PR A G M AT I C S O F N O N -C A N O N I C A L Q U E ST I O NS 7. Non-canonical questions as first moves in a dialogue 7.1 Request for information versus request for affirmation 7.2 Some syntactic consequences 7.3 Summary

119 121 133 136

8. Non-canonical questions as second moves in a dialogue

137

8.1 Answering and responding via questions 8.2 The syntax of responses and pragmatic inferencing 8.3 Summary

139 141 145

PA RT I V A N E W F R A M EW O R K F O R N O N -C A N O N I C A L Q U E ST I O NS 9. Emotional intensities in non-canonical questions 9.1 The expression of intensity in language 9.2 The expression of intensity in questions 9.3 Summary

10. Non-canonical questions, expressivity, and the model of communication 10.1 10.2 10.3 10.4 10.5 10.6

Two notions of expressivity Information structure and EB Information structure and EN Questions within a model of communication Questions and the syntax of indirect speech acts Conclusions

References Index

149 150 154 161

162 163 168 172 178 181 190

194 211

Acknowledgments This book would not have been possible without the invaluable contributions, support, and encouragement of numerous individuals and institutions. I am deeply grateful for the opportunity to express my heartfelt thanks to all who have played a role in bringing this book to life. First and foremost, I am indebted to the following colleagues who provided invaluable feedback, raised interesting questions, and collaborated with me on the topics discussed in this book: Josef Bayer, Andrea Beltrama, Lisa Brunetti, Agnès Celle, Anna Czypionka, Maria Teresa Espinal, Urtzi Etxeberria, Ricardo Etxepare, Werner Frey, Anastasia Giannakidou, Daniel Gutzmann, Manfred Krifka, Sergio Monforte, Laura Reimer, Sophie Repp, Jessica Rett, Lucia Tovena, Xavier Villalba, and Eva Wittenberg. Your insights, critiques, and encouragement have been instrumental in refining my thoughts and shaping the ideas presented in this book. I also thank the participants at the workshop Non-Canonical Questions in French, German, and Beyond (Konstanz, 2019), the members of the DFG Research Unit FOR 2111 Questions at the Interfaces, and the German Academic Exchange Service (DAAD) for supporting the project Surprise Questions from a Comparative Perspective (grant no. 57444809). Last but not least, I thank the SFB 1252 Prominence in Language (University of Cologne) for hosting me as a Visiting Professor and thus giving me the opportunity to write this book. Special thanks go to the team at Oxford University Press, particularly my editors, Julia Steer and Vicki Sunter, whose expert guidance has been indispensable in finalizing this enterprise. Additionally, I am grateful to the entire publishing team and to two anonymous reviewers who provided detailed feedback on this project.

Abbreviations ABS ACC addr Aff ALLOC AUX CG CL CP DAT DC EmpP ERG FinP Foc ForceP FUT INF MAS ModP Neg NOM PART PERF PL PRS PrtP Q QUD SAP SG SPrtP SUBJ TOP TP VP

absolutive accusative addressee affirmative allocutive auxiliary common ground clitic Complementizer Phrase dative Discourse Commitment Emphasis Phrase ergative Finiteness Phrase focus Force Phrase future infinitive masculine Modal Phrase negation nominative particle perfect plural present Particle Phrase question particle/marker Question under Discussion Speech Act Phrase singular Small Particle Phrase subject topic Tense Phrase Verb Phrase

PART I

IN TRODUCTION TO NON-C A NONIC A L QUESTIONS

1 Introduction This monograph is the first to present a comprehensive theory of noncanonical questions at the syntax–pragmatics interface. By ‘non-canonical’ questions, the book refers to question types that do not (only) request information from the addressee, but rather (additionally) tell us something about how the speaker conceptualizes the relation to the addressee, including different types of epistemic and/or emotional states. Prominent cases include so-called can’t-find-the-value questions, declarative questions, rhetorical questions, and surprise questions. All those types will be discussed in the book in more detail in the following chapters. At this point and at the very beginning of the book, I would like to highlight that the book deals with the empirical domain of non-canonical questions from a syntactic perspective. That does not mean that I will ignore the vast semantic literature and the more recent phonetic research on the topic. The phenomenon of ‘questions’ in general is a classical interface topic in linguistics, and one cannot look at this area of research in any comprehensive way without taking into account all relevant linguistic levels that constitute a question. However, the main parts of this book focus on the interplay between syntax and pragmatics—and in doing so, I argue, the book follows a more traditional path. To see this, let me highlight that the term ‘question’ refers to a speech act, and it is thus a pragmatic phenomenon in the first place. In the classical literature on speech acts, Searle (1969: 22) postulated canonical syntactic versions for performing speech acts such as assertions (Sam smokes habitually.), questions (Does Sam smoke habitually?), or orders (Sam, smoke habitually!). In those canonical versions, we see syntactic differences that help to distinguish between the performance of different speech acts. Searle (1969) was not concerned with any of the fine-grained phonetic differences, let alone with modeling the semantic differences between polar questions such as Does Sam smoke habitually? and assertions like Sam smokes habitually. There are rich traditions in both frameworks of formal semantics and (experimental) phonetics/phonology that are concerned with those differences (see Chapter 2 below).

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0001

4

INTRODUCTION

However, there is also a whole strand of linguistic research that deals with speech acts from an explicitly syntactic perspective, at least since Ross’s famous Performative Hypothesis (1970). Without going into too much detail at this point, it is fair to say that the standard view in this branch of the syntax– pragmatics literature is that most special or non-canonical interpretations of questions involve particular syntactic projections at or even above the level of illocutionary force. I will argue in this monograph that this approach is a mistake, and I will propose instead a new theory of the syntax of non-canonical questions where both their special pragmatics and their syntax can be derived solely from syntactic categories that are well known from more basic discourse interpretations, which, as soon as they are combined with question force, can yield non-canonical interpretations. This account simplifies the syntactic analysis of non-canonical questions quite dramatically, and it is also able to capture some unnoticed data in the discourse behavior of those question types. Moreover, it provides a new way of looking at functional categories that encode pragmatic meaning as a whole by rethinking the distinction between canonical and non-canonical speech acts. To be sure, one could argue against dedicated syntactic categories for encoding non-canonical question readings also on conceptual grounds only, given some recent conceptions and axioms of minimalist syntax (Newmeyer 2009; Trotzke and Zwart 2014; and Carlson et al. 2021 for a recent broader perspective on the issue). However, this is not what I am going to do in this book. In contrast to such criticism, my approach postulates that some basic concepts of pragmatics and discourse meaning should indeed be encoded in our syntactic representations, but it turns out that many of the non-canonical readings of questions that have been proposed in the literature can be reduced to more primitive discourse notions—and those notions are relevant for many other phenomena that have so far not been discussed in the context of non-canonical questions. Ultimately, some of the consequences of the approach presented in the book will change our understanding of the interface between syntax and speech acts as a whole. Given this broad outlook of the claims presented in this book, in the rest of this introductory chapter I will now focus on the notion of canonicity in grammar by discussing the distinction between canonical and non-canonical speech acts, and how this distinction relates to the common classification between major (declarative, interrogative, imperative) versus minor (e.g., exclamative, optative) clause types. I will introduce my driving hypothesis that specific illocutionary forces are neither ‘canonical’ nor ‘non-canonical’ per se, and that the illocutionary potential of different clause types (often referred to

1.1 CANONICAL CL AUSE T YPES AND SPEECH ACT SYNTAX

5

as ‘mood’) is reflected in their syntactic form only to a very limited extent. I will sketch the idea that the canonicity of a clause type—and all the discourse effects that follow from it—is determined by categories that are much broader than specific illocutionary forces, and on this basis I sketch an outline of the rest of the book.

1.1 Canonical clause types and speech act syntax In very general terms, the syntax–pragmatics interface is the interface between two essential aspects of language. Syntax refers to the set of rules that govern the formation of grammatical sentences in a language. Pragmatics, on the other hand, refers to the study of how people use language in context to achieve their communicative goals. The syntax–pragmatics interface is the point where these two aspects of language meet and interact. It is the point at which the grammatical structure of a sentence meets the context in which it is used to convey meaning. This particular interface is important, because it helps to explain how language users can interpret sentences with respect to the effect that utterances can have in a given social and discourse constellation. It is crucial to note that this pragmatic level of interpretation is deeply rooted in the grammar of each sentence—quite contrary to what could be expected when assuming that pragmatic meaning can be determined only by a concrete context or that pragmatic meaning is ‘meaning in context’, as many textbooks on the topic point out (e.g., Clark 2021; Cutting and Fordyce 2020; and the classical work by Levinson 1983). To see that, let me briefly illustrate that the situation is different in classical cases of ambiguity, where the context indeed is key for determining the intended interpretation. For example, consider the English sentence I saw her duck. This sentence is syntactically ambiguous, as it could mean ‘I saw the duck she owns’ or ‘I saw her quickly lower her head.’ In those cases, the context must determine which interpretation is correct. If the sentence is used in a conversation about farm animals, the first interpretation is more likely, whereas if the sentence is used in a conversation about a sports game, the second interpretation is more likely. The situation is different when we look at the interface between syntax and speech acts. A syntactic structure like Give me the salt! will most probably be interpreted as a command, no matter what the concrete context looks like. Considerations like that are a hot topic in the literature on the two

6

INTRODUCTION

theoretical frameworks of literalism and contextualism (see Recanati 2005 on the following). In a nutshell, the distinction between literalist and contextualist approaches pertains to how linguists and philosophers of language analyze sentence types and their meanings. These two approaches offer different perspectives on the relationship between the linguistic form of a sentence and the meaning or function it conveys in various contexts. The literalist approach posits that the meaning of a sentence is primarily determined by its syntactic and lexical components. According to this view, each sentence type (declarative, interrogative, imperative, and so on) has an inherent meaning or function that can be derived directly from its linguistic form. In other words, this approach suggests that the meaning of a sentence is relatively fixed and independent of the context in which it is used. For instance, a literalist might argue that declarative sentences inherently convey information or propositions, interrogative sentences inherently request information, and imperative sentences inherently express commands or requests. The contextualist approach, on the other hand, emphasizes that the meaning and function of a sentence depend significantly on the context in which it is used, very much like our example of structural ambiguity above. According to this view, linguistic forms alone are often insufficient to determine a sentence’s meaning or function, as the same sentence can take on different meanings or perform different speech acts depending on factors such as the speaker’s intentions, the listener’s expectations, and the broader conversational context. Contextualists would thus argue that the relationship between sentence types and their meanings or functions is more flexible and context-dependent than the literalist approach suggests. For example, a declarative sentence might be used to make a request (You could give me the salt!), express a command (You should give me the salt now!), or convey a question (You take some salt? Really?) in certain contexts. To be sure, there are several cases of ambiguity in the domain of speech act syntax that pose a challenge to both literalist and contextualist approaches. A prominent example are highly ‘conventionalized’ indirect speech acts like the interrogative Could you pass me the salt? (see Ruytenbeek 2021 for a comprehensive overview of the state of the art in that research domain). This sentence is most likely to be interpreted as a request, which means that its pragmatic function is to get the listener to perform a particular action. In this case, the action is to pass the salt—and not to answer the ability question by saying ‘yes’ or ‘no.’ Still, the syntax of the sentence is important in conveying the directive meaning, but it is the pragmatic function and the context of the speech act

1.1 CANONICAL CL AUSE T YPES AND SPEECH ACT SYNTAX

7

that ultimately determine its interpretation. I have already mentioned above that the imperative counterpart (Give me the salt!) might be a good example where the syntax is non-ambiguous. Another more extreme case can be found in explicit performatives like the sentence I promise to come to your party (see Recanati 1987 for seminal work and Garcı´a-Carpintero 2013 for more specific considerations). This sentence is a commissive speech act, which means that its pragmatic function is to commit the speaker to a future action. The syntactic structure will most likely be interpreted as commissive because the speaker uses the speech act verb to promise and is thus very explicit about the intended interpretation. In this book, we will have a look at a variety of empirical phenomena that illustrate the whole scale of being indirect and implicit in the context of question speech acts, on the one hand, and being explicit in the domain of those speech acts, on the other hand. In particular, we will see that the syntax of a sentence provides the basic structure for conveying the relevant discourse meanings, but the pragmatic function of the speech act is in many cases an independent level, which interfaces with other linguistic levels such as phonology and the lexicon. Importantly, the main thread of this book will be that the syntax should reflect pragmatic interpretations that can be derived only from syntactic changes, and not from changes in other domains of grammar. In a way, this goes against the whole tradition in theoretical syntax originating in the so-called Performative Hypothesis. The Performative Hypothesis is a concept in linguistics proposed by Ross (1970), and it aims to provide an account of the underlying structure of sentences that involve speech acts or performatives. According to the Performative Hypothesis, all sentences, regardless of whether they explicitly involve a performative verb (such as to promise above) or not, have an underlying performative structure. In this structure, the main clause of a sentence is embedded under a higher, unpronounced performative verb, such as say or state. This higher verb is responsible for the illocutionary force of the utterance, which is the speaker’s intent or the effect the utterance has on the addressee. For example, consider the following cases (Ross 1970: 222–3): (1)

a Prices slumped. b I say to you that prices slumped.

Ross (1970: 223) claims that sentences like (1a) “must be analyzed as being implicit performatives, and must be derived from deep structures containing an explicitly represented performative main verb,” see example (1b). In

8

INTRODUCTION

other words, the sentence (1a) involves a syntactic deep structure that contains two sentences (1′), and the two sentences are derived from each other via transformational rules, which we cannot discuss in detail here. (1′)

a [S Prices slumped.] b [S I say to you [S (that) prices slumped.]]

The Performative Hypothesis was an influential idea in early transformational–generative grammar, as it aimed to unify the structure of various types of sentences and account for their performative aspect. However, its exact implementation and influence have waned over time; other linguistic theories have been developed that offer alternative explanations for speech act phenomena and sentence structure—especially in the context of theories of clause typing (see Siemund 2018 for a recent comprehensive overview). In this theoretical domain, some syntactic characteristics of speech acts have repeatedly been discussed, with the goal of a broad classification of the structure and grammar that we use when making statements, asking questions, giving commands, making promises, expressing gratitude, and performing other speech acts. For English, the most common basics in this context probably concern the speech acts of statements, questions, and commands (see Collins 2020). Prototypically, statements are declarative sentences that convey information or make assertions. The typical syntax for a statement in English is Subject–Verb–Object (e.g., The mouse eats the cheese.). Questions are interrogative sentences that ask for information or clarification. The syntax for a polar question in English is Auxiliary Verb–Subject–Main Verb–Object (e.g., Do you like pizza?). Last but not least, commands are imperative sentences that give instructions or orders. The syntax for an English command is Verb–Object (e.g., Clean your room!). Needless to say, those aspects of the illocutionary potential of certain word orders depend heavily on the individual language. However, based on such a basic understanding about how certain syntactic forms are connected to the level of speech acts, one can classify syntactic structures in terms of clause types. In addition to the basic types of main clauses indicated above, there are also complex and embedded clauses, which can have different functions and structures depending on their syntactic and semantic context as well (e.g., providing additional information about a nominal referent in the main clause, that was was directed directed by by Steven Steven Spielberg Spielberg..). Those clauses will as in I saw a movie that not be considered in detail in this monograph, but they are also a lively area of research in the field of clause types (see, e.g., Danckaert 2012 for a

1.1 CANONICAL CL AUSE T YPES AND SPEECH ACT SYNTAX

9

detailed generative study on different types of embedded clauses and their classification). The notion of clause type (and its potential for shedding light on regularities at the interface between syntax and speech acts) is not confined to a single linguistic framework. However, in generative theories of syntax it is one of the oldest research topics (e.g., Baker 1970; Cheng 1991; Rizzi 1990; and many more). In generative linguistics, clause typing is typically analyzed in terms of the functional projections that are associated with different types of clauses. According to versions of generative syntax that emerged in the 1990s, the structure of a sentence is composed of a set of functional projections that are associated with different levels of syntactic and semantic structure. Essentially, this concept goes back to Chomsky’s seminal suggestion (1986: 3) “to extend conventional notation for the lexical categories to the nonlexical ones.” In more recent versions of that framework, clauses are analyzed as the projection of a functional head called Force0 , which represents the encoding of the illocutionary force or intended speech act of a particular sentence (see Rizzi 1997 for early work in this domain). Again, this concerns not only the basic types of main clauses, but also complex and embedded clauses, which are analyzed in terms of the functional projections associated with their syntactic and semantic structure as well—and some of them can contain Force(-like) components too (Heycock 2017; Julien 2015; Woods 2016; and many more). In sum, generative theories of clause typing provide a formal and structured framework for analyzing the syntactic and semantic properties of different types of clauses and understanding the ways in which syntax is used to perform different speech acts. We will see this in great detail in Part II of this book. Another pragmatic level that is also relevant and commonly mentioned in contexts where we distinguish between canonical versus non-canonical clauses is the domain of information structure (see Erteschik-Shir 2007; Leino 2013; Neeleman and Vermeulen 2012; Shlonsky 2010; and Si and Rizzi 2021 for overviews of different fields and research traditions). This strand of research makes clear that the level of information structure also plays an important role at the syntax–pragmatics interface, and it concerns the ways that information is organized within a sentence or discourse, including notions such as topic, focus, and background. The pragmatic function of information structure is to help the speaker and listener to manage the flow of information in a conversation, to highlight new or important information, but also to convey the speaker’s attitude or stance towards the information being communicated.

10

INTRODUCTION

For example, consider the sentence John bought a car. The pragmatic function of the information structure in this sentence is to communicate the new or focal information, which is that John bought a car. In this constellation, the referent of John is typically considered given or backgrounded information, and the new information is the object a car. Now consider the sentence It was John who bought the car. The pragmatic function of the information structure in this sentence is to communicate a contrastive focus, which means that the speaker is contrasting John with someone else who may have been expected to buy the car. The information structure in this sentence can additionally convey the speaker’s attitude towards the information, which is that the fact that John bought the car is somehow surprising or unexpected (see Destruel et al. 2019 for recent work). In summary, the pragmatic level of information structure—just like the level of illocutionary force—plays an important role at the syntax–pragmatics interface, because it helps to organize information in a way that is meaningful and efficient for the speaker and addressee. This information structure of a sentence or discourse can additionally help to convey the speaker’s attitude towards the information being communicated. Crucially, the common understanding of those pragmatic aspects of syntax is that the level of information structure and the level of speech acts are related, but nevertheless distinct aspects of language. While there may be some overlap between the two levels, they operate at different levels of linguistic analysis (i.e., concern different syntactic domains of functional projections), because they serve different functions. In particular, speech acts refer to the intended communicative function of a particular utterance, such as making a request or expressing an opinion. As we have already indicated, speech acts can be classified into different categories, such as assertives, directives, and commissives. Information structure, on the other hand, refers to the way that information is organized within a sentence or discourse, and it includes elements such as topic, focus, and background. The pragmatic function of information structure is to help the speaker and listener to manage the flow of information in a conversation (e.g., to highlight new/important or even surprising information). While there may be overlaps between the two levels, the difference clearly is that speech acts are focused on the intended function of the whole utterance, thus crucially reflecting the speaker’s intention(s). Information structure, on the other hand, is focused on the organization and highlighting of information within a particular sentence, and thus it very often concerns only parts of the utterance. However, there are also similarities between the two levels. At a very general level, both speech acts and information structure play

1.2 NON-CANONICAL CL AUSE T YPES

11

important roles in conveying meaning in context. They both rely on pragmatic knowledge and contextual information to be fully understood, and they both help to manage the flow of information in a conversation. In the final parts of the present monograph, I will argue that those two domains actually instantiate the same syntactic domain of organizing discourse as to how a sentence is directed at an addressee, and that we do not need to distinguish between the two domains when we look at the relevant phenomena from a syntax–pragmatics perspective. Let us take stock. The distinction between canonical and non-canonical clause types refers to the differences in clause structure and word order between the most typical, expected form of a clause and the more deviant or unexpected forms of the same clause type. In linguistics, canonical clauses are those that feature a word order that is typical of the respective clause (e.g., declarative, interrogative, and so on). Non-canonical clauses, on the other hand, are those that deviate from this standard and unmarked word order, often because of the way in which the clause is used in context or to convey a specific meaning. The reason for this deviation from the standard word order can be due to a variety of factors, such as the context in which the clause is used, the emphasis the speaker or writer wants to place on certain elements of the clause, or the desire to convey a specific meaning or attitude, as we will see in the next section when we look at prominent cases of non-canonical clause types in more detail.

1.2 Non-canonical clause types and rethinking the role of syntax Non-canonical clauses, according to the ideas and distinctions we have discussed in the previous section, can be a variety of things. For example, a case of a non-canonical clause in English might be a question that begins with a prepositional phrase, such as From the store, did you buy bread? The sentence might sound somewhat less natural or less common compared to the more conventional way of forming the question: Did you buy bread from the store? Specifically, the canonical version of the question places the prepositional phrase from the store after the direct object bread, which is more typical in English word order. Nonetheless, both sentences are grammatical, and the choice between them depends on the speaker’s intended emphasis or stylistic preferences. That is, the non-canonical version From the store, did you buy bread? might be used when the speaker wants to emphasize the location (the

12

INTRODUCTION

store) or contrast it with another location. Non-canonical clauses are thus a way for speakers to convey specific meanings, attitudes, or emotions and to emphasize certain elements of the clause. They can also provide important information about the context in which the clause is used and can contribute to the overall meaning and interpretation of the discourse. Examples like that make clear that one important aspect for the distinction between canonical and non-canonical clauses is, again, the concept of information structure in linguistics. As already mentioned in the previous section, non-canonical clauses can be used to change the information structure of a sentence in order to emphasize different elements or to create a specific tone or emphasis. For example, a non-canonical clause might be used to place the focus of the sentence on an element that would normally be in the background, or to present new information in a more unexpected or attention-grabbing way. However, the distinction between canonical and non-canonical clause types is also very often applied by distinguishing between major clause types and minor clause types. The major clause types are those that are used most frequently and are considered to be the standard forms of clause construction in a language. As already sketched above, those major clause types are declaratives, which make a statement, interrogatives, which ask a question, and imperatives, which issue a command (see Collins 2020; Sadock and Zwicky 1985; and, again, Siemund 2018). It is important to note that, while there are typical word orders and functions of each of the major clause types in English (see Section 1.1), the exact form and function of clauses can vary greatly depending on the language and context. For example, in some languages, the subject may be omitted in declarative clauses, and, in others, the verb may appear at the beginning of the clause. Additionally, the distinction between declarative, interrogative, and imperative clauses can sometimes be a matter of interpretation, as the same words can be used to convey different meanings in different contexts. More specifically, there are indeed many languages where even the threefold distinction between the major types of declarative, interrogative, and imperative clauses is not always clear. Some languages do not have a clear morphological or syntactic distinction between these three clause types, and instead rely on context, intonation, or other cues to convey the intended meaning. For example, in many African languages the distinction between declarative and interrogative clauses is not always clear, as the same basic word order is used for both types of clauses, and intonation is used to signal the difference.

1.2 NON-CANONICAL CL AUSE T YPES

13

In these languages, an interrogative clause may be formed simply by using a rising intonation at the end of the sentence, rather than using a specific word order or any morphological marker. Let us look at the Bantu language Swahili, where the same basic word order is used for both declarative and interrogative clauses, and intonation is used to signal the difference. Here are some relevant examples from Swahili (Krifka 1985: 56): (2)

a Sarah amefika. Sarah has.arrived ‘Sarah has arrived.’ b Sarah amefika? Sarah has.arrived ‘Has Sarah arrived?’ c Je, Sarah amefika? PART Sarah has.arrived ‘Has Sarah arrived?’

The sentence in (2a), when said with a falling intonation, would be translated as a declarative sentence: ‘Sarah has arrived.’ If the same sentence were said with a rising intonation (2b), it would be translated as an interrogative sentence: ‘Has Sarah arrived?’ In this example, the word order Subject–Verb is the same for both the declarative and interrogative sentences, and the only difference is the intonation. Note that Swahili has an optional interrogative particle, je (2c), which would be an additional cue for indicating that the sentence should be interpreted as a question, apart from intonation (and interrogative pronouns for constituent questions; see Nicolle 2000). Another prominent example for a language where the distinction between the major clause types declarative and interrogative is not due to syntactic differences is Mandarin Chinese. Consider the following cases from Paul and Pan (2017: 53): (3)

a Tā huì shuō bāfálìyàyuˇ. 3SG can speak Bavarian ‘He can speak Bavarian.’ b Tā huì shuō bāfálìyàyuˇ↑ ? 3SG can speak Bavarian ‘Can he speak Bavarian?’ c Tā huì shuō bāfálìyàyuˇ ma? 3SG can speak Bavarian PART ‘Can he speak Bavarian?’

14

INTRODUCTION

Again, the same word order (SVO) is used in declaratives (3a) and interrogatives (3b), the only difference being the rising intonation in (3b). Interestingly, as in Swahili, we also find particles in Mandarin Chinese (ma in (3c)), which can signal the question reading. These examples illustrate that word order is not always a distinguishing factor between the so-called major sentence types in many languages. However, it is essential to remember that other cues, such as context, intonation, and particle elements, may still differentiate between the two sentence types. Crucially, as we will see, in English we often use the same word order for declarative and interrogative sentences as well, and rely on other cues, such as a rising intonation at the end of the sentence, to signal that the sentence is a question (so-called declarative questions, see Chapter 4). The difference between English and Swahili, from a syntactic point of view, is that, in English, the declarative word order in questions can be considered a marked and noncanonical choice because there is an unmarked, canonical version where the word order is different from the declarative pattern. In Swahili, the word order always remains the same, and intonation as well as particle elements are the only cues used to distinguish between declarative and (polar) interrogative sentences. These cross-linguistic observations already suffice to illustrate that canonicity cannot be defined in terms of word order in every language, even in the domain of the so-called major clause types. Given these cross-linguistic remarks, we now turn to what has been referred to as minor clause types in the literature. Typically, the minor clause types are less frequent and are used for more specific purposes, such as to express strong emotions or wishes. That is, the most famous examples of minor clause types include exclamatives, which express strong emotions such as surprise, and optatives, which express wishes or desires. In general, minor clause types are more likely to have non-canonical word orders and structures than major clause types (see Villalba 2017a for a Romance perspective and Grosz 2012 for a general theory of their syntax–semantics). For example, in English, an exclamative clause can have a verb-final word order and special wh-elements and/or operators such as what a in the sentence What a beautiful day it is! These non-canonical morphosyntactic features convey the emphasis and emotion expressed in the clause (see d’Avis 2016 for a cross-linguistic discussion). Optative clauses are used to express wishes, hopes, or desires. In some languages, they may be marked by a specific verb form or by the use of the subjunctive mood, and they can also feature non-canonical word order such as English If only I could speak French! (see Biezma 2011: Chapter 5; Grosz 2012). Again, it is important to note that

1.3 CANONICIT Y AND INDIRECT SPEECH ACTS

15

not all languages have specific word orders for exclamative and optative clauses, and, in such cases, the intended meaning must be conveyed through other means, such as intonation, context, or other lexical choices. Crucially, as we will see in the context of non-canonical questions, minor clause types can have a non-canonical information structure compared to the major clause types, such as declarative, interrogative, and imperative clauses. This noncanonical structure is often used to emphasize certain elements in the sentence or to convey a specific meaning or emotion. The distinction between major and minor clause types is not universally recognized across all languages, and some linguistic traditions do not make this distinction (see Croft 1994 for a seminal approach within cognitive linguistics). However, many languages do have constructions that can be classified as exclamative or optative clauses, and postulating the distinction between major and minor clause types is also relevant for understanding the development of clause types within a language and for analyzing historical variation (e.g., Beijering et al. 2019; Evans 2007). In the next section of this introduction, I would like to add another dimension to our discussion of canonicity of clause types—in addition to information structure and clause typing. In particular, a domain where we can also potentially distinguish between canonical and non-canonical mapping between syntax and pragmatics is the well-known phenomenon of indirect speech acts.

1.3 Canonicity and indirect speech acts In non-technical terms, indirect speech acts are speech acts where the speaker intends to convey a message that is different from the literal meaning of the words that are used. In the context of this book, let us briefly consider some prominent examples of indirect speech acts where the syntactic form is the typical form of a question. For instance, a speaker may use a question as a way to perform a request. That is, instead of saying Please pass me the salt, the speaker might ask: Do you know where the salt is? The question is not really a request for information, but it is intended to communicate the speaker’s desire for the addressee to pass the salt. Similarly, a speaker might use a question as a way of inviting someone to do something. For example, instead of saying I’d like you to come to my party, the speaker might ask: Would you like to come to my party? Again, the question is not really a request for information, but it is intended to communicate the invitation. A last example would be when a

16

INTRODUCTION

speaker uses a question as a way to make a suggestion. For example, instead of saying You should try the lasagna, the speaker might ask: Have you tried the lasagna? The question is not really a request for information, but it is intended to suggest that the listener try the lasagna. Given examples like these, an indirect speech act—according to Searle’s seminal paper (1975)—can be defined as follows: In indirect speech acts the speaker communicates to the hearer more than he actually says by way of relying on their mutually shared background information […] together with the general powers of rationality and inference on the part of the hearer. […] Unless a hearer has some inferential strategy for finding out when primary illocutionary points differ from literal illocutionary points, he has no way of understanding indirect illocutionary acts. (Searle 1975: 60–3)

In his analysis, Searle (1975) emphasizes the importance of context and pragmatic inferences rooted in Gricean implicatures (he refers to “mutually shared background information” and the “power of rationality,” which is clearly connected to Grice’s cooperation principle (1975)). However, in the literature that is more focused on linguistic aspects of speech acts, an indirect speech act has been characterized as follows: Indirect speech act A speech act realized by sentence type x whose illocutionary force in context type1 does not correspond to the illocutionary force that is prototypically (normally, standardly …) assigned to an utterance of this sentence type in context type2 . A deviation from prototypical speech act assignment may be signaled by indicators of indirectness. (Meibauer 2019: 79)

Meibauer (2019), similar to Searle (1975), underscores the significance of context; however, unlike Searle (1975), he also draws attention to the notion of sentence type and the potential existence of “indicators of indirectness.” This latter aspect of the indirect speech act concept enables linguistic frameworks like Meibauer’s to encompass utterance types that are unambiguous and, therefore, not solely reliant on pragmatic inference like Searle’s famous case Can you pass the salt? Interestingly, and based on this difference from Searle’s conception (1975), Meibauer (2019) categorizes cases often referred to as non-canonical questions as indirect speech acts. Consider the following example, which illustrates a German rhetorical question (Meibauer 2019: 75–6):

1.3 CANONICIT Y AND INDIRECT SPEECH ACTS

(4)

17

a Wer mag Merkel? who likes Merkel ‘Who likes Merkel?’ b Wer mag schon Merkel? [ASSERTION / QUESTION] who likes PART Merkel ‘After all, who likes Merkel anyway? (Nobody!)’

While (4a) seeks information, the version in (4b) with the German particle schon is unmistakably a rhetorical question (and thus, according to Meibauer and much of the rhetorical question literature, equivalent to an assertion speech act ‘Nobody likes Merkel.’ ). The crucial observation here is that, according to this notion of indirect speech acts, an utterance can be an indirect speech act while still explicitly indicating (through its linguistic form: e.g., dedicated particles) its ‘indirect’ interpretation. Meibauer (2019: 76) articulates this quite clearly: “I would not like to argue that [those examples] are, because of the modal particle, direct speech acts; instead, I want to treat the modal particles as dominating indicators and still respect the illocutionary force potential of the respective sentence types.” For us, the relevant part of this statement is the concept of ‘dominating indicators’, suggesting that we could look at illocutionary indicators in terms of hierarchies, meaning that the aspect of word order and sentence type is just one of them. Crucially, this is indeed what some of the most recent psycholinguistic literature on indirect speech acts has done, referring to the ‘degree of illocutionary force salience’ (see Ruytenbeek 2021: ch. 4 for an overview). In particular, the emergence of the field of experimental pragmatics at the beginning of the twenty-first century has spurred a wave of research on the processing of conventionalized versus non-conventionalized indirect speech acts like requests performed by using an interrogative (e.g., Can you pass me the salt?/ Is it possible to pass me the salt?; see Marocchini and Domaneschi 2022; Noveck and Sperber 2004; Ruytenbeek et al. 2017). A major puzzle in this context that has not been solved to date and that has been noted by Searle (1975) himself is the following: The biggest single problem with the foregoing analysis is this: If, as I have been arguing, the mechanisms by which indirect speech acts are meant and understood are perfectly general […] and not tied to any particular syntactical form, then why is it that some syntactical forms work better than others. Why can I ask you to do something by saying Can you hand me that book on the top shelf ? but not, or not very easily, by saying Is it the case that you at present have the ability to hand me that book on the top shelf ? (Searle 1975: 75)

18

INTRODUCTION

Searle (1975) did not provide a detailed attempt at solving this problem. In the end, his interests were philosophical in nature, and therefore he did not investigate the syntactic differences between the different versions of indirect speech acts discussed in his paper in any detail. But the problem remains and touches on our topic of canonicity: It is not clear why some syntactic forms in the domain of indirect speech acts are more canonical than others. In this book, I will adopt the hypothesis that the connection between the syntactic form of such indirect speech acts and the respective speech act meaning is not arbitrary and can thus in principle be accounted for in a theory of speech act syntax. Particularly in Part IV of the book, I will put forward the view that examples such as Can you pass me the salt? are not ‘indirect’ speech acts in any reasonable sense of the term ‘indirect’. I postpone the detailed argumentation to those later parts of the book, but let me mention already at this point that experimental studies have repeatedly shown that the processing of indirect speech acts expressing a request interpretation (like Can you pass me the salt?) is not characterized by additional costs when compared to the processing of direct speech act comprehension (e.g., Ruytenbeek et al. 2017), and that the distinction is thus a theoretical (probably artificial) one, which does not fit the psychological reality. As we will see, one way to deal with this problematic status of the term ‘indirect’ speech acts is instead to refer to utterances like Can you pass me the salt? as ‘natural conventions’, as recently proposed by Simons and Zollman (2019) in their extensive discussion of the nature of linguistic conventions. To get a rough idea about the formal variation of those natural conventions, see Table 1.1, which provides an excerpt from their table illustrating natural ways of performing a request in English (cf. Simons and Zollman 2019: 13)—all of which are, however, mostly characterized as noncanonical requests because they are performed not by means of an imperative, but by means of interrogatives. As Searle (1975) has already noted, the main challenge is that even such a big set of possible forms is not arbitrary (e.g., alternative forms such as Searle’s Is it the case that you at present have the ability to hand me that book on the top shelf ? would be odd and most probably be interpreted as a polar question, and not as a request). Another challenge I see emerging from this type of literature on natural conventions is that forms such as Can you pass me the salt? seem to be the unmarked and neutral way of performing a request. The imperative form Give me the salt, (please)!, even if containing the marker please, requires a much more specific social context than the ability question—namely, one where less politeness is required. The idea presented in the final parts of this book will be that those observations are not due to an idiosyncratic quirk of the grammar of requests, but rather hold across speech acts.

1.4 OUTLINE OF THE BOOK

19

Table 1.1 Natural ways of performing a request in English, including a proposal of classification. means

form Asking about

Can you help

Would you be

Is it at all

Do you have a

ability

me with this?

able to help me

possible for you

moment to help

with this?

to help me with

me with this?

this? Asking about

Would you mind Do you mind

Are you willing

I’m hoping you

willingness

helping me with

helping me with

to help me with

wouldn’t mind

this?

this?

this?

helping me with this.

All in all, this book about non-canonical questions will take into account the common dimensions that are claimed to yield non-canonicity in the domain of questions: changes in their clause type (questions performed by declaratives), changes in their information structure (as in many emphatic and expressive question types), and changes in their speech act interpretation (as in indirect requests performed by interrogative syntax). The main thread will be to ask which of those dimensions and their respective empirical phenomena should be reflected in a syntax–pragmatics theory of non-canonical questions. I will argue that not all data usually discussed in the context of non-canonical questions are relevant for such a theory, because it is only a subset of noncanonical questions where syntax and its interface with pragmatics really play a role. With this in mind, let us now briefly outline the main structure of the book.

1.4 Outline of the book The book will start with a chapter on key concepts in the study of questions more generally. In particular, Chapter 2 will introduce the reader to the basic assumptions about question speech acts in both classical and more recent pragmatics, to some of the most prominent ways of how to deal with

20

INTRODUCTION

questions in formal semantics, and finally to some key observations in the field of phonology. Since the rest of the book will be about the syntax–pragmatics of questions, Chapter 2 will merely sketch basic assumptions that are discussed in those other subfields of linguistics. In Chapter 3, I will propose a typology of non-canonical questions based on three central notions that are very often used to analyze and classify noncanonical questions: the speaker’s bias, epistemic states (of both speaker and addressee), and the speaker’s emotion. The chapter will first introduce the three notions and then discuss relevant types of non-canonical questions, grouped into separate sections that are based on this threefold distinction. Chapter 4 on non-canonical word order in questions is the first chapter of the syntactic part of the book. I will first discuss key aspects of the interplay between word order, prosody, and information structure, and then turn to the phenomenon of wh-in-situ and echo questions, focusing on languages where this syntactic word order results in non-canonical readings. The next section in Chapter 4 will deal with declarative questions, another domain where the word order of questions in many languages differs from that of canonical question speech acts. The last empirical phenomenon we will look at in Chapter 4 will be non-canonical frontings in questions (i.e., non-canonical word orders involving the periphery of the clause). In Chapter 5, we will be investigating the syntax of non-canonical whelements in questions, and we will discuss various empirical phenomena such as so-called why-like-what questions, questions containing (extreme) degree readings of wh-elements, and some cross-linguistics observations about how come questions. Chapter 6 explores the domain of non-canonicity in questions that is expressed via modal particles. I will start with a short discussion of modal particles and speech act syntax, compare modal particles to similar syntactic elements (most notably, higher adverbs), and then elaborate on a crosslinguistic syntax of modal particles in questions, based on intriguing patterns in Basque and German. Chapter 7 is the first chapter of Part III of the book, which looks in more detail into the pragmatic properties of some types of non-canonical questions and asks whether some pragmatic aspects of those questions that can be observed in the context of dialogue sequences interface with the syntactic representations as well. In Chapter 7, I first look at the features of non-canonical questions as first moves in a dialogue, and I discuss the fundamental distinction between a request for information and a request for affirmation in this domain. Some expressive non-canonical questions share many aspects with

1.4 OUTLINE OF THE BOOK

21

exclamation speech acts, and we will explore to what extent they should also be analyzed along the lines of exclamations. Given this discussion, the last section of Chapter 7 summarizes some consequences for the syntactic analysis of some of the expressive types of non-canonical questions. Chapter 8 of the book then turns to non-canonical questions as second moves in a dialogue and, again, introduces an important distinction for our discussion—in this case the distinction between answers and responses. I highlight how some of the non-canonical question types discussed in previous chapters can indeed be used as responses to information-seeking questions, and I sketch a syntactic account of those findings in terms of recent approaches to the syntax of answers, also factoring in the important role of pragmatic inferencing. In Chapter 9, I start the final part of the book, where I provide a more general conceptual framework for the data and syntactic analyses proposed in the previous chapters. In particular, in Chapter 9 I first turn to a general characterization of the role of emotional intensities in non-canonical questions, and how they connect to key concepts in degree semantics. I will pay particular attention to proposition-level intensification strategies, because I hypothesize that those are the most relevant scope environments that are at play in the domain of non-canonical questions. Chapter 10 will then end the book with a new proposal for how to account for the phenomenon of non-canonical questions at the level of fundamental functions of human language. I will first elaborate on different notions of expressivity of language, and I will propose a distinction between two fundamental types of expressivity, allowing us to think of discourse phenomena like information structure, on the one hand, and emotional aspects of meaning, on the other hand, in a new and unified way. Given this new account of the expressive function of language, I will then turn to the addressee-oriented function of language, where language is viewed within the context of dialogues. I will discuss the hypothesis that both speech act meaning, but also social phenomena such as politeness, belong to that function, and I will explore to what extent this view allows us to account for non-canonical as well as indirect speech acts more generally.

2 Key concepts in the study of questions This chapter provides some theoretical background on the study of questions in linguistics. Questions have always been a prominent topic in the language sciences, but especially so in the subfields of pragmatics, semantics, and phonology. Section 2.1 will deal with the notion of questions in classical speech act theory and pragmatics and highlight the fact that questions form a distinct type of directive speech acts, featuring a dedicated set of felicity conditions. Section 2.2 will introduce basic semantic notions that are discussed in the literature to model the meaning of questions; in this context, I will introduce the standard proposition set approach, which analyzes questions in terms of the set of propositions that are answers to the question. Section 2.3 sketches the interface between question meaning and phonology by referring to some basic patterns, and by pointing out to what extent those patterns are language-specific or might hold cross-linguistically.

2.1 Questions in classical speech act theory and pragmatics Searle (1975), in his classical approach to speech acts, categorizes speech acts into five main types: representatives, directives, commissives, expressives, and declarations. Questions fall under the category of directives in Searle’s classification. The main characteristics of the speech act of questions, as discussed in classical speech act theory, are as follows (see Searle 1969, 1975, 1976; Searle and Vanderveken 1985). First, questions, as a type of directive, primarily have the illocutionary force of requesting information from the addressee. That is, by asking a question, the speaker aims to elicit a response from the addressee, usually in the form of an answer or clarification. Second, like other speech acts, questions are characterized by the speaker’s intention to achieve a specific goal, which in this case is to acquire information. The success of a question as a speech act depends on the speaker’s ability to convey this intention effectively and the addressee’s ability to recognize it. As for the so-called sincerity condition, for a question to be successful as a speech act the speaker must genuinely desire the information they are requesting. This means that the speaker should Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0002

2.1 QUESTIONS IN CL ASSICAL SPEECH ACT THEORY AND PRAGMATICS

23

not already know the answer or be asking the question for rhetorical purposes. Finally, we can look at the felicity conditions, which are the contextual requirements that must be met for a speech act to be successful. For questions, these conditions might include the addressee’s ability and willingness to provide the requested information, the appropriateness of the question in the given context, and the speaker’s adherence to social norms and politeness. There are also potential perlocutionary effects of a question, referring to the response or outcome it generates from the addressee. Those effects can include providing the requested information, seeking clarification, or engaging in further conversation. These are the main characteristics of question speech acts as highlighted in classical speech act theory. More recently, Farkas (2022) has revised the relevant aspects and rephrased these features in more modern terms, based on a notion of a conversational context. According to Farkas (2022: 297–8), canonical question speech acts come with the following contextual assumptions: (i) the speaker’s epistemic state remains neutral regarding the possible resolutions of the issue they bring up; (ii) the speaker assumes that the addressee possesses the information needed to resolve the issue they raise; (iii) the speaker assumes that the addressee will supply this information in the conversation’s immediate future as a consequence of the speaker’s speech act; (iv) it is assumed that the primary objective the speaker seeks when raising an issue is to have it resolved in the conversation’s immediate future. Farkas’s characterization includes some crucial implications, which she discusses in her paper (again, see Farkas 2022: 297–8 on the following): First, assumption (iii) about addressee compliance implies assumption (ii) about addressee competence, but not vice versa, because an addressee cannot be expected to provide information they lack, but a knowledgeable addressee might still choose not to share the requested information. Furthermore, it follows from the way the assumption about addressee compliance is formulated above that the speaker wants to obtain the information they lack. If one performs an act intended to result in a specific state of affairs, it is reasonable to assume that one desires that state of affairs to come about. Second, given addressee compliance, a cooperative and competent addressee should respond by offering the true answer to the question posed. Lastly, it is worth noting that, according to Farkas’s (2022) characterization of canonical questions, the participant presumed to supply the requested information is the addressee, which might hold for canonical questions, but may not hold in the strict sense for other question speech acts, as we will see in this book. Let me briefly indicate how the status of the central assumptions listed above can change.

24

KEY CONCEPTS IN THE STUDY OF QUESTIONS

Naturally, as Farkas (2022: 298) points out, not every question act falls into the canonical category. She refers to Searle (1969) and provides the example of a type of question commonly referred to as a ‘quiz’ question, as illustrated in (1): (1)

Context: During a linguistics class, Andreas asks a student: What does the category CP refer to?

Given the context of the utterance, all four assumptions above are not really applicable. In this situation, the speaker is presumed to be well informed, and his primary objective in posing the question most probably is to be determining whether his addressee possesses the same knowledge. Farkas (2022) highlights that, in the pragmatics postulated in Searle (1969), these kinds of questions possess distinct felicity conditions compared to information-seeking questions. She argues along these lines too because, according to her, ‘quiz’ questions are not considered canonical, since the specific contextual assumptions they are based on deviate from the ones above and supersede them. In other words, the interpretation of a question as either information-seeking or as a ‘quiz’ question depends on the assumptions made by the discourse participants regarding the context they find themselves in and the motivations behind the speaker’s choice to utter the interrogative configuration. In a related paper, Farkas (2020) has formulated these basic pragmatic characteristics of question speech acts in more technical terms, based on a dynamic model of conversation and discourse effects (see Farkas and Bruce 2010). In particular, Farkas (2020: 13) points out that canonical questions introduce an inquisitive issue P into the set of issues that await resolution in a given discourse (and not an information issue, as in declaratives). Questions furthermore add a minimal speaker commitment, and guide the conversation towards a group of states, with one for each alternative in P. In each of these alternatives p ∈ P, there is a projected state where the commitment anchor in ps commits to p. Farkas (2020: 13) points out that this projected commitment is considered a strong one, and the commitment anchor is the addressee. Given that the primary aim of the conversation is to gain information related to P, some of the four pragmatic assumptions associated with question speech acts already introduced above can be rephrased as follows (see Farkas 2020: 13 on the following): (i) the speaker assumes that the issue they introduce into the set of active issues in the discourse is not yet resolved in the input context, meaning that there is not any alternative p ∈ P where p ∈ cgi (i.e., is already part of the c[ommon] g[round]); (ii) the speaker portrays themselves as being in an epistemic state that does not allow any

2.2 QUESTIONS IN SEMANTIC THEORY

25

commitment to one of the alternatives in P; (iii) the speaker presents themselves as believing that the addressee’s epistemic state enables their commitment to the true alternative in P—that is, the alternative containing wa ; (iv) the speaker presents themselves as expecting that the addressee will settle the issue by publicly committing to the true alternative in P. The assumption in (i), according to Farkas (2020), stems from the default belief that the conversation’s driving force behind the speaker’s move is to increase information relative to P. This assumption is incompatible with considering P as already resolved in the input context. The assumption in (ii) arises from the fact that, if the speaker believed they knew the true alternative in P, asserting a declarative sentence that publicly commits them to that alternative would be a more efficient way to increase information. Assumption (iii) is derived from the addressee being the default anchor for projected commitments and the fact that the speaker’s conversational move guides the conversation towards a state where the addressee resolves the issue. Consequently, according to Farkas (2020), it is logical for the speaker to assume that such a state can be achieved. Finally, according to Farkas’s approach, the assumption in (iv) follows from the speaker directing the conversation towards states where the addressee supplies the necessary information. Without this assumption, as Farkas points out, the speaker’s speech act would be pointless in terms of information increase. In this dynamic conversational model, one could say that the addressee is put under pressure in a canonical question. The speaker anticipates that the addressee will commit to one of the alternatives related to the issue the speaker has just introduced, initiating the process of reaching a mutually agreed resolution. In this process, the addressee must offer new information. By asking a canonical question, the speaker assumes the addressee possesses greater knowledge regarding P than the Speaker does. In sum, by adopting some of the concepts in classical work on speech acts by Searle (1969 et seq.), more recent accounts in pragmatics have tried to refine the notions that are needed to model the conditions or ‘pragmatic assumptions’ that must hold when we speak of canonical (i.e., information-seeking) questions. Let us now turn to some of the key proposals in formal semantics.

2.2 Questions in semantic theory There has been a lot of recent research on the semantics of both canonical and non-canonical questions (e.g., Benz et al. forthcoming; Dayal 2016; Giannakidou and Mari 2022; Goodhue 2018; Krifka 2011; and many others), and

26

KEY CONCEPTS IN THE STUDY OF QUESTIONS

there are various—and often diverging—approaches to modeling the meaning of questions. Needless to say, the following paragraphs can be only a brief sketch of the basic concepts discussed in the numerous works on the semantics of questions. As a first fundamental background to analyzing questions in semantics, I would like to mention Montague Semantics. In Montague Semantics, the meaning of sentences, including questions, is represented using logical formulas, and the compositional nature of language is captured through a system of rules that combine the meanings of individual words and phrases to derive the meaning of larger expressions (see Montague 1970; Partee 1984). Modeling questions in Montague Semantics involves representing them as logical expressions, just like declarative sentences. However, the main difference lies in the way questions are interpreted. While declarative sentences are usually modeled as truth-functional expressions that can be assigned a truth value (i.e., true or false) based on a given model, questions are modeled as expressions that denote a set of possible answers or a function that maps possible answers to truth values (see below for more on this concept of a set of answers). For wh-questions, such as Who is happy?, Montague Semantics often employs generalized quantifiers and lambda abstraction to model the meaning of the question. For instance, the question could be represented as a formula like Q(λx.Happy(x)), where Q is a question operator, λx is a lambda abstraction introducing a variable x representing an individual, and Happy(x) is a predicate denoting the property of being happy. The meaning of the question can then be interpreted as a set of individuals that satisfies the predicate Happy(x), or as a function that maps these individuals to truth values. In summary, Montague Semantics models questions as logical expressions that involve question operators, variables, predicates, and lambda abstraction. The interpretation of questions is based on sets of possible answers, possible worlds, or functions that map answers to truth values, reflecting the information-seeking nature of questions (see also Section 2.1 above) and allowing for a systematic analysis of their meaning and interaction with other linguistic expressions. Let us now briefly look at modeling questions in so-called Dynamic Semantics. Dynamic Semantics is an approach to natural language meaning that focuses on how the meaning of linguistic expressions affects the context in which they are used, and how this context evolves during a conversation. Unlike approaches that treat meaning as static and context-independent,

2.2 QUESTIONS IN SEMANTIC THEORY

27

Dynamic Semantics recognizes that meaning is often context-dependent and subject to change as new information is introduced or updated. Modeling questions in Dynamic Semantics involves analyzing how questions contribute to the update of the discourse context and the common ground between the speaker and the addressee (see Groenendijk and Stokhof 1991 on fundamental background on dynamic predicate logic). In this respect, this approach overlaps with modeling the pragmatics of questions, because the formal analysis of the detailed pragmatic assumptions relevant for the question speech act is couched in terms of Dynamic Semantics as well (see Section 2.1 above). The key concepts in Dynamic Semantics for modeling questions include context change potentials, information states, and context update functions. Let us first turn to context change potentials (CCPs); see Heim (1982) on the basic concept. In Dynamic Semantics, linguistic expressions, including questions, are associated with CCPs, which are functions that describe the potential effect of the expression on the discourse context. The CCP of a question represents the information-seeking intent of the speaker and the possible ways in which the question might lead to updates in the common ground. The second component needed for the modeling of questions comprises information states. Information states represent the knowledge or beliefs shared by the speaker and the listener during a conversation. In the context of questions, information states help capture the current state of knowledge about the subject matter and the possible answers that are under consideration. Finally, we need a context update function, which is a formal mechanism that captures how the discourse context is updated in response to a linguistic expression, such as a question or an answer. For questions, the context update function specifies how the information state is updated to accommodate the information-seeking intent of the question and the possible answers it raises. Modeling questions in Dynamic Semantics typically involves specifying the CCP of the question, the initial information state, and the context update function that maps the information state to a new state after the question has been asked. The updated information state reflects the speaker’s intent to acquire new information, the addressee’s understanding of the question, and the possible answers that are under consideration. In summary, Dynamic Semantics models questions by analyzing their context change potentials, information states, and context update functions. This approach provides a dynamic and context-sensitive perspective on question meaning, capturing the information-seeking nature of questions and their role in updating the discourse context and the common ground between interlocutors.

28

KEY CONCEPTS IN THE STUDY OF QUESTIONS

However, the most influential approach to modeling questions is the proposition set approach. This account, also known as the partition theory or the possible answers approach, is a framework in semantics that models the meaning of questions by representing them as a set of possible propositions or answers (see Hamblin 1976 for basic concepts, and Karttunen 1977 for a prominent version). The approach aims to capture the information-seeking nature of questions by identifying the alternative answers that a question raises. In the proposition set approach, a question is considered to be well formed if it can be understood as partitioning the logical space into mutually exclusive and exhaustive possibilities. Each possibility corresponds to a complete answer that resolves the question, and the meaning of the question is represented by the set of all these possible answers. For example, consider the simple yes–no question: Is it raining? In the proposition set approach, this question can be modeled as a set of two possible propositions: {It is raining, It is not raining}. The question effectively asks the addressee to determine which of these two propositions is true, and the addressee’s answer will provide the requested information. Of course, the proposition set approach can also be applied to more complex questions, such as those involving wh-words (e.g., who, what, where, when, why, how). For example, consider the question: Where is Elio? This question can be modeled as a set of propositions that correspond to all the possible locations where Elio might be: {Elio is at home, Elio is at kindergarten, Elio is at the park, …}. The exact set of possibilities depends on the context and the addressee’s knowledge of the relevant locations. The main advantage of the proposition set approach is that it provides a clear and systematic way of representing the meaning of questions in terms of their possible answers. This allows for a more precise analysis of the relationship between questions and answers, as well as the interaction between questions and other linguistic expressions, such as quantifiers and modals. However, the approach also has some limitations, such as the potential for generating an infinite or unmanageably large set of propositions for certain types of questions, and the difficulty of accounting for some pragmatic aspects of question meaning, such as presuppositions and implicatures (see Krifka 2001 for further shortcomings). We will turn to an application of this approach to some types of non-canonical questions in Section 3.3 below, because in those types the size of the set of propositions plays a crucial role for non-canonical interpretations. All in all, it is worth noting that the approaches sketched above are highly formal and technical, and this brief section can therefore provide only a rough

2.3 QUESTIONS IN PHONOLOGY

29

overview of the basic ideas. However, and needless to say, the use of formal formulas is crucial for capturing the precise meaning of questions and for making rigorous and systematic arguments about the interaction between the linguistic form of the question and the context in which it is used. Before we dive into the domain of the syntax of questions, let us now finally turn to some aspects of questions that have been discussed in the field of phonology.

2.3 Questions in phonology Questions can be analyzed in phonology in terms of their characteristic prosodic and intonational features that distinguish them from other types of sentences, such as declaratives or imperatives. Key concepts and insights in the phonological analysis of questions include intonation contours, pitch accents, and boundary tones (e.g., Dehé and Braun 2020; Hedberg and Sosa 2002; Hedberg et al. 2017; Petrone and Niebuhr 2014; and many more). In broad terms, intonation contours refer to the overall melodic pattern of a sentence, which is shaped by variations in pitch, stress, and duration. Questions often exhibit distinct intonation contours compared to declarative sentences. For instance, in many languages, polar aka yes–no questions are characterized by a rising pitch pattern towards the end of the sentence, while declaratives usually have a falling pitch pattern. Pitch accents are prominent pitch movements that occur on specific syllables within a sentence, typically associated with the focus or new information. In questions, pitch accents can play a crucial role in signaling the information-seeking intent of the speaker and highlighting the part of the sentence that requires a response. For example, in English, the pitch accent in a question might fall on the interrogative word (e.g., WHO has seen this guy?) or on (the stressed syllable of ) a content word that is being questioned (Who has seen this GUY?). Last but not least, boundary tones are phonological features that occur at the edges of prosodic units, such as phrases or sentences. They contribute to the overall intonation contour of a sentence and can help signal its type, such as a question or a declarative statement. In the case of questions, the final boundary tone is often associated with a rising pitch movement (in polar questions) or a specific pitch pattern that indicates the question status of the sentence. Phonological analyses of questions can provide insights into cross-linguistic patterns and language-specific variations in the prosodic and intonational features of questions. Such analyses can reveal commonalities and differences

30

KEY CONCEPTS IN THE STUDY OF QUESTIONS

between languages in terms of how questions are phonologically marked and how these features interact with other aspects of the language system, such as syntax and semantics. Let me briefly mention some interesting cross-linguistic differences in this context. In Mandarin Chinese, polar questions often have a rising intonation pattern similar to English and many other languages. However, the specific pitch contour may vary depending on the sentence structure and the presence of question particles, such as ma. Look at the cases from Paul and Pan (2017: 53) again: (2)

a Tā huì shuō bāfálìyàyuˇ↑ ? 3SG can speak Bavarian ‘Can he speak Bavarian?’ b Tā huì shuō bāfálìyàyuˇ ma? 3SG can speak Bavarian PART ‘Can he speak Bavarian?’

While the version without ma obligatory features rising intonation (2a), the version with the particle (2b) works without such an intonation pattern. In Russian, polar questions often have a rising boundary tone at the end, while wh-questions may have a falling boundary tone (see Duryagin 2020; Rathcke 2006; Svetozarova 1998). There are also languages and language families that exhibit rather unique features in the phonological marking of questions. Some of these features may be specific to a particular language or shared among a group of related languages. We have already seen in Chapter 1 above that in some Bantu languages, such as Swahili, the question status is indicated by the presence of specific question words or particles, and the intonation pattern may be less prominent as a cue. Since the intonation pattern of questions in Swahili is quite similar to declarative sentences, the presence of question markers serves as the primary indicator of the question status. Finally, Greenlandic, an Inuit language spoken in Greenland, marks questions by changing the mood of the verb (see Aikhenvald 2011). That is, polar questions are indicated by using the interrogative mood, which is marked by specific morphological endings on the verb. While the intonation pattern in Greenlandic questions might exhibit a slight rise, the primary marker of the question status is the verb morphology rather than the intonation pattern. These examples illustrate that, while many languages share some common features in the phonological marking of questions, such as rising intonation patterns or pitch accents on focused elements, there are languages and language families that exhibit unique features, such as the use of specific

2.4 SUMMARY

31

question particles, tone changes, or morphological markings. These features provide valuable insights into the diversity and complexity of the phonological marking of questions across the world’s languages.

2.4 Summary This chapter has provided some basic background about questions in pragmatics, semantics, and phonology. In particular, I have first introduced basic pragmatic assumptions about the speech act of questioning in Section 2.1 by referring both to Searle (1969 et seq.) and to a more recent attempt at capturing those assumptions. In Section 2.2, I have then sketched some of the basic ideas of how to model questions in formal semantics. Finally, we have seen in Section 2.3 that questions are also an interesting cross-linguistic topic in phonology. The rest of the book will deal with the syntax–pragmatics of questions, with a special focus on the huge and heterogenous domain referred to as noncanonical questions. To get a first idea what kind of data are grouped together under this general label, the final chapter of this introductory Part I of the book will now sketch a typology of non-canonical questions.

3 A typology of non-canonical questions After having introduced basic notions of the syntax–pragmatics interface in Chapter 1 and fundamental aspects of questions more generally in Chapter 2, I now start with the main topic of the book and sketch an empirical landscape of non-canonical questions. In particular, I will group the various forms of non-canonical questions that have been discussed in the literature into three basic types: questions where the speaker is biased towards or against the truth of one of the alternatives (e.g., negative polar questions; Isn’t he running fast?), questions where the speaker expresses that their ignorance is either extremely strong (e.g., can’t-find-the-value questions; Where on earth did I leave my keys?) or extremely weak (e.g., rhetorical questions; After all, who likes drinking cold coffee?), and questions where the speaker expresses a presupposition of emotion (e.g., What the hell is going on?). Before we turn to the data on noncanonical questions in detail, let me first introduce the three notions that will help us to group together the different types of non-canonical questions: bias, epistemic states, and the speaker’s emotional state.

3.1 Introducing bias, epistemicity, and the speaker’s emotion In this section, I would first like to sketch the three basic notions that will help us in our attempt at clustering the diverse types of non-canonical questions that can be found in the literature. Let us start with the notion of ‘bias’. In theoretical linguistics, particularly in semantics, the term ‘bias’ refers to the inclination or preference towards certain interpretations or meanings of words or phrases based on various factors, such as personal beliefs, cultural background, or contextual information (see Beukeboom and Burgers 2017 on fundamental concepts). Biases can influence how language is used and understood, and can affect communication and discourse. Let us look at some examples of bias, which an English speaker can express when using more or less explicit grammatical means to convey a certain bias in interpretation at various levels. For instance, the use of the passive voice can sometimes obscure the agent or actor responsible for an action. This can be

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0003

3.1 INTRODUCING BIAS, EPISTEMICIT Y, AND THE SPE AKER’S EMOTION

33

used to avoid attributing responsibility for an action, or to shift the focus away from the agent. For example, Mistakes were made is a passive construction that does not identify who made the mistakes, and can be used to deflect blame. Also, the use of gendered language can reinforce gender stereotypes and biases. For example, the use of he as a generic pronoun can imply that males are the default or norm, while the use of she can be seen as marking an exception or deviation. Similarly, using masculine terms to describe jobs or roles can imply that those jobs are meant for men. A rather explicit way to convey a bias is the use of hedges (e.g., I think, It seems like), but also the use of intensifiers (e.g., totally, absolutely), which can signal a lack of confidence or certainty, or, conversely, a high degree of confidence or certainty, and we will see in later parts of the book (Chapter 9) how those basic intensification strategies are often applied in non-canonical questions. Turning now to the notion of epistemic state, we can summarize this concept as referring to the speaker’s degree of certainty or knowledge about the information they are communicating. As such, it overlaps with the domain of speaker bias. Epistemic states convey the speaker’s mental state or attitude towards the information they are conveying, and can be expressed through various means, such as the use of modal verbs, adverbs, or intonation (see Nuyts 1999 for a seminal study). Epistemic state is closely related to the concepts of modality and evidentiality, which also concern the expression of the speaker’s perspective on information (see Degen et al. 2019 for recent experimental work). Modality refers to the speaker’s assessment of the likelihood or desirability of a proposition, while evidentiality refers to the source or quality of the evidence that supports a proposition (see De Haan 1999; McCready and Ogata 2007 on this connection). By indicating their epistemic state, speakers can signal to the addressees how confident they are in the information they are conveying, and can provide important contextual cues for interpretation (e.g., a speaker who says I’m not sure, but I think … is indicating a lower degree of certainty than a speaker who says I know for a fact that …). The addressee can then adjust their own interpretation of the information accordingly. Finally, let me mention the expression of the speaker’s emotion. In this domain, most of the linguistic work on the topic of emotions is concerned with the conceptualization of emotions. In particular, many researchers have investigated the so-called emotional lexicon of different languages, carrying out cross-linguistic comparisons to find out which conceptualizations might be universal and which are language- and/or culture-specific (e.g., Jackson et al. 2019; Majid 2019; and

34

A T YPOLOGY OF NON-CANONICAL QUESTIONS

Wierzbicka 1999 for seminal work). Within this research tradition, particular emphasis is placed on content words such as nouns (joy), verbs (enjoy), and adjectives (happy). And, indeed, interesting cross-linguistic variations can be observed. For example, Foolen (2012) points out that the German (and English for that matter) distinction between Eifersucht (‘jealousy’) and Neid (‘envy’) does not exist in Dutch, because in this language the noun jaloezie refers to both emotional concepts. We thus see that emotional lexicons can be organized quite differently, even in languages that are closely related (such as Dutch and German). Most of the previous work on language and emotions illustrates that we can speak about emotions such as happiness and anger in descriptive terms (I’m enjoying my time with you versus I’m angry at you). However, in the context of this book, those descriptive strategies will be less relevant, and we rather focus on which (formal) linguistic elements are used to express emotional meaning. The expression of emotions is a dimension at the language–emotion interface that has been studied less intensively than the conceptual dimension (see, however, recent work by Gutzmann 2019 and Trotzke 2017a). Seminal work by both Bu¨hler (1934 [2011]) and Jakobson (1960) has already highlighted the important role that emotivity plays when one tries to define the central formal characteristics of human language. As Jakobson (1960: 354) has put it, “[t]he emotive function, laid bare in the interjections, flavors to some extent all our utterances, on their phonic, grammatical, and lexical level.” In other words, the function of interjections such as ouch! and wow! (i.e., conveying emotions) can be found in many linguistic domains and at all complexity levels. To illustrate this very briefly, let me mention the two domains of function words and syntax. The domain of emotional function words is exemplified best so tall). Interjections are a prime by interjections (wow!) and intensifiers (so in so example of expressing emotions (Goddard 2013; Norrick 2009), and intensifiers are elements that are used to strengthen or ‘boost’ the value of content words and phrases (Beltrama and Trotzke 2019; Gutzmann and Turgay 2014; Van der Wouden and Foolen 2017). Last but not least, let me highlight that many languages of the world feature a distinct sentence type—the exclamative—to express a basic human emotion: ‘surprise’ (English How fast he was!); see also our discussion of minor sentence types in Chapter 1 above. Exclamatives are indeed a prominent topic in the syntax–semantics literature on the expression of emotions (Castroviejo Miró 2006; Rett 2011; Villalba 2003, 2008). All in all, it is fair to say that recent years have seen a huge interest in the linguistic expression of emotions. Particularly

3.2 NON-CANONICAL QUESTIONS AND SPE AKER BIAS

35

work in formal semantics, syntax, and pragmatics has investigated relevant phenomena by using the term ‘expressive meaning’ or ‘expressive dimension of language’ (Potts 2007a). With these three concepts (speaker bias, epistemic states, and the speaker’s emotion) in mind, let us now see to what extent prominent cases of noncanonical questions can be grouped around those notions.

3.2 Non-canonical questions and speaker bias The most popular utterance type discussed in the literature in the context of expressing speaker bias via questions are so-called negative polar questions (see Bu¨ring and Gunlogson 2000; Reese 2006; and many more). Negative polar questions are questions that contain negation and, being a polar question, can be answered with a simple yes or no. These questions typically seek confirmation or clarification about a negative statement or assumption. The structure, interpretation, and response patterns of negative polar questions can vary across languages. In English, for example, negative polar questions often involve negation of the auxiliary verb do (e.g., Don’t you like pizza?). In other languages, negation might be expressed using different means, such as negative particles, affixes, or other morphological markers. Negative polar questions can be used for conveying different speech act interpretations, depending on their context and intonation. In particular, they can be used to seek genuine information (e.g., Don’t you have any siblings?), express surprise or disbelief (e.g., Didn’t you know that?), or elicit agreement or sympathy (e.g., Isn’t it a beautiful day?). The response patterns for negative polar questions can be more complex than for positive polar questions, as they can involve double negation and may depend on the specific language (see Repp and Geist forthcoming; Servidio et al. 2018 on reacting to a negative polar question). In English, for instance, answering a negative polar question like Don’t you like pizza? with yes can be ambiguous, as it could mean either Yes, I do like pizza (agreeing with the positive) or Yes, I don’t like pizza (agreeing with the negative). To avoid ambiguity, responses to negative polar questions often include additional information or clarification (e.g., Yes, I do or No, I don’t). When we now turn to the notion of speaker bias, then it is easy to see that negative polar questions are probably the clearest case of question speech acts where the speaker is biased towards or against the truth of one of the alternatives being considered. This bias may be based on the speaker’s prior

36

A T YPOLOGY OF NON-CANONICAL QUESTIONS

knowledge, beliefs, or opinions. As the examples above make very clear, when a speaker asks a negative polar question, they often expect or assume that the addressee will agree with the negative statement embedded in the question. This expectation can influence the interpretation and response patterns of negative polar questions, as well as the interaction between the speaker and the addressee. As already indicated above, expressing bias in negative polar questions can be used in order to perform different communicative functions. For instance, the questions can be used to seek confirmation of a negative assumption. The speaker might believe that the addressee agrees with the negative statement and wants them to confirm it. For example, in the question Don’t you think it’s too late to start a new project?, the speaker might assume that the addressee also believes it is too late, and they are looking for confirmation of this belief. However, in some cases negative polar questions can also express surprise or disbelief about a situation. The speaker might have expected a different outcome or response from the addressee and is seeking clarification or an explanation. In this case, the expression of speaker bias overlaps with conveying the speaker’s emotion (see Section 3.4 below). Take an example like Didn’t you know that she was coming?. Here, the speaker might be surprised that the addressee did not know about the event and is questioning why they were unaware. Finally, negative polar questions can also be used to elicit agreement or sympathy from the addressee. The speaker might use the negative question to emphasize a shared opinion or belief, seeking validation or support. For example, in the question Isn’t it a terrible day outside?, the speaker might expect the addressee to agree with the negative assessment of the weather, reinforcing their shared perspective. It is important to note that the bias expressed in negative polar questions can be influenced by various factors, such as the context of the conversation, the relationship between the speaker and the addressee, and the prosodic features of the question (intonation, stress, and so on). In some cases, a negative polar question might not convey a strong bias and could be interpreted as a neutral request for information (see Goodhue 2022). However, the presence of bias in negative polar questions can lead to certain challenges in communication. As discussed earlier, responding to negative polar questions can be ambiguous, as a simple yes or no might not clearly indicate whether the addressee agrees with the positive or negative statement. This ambiguity can be exacerbated by the presence of bias, as the addressee might feel pressure to

3.2 NON-CANONICAL QUESTIONS AND SPE AKER BIAS

37

align with the speaker’s expectations. What is more, negative polar questions can sometimes be perceived as face-threatening acts, as they can challenge the addressee’s knowledge, beliefs, or opinions (Reese 2007). When a speaker asks a negative polar question, the addressee might feel that their competence or credibility is being questioned, leading to potential misunderstandings or conflicts. To summarize, negative polar questions are a clear case where speakers can express a bias towards the truth value of the proposition of a question. This bias can pragmatically result in different effects, such as seeking confirmation, expressing surprise or disbelief, or eliciting agreement or sympathy. In closing this section, I hasten to add that a phenomenon as pervasive as English tag questions might be considered as falling into the same category of biased questions (see Krifka 2015; Reese and Asher 2007 on semantic details). They can be considered a type of polar question, but they have a unique structure that sets them apart from standard negative polar questions. Tag questions consist of a declarative statement followed by a short question tag, which is often a negation of the statement’s verb. The tag is used to seek confirmation, clarification, or agreement from the addressee. While not all tag questions are negative (e.g., You don’t like pizza, do you?), many of them do contain negation (e.g., You like pizza, don’t you?), making them similar to negative polar questions in some respects. Very similar to negative polar questions, tag questions can have a variety of functions as well. For instance, tag questions can be used to seek confirmation of the speaker’s assumption or belief. By using a tag question, the speaker invites the addressee to confirm or deny the statement. For example, in the question It’s going to rain tomorrow, isn’t it?, the speaker might assume that the addressee agrees with the prediction and is seeking confirmation. On the other hand, tag questions can also be used to soften a statement or assertion, making it less forceful or direct. By turning a statement into a question, the speaker allows the addressee to confirm, deny, or modify the statement, reducing the potential for conflict or disagreement. For example, in the question You’re coming to the party tonight, aren’t you?, the speaker might be seeking clarification about the addressee’s plans while also expressing an expectation or desire for their attendance. In sum, negative polar questions (as well as tag questions) express a speaker bias, and on this basis they can serve special communicative purposes, such as eliciting agreement, seeking confirmation, and so on. Let us now turn to the second main concept that can be used for a typology of non-canonical questions: epistemic states.

38

A T YPOLOGY OF NON-CANONICAL QUESTIONS

3.3 Non-canonical questions and epistemic states Epistemic states are the speaker’s degrees of certainty, doubt, or belief about the truth of a proposition, and therefore the examples above about biased questions overlap with some of the features discussed in this section. Needless to say, there is a wide range of cross-linguistic variation in the ways that speakers can express their epistemic states in questions and how these states are modified in different languages. In this section, I will concentrate on two types that illustrate the whole scale of the speaker’s degrees of certainty. In particular, I will focus on a type of non-canonical question where the speaker expresses that their ignorance is either extremely strong (e.g., can’t-find-the-value questions; Where on earth did I leave my keys?), and, on the other hand, on a type where the speaker’s ignorance is extremely weak (e.g., rhetorical questions; After all, who likes drinking warm beer?). Let us start with the type referred to as can’t-find-the-value questions. In the following sketch, I will focus on German, because in this language, the can’t-find-the-value reading of questions is signaled by dedicated modal particles. Specifically, when the German modal particles nur or bloß are included in a question like (1), it gains an additional layer of meaning beyond simply seeking information. Often referred to as a wh-on-earth component, this extra meaning has been argued to create a unique type of non-canonical question (see Bayer and Obenauer 2011; Hinterho¨lzl and Munaro 2015, 2021). Essentially, the presence of the modal particles indicates that the speaker has made multiple attempts to find a value for the wh-variable, but has been unsuccessful. (1) Wer hat mein Buch {nur, bloß} gestohlen? who has my book PART,PART stolen ‘Who on earth has stolen my book?’ (1) has distinct use-conditional properties, compared to a neutral informationseeking question. More specifically, (1) is felicitous only when the speaker has already asked themselves the question multiple times and is genuinely perplexed, as they cannot identify a specific person who might have stolen the book. In earlier research, the meaning of modal particles in such questions has been described informally as conveying a kind of intensifying effect, as the use of nur/bloß as modal particles in these questions indicates that the speaker is highly engaged or even desperately seeking an answer. Thurmair (1989) refers to this as the ‘strengthening’ feature (translated from German: Versta¨rkung).

3.3 NON-CANONICAL QUESTIONS AND EPISTEMIC STATES

39

Evidence for this intensifying effect of nur/bloß can be seen in their inability to appear in questions that serve phatic communication, or casual small talk (examples adapted from Thurmair 1989: 180): (2) a

Wie geht’s Dir? / Wie geht’s Ihrer Familie? how goes.it you how goes.it your family ‘How are you?’ / ‘How is your family doing?’

b #Wie geht’s Dir {nur, bloß}? / #Wie geht’s {nur, bloß} how goes.it you PART,PART how goes.it PART,PART Ihrer Familie? your family ‘How on earth are you? / How on earth is your family doing?’ It is evident that, when posing questions like (2a), the speaker usually has little interest in obtaining a response. This is why the use of nur and bloß is infelicitous in such examples (2b). One feature that might seem contradictory to this heightened interest in receiving an answer (the strengthening effect) is the near-mandatory presence of the modal particles nur and bloß in German verbfinal wh-questions that convey a deliberative interpretation. Importantly, this type of question is generally self-oriented rather than focused on the addressee. As a result, the speaker does not anticipate a direct answer but is instead contemplating something that is currently at the forefront of their thoughts (Truckenbrodt 2013: 241): (3) a

Wen sie nur/bloß whom she PART/PART ‘Whom does she like?’

mag? likes

b #Wen sie mag? whom she likes (intended: same as [3a]) In this situation, it is noteworthy that the modal particles nur and bloß are frequently employed to make a rhetorical question sound more authentic, even though their presence is not essential for understanding it as a rhetorical question (unlike the mandatory nature observed in the deliberative instances mentioned above). To illustrate this point, consider the next example: (4) Was will der {nur, bloß} mit dieser Schrottkarre? what wants he PART,PART with this junk.car ‘What on earth does he want with that junk car? I don’t see the point (of him buying that car)!’

40

A T YPOLOGY OF NON-CANONICAL QUESTIONS

The question in (4) evidently does not require a response, which is similar to the circumstance in the deliberative case (3), as those questions are typically not directed towards the addressee. However, in our key example (1), the speaker may still pose the question to a recipient, asking for assistance in determining who has stolen their book. The question now is how to account for the effect of nur and bloß in wh-questions semantically and in terms of epistemic states. To answer that question, let us look at the seminal work of Den Dikken and Giannakidou (2002), who build on Karttunen’s influential perspective (1977) that a question consists of a set of propositions encompassing its true responses (see also Section 2.2 above; for an alternative viewpoint, see, e.g., Groenendijk and Stokhof 1984). According to this methodology, our non-canonical question in (1), just like the neutral information-seeking question without the particle, represents a collection of propositions in the form of ‘x stole the book’, with x being an individual. Den Dikken and Giannakidou (2002: 41) apply the condition p(w) to guarantee that only true propositions are pertinent: (5)

a Wer hat mein Buch gestohlen? ‘Who has stolen my book’? b ⟦who⟧(w) = ⟦somebody⟧(w) = λQ ∃x [person(x) (w) ∧ Q(x) (w)]

(6)

a ⟦Wer hat mein Buch gestohlen?⟧(w) = λp ∃x [p(w) ∧ person(x) (w) ∧ p = λw (steal(x) (book) (w))] b Q = {Eva has stolen my book, Luisa has stolen my book, Max has stolen my book}

If Eva, Luisa, and Max have stolen the book in the real world, then the trio of propositions outlined in (6b) form the accurate and comprehensive response to the query in (5a). But how about utterances where the modal particles nur or bloß turn a question like (5a) into a non-canonical can’t-find-the-value question? Den Dikken and Giannakidou (2002: 43) contend that, in the case of whon-earth/wh-the-hell questions, modifiers such as on earth/the hell broaden the domain of quantification (hereafter referred to as D). In other words, D encompasses not only a presupposed subset as seen in standard wh-questions (5/6), but also defines an open set that includes all people in the universe and all conceivable values for x (even the less probable ones): (7)

⟦Wer hat mein Buch nur/bloß gestohlen?⟧(w) = λp ∃x ∈ D [p(w) ∧ person(x) (w) ∧ p = λw (steal(x) (book) (w))]

3.3 NON-CANONICAL QUESTIONS AND EPISTEMIC STATES

41

The expansion of this domain precisely captures the sense of bewilderment or disorientation previously described as being the special epistemic state experienced by the speaker when uttering this type of non-canonical question. Since the search domain for identifying a value for x is significantly broadened by employing the modal particles nur and bloß, questions like (7) frequently convey that the speaker genuinely lacks any notion of what the answer might be. In this regard, the modal particles nur and bloß behave similarly to polarity items (such as any), which are also known for domain extension or ‘widening’ (Kadmon and Landman 1993). Nonetheless, nur and bloß also exhibit differences from wh-the-hell and similar expressions—and this is where their focus particle semantics comes into play: both nur and bloß are polysemes and can also be used as the focus particle ‘only’ in German (see Do¨rre 2018; Do¨rre and Trotzke 2019; and Reimer and Trotzke 2022). The distinction between our cases of non-canonical questions and the cases discussed by Den Dikken and Giannakidou (2002) lies in the fact that German wh-questions containing the modal particles nur and bloß neither express doubt about the existence of a value for x nor communicate the presupposition that the speaker has a negative attitude towards the value for x. Den Dikken and Giannakidou assert that such a presupposition is an integral part of the semantics of wh-the-hell. To observe this, consider the following example: (8)

Who the hell has stolen my book?

Without delving too deeply into the specifics, we can rephrase the presupposition of a negative attitude as a conditional statement: if there exists any person x who stole the book, then x should not have stolen it. This presupposition is also readily identifiable when using the German equivalents w-in-aller-Welt (‘wh-in-all-world’) and w-zum-Teufel (‘wh-to.the-devil’); for studies on these expressions, see Stefanowitsch (2011) and Wiltschko (1997). All in all, we can say that can’t-find-the-value questions can be signaled via modal particles in German or by using idiomatic expressions such as wh-onearth in languages other than German. As for epistemic states, the speaker conveys that the degree of ignorance and thus uncertainty is extremely high, often yielding a sense of frustration. With this in mind, I now turn to rhetorical questions, the type of non-canonical question where the speaker’s ignorance is extremely weak. In non-technical terms, rhetorical questions are a type of question that is not intended to elicit an answer from the addressee but rather to make a point, convey an opinion, or emphasize a certain perspective. Rhetorical questions

42

A T YPOLOGY OF NON-CANONICAL QUESTIONS

can serve various communicative functions, such as expressing surprise, sarcasm, disbelief, or criticism (see Caponigro and Sprouse 2007; Frank 1990). In this brief discussion, I will focus on the basic semantics of rhetorical questions and the extent to which they signal the speaker’s certainty about the truth value of a proposition. We will discuss the pragmatics of this prominent type of non-canonical questions in Chapters 7 and 8 in more detail. The semantics of rhetorical questions is distinct from that of regular information-seeking questions. In regular questions, the speaker seeks information, clarification, or confirmation from the addressee. In contrast, rhetorical questions do not aim to gather new information but instead express a presupposed or implied proposition. The speaker uses the rhetorical question as a device to convey their stance, opinion, or emotion regarding the subject matter. For example, consider the rhetorical question Is the Pope Catholic? The speaker is not genuinely asking about the Pope’s religious affiliation; instead, they are expressing that the answer is obvious and indisputable. The rhetorical question, in this case, implies a positive answer and highlights the speaker’s strong belief in the proposition’s truth. Accordingly, as for the domain of epistemic states, rhetorical questions can signal a high degree of certainty on the part of the speaker about the proposition in question. By posing a question that has an obvious or self-evident answer, the speaker implies that the proposition is indisputable and that any other perspective or interpretation would be absurd or illogical. In this context, it is worth mentioning that this extreme degree of certainty is often conveyed through the use of rhetorical devices, such as irony, sarcasm, or hyperbole (see, e.g., Oraby et al. 2017). Let me give some examples of rhetorical questions that signal the speaker’s extremely strong certainty about the proposition’s truth value: in the following cases, by highlighting the self-evident nature of the proposition’s truth and by using humor and exaggeration to underscore the certainty of the proposition, the speaker suggests that any doubt or disagreement would be unreasonable: (9)

a Do bears live in the woods? b Can birds fly? c Does a one-legged duck swim in circles?

However, I hasten to add that not all rhetorical questions signal that strong degree of certainty. Some rhetorical questions may express doubt, criticism, or disbelief, and their purpose might be to challenge or undermine a given proposition. In these cases, the rhetorical question may signal not the speaker’s strong

3.4 NON-CANONICAL QUESTIONS AND THE SPE AKER’S EMOTION

43

certainty but rather their skepticism or disagreement with the proposition. For example, consider a rhetorical question such as Who died and made you king? The speaker is not genuinely asking about a royal succession; instead, they are challenging the addressee’s authority or right to make a certain decision or judgment. The rhetorical question, in this case, signals the speaker’s disagreement with the addressee’s position, rather than expressing strong certainty about a proposition. In conclusion, rhetorical questions very often signal an extremely high degree of certainty on the part of the speaker about the proposition’s truth value in question, especially when used as rhetorical devices, as illustrated above. As we will see in later parts of the book, rhetorical questions can also convey the speaker’s emotion towards a proposition. In a question like Did you ever lift a finger to help me?, the speaker is most likely conveying a negative emotion. In the following section, we will briefly mention question types where this emotive component is a signature property of those cases.

3.4 Non-canonical questions and the speaker’s emotion In this section, we have a look at a prominent example where the noncanonicity of a question speech act is due to the expression of the speaker’s emotion. We will discuss many more cases in the following chapters (especially in Chapters 5 and 9), but the following example is a prominent question type for illustrating how questions can additionally convey presuppositions of emotions—plus, we have already introduced some aspects of this particular case above. As already mentioned in Section 3.3, wh-the-hell questions, also known as wh-the-heck, wh-the-devil, or wh-the-dickens questions, are a type of emphatic question that involves the use of any wh-question word (i.e., who, what, where, when, why, and how) followed—in English—by the idiomatic expressions the hell, the heck, the devil, or the dickens (analogous expressions can be found in many other languages in this type of question, see Celle et al. 2021). These questions are used to express strong emotions, such as surprise, disbelief, annoyance, or frustration (on surprise questions more generally, see Giorgi 2016, 2018). The addition of the idiomatic element serves to intensify the question and highlight the speaker’s emotional state (see Chapter 9 on more details about such intensification strategies). Examples of wh-the-hell questions include:

44

(10)

A T YPOLOGY OF NON-CANONICAL QUESTIONS

a b c d

What the hell is going on? Why the heck did you do that? Who the devil ate my sandwich? How the dickens did you get here so fast?

Wh-the-hell questions convey negative emotions to varying degrees, depending on the context, the speaker’s intention, and the specific emphatic expression used. For instance, wh-the-hell questions can be used to express shock or astonishment at an unexpected event or situation. The idiomatic element intensifies the question and underscores the speaker’s surprise (e.g., What the hell just happened? or How the heck did you manage to do that?). Wh-thehell questions can also convey the speaker’s frustration or discontent with a situation or someone’s actions. In those negative cases, the addition of the idiomatic expression signals that the speaker is not merely seeking information but also expressing their displeasure (e.g., Why the hell didn’t you tell me earlier? or Where the devil have you been?). In some cases, wh-the-hell questions can express very strong negative emotions, such as outrage or anger. In those instances, the wh-the-hell phrase serves to emphasize the speaker’s emotional state and communicate their dissatisfaction or disapproval (e.g., Who the hell do you think you are? or What the dickens were you thinking?). In sum, while wh-the-hell questions generally express negative emotions, the specific emotion conveyed and its intensity can vary based on the context. Den Dikken and Giannakidou (2002) present a comprehensive analysis of wh-the-hell questions, specifically focusing on the ‘aggressively non-D-linked’ aspect of the wh-phrases in those questions and their behavior as polarity items. Very briefly, in their analysis Den Dikken and Giannakidou argue that the hell (or other similar expressions) in wh-the-hell questions functions as a negative polarity item. Negative polarity items are linguistic expressions that are restricted to occurring in contexts with a certain type of negative or downward-entailing environment. The authors propose that wh-the-hell questions are ‘aggressively non-D-linked’ because they lack any specific referential properties or discourse connections, unlike typical D-linked wh-phrases (e.g., which book). The term ‘aggressively’ here emphasizes the extreme lack of referentiality in wh-the-hell phrases. According to Den Dikken and Giannakidou, the non-D-linked nature of wh-the-hell questions makes them sensitive to the polarity of their environment. They observe that wh-the-hell questions often occur in contexts that license negative polarity items, such as negation (11a), conditional clauses (11b), and, important in our context, questions (11c):

3.5 SUMMARY

(11)

45

a I don’t know what the hell you’re talking about. b If you find out where the hell he is, let me know. c What the hell are you doing?

In their paper, they argue that the presence of the hell or similar expressions within the wh-phrase acts as a licenser of negative polarity items. They propose a syntactic account in which the the hell component is generated in a position associated with the wh-phrase’s polarity-sensitive behavior. The paper also discusses cross-linguistic data, demonstrating that similar polarity-sensitive behavior can be observed in other languages’ equivalents of wh-the-hell questions (again, see Celle et al. 2021 for such cross-linguistic observations). Given more recent accounts about expressive meaning components (Potts 2007a), one could also analyze the negative presupposition of wh-the-hell questions as an expressive presupposition in the sense of Schlenker (2007). Expressive presuppositions refer to the underlying assumptions or attitudes that a speaker conveys through their use of specific linguistic forms or expressions. In the case of wh-the-hell questions, the expressive presupposition is the speaker’s negative emotions or attitude towards the situation or the topic of the question. The expressive presupposition of wh-the-hell questions is closely related to their pragmatic function. In particular, those questions are, pragmatically, often used to emphasize the speaker’s emotions or attitude, rather than to seek new information or clarification. The study of expressive presuppositions is a relatively recent area of research in semantics and pragmatics, and there is ongoing debate about the best way to model them. Some scholars argue that expressive presuppositions should be treated as part of the semantic content of the utterance, while others argue that they are part of the pragmatic meaning and arise from the speaker’s use of language in context, and this is why one could also frame the interpretation of wh-the-hell questions as a type of use-conditional meaning (see Gutzmann 2015 and our detailed discussion about expressivity in Chapter 10). Given this brief illustration of a non-canonical question type where the expression of the speaker’s emotion is the prominent semantic–pragmatic feature, let me now summarize the whole chapter.

3.5 Summary In this chapter, I have introduced some common forms of non-canonical questions that have been discussed in the literature by focusing on three basic interpretative components that are associated with non-canonical

46

A T YPOLOGY OF NON-CANONICAL QUESTIONS

readings: questions where the speaker is biased towards or against the truth of one of the alternatives (e.g., negative polar questions; Isn’t he running fast?), questions where the speaker expresses that their ignorance is either extremely strong (e.g., can’t-find-the-value questions; Where on earth did I leave my keys?) or extremely weak (e.g., rhetorical questions; After all, who likes drinking warm beer?), and questions where the speaker expresses a presupposition of emotion (e.g., What the hell is going on?). Needless to say, this typology lacks many further phenomena in the domain of non-canonical questions, which we will discuss in the following syntax–pragmatics chapters, in addition to the types mentioned in this introductory Part I of the book. However, this chapter has already made clear which interpretative components—bias, epistemic state, and emotion—we very often use to characterize non-canonicity in the domain of question speech acts.

PART II

THE S Y N TA X OF NON-C A NONIC A L QUESTIONS

4 Non-canonical word order In Part I of the book, I introduced some aspects of the interface between syntax and speech acts (Chapter 1), the basic approaches to the meaning and phonology of questions more generally (Chapter 2), and some core domains and empirical phenomena illustrating typical non-canonical questions (Chapter 3). This part of the book now turns to the syntactic triggers of non-canonical interpretations of questions. In this first chapter of Part II of the book, we will first explore the aspect of non-canonical word order in questions. In particular, we will discuss non-canonical effects of optional wh-in-situ configurations and so-called echo questions in languages such as English, (colloquial) French, and German. After that, we will focus on the phenomenon of declarative questions (He is running fast? Really?) and non-canonical frontings in polar questions (i.e., object fronting in questions in German and Italian). Again, the chapter concludes with a summary and in this case with some first conclusions as to why not every type of non-canonical questions can be explained at the syntax–pragmatics interface. Before we turn to the individual phenomena illustrating where word order plays a role in the domain of non-canonical questions, let me briefly sketch the interaction between word order—both canonical and non-canonical— prosody, and information structure in more general terms.

4.1 Word order, prosody, and information structure Word order and prosodic patterns are two essential components of language that work together to convey meaning. By word order, I refer to the syntactic arrangement of words in a sentence. Word order varies across languages, with some languages having strict word-order rules and others being more flexible. The most common word orders are Subject–Verb–Object (SVO) as in English, Subject–Object–Verb (SOV ) as in Japanese, and Verb–Subject–Object (VSO) as in Irish. The arrangement of words in a sentence serves to establish grammatical relationships and convey meaning at different levels. On the other hand, prosodic patterns pertain to the rhythm, stress, and intonation of spoken Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0004

50

NON-CANONICAL WORD ORDER

language. These patterns can influence how a listener interprets a sentence, as they can provide cues to a speaker’s emotional state, focus, or intentions. Prosodic features include pitch, duration, intensity, and pauses. Importantly, they often work in tandem with word order to create well-formed utterances. Let us delve into the interplay between word order and prosody by looking at some basic patterns. Word order and prosodic patterns interact in several ways, but they can also be disconnected from each other. For instance, prosody can be used to shift the focus or emphasize certain words in a sentence, even when the word order remains unchanged. By stressing a particular word or using a specific intonation pattern, a speaker can draw attention to a specific part of the sentence. For example, in the English sentence I didn’t take the book, stressing different words can change the meaning: (1) a I didn’t TAKE the book [implies that the speaker didn’t take the book, but perhaps did something else with it] b I didn’t take the BOOK [implies that the speaker didn’t take the book, but maybe took something else] Since prosody can thus act rather independently, it can help clarify the meaning of a sentence with ambiguous word order. In languages with flexible word order, prosodic cues can signal the intended grammatical relationships among words. For instance, in Russian, word order is relatively free, and prosodic features such as stress and intonation can help listeners determine the correct interpretation of a sentence (see Bailyn 2012 on the syntax and flexible word order in Russian). As we will see in the following sections of this chapter in more detail, this autonomy of prosody also plays an important role in the domain of questions. For example, questions in English can be formed by altering the intonation of a sentence without changing the word order. A rising intonation at the end of a sentence can turn a declarative sentence into a yes–no question: You’re going to the store? (see Section 4.3 below). An important domain where word order and prosody often interact is information structure, where non-canonical word orders can have an effect on both prosody and discourse interpretation. In a very basic sense, non-canonical word order refers to arrangements of words that are marked and deviate from the typical word order of a specific language mentioned above. These deviations can be used for various purposes, such as emphasizing specific elements,

4.1 WORD ORDER , PROSODY, AND INFORMATION STRUCTURE

51

highlighting new or contrasting information, or accommodating the requirements of a particular discourse context. The interplay between non-canonical word order, prosody, and information structure can be observed in several ways. For instance, non-canonical word orders are often used to emphasize or focus on specific elements in a sentence. This rearrangement can be accompanied by prosodic changes, such as stress or pitch accents, further to highlight the focused element. For example, in English we can say The book, she took instead of the canonical She took the book to emphasize the object the book. In this case, prosodic cues, like stress on the book, also contribute to the emphasis. Some languages use non-canonical word orders to mark the topic of a sentence or to front a particular element to the periphery of the clause. In these cases, prosody can work together with word order to convey the intended information structure. For example, in German, topicalization can involve moving an element to the beginning of a sentence, often accompanied by a distinct intonation pattern to signal the topic status. Let us look at this case in more detail to see that the interaction between word order, prosody, and information structure requires a close look at the relevant data, because not any type of discourse meaning in the domain of information structure interfaces with syntax. To see this, we must first differentiate between pragmatically significant and non-significant instances of moving a constituent to the German left periphery. In German, elements such as subjects or high adverbials (e.g., leider ‘unfortunately’) appear in the left periphery of a clause without acquiring any unique pragmatic interpretation. However, as Frey (2004a) highlights, the left periphery can also accommodate topical content, as demonstrated in (2): (2)

Let me tell you something about Eliud. Den Eliud muss jeder Konkurrent fu¨rchten the.ACC Eliud must every competitor fear ‘A s for Eliud, every competitor must fear him.’

Importantly, as Frey maintains within his generative approach, topical content has already undergone a movement operation before being relocated to the left periphery. In particular, den Eliud acquires a topical interpretation while still in the middle field of the German clause (for more details on this topic position, see Frey 2004b):

52

NON-CANONICAL WORD ORDER

(3)

a

… dass (‘that’) den Eliud jeder Konkurrent fu¨rchten muss.

b

CP

TopP

dass

TP

den Eliud

jeder Konkurrent

Tˈ

T0

VP

den Eliud fürchten muss Frey posits that the constituent [den Eliud] is nearest to the specifier of CP and, under this assumption, is the primary (and, given certain assumptions about derivational economy, also the sole viable) target for so-called Formal Movement (for this concept, see Fanselow 2002), as depicted in (4):

(4)

CP

den Eliud

Cˈ

muss

TopP

den Eliud

TP

jeder Konkurrent den Eliud fürchten muss

4.1 WORD ORDER , PROSODY, AND INFORMATION STRUCTURE

53

Consequently, movement to the left periphery in this instance “appears to be unrelated to any semantic or pragmatic attributes and seems to be purely formal” (Frey 2004a: 8). Let us now examine a case that, as per Frey, is connected to a unique discourse interpretation in the left periphery and also goes along with a special prosody. Frey posits that the movement of a constituent from an embedded clause (‘long movement’) to the left periphery inevitably leads to contrast (5), while the short movement of the same element, as observed earlier (2), does not: (5)

Let me tell you something about Eliud. [Den ELIUD]i denkt Andreas, dass jeder Konkurrent ti the.ACC Eliud thinks Andreas that every competitor fu¨rchten muss. fear must.

What is crucial in our context is that this observation also applies to nonembedded cases such as resultative predicates; observe the following adjectival example: (6)

STOLZ wird Eliud seine Fans machen (und sie proud will Eliud his fans make (and them nicht entta¨uschen). not disappoint) ‘Eliud will make his fans proud (and not disappoint them).’

In (6), the constituent stolz is emphasized with focal stress and is interpreted contrastively. Crucially, stolz is neither base-generated at the highest position of the TP-zone (7a) nor can it be relocated to the position that accommodates topical content, as illustrated in (7b); see Frey (2004a: 19): (7)

[CP [C’ dass (‘that’) [TopP [Top’ Top0 [TP stolz [TP Eliud [T’ T0 [VP seine Fans machen wird]]]]]]]] b ∗ [CP [C’ dass (‘that’) [TopP stolz [Top’ Top0 [TP Eliud [T’ T0 [VP seine Fans stolz machen wird]]]]]]]]

a

∗

As (7) represents, (6) cannot result from Formal Movement and thus stolz has reached its position via another type of movement (‘long Ā-movement’), which can be depicted as follows:

54

NON-CANONICAL WORD ORDER

(8)

KontrP

STOLZ

Kontrˈ

wird

TP

Eliud seine Fans stolz machen wird As depicted in (8), this form of movement is syntactically distinct from other methods used to fill the German left periphery. Importantly, as pointed out by Frey as well, it is the sole type of marked word order that is obligatory linked to focal stress on the moved element in the domain of prosody and a unique pragmatic interpretation in the domain of discourse. Consequently, Frey suggests that stolz targets the specifier position of a functional projection designated for contrastive elements (KontrP). Those facts about German syntax illustrate that non-canonical word orders can have a significant impact on both the prosody and the discourse interpretation. By deviating from the typical word order, speakers can emphasize, focus, or contrast specific elements, and prosodic patterns can further enhance these effects—and sometimes they do so obligatorily, as in the German example of left-peripheral frontings. All in all, we have seen that the interaction between non-canonical word orders, prosody, and discourse interpretation can be more or less direct, and not every change in the word order comes with a special prosody and/or pragmatic interpretation. With those basic aspects of non-canonical word order in mind, we now again turn to non-canonical word order in questions.

4.2 wh-in-situ and echo questions Structures referred to as wh-in-situ questions are a type of interrogative construction in which wh-words (such as who, what, where, when, why, and how) remain in their original, or ‘in-situ’, position in the sentence, rather than being fronted or moved to the periphery of the sentence, as in

4.2 WH -IN-SITU AND ECHO QUESTIONS

55

English (see Bayer and Cheng 2017; Cheng 2009 for general overviews of this phenomenon). In many languages, those questions instantiate the neutral (i.e., ‘canonical’) way to ask a question. In particular, in wh-in-situ languages the wh-word remains in the position where the corresponding syntactic phrase would appear in a non-interrogative sentence. Here is an example from Mandarin Chinese (Cheng 2003: 103): (9)

Hufei mai-le shenme Hufei buy-PERF what ‘What did Hufei buy?’

Interestingly, some languages also employ a mix of wh-movement to the peripheries and wh-in-situ strategies, depending on the context or specific whwords involved. For example, certain wh-words in French, such as pourquoi (‘why’) and comment (‘how’), can appear in situ, while others, such as qui (‘who’) and que (‘what’), typically undergo wh-movement (see Adli 2006; Chang 1997; Glasbergen 2021; Shlonsky 2012). More specifically, wh-in-situ questions in French are relatively rare compared to canonical wh-questions involving wh-movement, but they do occur in some informal, conversational contexts. Here are some examples of wh-insitu questions in French (10a, 11a), as compared to the left-peripheral options (10b, 11b); examples are constructed based on personal knowledge: (10)

a Tu fais quoi? you do what ‘What are you doing?’ b Qu’est-ce que tu fais? what.is.it that you do What are you doing?

(11)

a Il habite où? he lives where ‘Where does he live?’ b Où est-ce qu’il habite? where is.it that.he lives? ‘Where does he live?’

As mentioned earlier, certain wh-words in French can appear more readily in wh-in-situ constructions, while others typically undergo wh-movement. For example, pourquoi and comment often appear in situ (12), while qui and que

56

NON-CANONICAL WORD ORDER

usually undergo wh-movement (13); examples constructed based on personal knowledge: (12)

a Tu pleures pourquoi? you cry why ‘Why are you crying?’ b Tu t’appelles comment? you are.called how ‘What’s your name?’

(13)

a Qui a fait ça? who has made this ‘Who did this?’ b Que veux-tu manger? what want.you eat ‘What do you want to eat?’

I have already indicated above that wh-in-situ questions in French are generally considered more informal and conversational than their canonical counterparts. They are more common in spoken language and colloquial speech, while the canonical wh-questions involving wh-movement are more typical in written language and formal contexts. Crucially, the interpretation of wh-in-situ questions in French is similar to that of their canonical counterparts, although they convey a more informal tone. In other words, wh-in-situ questions in French are not really an instance of a non-canonical question type, according to the main concepts we have introduced in Chapter 3: they are not expressing any speaker bias, a special epistemic state, or any particular emotion. In contrast to French, wh-in-situ questions in English are relatively rare and occur only in special contexts, because they indeed feature a non-canonical and special interpretation. Examples of wh-in-situ questions in English include colloquial cases like: (14)

a You saw who? b He went where? c She did what?

In terms of their interpretation, wh-in-situ questions in English typically convey the same basic question meaning as their canonical counterparts (i.e., Who did you see?, Where did he go?, What did she do?). However, they additionally have a slightly different pragmatic effect, often suggesting surprise, disbelief, or a more informal and conversational tone (see Cohen 2007; Repp and

4.2 WH -IN-SITU AND ECHO QUESTIONS

57

Rosin 2015). In some cases, these questions can also convey a sense of emphasis or contrast, particularly when combined with prosodic cues like contrastive stress. It is worth noting that the meaning of wh-in-situ questions in English can be captured only by looking at their conversational context, as in (15): (15)

A: I met the new manager yesterday. B: You met who?

In this case, the wh-in-situ question (You met who?) serves as an echo question, with Speaker B seeking clarification or expressing surprise regarding the information provided by Speaker A. In conclusion, wh-in-situ questions in English do involve a non-canonical word order, and their interpretation features a pragmatic meaning component that restricts their use to particular contexts in a conversation. In the context of this chapter, the question for us now is whether we should postulate any dedicated syntactic structure or movement operation in our syntactic analysis to account for this special interpretation. We have seen in Section 4.1 above that, sometimes, a syntactic operation can result in both a special prosody and a non-canonical discourse interpretation. Interestingly, in the case of wh-in-situ questions in English, it is the lack of a syntactic operation (i.e., no movement to the left periphery) that results in a non-canonical interpretation. Recall that the main thread of this chapter is the interaction between word order, prosody, and discourse interpretation. Given this background, we observe that in all cases of English wh-in-situ questions above, the wh-element would receive heavy stress (e.g., You met WHO?). One hypothesis could therefore be that this type of prosody blocks movement of the whelement and yields the non-canonical interpretation. This is the claim I would like to argue for here, and I will provide evidence in favor of this account further below. Note that the other theoretical alternative would be to postulate an additional syntactic category in the lower projections of the clause, which would encode the special discourse interpretation. I consider this alternative as theoretically unsound, because it is just restating the facts in terms of a dedicated syntax–semantics mapping. One piece of evidence corroborating my claim that the special readings of wh-in-situ constructions in languages like English do not involve any dedicated syntax comes from a type of non-canonical question that we have already introduced in Chapter 3 above in great detail: wh-the-hell questions. It is a crucial feature of those questions that they display a root/non-root asymmetry. Specifically, English wh-the-hell elements seem to be restricted to the left periphery of the clause—the ‘root’ of the syntactic tree. That is, as initially

58

NON-CANONICAL WORD ORDER

observed by Lasnik and Saito (1984), English seems to prohibit a wh-phrase containing the hell from remaining in situ; observe the following examples: (16)

a ∗ You ate what the hell? (cf. What the hell did you eat?) b ∗ You went where the hell? (cf. Where the hell did you go?)

Crucially, as we will see in more detail in later parts of the book, the same applies to other types of non-canonical questions such as surprise–disapproval questions (see Chapter 5 below). Very briefly, in the German version of those questions, the wh-element fails to cover a set of alternative values for reasons. Unlike the German wh-item warum (‘why’) in (17a), the why-like was (‘what’) in these configurations is not focusable and, importantly, cannot appear in situ in multiple questions (17b); see Bayer and Trotzke (2015): (17)

a

∗

Wer lacht WAS?! (cf. Was lacht der?!) who laughs what Intended: ‘Who is laughing why?’ b Wer lacht WARUM? (cf. Warum lacht der?) who laughs why ‘Who is laughing why?’

We can conclude from the examples above that, in those cases where whelements are confined to the expression of non-canonical readings (e.g., whthe-hell and why-like-what), they must move to the left periphery in languages such as English and German and cannot stay in situ. Since, in the wh-in-situ cases illustrated above (e.g., You met WHO?), the wh-element is not attracted to the left periphery, I submit that those elements do not interact with any of the discourse-related projections in the left periphery involved in other types of non-canonical questions, and that their special interpretation is due to their non-standard prosody only. Let us now turn to another case for exploring the interplay between syntax, prosody, and interpretation in the domain of questions: so-called declarative questions.

4.3 Declarative questions Declarative questions are a type of interrogative construction that have the grammatical form of a declarative statement but function as a question. These questions are used to seek confirmation, clarification, or agreement from the listener while presenting the information in the form of a statement. Declarative questions often rely on prosodic cues to convey their interrogative nature

4.3 DECL ARATIVE QUESTIONS

59

(see Gunlogson 2003 and Safárová and Swerts 2004 on a seminal prosodic study; Stivers 2010 on their pervasiveness in American English). In particular, this question type typically employs a rising intonation pattern to indicate that the sentence is a question. This rising intonation pattern is particularly important in distinguishing declarative questions from statements, as the syntax remains the same, as the following English examples illustrate: (18)

a You’re coming to the party tonight? b She’s your sister? c They already left?

Declarative questions are, of course, not exclusive to English; they can be found in many other languages as well (see Beun 1990; Henriksen 2012; Seeliger and Repp 2018; and many more). The exact prosodic cues and intonation patterns used to mark declarative questions may vary across languages, but the general concept remains similar. Declarative questions often serve a rather specific (hence non-canonical) pragmatic function in conversation. They can be used to seek confirmation or clarification, to express disbelief or surprise, to convey politeness or indirectness, or to check the addressee’s understanding or agreement. Let us briefly illustrate some of those functions. First, declarative questions can be used to seek confirmation or clarification of information, implying that the speaker has some prior knowledge or assumption but needs validation from the addressee. An example for that particular use would be the following¹: (19)

You’re from New York(, right)?

Declarative questions can also convey politeness or indirectness by avoiding a direct question, which may be perceived as too intrusive or demanding. By using a declarative question, the speaker in these cases implies that they expect the addressee to agree, allowing the listener to provide information voluntarily: (20)

You wouldn’t happen to know the time(, would you)?

A third interpretative aspect of declarative questions is that they can be used to express disbelief, surprise, or doubt regarding the information provided by the addressee. In these cases, the declarative question serves as a rhetorical device, with the speaker seeking confirmation of unexpected information: ¹ Although tags are often used in declarative questions, they are not obligatory; in what follows, I sometimes added those elements to signal the intended reading clearly.

60

(21)

NON-CANONICAL WORD ORDER

She’s your sister?

Last but not least, declarative questions can also function as a way to check the addressee’s understanding or agreement with the speaker’s statement. These questions often invite the addressee to confirm or elaborate on the speaker’s understanding of a given topic. (22)

So, the meeting is at 3 pm(, correct)?

In summary, declarative questions can serve various pragmatic functions, including seeking confirmation or clarification, expressing politeness or indirectness, conveying disbelief or surprise, and checking understanding or agreement. One way to go about their syntactic analysis would be to assume that the surface form of declarative questions resembles declarative statements, but their underlying syntactic representation involves certain covert syntactic categories that distinguish them from simple statements. In particular, one approach to analyzing declarative questions in generative syntax is to posit that they involve a covert question operator (e.g., Krifka 2015). Additionally, some approaches postulate that the special reading of declarative questions is represented in the syntactic structure as well, within the domain of speech act phrases that concern the role of the speaker and the hearer (e.g., De Clercq 2017). We will come back to this particular approach within recent speech act syntax in Chapter 10. In this context, I merely wanted to highlight that those approaches assume covert (syntactic) structure to account for the various interpretative properties of declarative questions, although, at the surface level, the word order remains unchanged. Another approach to analyzing declarative questions would be to focus solely on the role of prosodic and intonational features in conveying the interrogative nature of these constructions and to say that the syntax of declarative questions is exactly the same as that of declarative statements, with the crucial difference being the intonation pattern associated with the sentence (e.g., a rising intonation at the end of the sentence in English). In this view, the interrogative interpretation of declarative questions arises from the interaction between syntax and prosody only, rather than from covert syntactic operations. In other words, this approach does not postulate any dedicated structures and/or movement operations that hold for declarative questions. As I have also argued in the context of wh-in-situ questions, my approach to non-canonical questions is in line with the latter account and thus does not stipulate additional syntactic structure, unless there is evidence for any

4.4 NON-CANONICAL FRONTING IN QUESTIONS

61

dedicated syntactic operation that would support such stipulation. To see the difference where we cannot do without additional syntactic projections and movement, we now turn to non-canonical frontings in questions.

4.4 Non-canonical fronting in questions We have already illustrated a case of non-canonical fronting in Section 4.1 above. In general terms, we can say that non-canonical frontings refer to movement of constituents to the front of a sentence or clause for reasons of emphasis, contrast, or focus. As we have seen for German declaratives in Section 4.1, non-canonical frontings often serve very specific pragmatic functions. In questions, non-canonical frontings can result in constructions that deviate from the expected interrogative word order. Let us first consider some English examples of non-canonical frontings in questions. In English, cleft constructions instantiate one type of such non-canonical frontings in a question speech act: (23)

It’s John whom you saw at the store?

In this example, the cleft construction involves non-canonical fronting of the focused constituent John to express marked discourse interpretations such as contrastive focus. While examples like (23) can be analyzed analogously to declarative cleft constructions (e.g., Kiss 1999), an interesting phenomenon that requires a special syntax for non-canonical questions is non-canonical fronting of question particles in languages other than English. Let us consider German to illustrate such frontings (for the same type of fronting in Basque, see Chapter 6 below). In German, this phenomenon is unique, because it deviates from a robust generalization that holds for the syntactic position of particles in all other speech acts. In particular, the general observation is that German modal particles, unlike adverbs, remain in the middle field of the clause, and thus they do not occur in the left periphery. To see this, examine the following minimal pair involving the polysemous adverb/particle vielleicht (literally, ‘perhaps’): (24)

a Eliud ist vielleicht SCHNELL. Eliud is perhaps fast ‘Eliud is perhaps fast.’ b Vielleicht ist Eliud SCHNELL. ‘Perhaps, Eliud is fast.’

62

(25)

NON-CANONICAL WORD ORDER

a ELIUD ist vielleicht schnell! Eliud is PART fast ‘My god, Eliud is so fast!’ b ∗ Vielleicht ist ELIUD schnell! (intended: same as [25a])

When acting as an adverb, vielleicht can be fronted as seen in (24b); however, when functioning as a modal particle, this is not possible, as demonstrated in (25b). Even though positional rigidity is a highly dependable characteristic of modal particles, wh-questions with particles appear to be an exception. Take a look at (26a, particle in bold) and the alternative in (26b); see Thurmair (1989) and Abraham (1991) for early discussion of this phenomenon: (26)

a Wie hat er nur den Rekord how has he PART the record b Wie nur hat er den Rekord how PART has he the record ‘How on earth did he break the record?’

brechen break brechen break

ko¨nnen? could ko¨nnen? could

In (26b), the combination of wh+modal particle must form a constituent, since German, as a V2 language, typically does not allow more than one constituent to the left of the finite verb (SpecCP) in the main clause. The syntax of this type of non-canonical question has often been discussed in the literature (Bayer and Obenauer 2011; Bayer and Trotzke 2015; Meibauer 1991, 1994; Thurmair 1989). In Trotzke (2017a), I have argued that the (26b) variant conveys a higher degree of frustration, given the core semantics of can’t-find-the-value questions illustrated in Chapter 3 above. Accordingly, I have claimed that this construction demonstrates the interpretative effect of ‘emphasis for intensity’, which is also supported by a related production study that explores whether the construction, when compared to the nonfronting version in (26a), exhibits core phonetic characteristics of expressive and attitudinal evaluation (Trotzke and Turco 2015). I cannot go into the details of the proposal in Trotzke (2017a) here. I merely want to sketch the idea that this type of fronting of the [wh+Prt]-constituent indeed requires a dedicated movement type, because the interpretation is different from both the version with the particle in situ (26a) and a version with fronting of a regular wh-constituent that contains no particle. In technical terms, the regular derivation for questions needs to be expanded by incorporating an additional step that involves moving the [wh+Prt]element to the specifier of a functional projection that encodes this specific

4.4 NON-CANONICAL FRONTING IN QUESTIONS

63

type of emphasis (EmpP). As illustrated in Trotzke (2017a), the operator that encodes emphasis for intensity is situated below the illocutionary layer, because many phenomena support the idea that this effect can be present both in assertions (indicated in Section 4.1) and in non-assertive speech acts such as emphatic [wh+Prt]-questions. The final movement target for [wh+Prt] is ForceP, given that the wh-element within the [wh+Prt]-constituent possesses an active Q feature requiring valuation. Since the emphatic effect in [wh+Prt]questions can be analyzed as a general intensification strategy, highlighting the lowest or highest point on a scale, the interpretative result of this intensity effect in can’t-find-the-value questions is that the speaker emphasizes that it is extremely unlikely that the question will be answered—hence the sense of frustration. The relevant derivation can be summarized as follows: (27)

[ForceP wh+Prt [Force’ Force0 /Q [EmpP wh+Prt [Emp’ Emp0 [TP … wh+Prt…]]]]]

Crucially, this type of syntactic derivation can also be employed for an analysis of non-canonical frontings in polar questions. In this context, let me once again highlight an important aspect of the analysis suggested above—namely, that the operator responsible for the emphasis for intensity effect should be situated below illocutionary force, as indicated above for emphatic [wh+Prt]questions. Besides the evidence from [wh+Prt]-questions, I would like to point out at the end of this section that we observe similar effects of unexpectedness owing to marked word order in utterance types that should be classified as polar questions at the level of illocutionary force: (28)

a Du gehst morgen an die Uni? you go tomorrow to the university ‘Are you going to the university tomorrow?’ b An die UNI gehst du morgen? to the university go you tomorrow ‘Are you going to the university tomorrow? Really?’

Following a syntax–semantics account of parallel cases in Italian (Bianchi and Cruschina 2016), let us consider that yes–no questions represent a set made up of the proposition p and its opposite {p, ¬p}. It is important to note that the emphatic effect in (28b) is interpreted in relation to the focus alternatives of proposition p, rather than its opposite ¬p. Specifically, the fronted version in (28b) above implies that there are likelier alternatives in the form of ‘x goes to y tomorrow’. Since the polar set is created by a Polar Question operator that takes proposition p as input, it is necessary to posit that the Emphasis operator

64

NON-CANONICAL WORD ORDER

is situated below the Question operator. This is because the Emphasis operator must be at a compositional level where the opposing proposition ¬p is not yet present. Using Sicilian as the primary example, Bianchi and Cruschina (2016) furthermore point out in their work that this effect of emphasis can be either a mirative implicature or a double-checking interpretation. They also add that it remains uncertain how these interpretations, and potentially others, for focus fronting in declarative polar questions are distributed among various languages. In sum, we have seen in this section that some types of non-canonical questions indeed require an additional syntactic layer and a corresponding movement operation. My claim is that this structural requirement can be accounted for by postulating a functional projection of emphasis, which, importantly, is located at a compositional level below illocutionary force.

4.5 Summary In this chapter, we have looked at a variety of cases where we observe untypical word orders in questions. In particular, after a section on general aspects of the interaction between word order, prosody, and interpretation, I have sketched the two phenomena of optional wh-in-situ configurations and declarative questions. For both question types, I have argued that prosody— although interacting with the respective word orders—is the main cue for the non-canonical interpretation and that, therefore, we have no methodological reason to postulate any dedicated syntactic structures for the non-canonical effects in those two question types. I have then elaborated on a domain where I think the assumption of additional syntactic structure is inevitable: noncanonical frontings to the left periphery of questions. Specifically, I have indicated that, for frontings in both wh-questions and polar questions, we need an abstract encoding of an intensification strategy, which, based on previous work, I have referred to as ‘emphasis’. Crucially, for these phenomena I have claimed that the special interpretations of the respective questions cannot— syntactically speaking—be located at the level of illocutionary force of the clause. Given this general architecture of the clause, we will now look at some operator elements in non-canonical questions in the next chapter, and see where they might fit in, given our syntactic analysis above.

5 Non-canonical wh-elements After our discussion of the domain of non-canonical word order in questions, we will now focus on a set of syntactic wh-constituents that yield non-canonical interpretations. Specifically, we will have a look at so-called why-like-what elements, at extreme degree readings of wh-elements in utterances with interrogative syntax, and at cross-linguistic patterns of so-called how come questions in English, French, and Japanese. To be sure, the label ‘non-canonical wh-elements’ could potentially refer to a range of further elements such as wh-ever words (e.g., whoever, whatever, whenever, wherever); these words are known to feature a special syntax– semantics as well, because they can function as free relatives (1a) or indefinite pronouns (1b), conveying certain readings of generality, indifference, or inclusiveness: (1)

a Whoever wants to join the party can come. (Free relative) b I’ll do whatever it takes to succeed. (Indefinite pronoun)

However, in the following sections we are interested only in those whconstituents that are associated with question force, and our notion of noncanonical wh-elements is thus more restricted.

5.1 why-like-what In this section, we explore a unique category of non-standard questions that do not seek information but instead convey the speaker’s astonishment and/or disapproval concerning the event represented by the question’s propositional component. Accordingly, this category is another example of non-canonical questions where the expression of the speaker’s emotion is the central feature (see Section 3.4 above). Munaro and Obenauer (1999) were the first to examine this type of question in depth and from a cross-linguistic perspective. They discuss the general aspects of structures that combine interrogative syntax with a surprise– disapproval interpretation using examples such as:

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0005

66

NON-CANONICAL WH -ELEMENTS

(2)

a Cossa zìghe-tu?! [Pagotto; a sub-dialect of Bellunese] what shout-CL ‘Why are you shouting?!’ (Munaro and Obenauer 1999: 191) b Was lacht der denn so blo¨d?! [German] what laughs he PART so stupidly ‘Why is he laughing so stupidly?!’ (Munaro and Obenauer 1999: 238)

The key observation in these examples is that the wh-element ‘what’ (i.e., cossa in Pagotto; was in German) is employed not to indicate a syntactic argument (as in What is he buying? [He is buying a car]), but rather to convey a meaning akin to ‘why’. This is especially evident in instances like (2b). In this case, only a non-argumental interpretation of ‘what’ is viable owing to the intransitive verb lachen ‘to laugh’—an argumental interpretation of the wh-element (as in What is he buying? [He is buying a car]) would lead to ungrammaticality (∗ He is laughing x). This type of question construction is typically found in informal spoken language and is more colloquial in nature. The use of ‘what’ in such questions is not the standard way of asking for a reason in German and Pagotto, which would involve the respective question words for ‘why’. Importantly, when interrogatives display this why-like-what interpretation of the wh-element, Munaro and Obenauer (1999: 237–8) emphasize that it necessarily conveys an attitude from the speaker that spans from mild surprise to strong disapproval. In essence, these structures express a breach of the speaker’s expectations and can be classified as exclamations at the level of illocutionary force (i.e., expressive speech acts that communicate a violation of the speaker’s expectations). As the literature on exclamations highlights, various syntactic forms can convey them, interrogative syntax being just one of them. Consider these English examples from Rett (2011: 412): (3)

a b c d

(Wow,) John bakes delicious desserts! (My,) What delicious desserts John bakes! (Boy,) Does John bake delicious desserts! (My,) The delicious desserts John bakes!

Upon examining the data in (3), it is evident that exclamation as a speech act does not correspond to a specific sentence type. Instead, the concept of exclamation may refer to particular uses of several sentence types. In this view, the interrogative examples in (2) could simply represent another way to express

5.1 WHY -LIKE-WHAT

67

an exclamation, suggesting that these cases no longer function as proper (i.e., information-seeking) questions, asking for a reason in this case. However, Munaro and Obenauer (1999), in their original proposal, took a different approach (as did the syntactic literature that adopted their proposal; see Giorgi 2016, 2018; Obenauer 2004, 2006; and many more). According to Munaro and Obenauer (1999), surprise–disapproval questions should be distinguished from both information-seeking questions and exclamations at the speech act level (after all, utterances such as (2) do still have question force). They articulate this idea within a cartographic approach to syntax, which represents discourse-related meaning components as left-peripheral syntactic projections (see Rizzi 1997, 2014; and our proposal in Chapter 4 above). A central tenet of this approach is to incorporate illocutionary components of utterance meaning within sentence grammar (see also Krifka 2014, 2023). Adopting this perspective on syntax and pragmatics, Munaro and Obenauer (1999) suggest that the surprise–disapproval effect in cases like (2) is encoded in a functional projection separate from interrogative force. Munaro and Obenauer’s decomposed left periphery (1999: 217–18), as seen in (4), presents their ordering of functional projections; see also Obenauer (2006: 266):

(4)

ExclCP

Spec

ExclCˈ

ExclC

EvCP

Spec

EvCˈ

EvC

InterrForceP ( ... )

As depicted in (4), there are two additional functional projections above the projection encoding the interrogative force of the clause. Specifically, SpecExcl(amation)CP serves as the landing site for wh-phrases in whexclamatives (e.g., How fast he is!), while SpecEv(aluation)CP acts as a position for wh-phrases in non-canonical questions like (2) above. Consequently,

68

NON-CANONICAL WH -ELEMENTS

Munaro and Obenauer argue that special questions, such as surprise questions (as well as rhetorical questions and others), require incorporating functional structures on top of the structures encoding the question force of a clause (see also Nye 2009: ch. 2 on relevant pseudo-questions in English and a similar proposal). We will come back to why-like-what questions in Chapter 7 below when we look at their pragmatics in more detail. For now, and in the context of this syntactic analysis, I would like to point out that the structure in (4) deviates from my cartographic proposal for non-canonical frontings in questions in Chapter 4. In particular, I have argued that emphatic components of questions (such as expressing surprise in polar questions) can be accounted for by postulating a functional projection below illocutionary force. Munaro and Obenauer (1999), and others adopting their approach, instead claim that such a projection should be on top of force. I agree with them that this non-canonical use of ‘what’ yields an additional interpretation that might be accounted for in terms of surprise and disapproval on the part of the speaker. However, my central hypothesis is that this meaning component is something we can find across illocutionary forces, very similar to information-structural interpretations. With this in mind, let us now turn to the next empirical domain and look at special degree readings of wh-elements in questions.

5.2 Degree readings of wh-elements in questions Put in general terms, degree readings of wh-elements in questions refer to a type of question where a wh-word is used to ask about the extent, amount, or intensity of a property, state, or action. These questions often involve wh-words such as how in English, which may be combined with adjectives or adverbs to inquire about the degree to which a certain property, state, or action is realized. In particular, in a question such as How tall is he? how is combined with the adjective tall to ask about the height of a person. The answer is expected to be an expression of degree (e.g., He is 6 feet tall). In How quickly can you finish this task?, on the other hand, how is combined with the adverb quickly to inquire about the speed at which someone can complete a task. The answer might be an expression of degree or duration (e.g., I can finish it in thirty minutes). While those uses of the wh-element how are canonical, we also observe uses that are non-canonical in the context of questions, and are more similar

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

69

to using how in exclamation speech acts. The utterance How cool is that? is an example of an exclamatory question or an interrogative exclamation in English. It is a special question, because it combines the degree question word how with the adjective cool to express amazement, admiration, or excitement about something. The speaker does not expect a literal answer to this question. Instead, the utterance serves to emphasize the speaker’s positive reaction to the situation, object, or event being discussed. In those questions, the word how functions as an intensifier that emphasizes the speaker’s emotional reaction to a situation, object, or event. It is used to convey a strong degree of the quality expressed by the adjective that follows—in this case, cool. In other words, the meaning of how in this type of utterance is different from its use in regular degree questions. In a degree question, such as How tall is he?, how is used to inquire about the degree or extent of a certain property (e.g., height; see above). In contrast, in an exclamatory question like How cool is that?, how serves to intensify the speaker’s emotional response to the situation, and one could have the impression that the question is rhetorical, not seeking a literal answer. We can thus conclude that how is not the regular degree question word in those questions, and I would now like to look at its syntax–semantics interface in more detail. Intuitively, as already indicated above, when speakers say How cool is that?, they convey that something has surpassed a certain level of coolness they had anticipated. In other words, these structures express the speaker’s surprise because of a breach of their expectations. Similarly to what we have discussed in the previous section about why-like-what, we could thus say that How cool is that? can be considered an exclamation at the level of illocutionary force, which is an expressive speech act that communicates that a particular state of affairs has violated the speaker’s expectations. As the literature on exclamations demonstrates, they can be expressed through various syntactic forms (see examples in (3) above). One could thus hypothesize that exclamation as a speech act corresponds not to a specific sentence type, but rather to particular uses of several sentence types. Given the idea of derivational approaches to categorize sentence types according to their (morpho)syntactic features (see Chapter 1), we could conclude that there is no dedicated sentence type for exclamations; potentially, any sentence type can receive an exclamatory interpretation owing to nonsyntactic features like intonation, which are typically not relevant for defining sentence types in most derivational approaches.

70

NON-CANONICAL WH -ELEMENTS

On the other hand, we have already seen in Chapter 1 that influential accounts propose the existence of a minor sentence type called ‘exclamative’. For example, exclamatives can have a distinct syntactic signature, as in What a delicious dessert John bakes!, where we observe the special configuration what a and verb-final word order. These elements can be considered markers of exclamatives, and their distribution can be easily accounted for by postulating the category of exclamatives (e.g., Grimshaw 1979). Our main example raises intriguing questions regarding these issues, as it functions as an exclamation with the syntactic form of a whinterrogative, adding another option to the inventory presented in (3). In previous work, I have elaborated on the claim that this additional option can be characterized—from a pragmatic perspective—as a unique utterance type, namely ‘addressee-oriented exclamations’ (A-EXCs); see Trotzke (2020a). In the context of our syntactic discussion in the present chapter, let us now first discuss the theoretical possibility of postulating structural differences in the clause’s left periphery that might distinguish A-EXCs from their information-seeking wh-counterparts. To evaluate this possibility, we have to sketch some basic syntactic concepts in the literature on exclamations. Based on the English inventory of exclamation speech acts in (3) above, let us assume, for the sake of argument, that at least forms like (3b)—What delicious desserts John bakes!—can be categorized as a separate sentence type because they cannot function as regular wh-questions owing to their verb-final configuration. This is in accordance with most of the literature on English exclamatives (see Siemund 2015). In particular, one could say that wh-exclamatives have a dedicated syntactic structure in English and can clearly be distinguished from wh-interrogatives (examples and judgments from Repp 2013: 66): (5)

a b c d

[CP How long [TP this train journey could take]]! [CP How long [C’ could [TP this train journey could take]]]? ∗ [CP How long [TP this train journey could take]]? ∗ [CP How long [C’ could [TP this train journey could take]]]!

As far as the English language is concerned, our key example How cool is that? clearly features interrogative rather than exclamative word order. In terms of word order, it cannot count as an exclamative and can be used for both conveying a request for information (6b) and performing the speech

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

71

act of an exclamation (6a); in what follows, ‘#’ indicates pragmatic deviance in a dialogue: (6)

a A: How cool is that! [intended: exclamation] B: #I don’t know. b A: How cool is that (e.g., on a scale [intended: question] from 1 to 10)? B: To my mind, it’s very cool (I’d rate it a 9).

Both (6a) and (6b) have the syntactic form of an interrogative, but only (6b) is interpreted as a question speech act. Example (6a), by contrast, is interpreted as an exclamation. The utterance expresses that the degree of coolness is not in accordance with the speaker’s expectations, and the speaker conveys that the propositional content (that a salient entity in the ongoing discourse is very cool) is true. This propositional part of the utterance is the reason why Speaker B’s response is odd because, in the response given in (6a), Speaker B addresses an issue that is already backgrounded in the ongoing discourse. For many, this ‘backgrounding’ is a signature property of exclamatives and can technically be characterized as a factivity presupposition (see below). In sum, there is a clear difference in meaning between the two readings paraphrased in (6). However, there is no difference in their overt syntactic structure (i.e., word order). One theoretical possibility to distinguish between the two readings (question versus exclamation) on syntactic grounds is to refer to ‘covert’ syntactic structure that has been claimed for the left periphery of exclamatives. One prominent approach in this regard is the theory by Zanuttini and Portner (2003), who refer to the backgrounding of descriptive content already mentioned above in their syntactic proposal. More specifically, Zanuttini and Portner (2003), among others, have argued that exclamatives have a factive operator that interrogatives like (6b) lack. In their work, they focus on Paduan, a Romance variety spoken in the Italian city of Padua. They argue that this variety provides direct evidence that exclamatives contain an extra CP layer of structure. Without going into too much detail here, let me briefly illustrate their analysis for a wh-exclamative like (7a); see Zanuttini and Portner (2003: 64): (7)

a

Che alto che l what tall that SUBJECT.CLITIC ‘How tall he is!’

ze! is

NON-CANONICAL WH -ELEMENTS

72

b

CP

che alto

Cˈ

Ø

CP

FACT

Cˈ

che

IP

The primary takeaway from the data and analysis in (7) is that a pair of CP specifier positions is required to create a phrase-structural position for both the factive operator (also necessary for embedded contexts initiated by che ‘that’) and the wh-phrase che alto (‘what tall’). Zanuttini and Portner (2003) argue that, owing to evidence indicating that exclamatives consistently possess a factivity presupposition, their analysis should be proposed for other languages as well. According to their proposal, exclamatives are the sole unembedded structures exhibiting a factivity presupposition, which allows the factive operator to handle clause typing without the need for additional projections like ForceP or more specific ones such as ‘ExclamativeForceP’. We have already illustrated the theoretical alternative of Split-Force approaches to exclamatives and other structures expressing speaker surprise when we discussed Munaro and Obenauer’s account (1999) in Section 5.1 above. Although various approaches dispute the extent of functional structure required to explain the syntax of exclamatives, many theories follow Zanuttini and Portner’s general account (2003) and share the core assumption that covert syntactic structure must be postulated in languages like English to differentiate syntactically between interrogatives and exclamatives. Moving forward, I will present an alternative approach in this chapter and briefly outline its fundamental syntactic ideas. As a starting point, let us examine the exclamative version of our main example How cool is that! in (8): (8)

How cool that is!

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

73

I argue that the overt syntactic structure (without V-to-C movement) does not function as a syntactic feature to differentiate exclamatives from other sentence types. Despite the presence of a wh-phrase, the word order in the whexclamative in (8) actually reflects the syntax of a declarative, with no subject inversion (as noted by Ginzburg and Sag 2000: 224 regarding this parallel):

(8ˈ)

CP

[how cool]

Cˈ

Ø

IP

that

Iˈ

is

(...)

Thus, examining only the overt structure (word order), we observe that exclamatives possess the syntax of declarative assertions. The only remaining structural distinction in this area is based on covert structure, specifically the presence of a factive operator in a Split-CP approach, as demonstrated in (7b) earlier. However, what if the evidence supporting such a factive operator is diminished? Discussing all the semantic and pragmatic nuances found in the literature on exclamatives that underpin the notion that exclamatives do not exhibit assertive force but convey their descriptive content through factivity presuppositions would be beyond the scope of this chapter. Nonetheless, two frequently cited facts (originally presented by Grimshaw 1979) support this idea. First, exclamatives cannot be employed as responses to informationseeking questions; second, the descriptive content of exclamatives cannot be refuted. Let me provide a brief critical assessment of these two aspects, beginning with the latter one. The fact that the descriptive content of exclamatives cannot be denied has consistently been based on introspective judgments like those presented in (9) and (10), taken from Rett (2011: 414):

74

NON-CANONICAL WH -ELEMENTS

(9)

A: B:

(Wow,) John bakes delicious desserts! No (he doesn’t), these are store-bought. John’s actually a terrible cook.

(10)

A: (My,) What delicious desserts John bakes! B: ?No (he doesn’t), these are store-bought. John’s actually a terrible cook.

In the literature, it has been argued that the descriptive content of exclamatives (10) cannot be denied, as opposed to declarative assertions (9). However, recent experimental work on Spanish (Villalba 2017b) and German exclamatives (Trotzke 2019) contradicts this claim, demonstrating that the descriptive content of wh-exclamatives is indeed ‘at issue’ and can be denied. One can consult these papers for evidence that judgments like the one in (10B) do not hold up under empirical investigation, which assesses the acceptability of these patterns. Very briefly, the acceptability study (n = 112) by Trotzke (2019) demonstrates that there is no difference between wh-exclamatives and other forms of exclamations (notably declarative exclamations) when testing the felicity of different rejection strategies. More specifically, participants in this study had to rate the acceptability of Speaker B’s denials on a scale ranging from 1 (= very bad) to 6 (= very good). Crucially, all judgments of exclamation items were at ceiling (ranging from 5.2 to 5.7) and thus in accordance with filler items presenting perfect mini-dialogues (e.g., wh-question + corresponding declarative response); see Trotzke (2019) for full set of German items and detailed statistics.¹ Table 5.1 summarizes some of the results relevant in the context of this chapter. Table 5.1 Summary of some ratings from Trotzke (2019) strong denial (e.g., No!… German: Nein!…)

weak denial (e.g., Not really,… That’s not quite true…; German: Nicht wirklich, … Das stimmt nicht ganz …)

5.55 5.28

5.70 5.46

Exclamation type declarative wh-exclamative

As Table 5.1 indicates, both utterance forms allow rejection, and prefer the subtle strategy labeled as weak denial (e.g., not really etc.). This indicates that ¹ In a nutshell, a two-way ANOVA revealed significant main effects of both EXCLAMATION TYPE (p < .001) and DENIAL STRATEGY (p < .001), but there was no significant interaction (p > .05); see Trotzke (2019: 530–1) for further details.

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

75

the descriptive content is indeed in a way backgrounded in exclamatives, but this backgrounding is the same in declarative exclamations signaled by the overall preference for subtle rejection in this study. Let me now briefly discuss the second claim mentioned above—namely, that exclamatives cannot be used as responses to questions. According to this line of reasoning, while declarative assertions can serve as answers to informationseeking questions (11B), exclamatives are infelicitous, even when expressing the same descriptive content as their declarative counterparts (11B′). It is thus concluded that the descriptive content of exclamatives is a factivity presupposition and not asserted, and the pattern in (11) aligns with the broader observation that a question cannot be answered by a sentence that presupposes the answer (12). The original data and judgments of Grimshaw (1979: 321) to support this general claim are presented below: (11)

A: How tall is John? (12) A: Did John leave? B: John is very tall. B: #It’s odd that he did. B′: #How tall John is! B′: #I’d forgotten that he did.

My counterargument to the above reasoning can be summarized as follows. To comprehend what is happening in (11), I propose that the following example (13B) is a felicitous discourse move because it aligns with the constituent question in (13A). Conversely, (13B′) generates a mismatch at the information structure level. (13)

A: How fast was Eliud Kipchoge? B: Eliud Kipchoge was [very]F fast. B′: #[How fast Eliud Kipchoge was!]F

While (13B) serves as an ideal response to a narrow-focus question (13A), (13B′) is pragmatically not ok, as noted in the previously mentioned literature. However, when the information-structural context changes by using the two different utterance forms (i.e., exclamative versus declarative) as answers to a broad-focus question, the exclamative (14B′) improves significantly: (14)

A: Tell me, how did Eliud Kipchoge do in the race? B: My god! [He was very fast!]F B′: My god! [How fast he was!]F

I argue that (14B) is also acceptable in this context because it no longer expresses narrow focus, as in (13B), but, owing to different intonation, can be interpreted as an all-focus declarative. If something along these lines holds true, then exclamatives can be used as responses to information-seeking questions and exhibit assertive force.

76

NON-CANONICAL WH -ELEMENTS

Therefore, a covert factive operator is a theoretically unnecessary assumption, and exclamatives in English might be better analyzed as configurations featuring declarative word order (lacking subject inversion) and conveying an all-focus interpretation. From this viewpoint, exclamatives belong in the same category as other declarative configurations expressing surprise on the part of the speaker. Recent syntactic literature has analyzed these configurations as ‘mirative focus’ (Authier and Haegeman 2019; Cruschina 2012, 2019) or cases of ‘emphasis for intensity’ (Beltrama and Trotzke 2019; Trotzke 2017a), and some recent experimental work indicates that their interpretation can be distinguished from other information-structural readings (Destruel et al. 2019). Regarding the syntactic derivation, this implies that the [wh+Adj]constituent in wh-exclamatives targets the left periphery to express the speaker’s surprise, just as other phrases can be moved there to instantiate mirative fronting in declaratives (see literature cited above and our discussion of syntactic movement to the left periphery in Section 4.1 above). Given this situation, let us adopt the syntactic implementation that we have already introduced in the context of non-canonical frontings in questions in Section 4.4. Again, the respective operator Emp is hosted in a functional projection below the question/assertion operators that, according to cartographic syntax (e.g., Rizzi 2014), are situated in the domain of ForceP. Emp takes scope over the propositional core of the clause (FinP) in terms of all-sentence focus, as mentioned earlier:

(15)

ForceP

Force0

EmpP

Emp0

FinP

Fin0

(...)

When we apply the approach in (15) to exclamatives, then this analysis differs from other recent accounts that analyze the exclamation component as a modifier of illocutionary forces (e.g., Taniguchi 2017: 141–51). Although the

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

77

approach presented here shares the basic assumption with Taniguchi (2017) that exclamatives do not possess a force of their own, it conceptualizes the exclamation component along the lines of so-called mirative focus and thus locates the Emphasis operator in the information-structural layer of the clause. We will talk more about the parallels between expressive readings and information structure in Chapter 10. At this point, I would merely like to point out that, according to this approach, exclamatives are not unique and are just another type of emphatic assertions, similar to declaratives with exclamative intonation, lexical elements that mark unlikelihood like English even, or instances of mirative focus fronting. To summarize, up to this point we have seen some evidence demonstrating that there are valid reasons to question the postulation of covert syntactic structure to differentiate between questions (16), on the one hand, and exclamations, either exclamatives (17b) or non-canonical questions (17a), on the other hand. Specifically, both recent empirical work and recent theoretical proposals provide alternative ways to consider expressive configurations like (17). (16)

How cool is that?

[question]

(17)

a How cool is that! [exclamation] b How cool that is! [exclamation]

As a result, the onus is on those who suggest that there is covert structure enabling the grouping of (17a) and (17b), which exhibit different overt structure (word order), into a single syntactic category. I argue instead that the exclamation component of (17a) cannot be accounted for by a dedicated leftperipheral syntax. The next intriguing question might be how the inversion syntax (17a) contributes to the overall reading of configurations like How cool is that! But before delving into this question in greater detail, let us first summarize the special nature of the wh-constituent how in those constructions and what it tells us about the internal syntactic structure of the wh-phrase. Look at the following pair again, repeated here for convenience: (18)

a A: How cool is that! [intended: exclamation] B: #I don’t know. b A: How cool is that (e.g., on a scale [intended: question] from 1 to 10)? B: To my mind, it’s very cool (I’d rate it a 9).

78

NON-CANONICAL WH -ELEMENTS

An interrogative such as (18b) assumes that something possesses a certain level of coolness. The corresponding exclamation (18a) shares this assumption. However, (18a) conveys that the value of the undefined variable lies at a high point on a given scale, while the interrogative indicates that the addressee should provide the value. Despite a conceptual functional similarity, we observe that the semantic role of the lexical item how within utterances like (18a) differs from its role in regular wh-interrogatives; specifically, in (18a) it corresponds to ‘so’, the typical extreme degree element. This degree use of whmorphology is a realization of an abstract extreme degree morpheme that one could call SO-degree. For instance, That is so cool! has precisely this intense degree interpretation, and also lacks an apparent degree modifier such as very, extremely, and the like (for a detailed discussion of such extreme degrees, see Morzycki 2012). Accordingly, the meaning of the wh-phrase in non-canonical or ‘exclamatory’ questions such as (18a) can be summarized as follows: (19)

⟦SO Adjective⟧ = the property of the Adjective defined to an extreme degree by the speaker.

This is a widely recognized characteristic of exclamations more generally: they can include extreme degree uses of specific wh-elements. To illustrate, let us look at an approach by Krifka (2011) concerning question speech acts and presume that, while the degree of the relevant property is perceived as extraordinary in exclamations, it remains unspecified in interrogatives. Thus, akin to ordinary questions, utterances like How cool is that! represent a set of alternatives (specified in various ways, such as by a function or a proposition set; see Chapter 2). Moreover, Krifka (2011) posits that these alternatives are ordered along a degree scale. This order can be expressed as follows in a proposition set analysis (see Krifka 2011: 1775): (20)

{⟨λi[COOLi (THAT ‘X’)(d)], λi[COOLi (THAT ‘X’)(d′)]⟩ | d < d′)}

Phrases like how cool in How cool is that! are, therefore, similar to exclamatives in that they are constructed over a variable that spans various degrees, such as ‘coolness’ in this instance. At times, the degree usage of wh-elements formalized in (20) is instantiated by dedicated morphosyntactic means. For example, what serves as the head of a noun phrase (e.g., What did you read?) or as a determiner (e.g., What book did you read?) in interrogatives. However, in exclamations, what can act as a predeterminer (21a) and cannot appear in interrogatives in this form (21b); see Elliott (1974: 235).

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

(21)

79

a What a fast car he is driving! b ∗ What a fast car does John drive?

This distinction in terms of phrase structure is also applicable to other languages. It has been suggested that the degree reading of ‘what’ and ‘how’ possesses head status in exclamatives. Brandner (2010) presents an argument in this vein. Regarding German instances that could be interpreted as either an interrogative or an exclamation (22a), she contends that wie (‘how’) functions as a head element in the exclamation interpretation, behaving similarly to was (‘what’) in (22b); examples from Brandner (2010: 108): (22)

a Wie bist du groß how are you big b Was bist du groß what are you big ‘How tall you became!’

geworden!/? become geworden!/∗ ? become

Brandner (2010) provides a wealth of empirical evidence supporting the notion that both was and wie correspond to the degree element ‘so’ and that these elements adjoin to the finite verb in the C-position in a clitic-like manner: (23)

[CP [C ′ was/wie bist [IP du groß geworden]]] what/how are you big become ‘How tall you became!’

The question in our context now is whether the non-degree and degree use of wh-elements (and the related phrase-structural differences discussed in the literature) help us distinguish between the question and the exclamation reading of How cool is that? Focusing solely on the phrase-structural status, I argue that these differences cannot account for the ambiguity of the utterance How cool is that? Specifically, it is evident that how + adjective sequences (due to how’s modifying function) possess a clear phrasal status and, as a result, align with the general phrasal makeup observed in a typical wh-question. This means that the wh-element, along with the adjective, occupies the Spec-CP- and not the C-position in such questions: (24)

[CP [AP how cool] [C’ is [IP that]]]

Upon examining this level, we can therefore say that wh-questions (18b) and exclamations like How cool is that (18a) do not exhibit structural differences: both display T-to-C movement, and Spec-CP is occupied by a phrasal constituent containing a wh-element.

80

NON-CANONICAL WH -ELEMENTS

Naturally, there are structural distinctions when examining the whconstituent in more detail—differences that have been extensively studied in the literature (including some of the works cited above). As mentioned earlier, the wh-element in the exclamation reading of How cool is that! possesses a degree reading that corresponds to the degree marker ‘so’ and can be formalized as follows (‘≫ θ’ = ‘going beyond a threshold’); for a detailed discussion of the relevant degree semantics, see Kennedy and McNally (2005) among others: (25)

⟦HOW COOL⟧ = λx.cool(x) >> θcool

It is intriguing to observe that the [wh+Adj]-constituent with the reading in (25) displays a root/non-root asymmetry, which has also been noted in the licensing of other emphatic wh-constituents often used in surprise contexts (see our discussion of wh-the-hell and similar configurations in Chapter 3). Specifically, while wh-in-situ in the degree reading outlined in (25) is not possible at all (26), it is perfectly fine in the question reading (27); examples and judgments from Nye (2009: 19): (26)

a How cool is that! b ∗ That’s how cool!

(27)

a How cool is that? b That’s how cool?

We can therefore say that [wh+Adj] in its degree reading must access the clausal left periphery. To analyze this situation, we can again apply the approach already illustrated above. The wh-constituent in (25) carries a rootsensitive feature of emphasis and thus must be preposed to the relevant left-peripheral projection. Importantly, this projection—as I argue throughout the book—cannot be at the level of illocutionary force. As previously mentioned, potentially any clause type can feature exclamatory interpretations, so the relevant phrasal constituents are moved to the specifier of EmpP before further movement to Force is required for clause-typing reasons. The difference between (26) and (27) is that (27) lacks the extra movement to EmpP, and the [wh+Adj]-constituent in that case simply moves to the left periphery to form a (non-echoic) question. In summary, the differing interpretations (and thus the different phrasal compositions within the wh-phrases) can be attributed to the left-peripheral options that are not shared between exclamatory questions and whexclamatives, on the one hand, and interrogatives, on the other hand.

5.2 DEGREE RE ADINGS OF WH -ELEMENTS IN QUESTIONS

81

Instead, these options are shared between any emphatic-fronting configuration (including declaratives) and any non-emphatic speech act interpretations (lacking the intensifying effect of scalar readings), on the other hand. Overall, the differences in the wh-constituents in the question and exclamation readings of strings like How cool is that can be accounted for by referring to a general strategy of emphasis in the left periphery, rather than referring to a distinct feature of wh-exclamatives shared with exclamatory questions. This account can also be applied to counterparts of English How cool is that in other languages, as shown by Trotzke (2020a) for analogous German constructions. Based on these assumptions and observations regarding the exclamation component of those non-canonical questions, we can represent the derivation of questions containing the extreme SO-degree in its wh-constituents as follows:

(28) [how+cool] Force0/is [how+cool] Emp0 [TP... that is [how+cool]...]

To summarize, the analysis in (28) essentially represents the claim that utterances like How cool is that are interrogatives that contain an additional exclamation component. This exclamation component can be accounted for in a similar way to other cases where phrasal constituents receive an intensity reading by being moved to the specifier of EmpP before they must move further for reasons related to clausal typing. Again, it is important to note that EmpP is a syntactic layer that is distinct from the function of clausal typing and/or illocutionary force. Given this general approach, I now turn to the next empirical domain in the context of non-canonical wh-constituents in questions: so-called how come questions.

82

NON-CANONICAL WH -ELEMENTS

5.3 How come questions How come questions are a type of interrogative structure in English used to inquire about the reason or cause behind a situation, event, or action. They are very similar in meaning to why-questions, but the interpretation may be slightly different. The phrase how come is derived from the longer expression how comes it that or how does it come that. Over time, the phrase was shortened and became more idiomatic, leading to the modern usage of how come as a stand-alone interrogative structure (see Kim and Kim 2017 on the diachronic development). Here are some examples of how come questions in English: (29)

a How come you didn’t call me last night? b How come she always gets the best assignments? c How come the store is closed today?

In each of these examples, how come introduces a question about the reason or cause behind a specific situation, event, or action. The syntax of how come questions is similar to that of declarative sentences, with SVO order and the [how come]-constituent preceding the subject. Importantly, the meaning of how come questions is essentially the same as why-questions, but how come questions imply a sense of surprise, disbelief, or curiosity about the situation being questioned. This extra layer of meaning can also make how come questions sound more informal and conversational than their why-counterparts. In what follows, I will focus on recent syntactic work on those questions by Endo (2018). In this work, he discusses the problem concerning the syntactic position of how come and its relationship with the complementizer that, and he also demonstrates the varying usage of how come and the complementizer that among speakers. Endo (2018) starts his discussion with Zwicky and Zwicky’s observation (1973) that how come can be interpreted only in conjunction with the main clause, as shown in example (30); see Endo (2018: 304): (30)

How come you say that John is mad? [interpretation acceptable in the main clause, ∗ in the embedded clause]

This pattern could be explained by assuming that how come is generated at the base level within the C head of the main clause (Collins 1991). As a result, how come cannot appear in the embedded clause or undergo long-distance movement into the main clause, since head movement is typically restricted to individual clauses. Endo (2018) refers to Shlonsky and Soare (2011) when

5.3 HOW COME QUESTIONS

83

he identifies a potential issue with this analysis of how come as a C head of the matrix clause: He notes that how come behaves like a phrasal element such as why, rather than a head element like if or whether, which do not participate in sluicing; see Endo (2018: 304): (31)

They thought John left early, but they didn’t tell me why/how come/∗ whether/∗ if φ.

Based on this observation, Shlonsky and Soare (2011) propose that how come is generated at the base level in a specifier position. More specifically, they apply Rizzi’s notion (2001) that the Italian equivalent of how come, come mai, is generated at the base level in the specifier of Int(errogative)P within the CP system in (32):

(32)

ForceP

Ø

Forceˈ

that

[how come]

IntP

Intˈ (...)

Since Rizzi (1997), it has been a common assumption that the complementizer that occupies the head position of ForceP. As seen in (32), ForceP—in this type of approach—is situated above IntP, leading to the incorrect expectation that the complementizer that does not follow how come, which is found in the specifier of IntP. Endo (2018: 305–7) addresses this derivation problem by pointing out in this context that, while many speakers permit how come to be followed by the complementizer that, a small set of English speakers (n = 20) he consulted do not accept how come directly followed by that. To my mind, this line of argumentation is rather weak, and this is why I would like to suggest a different account—adopting the clausal architecture already suggested for some of the phenomena in the previous sections. In particular, we can analyze how come questions as another type of emphatic wh-questions. According to the analysis in (33), [how come] is basegenerated as an adverbial constituent inside the VP, but carries features of both

NON-CANONICAL WH -ELEMENTS

84

emphasis and clausal typing (i.e., a Q feature). Therefore, it must move to the specifier of EmpP, and then moves further to ForceP:

(33) [how come] Force0/(that) [how come] Emp0 [TP... you are tired [how come]... ]

From a cross-linguistic perspective, there is nothing special about the optional that in utterances like how come questions. Look at the following case of a declarative involving mirative focus fronting in French (Authier and Haegeman 2019: 7–9): (34)

Trois heures (qu’)elle a passé chez son avocat. three hours (that)-she had spent at her lawyer ‘Three hours she spent at her lawyer’s office.’

Here too, the complementizer ‘that’ is optionally added, and this strategy is known to occur in exclamatory contexts in Romance more generally (see Trotzke and Villalba 2021). Talking about Romance and French, let me point out that there have been some recent works discussing equivalent structures to how come questions in French that serve a similar function of inquiring about the reason or cause, while also conveying a sense of surprise, disbelief, or curiosity. For example, Brunetti et al. (2021) provide a detailed discussion of those constructions in French featuring the element comment (‘how’). Comment questions typically allow for both manner and reason interpretations. However, they can contain functional and lexical elements that reduce potential ambiguity and make the reason interpretation more strongly favored. Based on Sadock’s list (1974) of several factors that encourage a rhetorical interpretation of a question, Desmets and Gautier (2009) were the first to have a look at those relevant features in the context of reason-comment questions. In particular, a key characteristic of reason-comment questions is that the verb is often accompanied by a modal, such as pouvoir ‘can’, vouloir ‘want’,

5.3 HOW COME QUESTIONS

85

or a semi-modal like oser ‘dare’. However, Tovena’s more recent corpus study (2022) reveals that reason-comment questions easily combine with infinitival questions that lack an explicit modal, as demonstrated in the following corpus example (Tovena 2022): (35)

Comment refuser une telle proposition? how to.refuse a such proposition ‘How to refuse such a proposition?’

Accordingly, we can conclude that potential licensing conditions in terms of co-occurrence with specific modals in French are rather weak, if existing at all. As for the interpretation, Brunetti et al. (2021) summarize that the purpose of reason-comment questions seems to be to emphasize the speaker’s disbelief regarding the truth of the prejacent. Accordingly, this disbelief is what motivates the speaker to ask a reason-comment question initially. Notably, Brunetti et al. also point out that the prejacent is not presupposed. That is, reasoncomment questions are similar to regular yes–no questions in this regard, as they inquire about the truth of the proposition expressed in the interrogative without presupposing it. When posing reason-comment questions, the speaker, although doubtful of the prejacent’s truth, does not ask about it directly. Instead, the speaker requests the addressee’s assistance in making the prejacent plausible in accordance with their epistemic state. A syntactic factor distinguishing reason-comment questions from comment questions with canonical manner readings is the absence of subject–aux inversion, similar to what we have observed for English above. Brunetti et al. point out that subject–aux inversion is technically required in formal written French questions, but it has become optional in contemporary French. Other interrogative forms without inversion exist in the language and are more prevalent in everyday speech. It is therefore challenging to determine if the limited number of subject–aux inversions in reason-comment questions is due to the specific reading or to a matter of register (see, however, Tovena 2022 for critical discussion). A final issue at the syntax–pragmatics interface of how come questions is whether those questions can be interpreted as rhetorical questions. Brunetti et al. (2021) point out in this context that rhetorical questions are often thought to be associated with no degree of uncertainty regarding the answer on the part of the speaker (see our discussion in Chapter 3 above; and, e.g., Biezma and Rawlins 2017; Caponigro and Sprouse 2007). Under this view, the speaker

86

NON-CANONICAL WH -ELEMENTS

assumes that the hearer already knows what is being stated, and the proposition corresponding to the implicit answer is considered to be within the common ground. For French and the reason-comment questions presented above, Brunetti et al. (2021) claim that the speaker’s epistemic state and attitude are entirely different. That is, the implicit answer is not assumed to be within the Common Ground. In particular, the speaker does not request explicit acceptance from the hearer of the implicit answer, but rather asks the hearer to assist in revising her expectations and eventually accepting the prejacent’s possibility. The addressee needs to provide explanations or justifications, which will resolve the apparent contradiction with the speaker’s expectations. Brunetti et al. thus conclude that, while some reason-comment questions with specific morphosyntactic or semantic properties may be interpreted as truly rhetorical, most are not, and we can therefore say that how come questions more generally are speech acts with regular directive question force. The rhetorical usage cannot be considered the basic use of these questions. In sum, I would like to claim that there are no empirical data contradicting a simplified analysis along the lines of (33) above. Specifically, this analysis both captures the surprise–disbelief component and maintains that a complex wh-constituent moves to ForceP to make sure that the whole utterance is nevertheless interpreted as a question speech act. In both English and French (and further languages, see Endo 2018, 2022), the configuration of how come questions allow speakers to inquire about the reason or cause behind a situation, event, or action, while also conveying a sense of surprise, disbelief, or curiosity. Crucially, those questions too can be analyzed in my framework of non-illocutionary emphasis, where the functional hierarchy of the clause includes one single projection for intensification and scalar readings that feature propositional scope, but are not affecting the respective illocutionary force of an utterance.

5.4 Summary In this chapter, I have discussed a set of non-canonical wh-constituents that yield non-canonical interpretations in questions. Specifically, I have first illustrated why-like-what elements and demonstrated that they differ from regular why-questions. After that, I have explored extreme degree readings of whelements in utterances with interrogative syntax such as How cool is that? In this context, I have also proposed a syntactic analysis that can do without

5.4 SUMMARY

87

extra functional projections at or even above the level of illocutionary force. In particular, I have argued for an approach where exclamatory readings can be found across speech acts, analogous to other discourse-relevant interpretations (e.g., information structure) that can be found below illocutionary force in the functional hierarchy of the clause. Finally, I have presented some cross-linguistic data on so-called how come questions and proposed that, in this domain too, we do not need to assume any additional functional structure to explain the patterns observed in the literature. In the next chapter, we now turn to the final empirical domain of syntactic categories that can yield non-canonical readings in questions.

6 Non-canonicity via modal particles This final chapter of Part II of this book deals with the expression of noncanonical question interpretations via the syntactic category of modal particles. After an introduction to the interactions between modal particles and speech acts and to crucial differences from similar syntactic categories, I will illustrate how elements of this particular category function as cues for noncanonical meanings in questions by bringing data to bear from languages as diverse as Basque and German. At the end of this chapter, I propose a crosslinguistic syntax for modal particles in questions that represents them as functional heads in the clausal architecture and captures both their clause-medial and left- and right-peripheral occurrences.

6.1 Modal particles and speech act syntax Modal particles, which can be found in numerous languages and display specific syntactic behavior, offer a valuable opportunity to examine and determine the nature of the relationship between illocutionary-meaning components and syntactic mechanisms. Modal particles have been documented in a diverse range of languages, such as Dutch, German, Greek, Finnish, Hungarian, Mandarin, Japanese, and many others (for overviews, see Grosz 2020; Zimmermann 2011). Consider the following example. In English, sentence (1) can be interpreted as either an information-seeking or a rhetorical question. When only examining the surface syntactic string, it appears that disambiguation arises from pragmatic inferencing and the context only: (1)

Who likes camping when it’s raining?

In German, a language known for its extensive inventory of modal particles, it is relatively easy to distinguish between an information-seeking question and a rhetorical one, as German incorporates particles that enforce the rhetorical reading. Compare sentences (2a) and (2b):

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0006

6.1 MODAL PARTICLES AND SPEECH ACT SYNTAX

(2)

89

a Wer zeltet gerne im Regen? who camps with.pleasure in.the rain ‘Who likes camping when it’s raining?’ b Wer zeltet schon gerne im Regen? who camps PART with.pleasure in.the rain ‘After all, who likes camping when it’s raining? (Nobody!)’

Owing to the presence of schon (literally, ‘already’), (2b) cannot be interpreted as an information-seeking question. This demonstrates that modal particles can play a crucial role in forming utterance meaning, transforming basic sentence types such as questions into more nuanced types such as rhetorical questions and others. One could argue that elements like schon do not reveal much about the connection between syntactic operations and illocutionary force. A significant body of research in semantics and pragmatics posits that modal particles are only indirectly related to syntax. Many approaches assume that these particles simply constrain the context in which an utterance is appropriate by functioning as presupposition triggers (e.g., Egg and Zimmermann 2012; Grosz 2014; Zeevat 2004) or as meaning contributions at a separate ‘expressive’ or ‘useconditional’ level (Gutzmann 2015, 2017; Kratzer 1999, 2004; Potts 2007a; Rojas-Esponda 2014). In this chapter, I will illustrate that, despite the substantial progress made in semantic and pragmatic approaches to modal particles, a syntactic perspective is still worth exploring. According to such a syntactic view, modal particles are modifiers of a left-peripheral operator—be it a speech act operator (e.g., Haegeman 2014; Waltereit 2001) or a sentence-type operator (e.g., Zimmermann 2008). Those approaches assuming an interaction between force-related categories and particles make sense based on observations like the following German examples, which demonstrate that the denotation of the assertive particle ja (literally, ‘yes’) is incompatible with question force: (3)

a Andreas zeltet ja im Regen Andreas camps PART in.the rain ‘(As you and I already know,) Andreas is camping in the rain.’ b ∗ Warum zeltet Andreas ja im Regen? why camps Andreas PART in.the rain

Data patterns like (3) have led to proposals for modal particles that model the connection between force/sentence mood and the particles by assuming illocutionary features like [+assertive]. Bayer and Obenauer (2011), for example, put forth an analysis that keeps the particle in situ (in the so-called

90

NON-CANONICIT Y VIA MODAL PARTICLES

‘middlefield’/ TP-zone of the clause) and relies on probe-goal agreement at a distance (Chomsky 2000, 2001). Crucially, the syntactic mechanism proposed by Bayer and Obenauer (2011) is agreement at a distance, because, according to their approach, the particle has a fixed position and does not covertly move (i.e., at the level of Logical Form) to the Force domain of the clause. Several empirical facts support this approach, one of which is the informationstructural role of modal particles (see González López and Trotzke 2021; Grosz 2014; Trotzke 2018). To provide a brief sketch of this approach, consider the assertion in (3a). Here, assertive force is independent of the modal particle ja. In other words, the particle contributes to or modifies the illocutionary reading (signaling the ‘uncontroversiality’ of the assertion in this case) but does not constitute the illocutionary force. Instead, it can serve only as a ‘communicative cue’ (Grosz 2014), alongside intonation and potentially other features. The connection between any type of force/sentence mood (in this case, assertive) and the particles can be technically accounted for by adopting a feature-sharing version of Agree (Pesetsky and Torrego 2007). This allows for a mechanism where the functional head Force0 (e.g., ASSERT) does not have a Prt feature, but the respective particles are likely to have a feature matching the Force. This mechanism is necessary, because many other particles without an assertive feature (e.g., question particles to be discussed further below) are ruled out in assertions. Consider the following representation, where an interpretable feature probes an uninterpretable matching feature; adopting a notational convention, in (4c) agreement is expressed by an arbitrary value that fills the empty slot in [ ]:

(4) a [ForceP Force0 iASSERTForce [] [TopP ... [Prt uASSERTForce [] ... ]]] FEATURE PROBING b [ForceP Force0 iASSERTForce [] [TopP ... [Prt uASSERTForce [] ... ]]]

AGREEMENT c [ForceP Force0 iASSERTForce [4] [TopP ... [Prt uASSERTForce [4] ... ]]]

6.1 MODAL PARTICLES AND SPEECH ACT SYNTAX

91

Through agreement, Prt becomes part of the left periphery and its illocutionary components, such as ASSERT, Q(uestion), and IMP(erative), as suggested by Bayer and Obenauer (2011) and subsequent research (Bayer and Trotzke 2015; Trotzke and Monforte 2019, among others). It is important to note that the Force0 head and the particle are often not adjacent, as topical material can intervene. This has been referred to as the ‘watershed’ function of modal particles by Grosz (2016), marking the boundary between topical and focal material in a clause. Consider example (5), where ‘at the festival’ features a topical and ‘warm beer’ a focal interpretation: (5)

Wer [Force0 trinkt] auf dem Festival [Prt schon] who drinks at the festival PART gerne warmes Bier? with.pleasure warm beer ‘After all, who likes drinking warm beer at the festival? (Nobody!)’

The Force head and the particle can even be separated by CP-boundaries, as shown by experimental work in Bayer et al. (2016). In this context, the analysis in (4) presupposes a locality constraint on agreement and predicts that Force0 and Prt must belong to the same clause. This aligns with the traditional observation that modal particles are usually root phenomena. The distribution of modal particles in embedded clauses supports this analysis. Coniglio (2011) demonstrated that modal particles are limited to embedded environments considered embedded root contexts (see Heycock 2017; Hooper and Thompson 1973; and Haegeman 2002 for a cartographic implementation). In other words, only embedded clauses with a root-like CP-domain (including ForceP, in cartographic terms) allow the presence of modal particles. Given the syntactic requirement for particles to connect to the left periphery of the clause, it is not surprising that many languages other than German have their base position for modal particles in the left periphery. This means that there is no evidence of particles merging in the IP/TP-zone in these languages. Left-peripheral base generation (and thus surface occurrence) of particles has been suggested for languages such as Cantonese (Kuong 2008), Mandarin (Paul and Pan 2017), Japanese (Kuwabara 2013), and Indo-European languages like Romanian (Coniglio and Zegrean 2012). Consider the Romanian question particle oare. In (6a), oare precedes a left-dislocated element, and, in (6b), it is situated to the left of a left-peripheral wh-element (examples from Coniglio and Zegrean 2012: 240):

92

NON-CANONICIT Y VIA MODAL PARTICLES

(6)

a Oare și mașina și=a vˆandut=o Ion pˆană PART also car-ART REFL=have.3SG sold=it Ion until la urmă? at end ‘Has Ion sold his car, too, in the end (I wonder?)’ b Oare unde va pleca Ion mˆaine? PART where FUT leave Ion tomorrow ‘Where will Ion leave tomorrow? (I wonder)?’

Similarly, Hill (2002) posited that particles like oare in Romanian are merged as optional complementizers in the C-domain. Importantly, the CP-position of particles like oare must be differentiated from particles argued to merge outside the CP. For example, Japanese has been suggested to have higher sentential particles that occur in a projection above the CP-zone (Nasu 2012). However, these elements can be clearly distinguished from Japanese particles and speech act morphemes within the CP (see Sauerland and Yatsushiro 2017 for recent work) because particles believed to merge outside the CP can appear in various sentence types, not exhibiting the typical sentence-type restrictions of modal particles illustrated above (that also holds for oare: this particle can occur only in interrogatives). In Section 6.3 below, I will focus on Basque, because this language exhibits both positions for modal particles, meaning it has both a TP-internal and a CP-based position for those elements. Also, the functional composition of the Basque clause can be fruitfully connected to recent observations on the syntax of modal particles in questions—and I therefore try to propose a cross-linguistic syntax for modal particles in questions. But before we examine the syntax of Basque modal particles and their occurrences in questions in greater detail, let me briefly point out syntactic differences of modal particles as compared to elements that are very similar in their semantic interpretation.

6.2 The syntax of modal particles versus similar elements Much of the syntactic research on German (and also Italian) has concentrated on the (historically) close association between modal particles and sentence adverbs, proposing that modal particles can be categorized as ‘deficient’ sentence adverbs (see, e.g., Cardinaletti 2011, 2015a). This perspective suggests that modal particles can be derived from more general constraints that also apply to sentence adverbs. At a first glance, this analysis is supported by specific characteristics of modal particles frequently discussed in syntactic studies on

6.2 THE SYNTAX OF MODAL PARTICLES VERSUS SIMIL AR ELEMENTS

93

particles. In particular, particles exhibit similarities with higher adverbs in that they cannot appear within the surface scope of sentential negation (7) and can combine with other elements of the same category (8); the following German examples illustrating these properties are taken from Grosz (2016: 4): (7)

a Das ist {ja} nicht {∗ ja} seine Schuld. that is PART not PART his fault ‘That isn’t his fault (as you and I know).’ b She {probably} hasn’t {∗ probably} left.

(8)

a Da hat ja wohl der Lehrer versagt. there has PART PART the teacher failed ‘This is a case where the school teacher failed.’ b Unfortunately, I obviously didn’t do that.

Nonetheless, there are distinctions that separate German modal particles from sentence adverbs when considering solely syntactic properties, and some of these distributional variations are indeed shared with clitic elements such as ‘weak’ pronouns. For example, as demonstrated in (9), modal particles are not subject to coordination (9a), while sentence adverbs (in this case, German leider and unglu¨cklicherweise) do not follow this restriction (9b); the example is sourced from a Google search (last accessed on March 15, 2023): (9)

a ∗ Er hat halt und ja im Regen gezeltet. he has PART and PART in.the rain camped ‘He has camped in the rain.’ b Leider und unglu¨cklicherweise ist mir perso¨nlich unfortunately and unfortunately is myself personally ein Fall in meiner Familie bekannt. a case in my family known ‘Unfortunately, I myself know a case in my family.’

Another distinguishing syntactic characteristic of modal particles, which sets them apart from sentence adverbs (and is also relevant for what I will propose in Section 6.3 below) is illustrated in examples (10) and (11). In German, modal particles are typically thought to be limited to the so-called middle field (that is, the TP-domain of the clause). To understand this, consider the following pair of examples from Chapter 4, repeated here for convenience,

94

NON-CANONICIT Y VIA MODAL PARTICLES

featuring the adverb/particle polyseme vielleicht (literally ‘perhaps’); capital letters represent heavy stress: (10)

a Eliud ist vielleicht SCHNELL. Eliud is perhaps fast ‘Eliud is perhaps fast.’ b Vielleicht ist Eliud SCHNELL. ‘Perhaps, Eliud is fast.’

[adverb reading]

(11)

a ELIUD ist vielleicht schnell! Eliud is PART fast ‘My god, Eliud is so fast!’ b ∗ Vielleicht ist ELIUD schnell! (intended: same as [11a])

[particle reading]

It becomes clear that only the adverb vielleicht (10) can appear in the left periphery of the clause, while the corresponding particle (11) cannot. As a particle, vielleicht is part of the group of exclamative particles, which, in conjunction with the exclamative stress on the noun in this case, indicates that the declarative structure in (11a) should be interpreted as a sentence exclamation, rather than as a plain assertion as in (10a). In summary, even when focusing solely on German modal particles, there are valid reasons to believe that their syntactic behavior cannot be explained by the same syntactic constraints that apply to closely related word classes. Instead, modal particles appear to have their own unique syntactic distribution requirements. It is essential to emphasize that, although modal particles may sometimes share meanings with other modal elements, they should not be confused with these alternate modal devices. From a syntactic perspective, modal particles should be represented in a distinct functional domain of modality, separate from what Cinque (1999) proposed for modal adverbs. This observation is significant, since modal particles, in different languages, often express similar meanings compared to other modal expressions, whether in the epistemic (e.g., Degen et al. 2019; Meisnitzer 2012) or evidential domain (e.g., Wiemer and Stathi 2010). However, key differences exist in their scope-taking behavior, which is why they cannot be the heads of the modal projections that Cinque (1999) described for adverbs. This distinction has been demonstrated extensively in the context of German, and I would like briefly to highlight one example, focusing on the epistemic particle wohl, a prominent example in the literature on epistemic modality in German.

6.2 THE SYNTAX OF MODAL PARTICLES VERSUS SIMIL AR ELEMENTS

95

It is argued that wohl diverges from other seemingly synonymous modal devices, such as epistemic must and epistemic adverbs, in terms of scopetaking during question formation and structured propositions (see Zimmermann 2008, 2011). For example, consider the differences in scope-taking during question formation. The formal sketches in (12), taken from Zimmermann (2011: 2021), show that using must in a question leads to asking whether Max necessarily has to be at sea, while using probably asks whether there is reason to believe that Max is at sea: (12)

a ⟦Must Max be at sea?⟧ = ?{Max must be at sea, ¬(Max must be at sea)} b ⟦Is Max probably at sea?⟧ = ?{ASSUME(x, Max at sea), ¬ASSUME(x, Max at sea)}

The key point here is that, in both cases, the semantics of these modal devices contributes to the alternatives under discussion, as stated by Zimmermann (2011). Both the modal adverb and epistemic must, therefore, contribute to the propositional (and consequently, the truth-conditional) content of the utterance. With this understanding, let us now examine an example provided by Zimmermann (2011: 2020), which demonstrates the scope-taking behavior of the epistemic particle wohl in questions: (13)

a Hat Hans wohl Maria eingeladen? has Hans PART Mary invited ‘What do you reckon: Has Hans invited Mary?’ b WANT (S, A, know (S & A, ASSUME {Hans invited Mary, Hans did not invite Mary}))

In (13b), wohl is depicted as the operator ASSUME, which scopes over the set of alternative answers ‘Hans invited Mary, or he didn’t’. This binary set is brought up by an interrogative operator for polar questions, which is modeled here based on Truckenbrodt’s theory (2006), which views “all sentential speech acts as volitional on the part of the speaker […] S wants something, wishes for something, invites A to do something etc.” (Truckenbrodt 2006: 263–4). Importantly, both the translation (13a) and Zimmermann’s (2011) formal sketch in (13b) demonstrate that the semantics of wohl does not form part of the alternative propositions, unlike what we have seen in (12) for other modal devices. Specifically, the question in (13a) is not asking whether there is a lack of commitment towards the proposition or not. Instead, by using wohl, the speaker encourages the addressee to make their best guess regarding the alternative answers, which can be conveyed by the operator ASSUME,

96

NON-CANONICIT Y VIA MODAL PARTICLES

taking scope over the alternative answers rather than being part of them. Consequently, given plausible assumptions about the syntax–semantics interface, the modal meaning of modal particles cannot be located in the same functional projection of the clause as other modal means that might express similar meanings (see also Degen et al. 2019 for experimental evidence that further supports this assumption).

6.3 A cross-linguistic syntax of question particles Recall from Section 6.1 that we find modal particles in many other languages than German, especially so in the domain of questions. One significant crosslinguistic feature of the syntax of modal particles is that their base position can be located both clause-internally and at the peripheries of the clause. In German, those particles are clearly based clause-internally (in the TP-zone, see our discussion in the previous section). Languages that feature clause-peripheral modal particles do not usually have clause-internal ones (e.g., Romanian, see Section 6.1 above). In this section, I will focus on Basque, because in this language we find both positions for modal particles—and, even more importantly, we find this co-occurrence of syntactic positions in the speech act of questions (see Trotzke and Monforte 2019). Accordingly, based on Basque I hope to be able to draw some general conclusions for a comprehensive account of the syntax of question particles. Let me first mention some basic facts about Basque more generally, and about its syntax more specifically. Basque is renowned for being a unique pre-Indo-European language in Europe. It is currently spoken by approximately 800,000 people and is located around the Gulf of Biscay in northern Spain and southwest France. Administrative or political boundaries within this region do not coincide with dialectal borders (Camino 2019), and the literature recognizes five dialects: the western dialect, the central dialect, Navarrese–Lapurdian, Navarrese, and Souletin dialects. With many of these dialects having further sub-dialects, the Basque language is highly valuable for researchers investigating syntactic microvariation, as demonstrated in syntactic work on focus and negation (e.g., Etxepare and Uribe-Etxebarria 2008; Laka 2019; and many more), datives, differential object marking, and argument incorporation (e.g., Fernandez and Rezac 2016; Odria 2014), and on many more phenomena. Let us now turn to one syntactic detail that will become relevant for our analysis of Basque modal particles. One intriguing aspect of Basque syntax is that the finite verb can agree with the addressee, given the appropriate context.

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

97

This phenomenon is referred to as allocutivity (see Alok and Haddican 2022 for a recent overview). Let us briefly illustrate this feature in Basque. In Basque, when addressing a close friend or someone who would be referred to using the pronoun hi (familiar or informal ‘you’) rather than the neutral zu, the finite verb in Basque necessitates an additional morpheme corresponding to the overt addressee, which is not an argument selected by the verb (e.g., Alberdi 1994):¹ (14)

Hi, Alemanian egon nau-k/-n /∗ naiz you, Germany.IN be AUX.1SG.ABS-ALLOC.M/-F AUX.1SG.ABS ‘Dude, I was in Germany.’

In Basque, another situation where lexical verbs and auxiliaries are not positioned next to each other arises when the clause contains a modal particle. It is widely accepted that Basque grammar includes a range of evidential, epistemic, and interrogative particles, and I will discuss this in detail further below. For now, let us look at (15) and (16), where these particles (in this case: omen and ote) appear to the left of the finite verb: (15)

Irratia entzuten omen du radio.ABS LISTEN.IPFV PART AUX ‘S/he is said to listen to the radio.’

(16)

Egia esan ote du? truth.ABS say PART AUX ‘Did s/he say the truth (OTE)?’²

Let us now delve deeper into Basque modal particles and their presence in questions. Traditional Basque grammars categorize a group of elements that appear adjacent to the inflected verb as ‘modal’ or ‘discourse’ particles (e.g., de Rijk 1969; Elordieta 1997). The particles included in this category are ahal, al, bide, ei, ote (ete in western dialects), and omen. These particles are further divided into distinct semantic groups based on their general meaning contributions (e.g., evidential versus non-evidential particles). Alternatively, they can be grouped together based on their sensitivity to sentence type, meaning the types of sentences in which they predominantly appear.

¹ If not indicated otherwise, the following Basque examples and judgments were provided by my collaborator Sergio Monforte as part of my IDEX professorship at IKER in 2018/19 (ANR grant No.10-IDEX-03-02). I further thank Urtzi Etxeberria and Ricardo Etxepare for advice and support. ² Since the exact meaning contribution of ote depends heavily on the pragmatic context (see also my claims further below), I will not provide specific paraphrases in decontextualized examples like (16).

98

NON-CANONICIT Y VIA MODAL PARTICLES

All these particles can undergo the standard tests for modal particles discussed in the context of Romance languages, which demonstrate that modal particles are indeed syntactic heads and cannot be located in a specifier position (see Munaro and Poletto 2002, 2009). Specifically, particles like omen or ote cannot be modified (17) or focalized (18); cf. Munaro and Poletto (2002: 92) for these diagnostics: (17)

a Xabier Konstanzera etorri (∗ oso) omen da. Xabier Konstanz.ADL come very PART AUX ∗ ‘( Very) Reportedly, Xabier came to Konstanz.’ b Egon (∗ nahiko) ote da Asier Konstanzen? be enough PART AUX Asier Konstanz.IN ‘I’m (∗ enough) wondering whether Asier has been to Konstanz.’

(18)

a ∗ OMEN dator etxera. PART come home.ADL ‘S/he’s coming back home (reportedly).’ b ∗ OTE daki egia? PART know truth.ABS ‘Whether s/he knows the truth (I’m wondering).’

In the context of this book, I will concentrate on the category of question particles, specifically examining ote and al. Additionally, I will briefly discuss another element attached to the inflected form that has received less attention in grammars so far: the particle -a, which is used in eastern dialects. When mentioned, it has been identified as an interrogative marker (Euskaltzaindia 1987; and, more recently, Monforte 2018). To begin, let us first examine some basic characteristics of both wh- and polar questions in Basque before delving into the syntactic behavior of these particles in more detail. Extensive research has been conducted on the syntax of Basque questions, with multiple contributions from Ortiz de Urbina (1989 et seq.) on this topic. At a general level, Basque is an SOV language, which also exhibits a V2 effect observable in wh-questions. Example (19) demonstrates that the wh-element moves to SpecCP in Basque wh-questions, and the verbal complex is shifted to the C0 position by triggering head-to-head movement of the lexical verb to C0 . This movement is motivated by two factors (Ortiz de Urbina 1994): first, the lexical verb moves to I0 /T0 , since the finite verb must be lexicalized; then, the combination of both lexical and finite verbs (i.e., the verbal complex) triggers movement to C0 in order to check features in SpecCP.

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

(19)

99

Zer erosi du Nereak? what buy AUX Nerea.ERG ‘What did Nerea buy?’

(19ˈ)

CP

zer

Cˈ

TP

erosi du

Nereak

Tˈ

erosi du

VP

Nereak zer erosi du When an extra element is placed in front of the verbal complex located in C0 , the sentence becomes ungrammatical: (20)

∗

Nora Maitena joan da? where.ADL Maitena go AUX ‘Where did Maitena go?’

For polar questions, we also see the fronting of the constituent that consists of the lexical and the inflected verb. Ortiz de Urbina (1994) suggests that in yes– no questions without a focal constituent, a null operator is situated in SpecCP. This operator is responsible for the movement of the verbal complex, as both verbal components must check features in a head-specifier relation. (21)

Erosi du Mikelek janaria? buy AUX Mikel.ERG FOOD.ABS ‘Did Mikel buy the food?’

100

NON-CANONICIT Y VIA MODAL PARTICLES

(21ˈ)

CP

Ø

Cˈ

TP

erosi du

Mikelek

Tˈ

erosi du

VP

Mikelek janaria erosi du In the following discussion, we will observe that both wh-questions (19) and polar questions (21) can include modal particles in Basque. Let us first examine wh-questions and the relevant particle ote in this context. Take a look at the following wh-question once more: (22)

Zer erosi du Nereak? what buy AUX Nerea.ERG ‘What did Nerea buy?’

Wh-questions such as (22) can include the particle ote. This particle can serve multiple purposes in wh-questions. It can express the speaker’s doubt about the addressee’s ability to answer the question or be employed to form a rhetorical question (see Garmendia 2014; and Monforte 2020a for a detailed syntactic study). Additionally, ote is suggested to convey a mirative interpretation in the context of surprise–disapproval questions (cf. Alcázar 2016, 2017 for corpus examples). In the following analysis, I will not focus on the semantic interpretations of these various readings or propose a shared semantic basis for them; I will simply acknowledge that ote in wh-questions is utilized to create what can be considered a non-canonical question. Our primary focus will be on the syntactic behavior of ote, which appears to be independent of its different readings. As previously mentioned, since the precise meaning contribution of

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

101

ote is heavily reliant on the specific discourse context, I will not provide exact paraphrases in the following examples. Turning now to syntactic observations, we see in (23) that, when ote is present in a wh-question, the particle moves in conjunction with the inflected verb to CP, forming the wh-question: (23)

Zer erosi ote du Nereak what buy PART AUX Nerea.ERG ‘What did Nerea buy OTE?’

(23ˈ)

CP

zer

Cˈ

erosi ote du

TP

Nereak zer erosi ote du Observe that, in Basque questions constructed with zergatik (meaning ‘why’), the fronted wh-word can be moved individually, leaving the verbal complex in its original position (24), or the verbal complex can be shifted to C0 as demonstrated in the wh-questions previously discussed. In the latter case, the particle once more moves in tandem with the verbal complex (25): (24)

Zergatik Peiok hori galdegin ote why PETER.ERG that ask PART ‘Why did Peter ask me that OTE?’

(25)

Zergatik galdegin why ask dit? AUX

ote PART

dit AUX

Peiok Peter.ERG

dit? AUX

hori galdegin that ask

ote PART

From these observations, it is reasonable to deduce that the particle ote in Basque wh-questions connects to the inflected form of the verb. When that segment of the verbal complex moves, the particle is required to move as well.

102

NON-CANONICIT Y VIA MODAL PARTICLES

In fact, there is significant evidence to support this analysis (refer to Arregi and Nevins 2012; Elordieta 1997). For example, there are instances where I0 /T0 is moved to the periphery, leaving the lexical verb V0 in its original position, but still bringing the particle along, as seen in negative main clauses (26) containing the negation word ez (meaning ‘no’ or ‘not’): (26)

Baina zure alaba, non ez ote but your daughter where not PART oporretan egon ez ote da?! holidays.IN be not PART AUX ‘But your daughter, where hasn’t she been to OTE?!’

da AUX

non where

Additionally, in embedded questions, not only is the wh-word moved to the front, but the verbal complex also moves, with the particle ote once more appearing alongside the finite verb: (27)

Non utzi where leave

ote PART

duen non egunkaria AUX.C where newspaper.ABS

utzi leave

ote PART

duen jakin nahiko nuke. AUX.C know want.FUT AUX ‘I would like to know where s/he may have left the newspaper OTE.’ In eastern dialects, while a focal constituent situated in CP/FocP attracts the verbal complex (i.e., the lexical and finite verbs) in Standard Basque, there is also an option to license the focal element by fronting only the finite verb. Again, and as anticipated, the particle moves with the inflected form: (28)

ESTIK ote du ESTIK hori Esti.ERG PART AUX Esti.ERG that.ABS ‘Was it Esti who said that?’

erran say

ote PART

du? AUX

Note also that the presence of modal particles is contingent on the existence of TP. As demonstrated in (29), if the verb is not inflected, no particle can appear within the clause: (29)

Ez dakit nora joan (∗ ote) not know where.ADL go PART ‘I don’t know where to go.’

Lastly, it can be argued that the particle ote is an element attaching to I0 /T0 based on the fact that the particle does not impede synthetic verbs from occupying the clause-initial position. This implies that ote and inflected verbs form a single constituent (cf. Elordieta 1997):

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

(30)

∗

103

(Ba) ote daki zer eskatu behar duen? BA PART know what ask must AUX.C ‘Does s/he know what s/he must ask for OTE?’

In summary, evidence like the examples provided has led to the account that modal particles in Basque should be viewed as the head of a Mod(al) Phrase situated between VP and TP (31). This description is derived from their clausal position in positive declarative clauses (e.g., Elordieta 2001): (31)

[TP [ModP [VP …V ] Part] T]

Some researchers have positioned the ModP hosting the particle above TP based on the scope of particles (Haddican 2008) or due to morphological observations (Arregi and Nevins 2012): (32)

[CP [ModP [TP [VP…V ] T] Part]]

Both approaches offer valid points that are, however, based on different considerations and empirical areas (surface position of the particle, scope-taking position, and so on). I cannot determine which of these differing theories is correct, and in the context of this section, I simply want to emphasize that, according to prevalent theories on modal particles in Basque, particles rely on the structural representation of a TP-domain, and the phrasal projection hosting modal particles is adjacent to the TP-domain—whether it is directly above or below within the hierarchy of functional projections. With this general background in mind, let us now revisit the syntax of Basque ote and examine a new empirical observation that presents an exception to what we have claimed earlier and that has first been pointed out by Trotzke and Monforte (2019). I previously mentioned that, in some cases, the particle does not need to move along with the inflected verb, particularly in why-interrogatives. We will now delve into this interesting exception in more detail, which is a new finding and has been confirmed by twenty interviews conducted in Labourd and Low Navarre from January 2017 to August 2017 (see Monforte 2020b; Trotzke and Monforte 2019). In Basque why-questions, the particle may also appear adjacent to the wh-element in some dialects, primarily in Navarrese–Lapurdian (33). The option of fronting the verbal complex is also available in this configuration (34), which, however, does not alter the fact that the particle attaches to the wh-element rather than the finite verb: (33)

Zergatik why

ote PART

Peiok Peter.ERG

hori galdegin dit? that ask AUX

104

(34)

NON-CANONICIT Y VIA MODAL PARTICLES

Zergatik why

ote PART

galdegin dit ask AUX

Peiok Peter.ERG

hori galdegin that ask

dit? AUX

Indeed, the particle ote can also be found attached to other wh-elements besides ‘why’; consider the following example: (35)

Non ote utzi dut non ote where PART leave AUX where PART dut? AUX ‘Where did I leave the newspaper OTE?’

egunkaria newspaper.ABS

utzi leave

This observation supports the idea that, at some stage in the derivation, the wh-element and the particle need to merge, forming a single constituent. As we have previously discussed, Basque wh-questions exhibit a V2 effect, which means that only one constituent can appear in front of the verbal complex occupying the C0 position. Consequently, it is reasonable to hypothesize that the combination of the wh-word and particle must form a single constituent in configurations like (34). Additional support for the notion of the wh-element and particle forming a single constituent (and arguing against the finite verb and particle being co-constituents) can be observed in embedded questions (36) and negated wh-questions (37). In these instances, it is evident that the particle is not immediately adjacent to the verb: (36)

Ez dakit non ote non ote egunkaria utzi not know where PART where PART newspaper.ABS leave dudan. AUX.C ‘I don’t know where I may have left the newspaper OTE.’

(37)

Baina zure alaba, non ote ez da but your daughter where PART not AUX bakantzetan izan ez da? holidays.IN be not AUX ‘But your daughter, where hasn’t she been to OTE?!’

non ote where PART

Significantly, the co-constituency of wh-elements and modal particles is a syntactic option that can also be observed in other languages with modal particles in wh-questions. Let us look at some data from German that we have already mentioned in Chapter 4 of this book.

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

105

We have already emphasized that positional rigidity is a highly reliable characteristic of modal particles in German, and that this property sets them apart from other modifiers that express similar meanings, such as higher adverbs (see Section 6.2). However, modal particles in German wh-questions seem to present an exception. Consider the examples in (38a) and the alternative word order choice in (38b); examples from Abraham (1991: 237): (38)

a Wie habe ich nur den Schlu¨ssel how have I PART the key

verlieren ko¨nnen? lose could

b Wie nur habe ich den Schlu¨ssel verlieren ko¨nnen? how PART have I the key lose could ‘How on earth could I lose the key?’ Similar to the Basque examples we discussed earlier, the wh-element and particle must form a constituent in (38b) since German is a V2 language, typically restricting more than one constituent to the left of the finite verb (i.e., in SpecCP) of the main clause. Bayer and Trotzke (2015) offer evidence from various classic constituency tests that confirm the wh-element and particle in these constructions should indeed be regarded as a single constituent. Previously, Bayer and Obenauer (2011) identified this constituent as a so-called Small Particle Phrase (SPrtP). Returning to Basque examples like (39) below, the primary concept is that non+ote form a single constituent derived in a separate derivation layer. An essential computational rationale supports this approach. If the wh-element non were to merge with ote after the particle has been merged in ModP, the wh-element would strip the structure of the scope position that ote seems to occupy, essentially ‘decapitating’ ModP. Moreover, this operation would go against the Extension Condition of syntactic derivations. That is, in accordance with minimalism’s cyclic organization of grammar, “Merge always applies at the simplest possible form: at the root” (Chomsky 1995: 248). Consequently, it is evident that, at the derivational stage shown in (40), the wh-phrase can be merged only with the root (= [ModP ote [VP …]]) and not with ote alone: (39)

Non ote utzi dut non ote where PART leave AUX where PART dut? AUX ‘Where did I leave the newspaper OTE?’

(40)

non [ModP ote [VP … non…V ]]

egunkaria newspaper.ABS

utzi leave

106

NON-CANONICIT Y VIA MODAL PARTICLES

In light of the derivational challenges mentioned above, I propose that the constituent [non+ote] is created in a separate derivational workspace, and [non+ote] is then integrated into the unfolding V-projection, after which successive-cyclic movement of [non+ote], similar to wh-movement, may occur. Both ote and the wh-element are operators that need to be licensed in a scope position. As a result, in addition to functioning as an argument, [non+ote] must raise to the specifier of ModP, where ote can take scope over the proposition. Since ote is just one of the scope-taking operators involved, [non+ote] must move further to a position where the wh-element can check its relevant features. This derivational sequence is outlined in (41) and exemplified in (42): (41)

[CP [non+ote] [C’ C0 …[ModP [non+ote] [Mod’ Mod0 [… [non+ote]…]]]]]

(42)

Non ote utzi dut egunkaria? ‘Where did I leave the newspaper OTE?’

(42ˈ)

CP

non ote

Cˈ

utzi dut

... ModP

Modˈ

non ote

Mod0

... vP

non ote egunkaria utzi dut It is important to emphasize that deriving the constituent [non+ote] in a separate derivation is not unusual or exceptional, but instead aligns with standard minimalist approaches that assume interaction of subderivations (e.g., Trotzke 2020b; Trotzke and Zwart 2014), primarily in the form of generalized

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

107

transformations. Importantly, the Extension Condition on structure building mentioned earlier establishes that the derivation of complex subjects and adjuncts also implies that the computational system must handle multiple root syntactic objects simultaneously (Hornstein and Nunes 2014; Nunes 2012; Takano 2020). Van Craenenbroeck (2005) provides an alternative explanation for similar constructions, although his focus is limited to sluicing and other fragments. In his account, the wh-element and the rest do not constitute a single unit. Instead, the two parts are distributed across various CP-shells. However, in our cases, it is crucial to ensure that the wh-element and the particle form a constituent that moves together because this alignment allows us to make accurate predictions regarding V2 effects in both Basque and German. Given this analysis, let us now examine an aspect that differentiates cases where ote is connected to a wh-element from instances where it attaches to the finite verb and moves in tandem with verbal components (43), as opposed to moving with a wh-element (44): (43)

[CP Zer [C ′ erosi ote du what buy PART AUX ote du?]]] PART AUX ‘What did Nerea buy OTE?’

[TP Nereak Nerea.ERG

zer erosi what buy

(44) [CP Non ote [C ′ utzi dut [TP non ote egunkaria where PART leave AUX where PART newspaper.ABS utzi dut?]]] leave AUX ‘Where did I leave the newspaper OTE?’ As soon as [wh+ote] is introduced into the derivation, an additional pragmatic effect is added to the entire utterance, which is not present when ote attaches to the finite verb. Thus, apart from purely configurational reasons for proposing a separate derivation layer involved in generating [wh+ote], we also observe motivation from interpretative features suggesting a separate derivation for forming an interpretational unit [wh+ote]. This additional pragmatic effect can be explained in terms of emphatization (see Chapter 4 and Chapter 9), meaning that [wh+ote] is connected with an extra degree of the speaker’s emphasis. I rely on the concept of emphasis for intensity as proposed by Trotzke and Turco (2015) and Trotzke (2017a) for German configurations like the ones shown in (38). Therefore, I submit that

108

NON-CANONICIT Y VIA MODAL PARTICLES

the derivation provided in (41) and (42) must be further expanded by adding an extra step involving the movement of [wh+ote] to the specifier of EmpP before the wh-element has to move further for clausal typing reasons (see discussion in Section 6.1). It is important to note that EmpP represents a syntactic layer that cannot be identified with the function of clausal typing. More specifically, we have already seen in the context of non-canonical frontings in polar questions that emphasis in questions is a type of not-at-issue meaning whose operator must be below the Question operator for various empirical reasons (see Chapter 4):

CP

(45)

Cˈ

[wh+ote]

C0

EmpP

[wh+ote]

Empˈ

Emp0

...ModP

[wh+ote]

Mod0

Modˈ

...[vP ...[wh+ote]...]

The derivation in (45) illustrates the claim that [wh+ote] is merged into the developing V-projection. Since the particle ote, the wh-element, and the emphasis feature associated with the [wh+ote]-constituent are operators that require licensing in a scope position, [wh+ote] must first move to a position where the particle can take scope (specifier of ModP). Next, [wh+ote] moves to EmpP, where the Emphasis operator is located. Finally, [wh+ote] must move to a position where the wh-element can be licensed (i.e., SpecCP). In sum, this analysis can account for the syntax of clause-internal modal particles in Basque questions, and it has also proven to be an accurate syntactic approach to similar patterns in wh-questions in other languages than Basque.

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

109

With this analysis in mind, I now turn to Basque particles in polar questions and thereby look at a domain where particles in questions can also be base-generated in the periphery of the clause. In this context, as I will argue, Basque particles can function as clause-typing devices and are, in fact, used to express unmarked questions in Basque. More specifically, the particle ote represents a functional head that contributes not-at-issue/evaluative meaning at the level of modality and can thus be represented by a clause-internal head Mod0 . In contrast, particles like Basque al in polar questions occupy a head related to sentence mood and impose restrictions on the respective speech act to be performed (see Haddican 2001). Let us look at the details. In various Basque dialects, particles such as al (central dialect) and -a (eastern dialects) may be present in polar questions—and, if they occur in such questions, these particles are required to move along with the finite verb, similar to what we observed in the context of non-emphatic versions of wh-questions with ote above: (46)

Erosi al du Mikelek buy PART AUX Mikel.ERG du AUX ‘Did Mikel buy food?’

(47)

Erosi buy dua AUX.PART

dua AUX.PART

janaria? food.ABS

Mikelek janaria? Mikel.ERG food.ABS

erosi al buy PART (central dialect)

erosi buy (eastern dialects)

Now, let us delve deeper into the syntactic properties of the particle al (see also Monforte 2020b, 2021). I aim to argue that al is akin to ote in the sense that it also originates within the TP-domain (as discussed earlier). First, consider the fact that al can appear in embedded questions, and, in this context, it is compatible with the complementizer -(e)n. The hierarchical relationship between these particles and the complementizer -(e)n is that complementizers must c-command particles: (48)

Xabierri afaria egin al duen galdetu diot. Xabier.DAT dinner.ABS do PART AUX.C ask AUX ‘I asked Xabier whether he prepared the dinner.’

Additionally, the particle al can be utilized in situations where an allocutive morpheme is present in the inflected form. Recall from the outset of

110

NON-CANONICIT Y VIA MODAL PARTICLES

this section that these morphemes are considered to occupy the CP-domain (Haddican 2018; Miyagawa 2022): (49)

Egingo al di –k /-n afaria? do.FUT PART AUX-ALLOC.M/-F dinner.ABS ‘Is s/he going to prepare the dinner? (talking to, for instance, a close friend)’

The same applies to ote, which further supports the idea that ote is not located in competition with the allocutive morpheme in CP, but rather takes scope in a ModP projection inside the TP: (50)

Egingo ote di-k/-n afaria? do.FUT PART AUX-ALLOC.M/-F dinner.ABS ‘I’m wondering if s/he’s going to prepare the dinner (talking to, for instance, a close friend).’

As we can therefore hypothesize, both al and ote are particles originating in the TP-domain and are not left-peripheral (‘C’) particles like those found in other languages such as Romanian or Asian languages (see Section 6.1 above). However, there are also crucial differences between these particles. First, it is important to note that al is strictly limited to polar questions, meaning that it cannot appear in wh-questions: (51)

Noiz etorriko (∗ al) da afarira? when come.FUT PART AUX dinner.ADL ‘When is s/he coming to the dinner?’

Conversely, particles like ote can also be interpreted and used in polar questions: (52)

Etorriko ote da afarira? come.FUT PART AUX dinner.ADL ‘Is s/he’ coming to the dinner OTE?’

etorriko ote come.FUT PART

da AUX

Given this distribution, we can say that modal particles like ote can modify various sentence types, while al relies on the specific clausal configuration of a polar question. This suggests that they might be represented by a different functional head than Mod, as has already been proposed by Haddican (2001) and others. That is, particles like al instantiate a Mood head that, by definition, limits the class of sentence types the particles can appear in (e.g., Portner 2018). This also clarifies another fundamental difference between ote and al: unlike ote, al does not modify the speech act (in this case: a polar question). In fact, Basque polar questions containing al are the unmarked cases, aligning with

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

111

classic observations on polar questions and corresponding particles in many other languages (see, e.g., Cheng 1997; and our illustrations in Chapter 1 of the book). Consequently, it would be the omitting of al that can result in biased and non-canonical versions of the polar question, for example: (53)

Dena gainditu zenuen? all.ABS pass AUX ‘So, you passed all (your exams)? (Really?)’

As a result, I hypothesize that al represents a Mood head, which, based on our previous discussion, is part of the functional hierarchy within the TP-domain. However, there is another syntactic fact to consider at this point. Observe that al and ote display complementary distribution, implying that they cannot cooccur within a single clause: (54)

∗

Egin {al ote}/ {ote al} du do PART PART PART PART AUX ‘Did s/he prepare the dinner?’

afaria dinner.ABS

egin do

du? AUX

This is semantically logical and predicted. The particle al indicates unmarked or neutral questions, while the use of ote results in special or non-canonical interpretations, so al and ote cannot coexist in a single clause. From a syntactic perspective, we can conclude that TP-internal question particles in Basque occupy a fixed position within the functional hierarchy of the clause, which is associated with (at least) two distinct interpretations. In (55), let us label this fixed position as a PrtP, in line with our discussion in Section 6.2 on the differences between particles and related categories such as adverbs. In other words, the claim is that particles project their own functional projection featuring modality and mood interpretations, and thus do not take up the adverb-related positions within the TP proposed by Cinque (1999) and others: (55)

[TP [PrtP [VP …V ] Prt0 [mood]/[modality] ] T]

An interesting cross-linguistic outcome emerges from our examination and analysis of TP-internal Basque particles. In general, it has been argued that German (a language with TP-internal particles as well, see above) allows modal particles in questions to coexist and stack (Bayer and Trotzke 2015; Thurmair 1989): (56)

Wer zeltet denn schon gerne im Regen? who camps PART PART with.pleasure in.the rain ‘After all, who likes camping when it’s raining? (Nobody!)’

112

NON-CANONICIT Y VIA MODAL PARTICLES

It has been highlighted that there are ordering constraints (∗ schon denn) for stacked particles, which suggests the presence of individual particle heads with unique scope-taking properties. However, it has not yet been noted that German, akin to Basque, also displays complementary distribution of question particles. For instance, the particle nur, which expands the search domain for the variable x in wh-questions (see Chapter 3), and the particle schon, which introduces the reading of a rhetorical question (Meibauer 1986), are not able to co-occur in a wh-question: (57)

a Wer zeltet (∗ nur) schon (∗ nur) gerne im who camps PART PART PART with.pleasure in.the Regen? rain ‘After all, who likes camping when it’s raining? (Nobody!)’ b Wie habe ich (∗ schon) nur (∗ schon) den Schlu¨ssel how have I PART PART PART the key verlieren ko¨nnen? lose could ‘How on earth could I lose the key?’

This implies that we might also encounter modal particles in German that vie for the same structural position within the clause. As a result, Prt heads with intricate feature compositions, leading to complementary distribution, might be present in languages like German as well. Given these observations in the context of TP-internal particles, we will now examine the variant of Basque al used in eastern dialects, namely -a. Intriguingly, in this instance, we cannot explain the syntactic behavior of the particle by proposing structural representations like the ones discussed above. Instead, as I would like to argue, this particle involves a functional position outside the TP-zone, thereby representing a structural option observed in other languages such as Romanian and Asian languages. All in all, this will support the cross-linguistic claim that languages can feature two types of particles: one at the level of TP and another at the level of CP (see Del Gobbo et al. 2015). Prior research by Ortiz de Urbina (1993) and Laka (1990) has already suggested that the particle -a could occupy the head position within the CPdomain. However, neither author delves into its syntactic behavior in depth, and their assumption is based on its surface syntactic position and related analyses for question particles in other languages. In the following, I offer some syntactic evidence that reinforces their hypothesis (see also Monforte 2020b),

6.3 A CROSS-LINGUISTIC SYNTAX OF QUESTION PARTICLES

113

and I will then conclude by revisiting the other Basque question particles we discussed earlier. First, it is important to note that the morphosyntactic position of -a differs from the position of the other modal particles we discussed earlier, as -a acts as a suffix rather than a prefix: (58)

Egin du-a [> dia] afaria do AUX-PART dinner.ABS ‘Did s/he prepare the dinner?’

egin du? do AUX

This position is similar to the one occupied by complementizers, as they are also consistently attached to the right of the inflected verb; observe the following example: (59)

Xabierri afaria egin du-en galdetu diot. Xabier.DAT dinner.ABS do AUX-C ask AUX ‘I asked Xabier whether he prepared the dinner.’

Another observation worth mentioning is that -a cannot appear in embedded questions featuring, for example, the complementizer -en (60). In this regard, -a stands in contrast to the other particles ote and al, which can occur in such embedded contexts. (60)

Xabierri afaria egin du (∗ -a) -en(∗ -a) Xabier.DAT dinner.ABS do AUX -PART-C -PART diot. AUX ‘I asked Xabier whether he prepared the dinner.’

galdetu ask

Those observations imply that complementizers and -a display a complementary distribution, suggesting that -a occupies the C-position in the left periphery. Another important point to note is that, unlike TP-internal particles such as al and ote (as discussed earlier), the particle -a cannot appear in allocutive contexts. This is true even when discourse requirements necessitate the presence of the allocutive morpheme, which is generally believed to occupy the C-position (see literature cited above): (61)

Jinen da(∗ -n)-a hire laguna afarira come.FUT AUX-ALLOC.F.PART your friend.ABS dinner.ADL jinen da? come.FUT AUX ‘Dude, is your friend coming to the dinner?’

114

(62)

NON-CANONICIT Y VIA MODAL PARTICLES

Jinen du-n(∗ -a) hire laguna afarira come.FUT AUX-ALLOC.F.PART your friend.ABS dinner.ADL jinen da? come.FUT AUX ‘Dude, is your friend coming to the dinner?’

In summary, the particle -a serves as an example of left-peripheral modal particles, similar to those found in other languages. Taking into account the structural claims for ote and al, we can now provide a more comprehensive picture by incorporating a possible base position for question particles in the left periphery of the Basque clause. In eastern dialects, this position can be instantiated by -a:

CP

(63)

... TP

Prt0

T0

PrtP

Prt0[mood]/[modality] VP Importantly, the structural representation shown above correctly predicts that -a can coexist with other discourse particles like ote, as they do not compete for the C-position. This sets the eastern variant -a apart from the TP-internal particle al, which exhibits a complementary distribution with ote, as I have illustrated earlier. (64)

Egin ote du-a afaria do PART AUX-PART dinner.ABS ‘Did s/he prepare the dinner OTE?’

egin ote do PART

du? AUX

Similar patterns of particle co-occurrence, occupying different positions within the functional structure of a clause, have been documented for Germanic (e.g., Haegeman 2014) and Romance languages (e.g., Munaro and Poletto 2009), as well as Cantonese and Mandarin (see Li 2006). In conclusion, the variation in the use of modal particles in Basque questions provided us

6.4 SUMMARY

115

with insights into the syntax of modal particles in questions across languages more generally.

6.4 Summary In this chapter, our focus has been on the syntax of modal particles in non-canonical questions. We presented an analysis that establishes new connections between the syntactic behavior of modal particles in languages as diverse as Basque and German. Our discussion emphasized the importance of a syntactic perspective on modal particles, as they have mostly been studied in the context of semantics and pragmatics, with less attention given to formal syntax. In this regard, I highlighted that modal particles should be recognized as distinct syntactic elements, separate from their closely related counterparts (e.g., higher adverbs). For the detailed syntactic analysis of modal particles in questions, we examined the Basque language and its range of particles in both wh-questions and polar questions. For wh-questions, I have sketched some support for the idea that the particle ote occupies a TP-internal particle position, and, when combined with a wh-element, can create emphatic questions similar to those we have already seen in Chapter 4 of the book. Regarding polar questions, I further developed the concept that there are two distinct positions for modal particles in Basque: one within the TP-domain and another in the left periphery of the clause. Throughout this chapter, I have explored the hypothesis that modal particles are functional heads, significantly contributing to the functional architecture of the clause. In particular, rather than assigning a particle-specific head for each modal particle, which is theoretically unsound, I sought to understand the set of particle-specific projections by accounting for them in terms of a limited number of functional projections across languages. Overall, this chapter has demonstrated the value of pursuing a cross-linguistic syntax of modal particles, even considering the significant progress made in semantic–pragmatic approaches in this domain. With these fundamental syntactic mechanisms and distributions in the domain of noncanonical questions in mind, we can now delve deeper into the pragmatics of non-canonical questions and, on this basis, examine in more detail the extent to which the pragmatic behavior of non-canonical questions is reflected in their syntax.

PART III

THE PR AGM ATICS OF NON-C A NONIC A L QUESTIONS

7 Non-canonical questions as first moves in a dialogue In Part II of this book, I investigated several subtypes of non-canonical questions from a syntactic perspective. In particular, we first looked at noncanonical word-order patterns in wh-in-situ questions, in declarative questions, and in non-canonical frontings to the left periphery of the clause (Chapter 4). Given this discussion and a syntactic proposal for the analysis of those question types, we explored a variety of non-canonical wh-elements that can occur in non-canonical questions (Chapter 5): why-like-what, extreme degree readings of wh-elements, and so-called how come questions. Finally, Part II of the book presented a syntactic approach to analyzing non-canonicity in questions that comes about by using modal particles (Chapter 6), and I suggested a way of accounting for relevant patterns within a cross-linguistic syntax of question particles. This part of the book deals with basic discourse properties of non-canonical questions by looking at their pragmatic behavior as first moves and as second moves in a dialogue. I will start with first-move contexts in this chapter. In particular, I will introduce the difference between first moves that can be analyzed as a request for information (as in ‘canonical’ information-seeking questions) and a request for affirmation (as in many subtypes of non-canonical questions where the speaker wants the addressee to affirm or confirm their bias and their epistemic and/or emotional state). I then discuss the topic of at-issue versus not-at-issue content in the context of how an addressee might react to a particular non-canonical question, and I will explore at what level the addressee can affirm or object to the content(s) conveyed by the speaker’s question. As should already be clear from the previous chapters, non-canonical questions, such as rhetorical questions, biased questions, or expressive/emphatic questions are special in terms of their discourse behavior because they serve functions beyond the mere request for information that canonical questions

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0007

120

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

are intended for. These types of questions are used to convey a speaker’s attitude, perspective, or presuppositions, rather than to elicit information from the addressee in the first place. As we have already indicated, rhetorical questions, for example, are questions that are asked for effect, rather than to obtain an answer. They are often used to make a point, to challenge an addressee’s assumptions, or to express strong emotions. For example, the question Do I look like I care? is a rhetorical question that is not intended to elicit a response, but rather to convey the speaker’s indifference or irritation. The discourse behavior of non-canonical questions, such as rhetorical questions and biased questions, is special, because they are often used to convey a speaker’s attitude or perspective, rather than to elicit information. As such, they yield a different type of discourse effects than canonical questions and are an important aspect of pragmatic language use. In what follows, we will focus on discourse effects of non-canonical questions as first moves in a dialogue. In dialogue, the first move refers to the initial speech act or utterance that is made by one speaker to initiate the conversation. Needless to say, the first move sets the tone for the conversation and can influence the direction that the conversation takes. In principle, any speech act can be used as a first move in a dialogue. The choice of speech act used as the first move will depend on various factors, such as the speaker’s goals and intentions, the context of the conversation, and the relationship between the speakers (see Macagno and Bigi 2017). For example, an information-seeking question, a commissive speech act, an expressive speech act, or an assertion speech act can all be used as a first move in a dialogue. However, there are speech acts that might be less appropriate as first-move utterances. These speech acts may be less typical because they are more confrontational, or less relevant to the context of the conversation. For instance, while information-seeking questions are a common type of directive speech act used as a first move, more forceful or aggressive directives, such as commands or orders, may be less typical as first moves in a dialogue, because one would typically use more polite forms in order to initiate a veritable dialogue (see also our discussion in Chapter 10 of the book). Insults are speech acts that are used to attack or belittle the addressee. While insults can be used as a first move in certain contexts, such as arguments or debates, they are less typical as first moves in a dialogic situation, because they often do not aim at any kind of response. In other words, when we consider first-move utterances, we presuppose a dialogic context and thus the possibility of (at least) a second move, which we can label ‘response’ at a very general level. As we will discuss

7.1 REQUEST FOR INFORMATION VERSUS REQUEST FOR AFFIRMATION

121

in the next section, we must distinguish between certain subtypes of responses, namely—in the context of questions—between requesting information and requesting affirmation.

7.1 Request for information versus request for affirmation Broadly speaking, a request for information and a request for affirmation are two distinct types of speech acts that serve different pragmatic functions in conversation. Let us approach this distinction in simple terms. A request for information is a speech act in which the speaker seeks new information or knowledge from the addressee. As we have seen in Chapter 2 on the basic semantic and pragmatic features involved in this kind of speech act, this type of request typically comes in the form of a question, and the speaker genuinely does not know the answer (e.g., What time does the movie start?, How do I get to the train station?). The main purpose of a request for information is to obtain the desired information from the addressee. A request for affirmation, on the other hand, is a speech act in which the speaker seeks confirmation or validation of information they already believe to be true. This type of request often comes in the form of a question or statement, but the speaker is not seeking new information in the first place; instead, they want the addressee to confirm, agree with, or acknowledge the information or belief they already hold. Examples of requests for affirmation include examples like the following: (1)

a It’s going to rain tomorrow, right? b You’re coming to the party tonight, aren’t you? c This is the correct answer, isn’t it?

In these examples, the speaker is not looking for new information; rather, they want the addressee to confirm or validate their existing beliefs or understanding. The primary pragmatic distinction between a request for information and a request for affirmation thus lies in the speaker’s intention and the type of response they expect from the addressee. In a request for information, the speaker genuinely seeks new knowledge and expects the addressee to provide the desired information. In a request for affirmation, the speaker already holds a belief or understanding and expects the addressee to confirm or validate it. The request for affirmation often implies that the speaker has some

122

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

degree of confidence and/or a certain bias in their belief or understanding (see Chapter 3), but they want reassurance or acknowledgment from the addressee. With these broad ideas in mind, let us now discuss this fundamental distinction by looking in more detail at the difference between information-seeking questions and non-canonical questions that are not information-seeking. In this domain, I focus on questions that express the speaker’s surprise and/or disapproval with respect to the event denoted by the propositional part of the question. We have already introduced this type of question in Chapter 5, when we illustrated why-like-what elements in different languages. For conveniency, here is a German example of that type of question again (see Munaro and Obenauer 1999: 238): (2)

Was lacht der denn so blo¨d?! what laughs he PART so stupidly ‘Why is he laughing so stupidly?!’

It is important to remember that the primary focus of these data in Chapter 5 was to demonstrate that the wh-element ‘what’ does not refer to a syntactic argument (as in What is he buying? [He is buying a car]), but rather conveys a meaning similar to ‘why’. This is especially evident in examples where only a non-argumental reading of ‘what’ is possible, owing to the presence of intransitive verbs, such as ‘to laugh’ in (2). An argumental interpretation of the wh-element (as in What is he buying?) would result in ungrammaticality (∗ He is laughing x). We have also emphasized that, at the pragmatic level of speech acts, these examples could be classified as exclamations, which are expressive speech acts that convey a violation of the speaker’s expectations regarding a particular state of affairs. To investigate the illocutionary status of questions like (2) as first move in a dialogue, we can also ask how speakers can react to surprise–disapproval questions when they are used as first moves. Do speakers perceive those questions as requests for information in the first place? Or do they mainly interpret them as exclamation speech acts expressing the surprise of the speaker? In a recent experimental study, Trotzke and Czypionka (2022) have investigated the speech act status of non-canonical questions by testing the acceptability of different reactions that can follow questions conveying surprise and disapproval such as (2) in a mini-dialogue. The design of Trotzke and Czypionka’s study (2022) is based on some well-known diagnostics from the previous literature according to which questions like (2) should preferably be followed by an answer if they are predominantly perceived as directive speech

7.1 REQUEST FOR INFORMATION VERSUS REQUEST FOR AFFIRMATION

123

acts requesting an answer from the addressee. However, if cases such as (2) are mainly interpreted as expressive utterances conveying the (negative) surprise of the speaker, then they could also be felicitously followed by an affirmation (see original discussions in Grimshaw 1979: 321 for the answering patterns and in Zanuttini and Portner 2003: 47 for affirmations). Observe the following items from Trotzke and Czypionka’s experimental materials (2022: 243): Julia sieht, dass Marc sich mitten am Tag hinlegen mo¨chte. Dabei dachte sie, dass sie den Nachmittag miteinander verbringen. JULIA: „Was schla¨fst du denn jetzt?!“ [ANSWER] Marc: „Oh, das weißt du nicht? Ich bin gestern Abend spa¨t ins Bett gegangen, darum lege ich mich noch mal hin.“ [AFFIRMATION] Marc: „Du hast recht: Eigentlich hatte ich dir versprochen, dass wir den Nachmittag zusammen verbringen.“ Julia sees that Marc is going to lie down in the middle of the day. However, she thought that they were going to spend the afternoon together. JULIA: “What are you doing sleeping now?” [ANSWER] Marc: “Oh, you don’t know? I went to bed late yesterday, that is why I am lying down again.” [AFFIRMATION] Marc: “You are right, I promised you we would spend the afternoon together.”

In contrast to both information-seeking questions and exclamatives, Trotzke and Czypionka (2022) demonstrate in their experimental study that both the reaction ‘answer’ and the reaction ‘affirmation’ are acceptable in a minidialogue where those reactions follow surprise–disapproval questions. The results in Trotzke and Czypionka’s study (2022) indicate that surprise– disapproval questions are judged by participants as having both a question and an expressive interpretation at the level of speech acts. By contrast, information-seeking questions and exclamatives in their study display a clear preference for only one of the follow-up reactions (i.e., either answer or affirmation, respectively) and can thus be characterized as conveying only one speech act interpretation. All in all, we can conclude from this experimental study that surprise–disapproval questions can be considered a non-canonical speech act at the level of pragmatics, in the sense that they are instantiated by mixed utterance types, which combine pragmatic features of different speech act types (questions, on the one hand, and expressive speech acts, on the other

124

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

hand). Given these empirical data, let us now turn to the question of how we can deal with those question types theoretically. As widely acknowledged in the speech act literature on exclamations, and as we have already mentioned in Chapter 5 as well, they can be expressed through a variety of syntactic forms, including interrogative syntax. Consider again the following English examples from Rett (2011: 412): (3)

a b c d

(Wow,) John bakes delicious desserts! (My,) What delicious desserts John bakes! (Boy,) Does John bake delicious desserts! (My,) The delicious desserts John bakes!

Examining the data in (3), it becomes apparent that exclamation as a speech act does not correspond to a specific sentence type. Instead, the concept of exclamation may refer to particular uses of various sentence types. Under this perspective, the interrogative example in (2) might simply represent another means to convey an exclamation. Following this argument, we could assert that the case in (2) no longer functions as a genuine (i.e., information-seeking) question, but rather must be analyzed in formal pragmatics as featuring exclamation force. There are prominent theories in formal linguistics that posit a dedicated illocutionary force operator for exclamations. In fact, the ‘illocutionary force hypothesis’ that there is something like the exclamation speech act is shared by almost all of the existing approaches. To my mind, the most articulated and clearest version of the illocutionary force hypothesis has been formulated by Rett (2011), whose main point is to characterize ‘exclamation’ as “a speech act with a unique illocutionary force” (Rett 2011: 439). That is, when we consider various linguistic forms such as in (3) above (including their relevant intonation patterns) and how they convey expressive meaning, Rett’s driving hypothesis (2011: 413) is that “a characterization of exclamation as an independent speech act best captures this relationship.” Let us first illustrate this general idea by sketching very briefly the two formal approaches by Rett (2011) as well as the one by Grosz (2012), because those two approaches most explicitly postulated a dedicated illocutionary operator for exclamations and can thus illustrate the general thinking of the illocutionary force hypothesis in the context of our discussion. Crucially, both approaches explicitly build their proposals on a distinction between descriptive content, on the one hand, and expressive content, on the other (both Rett 2011 and Grosz 2012 explicitly refer to the original discussion and concepts in Kratzer 1999 and Potts 2005a; see also Chapter 10 below).

7.1 REQUEST FOR INFORMATION VERSUS REQUEST FOR AFFIRMATION

125

This dichotomy between descriptive and expressive content is an aspect of those two approaches that I think also lies at the core of many other prominent accounts, some of which I will discuss further below. For instance, Zanuttini and Portner’s claim (2003) that the descriptive content of exclamatives (i.e., their proposition) can be characterized as a factivity presupposition (and not as assertive content) is based on a general view according to which there is a “difference in force between exclamatives and declaratives like It is surprising that …, which assert closely related content” (Zanuttini and Portner 2003: 56). Crucially, this view implies that exclamatives are somehow ‘special’, because their main illocutionary point is ‘expressive’ and not the assertion of descriptive content. This general thinking is formally articulated in the approaches of Rett (2011) and Grosz (2012).¹ As already mentioned above, one of the most articulated and clearest versions of the illocutionary force hypothesis has been formulated by Rett (2011). Her illocutionary account is motivated by the intuition that, at the level of illocutionary force, there is no difference between exclamatives (e.g., How fast he is!) and declarative exclamations (e.g., He is so fast!). This is an intuition that I share (see also my discussion in Chapter 5 above). However, while the consequence of Rett (2011) is to postulate an additional operator that can capture the similarities between those different utterance types, my claim will be that we can account for the relevant similarities without adding a force operator, and instead manage with the illocutionary force of assertion that we must postulate in our syntax–semantics anyway. In (4), we see Rett’s formulation (2011: 429) of her illocutionary operator E-FORCE, which is based on Kratzer’s account (1999) of German modal particles as expressive items (see also Chapter 10 at the end of this book). (4)

E-FORCE(p), uttered by sC , is appropriate in a context C if p is salient and true in wC . When appropriate, E-FORCE(p) counts as an expression that sC had not expected that p.

The modeling of the illocutionary force of exclamations in (4) includes a speaker in a Context (sC ) and a world of utterance (wC ), and the force of an exclamation expresses that some proposition p was not expected by the speaker sC .

¹ Another reason why I focus on these two approaches is that their empirical domain is the class of exclamations in general, and not only the narrow class of exclamatives like in most of the existing literature. Since, in this book, I am interested in exclamation-like properties of questions (and not in the sentence type ‘exclamative’), I believe my focusing on Rett (2011) and Grosz (2012) is justified.

126

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

The main point of Rett’s analysis (2011) now is that the input for the EFORCE operator is different in exclamatives (How fast he is!) and declarative exclamations (He is so fast!). In other words, exclamatives and declarative exclamations differ in their content. Let us first look at Rett’s analysis (2011: 430) of declarative exclamations, as in (5): (5)

(Wow,) John won the race! a p = λw.wonw (john, ιx[racew (x)]) b E-FORCE(p), uttered by sC , is appropriate in a context C if p is salient and true in wC . When appropriate, E-FORCE(p) counts as an expression that sC had not expected that p.

(5) illustrates that exclamations such as John won the race! denote propositions (5a), and the interaction of the illocutionary operator and p is thus straightforward (5b). When we now turn to exclamatives, Rett’s claim (2011) is that they denote degrees, not propositions (for reasons already discussed in Rett 2008). Accordingly, she proposes the following modeling, where we look at a twostep analysis containing a proposition with an unbound variable (6b), which is then bound via existential closure in (6c) so that the proposition D(d′) can function as the input to the illocutionary operator (Rett 2011: 431): (6)

How tall John is! a λd.tall(john, d) b tall (john, d′) c E-FORCE(p) counts as an expression that ∃d′ such that sC had not expected that D(d′).

As mentioned above, and as already made very clear in previous chapters of this book, I share with Rett (2011) the goal to provide a unified analysis for cases such as (5) and (6). However, my claim will be that both (5) and (6) can be analyzed as assertions at the level of illocutionary force, and so we do not need any additional operator such as E-FORCE in (4) above. More specifically, regardless of the question of force, the distinctive feature of all the different forms of exclamations—including our case of non-canonical question in (2)— is that they convey intense emotion. Assertions, crucially, can also convey such emotion, and in this case we are looking at declarative exclamations such as (5) above. The declarative exclamations are typically marked by interjections and an exclamation mark in written form—Wow, that was an amazing meal!—or a simple addition of so: He runs so fast! The literature, including Rett’s account (2011) sketched above, typically dissociates wh-exclamatives

7.1 REQUEST FOR INFORMATION VERSUS REQUEST FOR AFFIRMATION

127

from those declarative cases as well as from non-canonical questions expressing surprise (2), but I claim that there is no good argument for doing so. But, before we discuss the detailed consequences for question cases like (2), let us have a brief look at another approach adopting the illocutionary force hypothesis, which puts an emphasis on the non-truth-conditionality of exclamatives and thus, similarly to Rett (2011), distinguishes exclamatives from declarative exclamations by not acknowledging the assertive character of exclamatives. Like that of Rett (2011), Grosz’s account (2012) is a version of the illocutionary force hypothesis that provides a unified illocutionary operator (his ‘EX’) for both exclamatives and declarative exclamations. However, and also along the lines of Rett’s analysis (2011), he proposes that EX combines differently in the semantic composition of exclamatives, on the one hand, and in the composition of declarative exclamations, on the other hand. Let us first look at how Grosz (2012) analyzes exclamatives. In the case of exclamatives, EX combines with a truth-conditional expression of type ⟨s, t⟩ (i.e., a truth-conditional argument of propositional type; i.e., functions from indices ‘s’ to truth-values, where ‘t’ is the type of truth-values) and maps this proposition onto felicity conditions that detail the speaker’s attitude towards the proposition. Crucially, the resulting denotation is not truth-conditional, but, according to Grosz, ‘felicity-conditional’. In particular, he claims that application of EX to a proposition yields a one-dimensional meaning of type E (defined as the type of expressive meaning). Consider the following example and the representation in (7′); see Grosz (2012: 118): (7)

(7ˈ)

Boy, is it raining!

EX(rain) : E

rain : ⟨st⟩

EX : ⟨st, E⟩

Grosz claims that EX removes its propositional complement from the level of descriptive at-issue meaning, and shifts it to the level of expressive meaning. By contrast, Grosz continues, if one utters a declarative exclamation like (8), one still expresses a truth-functional statement (i.e., that it is raining to an extreme degree): (8)

It’s raining so much!

128

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

To me, and many native speakers, the declarative exclamation in (8) sounds every bit as ‘expressive’ as the exclamative in (7); it is therefore not clear empirically that the two categories must be distinguished. The only difference (and the most crucial one for many approaches) is claimed to be that the expressivity in one case seems to be about the degree of rain (in (7)) whereas in (8) it seems to be about the fact that it is raining to an extreme degree. Other than that, declarative and exclamative sentences seem to be expressing an amazement stance and can be thought of as variants of the same kind of attitude. Yet, in the approach to exclamations sketched in (7) above, declarative exclamations are assigned a multidimensional meaning (with an expressive and a descriptive, truth-functional) component, whereas exclamatives are said to lack assertive meaning. Following Potts (2005a, 2007a), we could say that the descriptive propositional part ⟨st⟩ combines with the expressive part ⟨st, E⟩, by means of the composition operator ‘•’, which combines both meanings, passes the descriptive content up, and interprets it relative to the context modified by EX (see also Gutzmann 2015 on several applications of such a ‘multidimensional’ semantics). Formally, the composition operator ‘•’ yields the following result (Potts 2007a: 187): (9)

⟦EX⟧C • ⟦rain⟧C = ⟦rain⟧⟦EX⟧ (⟦rain⟧)C

According to this multidimensional analysis, the operator EX passes the descriptive content of ⟦rain⟧ unchanged—that is, it is at-issue and it can be asserted or questioned. However, its context of interpretation is altered by the content of the operator. So, in contrast to the representation for exclamatives illustrated in (7), the result for declarative exclamations, according to Grosz’s approach (2012), is the following two-dimensional semantic object:

(10)

rain: ⟨st⟩

EX(rain) : E

rain : ⟨st⟩

EX : ⟨st, E⟩

But, as we said before, is there independent undisputed empirical evidence that exclamatives and declarative exclamations differ substantially or at all? In the literature, we find two central data points that are cited in favor of

7.1 REQUEST FOR INFORMATION VERSUS REQUEST FOR AFFIRMATION

129

distinguishing between the two (see also Chapter 5 above): (i) the descriptive content of exclamatives is said not to be denied, but the descriptive content of declarative exclamations can be denied; and (ii) exclamatives, unlike declarative exclamations, have been claimed not to be used as responses to information-seeking questions. Both of these data points can be questioned on empirical grounds (see Trotzke 2019 on point (i) and Trotzke and Villalba 2020 on point (ii)). When we now turn back to our cases of exclamatory questions above, I submit that what is called ‘exclamation’ in the literature actually refers to the ubiquity of emotive attitudes on the part of the speaker across all speech acts and clause types. Hence, and in addition to questions expressing surprise like (2) above, examples of exclamation would also include utterances such as metalinguistic comparatives (11a), declarative questions (11b), and emotionally intense imperatives (11c): (11)

a I’d rather die than marry him! (Giannakidou and Yoon 2011) b You are camping in the rain?! (Really?) c Finish that book, damn it!

Clearly, exclamation does not reflect a specific clause type. Instead, I propose that the wide range of exclamation can be better captured as an attitude of emotion (specifically, a propositional attitude, see Chapter 9) rather than ahard-to-define illocutionary force that maps onto a specific clause type. It seems implausible that exclamation (the illocutionary force assumed for exclamative) is a specific, distinct speech act because it appears to combine with other speech acts (with assertions in (11a), with questions in (11b), and with imperatives (11c)). The common feature of all the examples above seems to be an emotive attitude on the part of the speaker to whatever the speech act is. One way to make sense of question examples such as (2) and (8b) would be to say that speech acts can also be combined, resulting in more than one illocutionary force per utterance (in this case: question + exclamation). And, indeed, an analysis in terms of multiple speech acts has been proposed for ‘mixed cases’ such as rhetorical questions, some of which can be characterized as being both a question and an assertion—as in the following patterns involving discourse markers such as after all and tell me (see Asher and Reese 2007: 14 and Sadock 1971, 1974 for the original observation). While after all can co-occur with assertions (12a), it is infelicitous with questions (12b). For the expression tell me, it is just the other way around: tell me is felicitous with questions (13a), but it cannot co-occur with assertions (13b): (12)

a After all, your adviser is out of the country. b #After all, is your adviser out of the country?

130

(13)

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

a Tell me, does John own a car? b #Tell me, John owns a car.

If we now turn to rhetorical questions, we observe that they are fine with both discourse markers (Asher and Reese 2007: 18): (14)

a After all, does John lift a finger to help around the house? b Tell me, does John lift a finger to help around the house?

This can be taken as evidence for the existence of multiple speech acts. If we stipulate that speech acts can indeed be combined, we could say that our examples of non-canonical questions expressing surprise exemplify combining the speech act of an exclamation with that of a question. However, what does it mean to stack speech act operators in a sentence? As far as I can tell, such a stipulation would have to be unique for these cases, and the notion of a speech ‘act’ would have to be reduced to a structure akin to embedding of attitudes— which is well motivated and attested independently. More consistently with Occam’s razor, then, instead of stipulating speech act stacking, I would like to take data such as surprise–disapproval questions to indicate that there is no specific, distinct speech act of exclamation in the first place. In other words, I argue that both surprise–disapproval questions (2) and exclamatives convey the same emotive content as a corresponding declarative assertion (I am surprised that you are laughing now!). With this approach in mind, let us now turn to an explicit proposal of how to account for this discourse contribution of non-canonical questions to the common ground (CG). The question is how we can analyze the discourse move of non-canonical questions that express surprise and/or disapproval (as I have illustrated above). In what follows, I will illustrate this based on the discourse model by Farkas and Bruce (2010) to account for these issues (see Chapter 2 on some basics on how to model questions within that model). Specifically, I will adopt the ideas by Chernilovskaya (2014) on the discourse contribution of exclamatives, which are proposed within this kind of discourse–semantic framework. However, I will add an important qualification: in contrast to Chernilovskaya (2014), I will claim that exclamatives and thus also non-canonical questions featuring a surprise component can consist of a combination of two operators and thus have two functions as far as contribution to the CG is concerned. In order to show that, let us first sketch the notion of the assertion speech act operator according to the framework by Farkas and Bruce (2010). Speech act operators in this discourse-oriented semantic framework can be defined in terms of how they change context states (in what follows: C).

7.1 REQUEST FOR INFORMATION VERSUS REQUEST FOR AFFIRMATION

131

These context-change effects then account for the contribution of the respective speech act to the CG. Given this conceptual background, look at the following representation of the assertion speech act operator (see Farkas and Bruce 2010: 92). For our purposes, we adopt the simplified notation by Chernilovskaya (2014: 100): A is Speaker A, QUD are ‘Questions under Discussion’ (see Roberts 1996, 2012), DC refers to ‘Discourse Commitment’, and CG∗ is the set of future common grounds, modeled as a projected set according to Farkas and Bruce (2010): (15)

ASSERT(A)(p)(Cinput ) = Coutput , where Coutput = Cinput , except: a QUDoutput = push({p},QUDinput ) b DCA,output = DCA,input ∪ {p} c CG∗ output = CG∗ input ∪{p}

In (15), we see that the assertive operator is defined as a three-place function: it takes Speaker A, the denotation of a sentence p, and the input context state. The result is then mapped to the output context state. The effect of ASSERT is then that p becomes a new issue under discussion (15a), Speaker A becomes committed to p (15b), and p is projected to future common grounds (15c). Now, how can we define the speech act operator for expressions that convey surprise such as surprise–disapproval questions? Chernilovskaya (2014: 117) proposes the following definition for exclamations: (16)

EXCL(A)(p)(Cinput ) = Coutput , where Coutput = Cinput , except: CGoutput = CGinput ∪{p}

The most important component of this definition is that the CG is directly updated with the content p expressed by the exclamation, and so there is no possibility for the addressee to object to it. In addition, and related to this central component, there is no update of the QUD component, which means that an exclamation, according to Chernilovskaya (2014), raises no issues that have to be settled. Also, the operator does not update future common grounds (CG∗ in (15) above) and so it does not require the addressee to react and/or to be committed to the content of the exclamation. Note that the content p conveyed by an exclamation is meant to be the expressive (and not also the presupposed/descriptive content) here. This distinction will become important for our claim on the multidimensional nature of non-canonical questions below. According to what we have said above, I now argue that the expressive component of non-canonical why-like-what questions instantiates assertive force and thus patterns with declarative exclamations such as I am surprised

132

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

that you are laughing now (of all times)! But how can we spell this out formally? Chernilovskaya (2014: 131–4) suggests the following combination for declarative exclamations: (17)

ASSERT – EXCL(A)(p, q)(Cinput ) = Coutput , where Coutput = Cinput , except: a QUDoutput = push({p},QUDinput ) b DCA,output = DCA,input ∪ {p} c CG∗ output = CG∗ input ∪{p} d CGoutput = CGinput ∪{q}

(17) makes an attempt at expressing that content p is asserted (thus proposing to update the CG), while, at the same time, the expressive content q conveys a noteworthiness evaluation and, according to the representation in (17) above, directly enters the CG. However, as noted by Chernilovskaya (2014) herself, there is a well-known problem with this combining of the two speech acts if one accepts the idea that the exclamation interpretation (here modeled as a noteworthiness evaluation) is factive. Then the proposition p (e.g., that the addressee is laughing now) must already be in the CG and cannot be part of the assertion, that is, of the proposal to update the CG. In other words, since q, which updates the CG directly (17d), presupposes p, an analysis that proposes that p is asserted does not hold. In what follows, and in contrast to Chernilovskaya (2014), I will propose that this major drawback of such an analysis disappears as soon as one adopts the general idea that surprise–disapproval questions are actually two-dimensional semantic objects. To see this point, we must also take into account data where non-canonical questions can be used as second moves in a dialogue. This will shed some light on the non-emotive (‘descriptive’) content they contribute to the common ground. In particular, since non-canonical questions often carry presuppositions that can potentially convey information asked for in a dialogue sequence, those questions can potentially be used as responses to other questions (e.g., A: Who came to the party? B: Who didn’t? Intended to convey: Everyone did). In the next chapter, I will connect these data points to my observations above and, based on the syntactic theory of other response strategies, develop an analysis that can account for the indirect ways in which non-canonical questions can be used as second moves in a dialogue. But, before we turn to second-move aspects, let us now discuss some of the syntactic consequences of the pragmatic observations sketched in this chapter.

7.2 SOME SYNTACTIC CONSEQUENCES

133

7.2 Some syntactic consequences Recall that we have sketched above that Trotzke and Czypionka (2022) discovered—quite expectedly—that the favored response to an informationseeking question is an answer, with affirmations being distinctly less preferred. The favored reaction to exclamatives is affirmation, which aligns with the theoretical literature on exclamatives demonstrating that, although an exclamative cannot be answered (19B), the descriptive content of the exclamative (which is presupposed) can be confirmed by the addressee (19B′); see Castroviejo Miró (2006); Chernilovskaya et al. (2012); Grimshaw (1979); Zanuttini and Portner (2003). The following pattern is adopted from Zanuttini and Portner (2003: 47): (19)

A How very tall he is! B ∗ Seven feet. B′ He really is! Indeed!

Intriguingly, Trotzke and Czypionka (2022) found that, while ratings for firstmove exclamatives with affirmations were considerably higher than those for exclamatives with answers, the latter still received higher ratings than anticipated. Notably, the exclamative-answer pattern implies that this response is perceived as more suitable than the information-seeking/affirmative condition. This could indicate a bias in the experimental paradigm towards an interrogative interpretation. An alternative explanation is that participants perceive the descriptive content of an exclamative as still being ‘at-issue’ and not presupposed, along the lines of what we have proposed in the previous section. In other words, participants might have believed that the speaker does not have a firm conviction about the proposition conveyed by an exclamative (e.g., ‘He bakes delicious pies’ in an exclamative like What delicious pies he bakes!), and that the proposition is thus still ‘at-issue’ (see Villalba 2017b for such an account) and not as far in the conversational background as traditionally assumed in the theoretical literature (e.g., Zanuttini and Portner 2003 and their analysis of the relevant proposition in terms of a factivity presupposition). The results for surprise–disapproval questions reveal that both types of reactions (i.e., answers and affirmations) are considered appropriate. Although there is a minor descriptive preference for answers over affirmations, the difference between the two reactions is only marginally significant. This suggests that, in the case of surprise–disapproval questions, there is no clear preference for either reaction type—neither for answers nor for affirmations.

134

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

This finding implies that surprise–disapproval questions may occupy an intermediate position between information-seeking questions and exclamatives. On the one hand, they can elicit answers similar to information-seeking questions, indicating that the speaker is looking for information. On the other hand, they can also elicit affirmations, which is more typical of exclamatives, reflecting the speaker’s emotional stance or attitude towards the situation. In summary, the study on the pragmatics of surprise–disapproval questions as first moves in a dialogue highlights the complexity and versatility of these questions in conversational contexts. They can function both as informationseeking questions and as expressions of surprise or disapproval, allowing for a broader range of appropriate reactions from the addressee. This underscores the importance of considering the various factors that may influence the interpretation and usage of such questions in real-life communication. The experimental results by Trotzke and Czypionka (2022) support the idea that surprise–disapproval questions are truly ‘mixed types’ in terms of their illocutionary status and syntactic properties. This suggests that we do not need to posit a separate illocutionary category or syntactic projection to account for them theoretically. Instead, these questions may exhibit properties of both information-seeking questions and exclamatives, which can be explained on syntactic grounds by broad illocutionary theories like Truckenbrodt’s account (2006). According to Truckenbrodt’s theory, the V-to-C movement property in German wh-questions results in a unique pragmatics that verb-final exclamatives lack. Recall that the why-like-what cases of surprise–disapproval questions feature V-to-C movement (e.g., Was lachst Du?! ‘Why are you laughing?’), while exclamatives do not (e.g., Was der lacht! ‘How he is laughing!’). This syntactic similarity between surprise–disapproval questions and information-seeking questions (e.g., Was liest Du? ‘What are you reading?’) might be the reason why surprise–disapproval questions are perceived as more felicitous with answers than wh-exclamatives. As for the syntactic analysis of such questions, recall that we have seen a similar case in Chapter 5 when we discussed English How cool is that?!, which also features interrogative syntax, while being interpreted as an exclamation. The derivation for why-like-what surprise–disapproval questions, according to the analysis in Chapter 5 for this parallel case, can thus be represented as follows:

7.2 SOME SYNTACTIC CONSEQUENCES

135

(20) was ‘why’ C0/lachst ‘laugh’ was ‘why’ Emp0[intensity] [TP... Du lachst was ‘you laugh why’]

Example (20) outlines the idea that surprise–disapproval questions featuring why-like-what can be considered as interrogatives with an additional exclamation component. This additional component can be analyzed in a similar way to other cases where phrasal constituents receive an intensification reading by fronting them to the specifier of EmpP (Emphasis Phrase) before further movement is required for clausal typing purposes. Again, according to my theory, it is important to note that EmpP represents a syntactic layer that is distinct from the functions of clausal typing and illocutionary force. The whylike-what originates in its adverbial position inside the IP/VP, because those questions are asking for a reason. However, owing to its intensity reading, why-like-what—just as [how cool] in our analysis in Chapter 5—must move to the left periphery to check its Emp feature and cannot stay in situ (cf. the ungrammatical Du lachst was?). This analysis also implies that why-like-what takes propositional scope, and this makes sense if we assume that the surprise reading of those questions instantiates a wide focus (also known as sentence focus) rather than a narrow focus as in the canonical/regular question reading (e.g., German Was liest Du denn jetzt? ‘What are you reading now?’ intended reading: ‘Why are you reading now [of all times]?’ and not: ‘Which book are you reading now?’). I think this claim provides a new and innovative theoretical account that indeed allows me to postulate that the same information-structural mechanisms hold for constructions featuring non-argumental ‘what’ and other options that might have a similar interpretation owing to non-syntactic factors such as intonation (e.g., in an English surprise question like WHY are you still in bed? the speaker’s surprise does not refer to the concrete reason why x is still in bed, but rather refers to the broader focus ‘that x is still in bed’).

136

NON-CANONICAL QUESTIONS AS FIRST MOVES IN A DIALOGUE

Given this syntactic proposal for surprise–disapproval questions and the relevant ‘mixed’ pragmatics we have sketched in this chapter, let me now summarize the previous sections.

7.3 Summary This chapter has looked at some discourse properties of non-canonical questions as first moves in a dialogue. In particular, I have first introduced the difference between a request for information (as in ‘canonical’ informationseeking questions) and a request for affirmation (as in many subtypes of non-canonical questions where the speaker wants the addressee to affirm or confirm their bias and their epistemic and/or emotional state). I have explored the topic of at-issue versus not-at-issue content in the context of how an addressee might react to a particular subtype of non-canonical questions— namely to surprise–disapproval questions. After a discussion of the pros and cons of postulating dedicated illocutionary operators for the exclamation speech act (representing surprise readings across speech acts), I have sketched a syntactic account of why-like-what surprise–disapproval questions that explains both their mixed behavior in the domain of discourse properties when used as first moves in a dialogue and their scoping over the whole proposition of the clause. I now go one step further and ask whether, and if so how, such non-canonical questions can also be used as second moves in a dialogue.

8 Non-canonical questions as second moves in a dialogue After our discussion of the pragmatic status of non-canonical questions as first moves in a dialogue, this chapter turns to the question to what extent they can be used as second moves in a dialogue. Intuitively, since non-canonical questions often carry presuppositions that can potentially convey information asked for in a dialogue sequence, those questions can potentially be used as responses to other questions (e.g., A: Who came to the party? B: Who didn’t? Intended to convey: Everyone did). I will connect these data points to the syntax of other response strategies (like anaphoric response particles) and develop an analysis that can account for the indirect ways in which non-canonical questions can be used as second moves in a dialogue via pragmatic inferences. Before we dive into the details, let us briefly look at the phenomenon of second moves in a dialogue more generally. Needless to say, possible second moves in a dialogue depend on the first move, the context of the conversation, and the goals and intentions of the speakers. The second move is the response to the first move, and it can take many different forms depending on the speech act used as the first move. Let me give some examples of possible second moves in a dialogue (see Lee 2012; Stivers et al. 2009 for comprehensive and cross-linguistic overviews). A typical second move in response to an opening statement like greeting someone is to reciprocate the greeting or to acknowledge the topic established in the opening statement. For example, Hi, I’m doing well, thanks for asking. How about you? or Yes, I’d love to talk about that topic. The second move in response to an information-seeking question of course typically is to provide the requested information. For example, The nearest coffee shop is just down the street or I’m from New York. The second move in response to a commissive speech act is to acknowledge the commitment or to provide further information about the action. For example, That would be great, thank you or Let’s discuss the details. There can also be responses to expressive speech acts.

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0008

138

NON-CANONICAL QUESTIONS AS SECOND MOVES IN A DIALOGUE

For instance, the second move in response to an expressive speech act could be to acknowledge the emotion or to reciprocate the feeling. For example, I’m happy to see you too or I’m sorry to hear that (see also our remarks on the pragmatic move of affirmations in the previous section). In principle, the possibilities for second moves in a dialogue are endless. In other words, there is no set or limited set of second moves that must be used in response to a particular first move. However, there are certain types of second moves that are more common or expected based on the first move or the context of the conversation. For example, and relevant in our context, in response to an information-seeking question, a typical second move might be to provide the requested information, and not to pose a counter-question, perform an imperative, and so on. There are several theories and frameworks in linguistics that are specifically concerned with second moves in a dialogue. Most notably, second moves in a dialogue are a prominent topic in conversation analysis, which is an approach to the study of natural conversation, focusing on the interactional structure and social context of talk (see Hall 2019; Liddicoat 2021 for recent overviews). Conversation analysis examines the organization of turns, sequences, repair mechanisms, and other conversational phenomena. In studying second moves in a dialogue, conversation analysis investigates how speakers manage turn-taking, respond to previous utterances, and establish coherence in conversation. Within this branch of linguistics, interactional sociolinguistics is an approach that combines insights from sociolinguistics and conversation analysis to study language use in social interaction. It examines how social identities, relationships, and norms are constructed and negotiated through linguistic practices (Rampton 2019). Analyzing second moves in a dialogue within this framework can reveal how speakers respond to each other’s utterances while also navigating social roles, power dynamics, and cultural expectations. Analyzing second moves in a dialogue has also become a prominent topic in computational linguistics. Specifically, in the context of dialogue systems, computational linguistics is concerned with developing algorithms and models to understand and generate appropriate responses in a dialogue (see Ginzburg and Fernández 2010; Jurafsky 2004; McTear 2020; and many more). Analyzing second moves in a dialogue can help improve the performance of natural language processing systems and develop more effective dialogue systems, such as chatbots and virtual assistants. Given these broad aspects of second-move utterances in a dialogue, let us now turn to non-canonical questions in this context.

8.1 ANSWERING AND RESPONDING VIA QUESTIONS

139

8.1 Answering and responding via questions In this section, I turn to some of our cases of non-canonical questions, especially those that convey not-at-issue content, as illustrated in the previous chapter for surprise–disapproval questions. I choose to look at those expressive question types, because it is much more trivial to see that rhetorical questions can be used as responses in a dialogue. Specifically, rhetorical questions are questions that are very often asked for asserting a proposition, rather than to elicit an answer. That is, since they usually serve to make a statement or convey a point, the speaker does not expect a direct answer to the question. In a dialogue, rhetorical questions can function as a way indirectly to express one’s opinion, challenge a statement, or prompt the other person to reflect on their statement or question. Let us briefly look at some examples: (1)

a A: I can’t believe he finished the project so quickly. B: Yeah, who needs sleep, right? b A: I’m thinking of buying that expensive car. B: Do you really need another car? c A: She said she doesn’t like the new policy. B: Well, who does?

In each of these examples, the rhetorical question serves as a response (hence a second move) in the dialogue, highlighting the speaker’s attitude or opinion about the situation being discussed. The use of rhetorical questions as responses can contribute to the conversational dynamics and enhance the expressiveness of the interaction. Crucially, the latter aspect of using non-canonical questions as responses is more pervasive in responses that have the form of emphatic questions, such as our cases of surprise–disapproval questions discussed in the previous chapter. To see this, we can compare their discourse behavior to the appropriateness of responses in specific dialogue sequences involving exclamatives. As in Chapter 7, we consider ‘response’ to include direct answers to informationseeking questions as well as other more indirect moves in a dialogue (e.g., through pragmatic inferencing or implicatures; see de Marneffe and Tonhauser 2019; Holtgraves 1998; Walker et al. 2011). This will become evident shortly. Grimshaw (1979: 321) famously claimed that exclamatives are consistently infelicitous as responses to questions. The following is her well-known example:

140

NON-CANONICAL QUESTIONS AS SECOND MOVES IN A DIALOGUE

(2)

A: How tall is John? B: #How tall John is! B′: John is very tall.

The wh-exclamative seemingly conveys the same descriptive content as the declarative assertion in (2B′) (though it might be more precise to say that it is equivalent to John is extremely tall, an end-of-scale degree). However, the argument states that, while declaratives are appropriate responses, exclamatives are not; thus, exclamatives lack assertive content. Zanuttini and Portner (2003) and numerous others further argued that the sole descriptive content of exclamatives is a factivity presupposition, and a presupposition cannot serve as a response offering new information. It is crucial to highlight that claims such as Grimshaw’s, based on (2), are centered on the syntactic form of an exclamative, rather than on exclamation speech acts in general, which would also include our cases of surprise– disapproval questions. Note, however, that an exclamatory declarative in the form of John is so tall! (2B′) is also an imperfect answer: (2′)

A: How tall is John? B: #How tall John is! B′: #John is so tall!

If we rely solely on Grimshaw’s account (1979), we might anticipate that the declarative form in (2B′), even when utilized as an exclamation, should be acceptable. However, it is unclear why the exclamative in (2B) should be ruled out, given that both the exclamative and the declarative exclamation can convey the same kind of expressivity towards the identical content. Crucially, I will argue, the capability or inability to respond to a question does not necessarily serve as evidence for assertive force or its absence. Many questions, for instance, can be addressed by other questions—often rhetorical—or imperatives: (3)

A: Who came to the party? B: Who didn’t? (Intended to convey: Everyone did). B′: Ask Mary. (Intended: She will tell you).

Being a response to a question and being an assertion are not the same thing, and Grimshaw’s initial argument needs to be seen in this light.¹ One could argue, of course, as Giannakidou and Mari (2022), Han (2002), and Sadock ¹ One might raise the issue whether the responses in (3) really constitute answers to the question. It seems clear that they do. In (3B), the content conveyed is that there were a lot of people, and in (3B′) that I do not know the answer to the question. The notion of content includes both what is asserted and what is implicated, and a response—as opposed to an ‘answer’—provides content. My point here is that, if we want to be consistent about the fact that responding to a question is not merely asserting a direct answer, we cannot use an unqualified notion of ‘answer’ as a diagnostic for assertion.

8.2 THE SYNTAX OF RESPONSES AND PRAGMATIC INFERENCING

141

(1971) do, that rhetorical questions like (3B) are equivalent to assertions semantically, in which case the clause type (question) does not determine the semantic value or discourse function. And, if that is the case, then the distinct clause type of exclamative does not necessitate a distinct semantic or pragmatic type either. Secondly, while the previous literature has mainly focused on the type of responses, I would like to explore the type of questions preceding the respective responses instead. When we do that, we find that not only exclamatives can be used to respond to information-seeking questions (see Trotzke and Villalba 2020 for experimental evidence), but also expressive questions—as long as they do not yield a mismatch at the level of information structure. Look at the following patterns: (4)

a A: B: B′: b A: B: B′:

Do you think it’s cool? This is [so cool!]F declarative #[How cool is that?]F expressive question What do you think? [This is so cool!]F [How cool is that?]F

The declarative (4a) is a perfect response to a narrow-focus question. The expressive question is odd in the same context, similarly to what has already been pointed out in the literature on exclamatives. Now, contrast (4a) with (4b), where the preceding question (i.e., the information-structural context) is not directly about the degree of coolness. We see in (4b) that, when the information-structural context is changed to a broad-focus wh-question, the expressive question in its use as a response has clearly improved. I claim that those questions are fine in such a context, because they typically express an all-focus interpretation. These patterns suggest that not only rhetorical questions, but also expressive questions can indeed be used as responses to information-seeking questions, provided that they occur in a relevant information-structural context. With this possibility in mind, let us now turn to some syntactic consequences of that observation.

8.2 The syntax of responses and pragmatic inferencing In this section, I will explore how the findings sketched in the previous paragraphs can be incorporated into a syntactic framework for analyzing dialogue moves. I would like to argue that the data above hold significance for the analysis of an empirical area at the intersection of syntax and pragmatics, often

142

NON-CANONICAL QUESTIONS AS SECOND MOVES IN A DIALOGUE

referred to as ‘the syntax of answers’ in recent research (Cheng 2022; Espinal and Tubau 2019; Holmberg 2013, 2015). I will now illustrate how the findings in the domain of expressive questions can contribute to this current branch of syntactic work. In recent research, numerous approaches have been proposed to represent the syntax of answers at the syntax–pragmatics interface. One of the most notable theories is Holmberg’s framework (2013, 2015) addressing the syntax of yes and no. Holmberg’s approach is particularly appealing, as it posits that the syntactic derivation of polar questions can be modeled in a way parallel to the derivation of wh-questions; accordingly, we can potentially say something more general about answering patterns by not being confined to just one of the possible contexts answers can appear in. His theory is grounded in the general framework of cartographic syntax, which we have already applied to some of the phenomena in previous chapters and which represents discourse-related meaning, such as focus, topic, and force, in the left periphery of the clause (Rizzi 1997, 2014). Although this model has been a point of contention even among advocates of generative syntax, I hypothesize that Holmberg’s model— along with its benefits mentioned above—is an appropriate framework for interpreting the empirical patterns presented in the previous section. In particular, Holmberg (2013) analyzes an answer—either affirmative (6a) or negative (6b)—to a polar question like Is he coming? as follows:

(5)

FocP

{yes[Aff ], no[Neg]}

Focˈ

PolP

Foc0

he

{is+[Aff ], isn’t+[Neg]}

Polˈ

TP

is he coming

8.2 THE SYNTAX OF RESPONSES AND PRAGMATIC INFERENCING

143

The fundamental idea in (5) is that the answer particles yes and no reside in the specifier of a focus phrase (providing the ‘new’ information), and, based on their polarity (affirmative or negative), they assign their value (Aff or Neg) to a corresponding sentence-internal polarity feature (Pol), resulting in either an affirmative or a negative assertion. The placement of answer particles in a focus phrase is the key aspect that allows Holmberg to model polar questions parallel to wh-questions. In the case of a wh-question like Who is coming?, the answer term (e.g., Mary) serves as the focus constituent. Consequently, the answer term Mary assigns its focus value to a variable within the propositional part of the utterance (for parallels and differences between wh-questions and other focus-marking constructions, see Bocci et al. 2021; Eckardt 2007): (6)

[FocP Mary[Foc] [TP Mary+[Foc] is coming]]

Keeping this information-structural method for the syntax of answers in mind, we can now explain the data discussed earlier. Let us examine the case of expressive questions discussed in Section 8.1 in more detail. We noted that these questions are favored as responses in a discourse when they come after broad wh-questions. Importantly, as already noted, this seems to contradict the parallel observation of Grimshaw (1979) that wh-exclamatives can never be employed as responses to wh-questions. Below are her classic examples again (Grimshaw 1979: 321): (7)

A: How tall is John? B: #How tall John is! B′: John is very tall.

While I concur with her basic observation, I have argued that Grimshaw’s example (7B) is not successful, not because it includes an exclamative per se, but owing to an information-structural mismatch. Observe again that her example has a narrow focus (‘how tall’), and the response in (7B) is meant to give a direct answer to this narrow-focus question. However, as soon as we use a wh-question with broader focus, the corresponding answers inferred from the exclamative responses can be less specific. In other words, the broad-focus questions in our cases discussed in Section 8.1 above permit the responses to be more indirect, which is why both exclamatives and expressive questions in those broader contexts can be used as responses to a question, contrary to Grimshaw’s general claim (1979). The indirect way of providing a direct answer using an expressive question can be illustrated like this:

144

NON-CANONICAL QUESTIONS AS SECOND MOVES IN A DIALOGUE

question: What do you think? (8) broad-focus broad-focus question: a How cool is that?! b ‘It’s extremely cool.’ [descriptive content conveyed by (8a)] In (8), it becomes evident that an expressive question (8a) conveys descriptive content (8b) from which the answer to the broad-focus question can be deduced. As a result, the alignment of the descriptive content of the expressive question with the broad-focus question is a pragmatic one (i.e., the answer can be deduced). Considering the syntax of answers we discussed previously (5)/(6), we can examine this difference by expanding Holmberg’s account to include our data. Specifically, I propose that the wh-element how in non-canonical questions such as How cool is that?! binds a (degree/d) variable within the proposition (9); the interpretation of the descriptive content of the question as a response to a broad-focus question involves pragmatic inferencing, as mentioned earlier. In particular, the relevant answer proposition can be inferred from the literal utterances in (9): (9)

broad wh-question: What do you think? I think it’s extremely cool. PRAGMATIC INFERENCING (inferred p)

question: [FocP[how cool][Foc] [TP is that d cool + [Foc] ]]

‘descriptive content/proposition: ‘It’s extremely cool.’ My theoretical suggestion is based on the observation that any end-of-scale degree reading is always a part of the descriptive content of the syntactic structure (see the following Chapter 9 on more details). As a result, I propose that the combination [how cool] expresses a degree reading (‘extremely cool’; see Chapter 5), and that this degree reading must be a part of the propositional content of the utterance, which, according to the syntactic approach I adopt here, is represented within the sentence-level phrase ‘T(ense)P(hrase)’. Consequently, elements like [how cool] in expressive questions bind a degree variable within the proposition. It is evident that our cases of expressive questions

8.3 SUMMARY

145

are not direct responses like It’s d cool or It’s not d cool. Instead, they enable pragmatic inferences that supply the relevant responses as shown in (9) above. The brief discussion in this chapter demonstrates that the perspective on non-canonical questions from a discourse standpoint must be expanded: First, it is not accurate to say that they are infelicitous as responses. Second, we can even observe intriguing new distinctions at the syntax–pragmatics interface when examining their response uses.

8.3 Summary Given the discussion of how non-canonical questions function as first moves in a dialogue in Chapter 7, this chapter turned to the question of how they can be used as second moves. I have first pointed out that non-canonical questions often carry presuppositions that can potentially convey information asked for in a dialogue sequence, and thus those questions—notably rhetorical questions—can potentially be used as responses to other questions (e.g., A: Who came to the party? B: Who didn’t? Intended to convey: Everyone did). I have then looked at the response use of expressive questions in more detail and connected the relevant data points to the syntax of other response strategies (like anaphoric response particles). On this basis, I have indicated an analysis that can account for the indirect ways in which non-canonical questions can be used as second moves in a dialogue via pragmatic inferences. Clearly, the examples discussed in this chapter are unconventional and indirect ways to provide an answer; however, they function similarly to typical cases of relevance implicatures already discussed in foundational work by Grice (1975) and more recently explored by Walker et al. (2011) and de Marneffe and Tonhauser (2019). Specifically, the content of the answer is deduced from the ‘at-issue’ meaning of the utterance through general conversational principles: Since the expressive questions in the particular contexts discussed above violate the so-called maxim of relevance (i.e., it is not directly relevant to Speaker A’s wh-question), it does not align with the broader cooperative principle: “I expect a partner’s contribution to be appropriate to immediate needs at each stage of the transaction” (Grice 1975: 47). From a purely pragmatic perspective, the specific syntactic forms of indirect responses that rely on such pragmatic inferencing should not significantly impact the response’s appropriateness. As a result, the primary concern should be the propositional content of the response (which generates the relevant implicatures) rather than the syntactic structure in which it appears. More

146

NON-CANONICAL QUESTIONS AS SECOND MOVES IN A DIALOGUE

broadly, I hope that, by incorporating unconventional ways such as noncanonical questions as responses into the syntax of answers, I have expanded the empirical scope of recent work in this area at the syntax–pragmatics interface. Additionally, my hope is to inspire further exploration into the distinctions between different syntactic forms used in dialogues by examining how they can be employed in conversational sequences.

PART I V

A NEW FR A MEWOR K F OR NON-C A NONIC A L QUESTIONS

9 Emotional intensities in non-canonical questions The closing part of the book draws some general conclusions for a new theoretical approach to non-canonical questions, and it tries to connect the various analyses for different syntactic phenomena discussed in the book by providing a general framework where they all might fit in. In particular, we have seen throughout the book that some subtypes of non-canonical questions indeed require a dedicated syntactic analysis because their special readings are not (only) due to non-syntactic factors such as prosody, but rather involve noncanonical word order, special wh-elements, and modal particles (Part II of the book). Moreover, we have observed in Part III of the book that the pragmatics of many types of non-canonical questions can be compared to the discourse behavior of exclamation speech acts, which means that those non-canonical questions are expressive speech acts, while still featuring directive illocutionary force. Given these observations, the following chapters will explore in more detail the role and notion of emotional intensities, expressivity, and question force within a model of human communication. In the present chapter, I first deal with the interplay between information structure and the expression of emotional potentials of utterances and argue for a theory where expressing emotion is a syntactic intensification strategy, often realized by certain movement and fronting patterns, but also by well-known intensification strategies such as augmentation, duplication, and iconicity. According to this view, the grammar of many of the subtypes of non-canonical questions discussed in previous chapters merely signals the emotional involvement of the speaker, and the specific emotion is conveyed by pragmatics and/or lexical means. As I have done for other domains investigated in this book, I will then summarize and conclude with the more general question whether those emotional interpretations should be represented in the syntactic architecture of the utterance. I will outline an analysis where the syntactic expression of information structure essentially refers to the same concepts needed to account for the expression of an emotional potential of an utterance. Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0009

150

EMOTIONAL INTENSITIES IN NON-CANONICAL QUESTIONS

9.1 The expression of intensity in language The linguistic expression of intensity is closely related to the expression of emotions, as intensity can be used to convey the speaker’s emotional state or the intensity of their emotional response to a current state of affairs. For example, the use of intensifiers such as very, extremely, or utterly can be used to convey the intensity of an emotion. Simply put, saying I am very happy conveys a stronger emotional state than saying I am happy. Similarly, the use of emphatic stress or tone can be used to convey the intensity of an emotion, which is known as expressive speech in the literature and can feature a designated set of prosodic properties (Niebuhr 2010). For us, it is interesting to note that the use of marked information structure constructions can also be used to convey the intensity of an emotion. By placing emphasis on certain elements of a sentence, speakers can convey the importance or significance of those elements in relation to their emotional state. For example, fronting the object of a sentence can be used to place emphasis on it, as in That cake, I really want to eat it, which can convey a strong desire for the cake. In what follows, I will illustrate to what extent the linguistic concept of ‘degree’ plays a crucial role in the context of language and emotion, and how it can refer to the intensity or strength of an emotion that is conveyed through language. Degree can be expressed through a variety of linguistic means, including the use of adjectives, adverbs, intensifiers, and other linguistic constructions. At first sight, it covers a different empirical area than the highlighting of linguistic elements by means of marked information structure. However, in this section I will argue that both domains are actually part of the same fundamental aspect of human language, which I will label ‘emphasis’. By using this concept, we will have a unified way to cover a variety of intensities in language. At a very abstract level, the term emphasis refers to the process by which a linguistic element is given special prominence, value, or importance in a context. There has been significant linguistic research exploring the various mechanisms used to convey emphasis across languages and constructions. A key area of focus for linguists has been instances where emphasis is achieved contrastively, meaning that an element is highlighted by contrasting it with and distinguishing it from a set of alternatives (Repp 2010, 2016; Wa¨lchli 2023). One classical example of achieving emphasis is through non-canonical word order: (1)

It was Searle’s book that I bought.

(Searle’s book ≠ {Chomsky’s book, …} )

9.1 THE E XPRESSION OF INTENSIT Y IN L ANGUAGE

151

When examining the emotional interpretations of numerous instances of non-canonical questions, we have already seen a distinct type of linguistic emphasis, which I referred to as emphasis for intensity in some of the previous chapters. I would now like to look at this concept in more detail. Specifically, I propose that emphasis for intensity differs from contrastive emphasis, because it highlights linguistic elements, not by directly contrasting them with alternatives, but by selecting them from the highest range of a scale. Look at the following examples from English and German: (2)

Eliud is extremely fast!

Scale = speed of x

(3)

Was lachst Du denn jetzt?! Scale = likelihood of p what laugh you PART now ‘Why are you laughing right now (of all times)?!’

Examples (2) and (3) already indicate that I argue that the semantics of lexical strategies of emphasis (2) and the syntactic strategies (3) basically work the same. To see that and to understand the core aspects of the relevant degree semantics, let us briefly look at emphasis for intensity by lexical means. Intensifiers make up a distinct group of expressions that communicate emphasis for intensity by compositionally choosing elements found in the upper portion of a scale. In what follows, we can differentiate between intensifiers that target a scale present in the denotation of a predicate and those that target a scale embedded in the speaker’s emotional or epistemic attitude toward the proposition (see Beltrama and Trotzke 2019). Take a look at the examples below: (4)

Eliud is extremely fast.

(5)

I’ll make the tuna salad, and you make the SALAD-salad. (Ghomeshi et al. 2004: 308)

(6)

Absolutely all the townspeople are asleep. (Lasersohn 1999: 534)

At first sight, the intensifiers above impact the adjacent predicate in various ways: increasing the degree to which a property is present (4), indicating its proximity to a prototype (5), or refining its interpretation (6). Despite their differences, these modifiers consistently cause a narrowing of the modified predicate’s denotation, excluding elements that would have been included otherwise. This is demonstrated by the asymmetric entailment relationship

152

EMOTIONAL INTENSITIES IN NON-CANONICAL QUESTIONS

existing between the intensified and non-intensified forms for all three constructions: If an individual is part of the intensified form’s denotation, it will necessarily be part of the non-intensified form, but the reverse is not true: (7)

a Eliud is extremely fast. ⇒ Eliud is fast. b You made a SALAD-salad. ⇒ You made a salad. c Absolutely all the townspeople are asleep. ⇒ All the townspeople are asleep.

(8)

a Eliud is fast ⇏ Eliud is extremely fast. b You made a salad. ⇏ You made a SALAD-salad. c All the townspeople are asleep ⇏ Absolutely all the townspeople are asleep.

Another similarity across these examples is that the denotation’s contraction is achieved by arranging its elements along a specific dimension, with the lowest-ranked elements being removed and the highest-ranked ones retained. Beltrama and Trotzke (2019) have proposed that this mechanism is what grants the retained elements exceptional prominence, generating emphasis; the distinctions among these constructions, on the other hand, lie in the types of orderings they target. Specifically, predicates such as fast are deemed linguistically gradable: their denotation is inherently ordered based on the degree to which each individual embodies the property. Intensifiers directly utilize this ordering to create a partition in the denotation, maintaining—and emphasizing—only the elements that exhibit the property to the highest degree. Intensification can also apply to expressions that are not linguistically gradable. Let us examine two examples: reduplication and precisification. The unique aspect of these cases is that intensification does not allow for an interpretation where the relevant property is present ‘to a high degree’. (9)

I’ll make the tuna salad, you make the SALAD-salad. ≠ a salad to a high degree

However, the above modifiers still convey emphasis by triggering a similar ordering-based partition of the denotation. In the case of reduplication in (9), expressions like salad are not typically viewed as linguistically gradable (see Morzycki 2009, 2016 for more discussion on gradability encoding in the nominal domain; and Hohaus and Bochnak 2020 for further proposals). Yet, these forms can be considered conceptually gradable: their denotation can be ordered based on how closely the objects in the denotation resemble a prototypical instance of the category (e.g., a ‘green salad’; Ghomeshi

9.1 THE E XPRESSION OF INTENSIT Y IN L ANGUAGE

153

et al. 2004). According to this perspective, reduplication narrows the denotation of the target expression to “a more precisely defined, more specialized range” (Ghomeshi et al. 2004: 308), which disregards conventional, unremarkable examples. By doing so, it produces a structurally similar contribution to intensifiers such as extremely, as mentioned earlier. We can conclude that reduplication is an important strategy for conveying emphasis for intensity. With this in mind, let us now turn to attitudinal scales. Consider the following: (10)

You should totally click on that link! It’s awesome. (Beltrama and Staum Casasanto 2017: 160)

(11)

Chris is SO next in line. (Potts 2005b: 130)

In the cases of next-in-line and click-on-a-link, which involve all-or-nothing properties, let us assume that totally and so still choose the higher portion of a target scale, maintaining an intensifying core as in the previous examples. The distinction between these cases and those above is that the ordering is based on the speaker’s attitude rather than the predicate’s denotation—in other words, it depends on how enthusiastic, confident, or astonished the speaker is about their statement. With regard to totally in (10), Beltrama (2018) and Beltrama and Staum Casasanto (2017) assert that the scale is based on the speaker’s commitment towards adding the anchor proposition to the common ground of the conversation—in this case, the suggestion that following the link is a wise decision. In some instances, the scale is associated with the speaker’s emotional engagement. For example, Irwin (2014) posits that the use of ‘drama so’ in (11) conveys an emotional, rather than an information-based, conviction of the speaker towards the assertion, offering ‘whole-heartedly’ as an interpretation. In the literature, such attitude-level intensifiers present two analytical challenges. First, it is important to understand the nature of the elements being organized. While degree intensification, reduplication, and precisification generally operate on individual-level components within the predicate’s denotation, attitudes often involve fully formed, proposition-level components. This distinction is crucial, as it correlates with a compositional difference: attitude-level intensifiers generally encompass the entire descriptive content of a statement. For instance, they are not affected by logical operators or direct denials and remain unresponsive to elements contributing to the at-issue dimension. The example below demonstrates this by comparing totally and extremely: the patterns in (13) and (14) indicate that attitude-level

154

EMOTIONAL INTENSITIES IN NON-CANONICAL QUESTIONS

uses of totally (14) cannot function under the scope of a negator and are not contestable with denials: (12)

Scale in the denotation a Eliud is fast, although not extremely fast. b A: Eliud is extremely fast. B: No! He’s just fast.

(13)

Attitude scale: a #You shouldn’t totally click on that link. [Intended: ‘I’m not certain you should click’] b A: You should totally click on that link. B: No! #I should click on that link, but you can’t be certain about giving this advice!

This compositional aspect is often explained by proposing that attitude-level intensifiers, unlike those discussed earlier in this section, apply to propositionlevel (e.g., Irwin 2014 for so) or speech-act-level objects (Beltrama 2018). Additionally, they are tied to a specific viewpoint (usually the speaker’s) and influence the appropriateness conditions of the statement, rather than its propositional content. Recall that we have proposed a proposition-level operator of emphasis for some of the cases of non-canonical questions in previous chapters. Given the semantic discussion about different sorts of scales, we can now say that, as for the relevant degree semantics, the emphatic readings of questions concern the level of attitude and are thus similar to attitude-level intensifiers at the lexical level. Let us now explore in more detail how this type of emphasis in questions can be conveyed through syntactic means.

9.2 The expression of intensity in questions As we have already seen in this book, and in contrast to previous literature on emphatic focus (see Chapter 4), we observe that emphasis for intensity at the clausal level can be found not only in assertive speech acts, but also in nonassertive ones like questions. Look at the following effects again that we find in marked polar questions with emphatic word order in both Italian (14a) and German (14b); see Bianchi and Cruschina (2016) and Trotzke (2017a): (14)

a Ma domani al mare andate al mare? but tomorrow to-the seaside go.2.PL to-the seaside ‘You are going to the seaside tomorrow? (I can’t believe it!)’

9.2 THE E XPRESSION OF INTENSIT Y IN QUESTIONS

155

b Ans Meer fa¨hrst du morgen ans Meer? to-the seaside go you tomorrow to-the seaside ‘You are going to the seaside tomorrow? (I can’t believe it!)’ In such instances, we have argued in Chapter 4—in accordance with Bianchi and Cruschina (2016)—that the corresponding operator responsible for the emphatic interpretative impact should be positioned below illocutionary force, but still scope at the propositional level. Specifically, if we assume that yes–no questions represent a set comprising the proposition p and its complement {p, ¬p}, then the emphatic effect in (14) can be understood in relation to the focus alternatives of the proposition p, rather than its complement, as argued by Bianchi and Cruschina (2016). In particular, the examples above convey that there are more probable alternatives in the form of ‘Addressee goes to x tomorrow’. This aspect of interpretation implies that emphasis for intensity within the domain of word order is not situated at the level of illocutionary force, and we should thus observe it across various speech acts and clause types—and this is indeed what we find, as I will now illustrate below. Crucially, emphasis, adopting this perspective, is a proposition-level modifier (see Section 9.1 above) and concerns the attitudinal level of meaning, which can be part of any speech act. Since, in many cases, the interpretation is one of surprise or ‘unexpectedness’, some have called this type of meaning a ‘mirative implicature’ (see Cruschina and Bianchi 2021). The term ‘mirativity’ originates from DeLancey’s research (1997) on the grammatical category of mirativity, which indicates that the propositional content a speaker asserts has been recently acquired and not yet integrated into their belief system. In the typological literature, there is a debate over whether mirativity constitutes a distinct grammatical category (e.g., Aikhenvald 2012) or represents a unique interpretation of certain evidential markers (e.g., Park 2020; Rett and Murray 2013). I will assume that the expression of mirativity may range from specific vocabulary to various syntactic constructions like the question cases given above. What is crucial in our context is the interpretative aspect of mirativity, and not its specific linguistic instantiations. Mirativity, as defined by DeLancey (1997), pertains to the grammatical representation of a proposition as new and surprising or unexpected information for the speaker. Peterson (2013: 3–4) convincingly demonstrates “that the primary meaning of mirativity is to register the speaker’s surprise, and that other meanings commonly associated with mirativity, such as unexpectedness, new information, etc., are entailed by surprise” (see also Beltrama and Hanink 2019). When examining the pragmatic interpretation of mirativity in

156

EMOTIONAL INTENSITIES IN NON-CANONICAL QUESTIONS

greater detail, we notice that, on the one hand, the mirative effect can be analyzed as a conversational implicature in instances where evidential expressions obtain a mirative interpretation owing to a Gricean quantity implicature. For an approach along these lines, see Peterson’s analysis (2010: 141–4) of the Gitksan language. In essence, Gitksan speakers can imply surprise through the use of an indirect grammatical evidential. This pragmatic inferencing strategy is absent in other languages like Romance or Germanic languages discussed above. On the other hand, mirativity can be explained as a conventional implicature in situations where mirativity is lexicalized, as in the Hare (DeLancey 1997) and Apache (de Reuse 2003) languages, or expressed through specific word orders or constructions. Focusing on this conventional aspect of pragmatic interpretation reveals a plethora of syntactic constructions that convey unexpectedness or a sense of surprise. For example, it can be proposed that exclamative clauses constitute a subtype of mirative constructions. In (14), we have already seen how mirativity can also be conveyed through particular word orders in questions. Those mirative frontings have initially been explored by Cruschina (2012) in his research on Romance languages. Consider the following Italian example provided by Bianchi et al. (2015: 6): (15)

[CONTEXT: Anna tells about a customer who complained for nothing] a Pensa te! Col direttore voleva parlare! think you with-the manager wanted speak.INF b Pensa te! Voleva parlare col direttore! think you wanted speak.INF with-the manager ‘Guess what! He wanted to speak to the manager!’

Concerning marked word-order choices like (15a), Cruschina (2012: 119) has claimed that they convey a mirative effect and that “mirativity defines a specific type of exclamatives, namely, exclamative sentences expressing surprise and unexpectedness, and not exclamatives in general […] Mirative Fronting is triggered by a speaker-related mirative operator.” In the context of the claims in this book and of the domain of non-canonical questions, it is important to note that the concept of mirativity or ‘mirative fronting’ may still be too limited to encompass all the emphatic instances where a fronted linguistic object gains prominence by being chosen from the upper end of an attitudinal scale. For example, we have seen in the context of non-canonical questions via modal particles (Chapter 6) that wh-questions can exhibit emphatic word orders in which the modal particles depart from their original positions and

9.2 THE E XPRESSION OF INTENSIT Y IN QUESTIONS

157

appear in the left periphery in V2 configurations (see Chapter 4 for German and Chapter 6 for Basque examples). Crucially, modal particles are functional categories that cannot be contrasted and, consequently, cannot be contrastively stressed (i.e., there are no semantic focus alternatives to the respective particle). As a result, it would not be logical to suggest that the particle is fronted to the left periphery for reasons of (contrastive) focus. However, it would also be unmotivated to hypothesize that the emphatic interpretation of the marked word orders involves any kind of mirativity or surprise. Look at the following examples from German and Basque again: (16)

a [Wie nur] habe ich den Schlu¨ssel [wie nur] how PART have I the key how PART ko¨nnen? could ‘How on earth could I lose the key?’

verlieren lose

b [Non ote] utzi dut [non ote] egunkaria? where PART leave AUX where PART newspaper ‘Where did I leave the newspaper?’ Although the readings are emphatic and intensified, they must be accounted for based on other concepts than mirativity. In particular, fronting of the particle can lead to an intensification of the ‘can’t-find-the-value’ interpretation (see Chapter 3 for a detailed semantics for that specific interpretation), where the speaker, assuming a scale exists for this particle interpretation, expresses a higher degree of frustration or desperation by opting for the marked fronting construction. In other words, in these instances, the meaning of the modal particles gains prominence by being chosen from the upper end of a scale corresponding to the interpretative domain (above: frustration or desperation in the context of ‘can’t-find-the-value’) that is lexically encoded by the respective particles. Still, the interpretation is one concerning the domain of the speaker’s attitude, although not necessarily mirativity. This can be seen very clearly in our cases of surprise–disapproval questions from Chapter 5. Recall that Munaro and Obenauer (1999) have demonstrated that a special use of ‘what’ in whquestions also functions to create questions that do not seek information but express the speaker’s evaluation of the proposition (surprise, disapproval, and so on). According to their approach, this is a cross-linguistic phenomenon related to the semantic underspecification of such wh-elements. Consider the following German example again (Munaro and Obenauer 1999: 238):

158

(17)

EMOTIONAL INTENSITIES IN NON-CANONICAL QUESTIONS

Was lacht der denn so blo¨d?! what laughs he PART so stupidly ‘Why is he laughing so stupidly?!’

Importantly, Munaro and Obenauer reveal distributional constraints of this special use of ‘what’ that exemplify a modal licensing of these constructions, and this indicates that the use of why-like-what is licensed in contexts that concern the speaker’s attitude. In particular, Munaro and Obernauer’s finding is that the occurrence of nonargumental ‘what’ in the left periphery is preferred when elements of a modal type are added (18b–d). It is worth noting that warum (‘why’), the actual meaning of ‘what’ in the following German examples, is perfectly acceptable in a context where such an evaluative modality is absent (18a): (18)

a

Warum lacht (d)er? why laughs he ‘Why is he laughing?’

b ?Was lacht der? what laughs he intended: ‘Why is he laughing (I can’t believe it!)?’ c Was lacht der denn?! what laughs he PART ‘Why is he laughing (I can’t believe it!)?’ d Was lacht der so blo¨d?! what laughs he so stupidly ‘Why is he laughing so stupidly (I can’t believe it!)?’ All in all, the presence of why-like-what in the left periphery necessarily leads to the expression of “an attitude of the speaker ranging from mild surprise to strong disapproval” (Munaro and Obenauer 1999: 237–8). Having indicated that adding linguistic material related to the domain of the speaker’s attitude is crucial for making those non-canonical wh-questions sound more acceptable and natural, we can conclude that the Emphasis operator we postulated for these questions in previous chapters is indeed similar to attitude-level modifiers, which, as discussed further above, operate at the propositional level. Given these considerations about mirativity, emphasis, and non-canonical questions, we can say that the special interpretations we find in non-canonical questions most likely cannot be analyzed in terms of mirativity, because the concept for left-peripheral options in this domain must be broader (see, e.g., our examples of can’t-find-the-value questions above).

9.2 THE E XPRESSION OF INTENSIT Y IN QUESTIONS

159

However, the relevant interpretations certainly all involve degree and intensification at the propositional level, similar to what we have seen in the context of attitude-level modifiers in Section 9.1. The remaining questions now are how these emphatic questions can be characterized in terms of more common information-structural concepts, and how we should characterize the emphatic interpretations at the level of semantics and pragmatics more generally (i.e., what kind of meaning is it?). Let us start with the first open question. To approach the first question, we first have to see that the implicature we referred to as emphasis for intensity is a focus-based implicature. In other words, it builds on a set of alternatives and conveys an ordering of entities among the set members. This reliance on possible alternatives is also applicable to emphatic wh-questions. Above, we have already observed that some special readings such as why-like-what are licensed through the expression of evaluative modality in the respective utterances. In these cases, however, the wh-element itself could not license the occurrence of the particle in the left periphery. This is because, in the case of non-argumental ‘what’, the whelement lacks any semantic alternatives. As a result, we find that the [wh+Prt]construction is incompatible with such question types, as demonstrated by Bayer and Obenauer (2011: 481–3): (19)

a Was lacht der denn so blo¨d?! what laugh he PART so stupidly ‘Why is he laughing so stupidly?!’ b ∗ [Was denn] lacht der so blo¨d?! what PART laugh he so stupidly intended: ‘Why is he laughing so stupidly?!’

In such instances, the wh-element does not establish a connection with a variable in the manner of a typical wh-interrogative. As previously discussed, the emphatic interpretation relies on a range of alternatives to convey a hierarchy among the elements within a set. For example, the emphatically marked wie in wie nur (translated as ‘how only’) from our key example (16) encompasses a variety of ways (attributes of events), and, by opting for the [wh+Prt]-structure, the speaker accentuates the extreme unlikelihood, if not complete impossibility, of pinpointing a specific way. When examining surprise–disapproval questions like those in (19a), it becomes evident that the wh-element is not covering a range of alternative values. Distinct from the corresponding wh-term warum (‘why’), the ‘what’ resembling ‘why’ cannot be emphasized and is unable to appear in situ in multiple questions:

160

(20)

EMOTIONAL INTENSITIES IN NON-CANONICAL QUESTIONS

a

∗

Wer who b Wer who

lacht laugh lacht laugh

WAS what WARUM why

so blo¨d?! so stupidly so blo¨d?! so stupidly

This shows that, although interacting with focus-based interpretations (as in the case of can’t-find-the-value questions), the emphasis reading cannot be reduced to such information-structural concepts. In what follows, I would like to argue that emphatic fronting in questions does not interact with the descriptive content of the utterance, much like the intensifiers examined in Section 9.1. This observation has also been made by others, including Frey (2010) and Bianchi et al. (2016), who contend that the relevant emphatic effect can be described as a conventional implicature in the sense of Potts (2007b). The following example, adapted for our discussion (cf. Frey 2010: 1422–4), illustrates this: (21)

a A: Ans Meer fa¨hrst Du morgen? to-the seaside go you tomorrow ‘You are going to the seaside tomorrow? (I can’t believe it!)’ B: Yes, but what’s so surprising/strange about that? b A: Was lachst Du denn jetzt? what laugh you PART now ‘Why are you laughing right now? (I can’t believe it!)’ B: Well, I think it’s funny, but I don’t understand why you’re so surprised about that.

The parallels in (21) suggest that emphasis for intensity at the propositional level of questions belongs to the category of non-descriptive content: The emphatic meaning can be disputed by the addressee independently of the descriptive content. This indicates that emphasis for intensity is typically conveyed through operators that are part of the non-descriptive content across subtypes of non-canonical questions and that can be characterized as a conventional implicature and, more generally, as syntactically encoded expressivity—a concept we will discuss in more detail in the next and final chapter of the book. But before we turn to this final chapter, let me briefly summarize the discussion above.

9.3 SUMMARY

161

9.3 Summary The picture that emerges from this chapter is that the syntax of emphatic versions of questions determines only that the respective questions contain an intensity component, and it does not determine any specific emotive interpretation such as mirativity or surprise. Specifically, I have argued that expressing emotion is, from a syntax–semantics perspective, a syntactic intensification strategy, often realized by certain movement and fronting patterns, but potentially also by other well-known intensification strategies such as augmentation, duplication, and prosodic prominence. According to this view, the grammar of non-canonical questions merely signals the emotional involvement of the speaker, and the specific emotion is conveyed by pragmatics only. I have indicated to what extent a clause-level operator of emphasis in questions resembles attitude-level modifiers that take propositional scope in the lexical domain. In the final chapter, we will try to develop a unified framework for this type of propositional-level expressivity and the domain of other linguistic strategies of prominence, notably in the area of information structure. We will also mention some further notions of canonicity in the context of questions that the previous chapters have not addressed in detail, and we will point out some possibilities of how to account for those cases in syntactic terms as well.

10 Non-canonical questions, expressivity, and the model of communication This final chapter of the book brings together the discussions about the syntax, the pragmatics, and the emotive components of non-canonical questions by summarizing the new syntax for non-canonical questions I have been arguing for throughout the book. Crucially, this syntax goes against the common claim in the literature that non-canonical questions involve a special grammar at (or even above) the level of illocutionary force. The theory proposed in this book demonstrates instead that those non-canonical interpretations of questions can solely be derived from a propositional-level operator of emphasis that does not affect the illocutionary level of utterances. To be sure, non-canonical questions, such as rhetorical questions, are a common feature of language and can be found in many, if not all, languages. However, the specific forms and functions of non-canonical questions—and hence also their syntax—may vary greatly within and across languages and cultures. That is, in many types of non-canonical questions, their special semantics–pragmatics is not due to syntactic factors, but is rather due to other features such as intonation, context, and so on. Let me also add that, from the view of cross-cultural pragmatics (see, e.g., Kasper 2004), some aspects of noncanonical questions may be more culture-specific, as they are influenced by the social and cultural norms of a particular language or community. Consider, for example, the domain of politeness. The ways in which speakers show respect or deference to others when asking a question may vary across cultures, with different languages and cultures having different norms for expressing politeness via questions. In conclusion, this means that any claim about the syntax of non-canonical questions must always be formulated in a way that does not concern languagespecific idiosyncrasies as to what counts as canonical versus non-canonical interpretation in a specific language. As a consequence, we can approach the syntax of non-canonical questions only in a cross-linguistic fashion by referring to categories and notions that are expressed in any language—be it via

Non-Canonical Questions. Andreas Trotzke, Oxford University Press. © Andreas Trotzke (2023). DOI: 10.1093/oso/9780192872289.003.0010

10.1 T WO NOTIONS OF E XPRESSIVIT Y

163

syntax or via other means such as prosody. In what follows, I will first try to suggest such an approach by rethinking the core notions of information structure and expressivity, which are probably relevant and part of the grammar of any of the languages of the world. My goal will be to show that we are essentially looking at only one domain, rather than at two separate phenomena. Given this redefining of those core notions, the syntactic analyses of non-canonical questions proposed in the previous chapters can be embedded within a conceptual framework for discourse meanings that holds cross-linguistically. At the end of the chapter, I will then turn to even broader issues that concern our basic concepts for a model of human communication and where non-canonical questions might fit in.

10.1 Two notions of expressivity At the outset of the book, we have already seen that non-canonicity in grammar is often defined in terms of the information structure of an utterance, and in what follows I will argue that this concept can be fruitfully connected to many of the cases of non-canonical questions discussed in the previous chapters. In particular, in this section I will explore the similarities, differences, and interactions between the domain of expressivity and a variety of discourse meanings commonly summarized in the literature as the domain of information structure. As we have also already seen in this book, both empirical domains are rather broad, and they have been investigated from very different perspectives. As for information structure, Krifka (2008: 243) has noted in a seminal paper that “[t]he basic notions of Information Structure […], such as Focus, Topic and Givenness, are not simple observational terms. As scientific notions, they are rooted in theory, in this case, in theories of how communication works.” Given that there are so many different theories of how communication works and thus also of information structure, I will focus on a special branch of the literature, which concerns, as I will argue, relevant connections between information structure and the domain of expressivity. Turning to the concept of expressivity, we observe that this term too has quite different readings in the literature. In order to review the connections between expressivity and information structure, I will therefore propose a distinction in this section between ‘expressivity—broad sense (EB)’ and ‘expressivity—narrow sense (EN)’. According to this distinction, EN covers a

164

THE MODEL OF COMMUNICATION

range of so-called use-conditional phenomena (more on this below) that also convey emotive meanings. By contrast, EB includes EN plus use-conditional meanings in general (probably the more common reading of ‘expressivity’ in the theoretical semantics–pragmatics literature). Crucially, the distinction between EB and EN will allow me to claim the following: while EB is always conveyed via phonology, only EN additionally triggers marked word order—a domain traditionally explored in information-structural terms. We have seen a lot of relevant examples for this type of expressivity throughout the book. The overall goal of this section will be to reconsider those examples and their respective analyses by showing that what is commonly known as the field of information structure can actually be reconceptualized as the study of expressivity across grammar. To see that, we first need a tentative and broad notion of expressivity. In a seminal paper, Kaplan (1999) discusses the differences between the English interjections ouch and oops and their descriptive counterparts I am in pain and I just observed a minor mishap, respectively. His main point was to distinguish between two types of information, descriptive information and expressive information, and how to account for them by means of formal semantics. In a nutshell, the descriptive statements I am in pain and I just observed a minor mishap can be accounted for in terms of their truthconditional contributions, whereas interjections such as ouch and oops can be analyzed only by looking at the felicity conditions of those linguistic expressions (i.e., how they can be used). For Kaplan (1999), ouch and oops thus convey expressive rather than descriptive information. While Kaplan in his discussion focused on interjections and items such as damn or bastard, Kratzer (1999) suggested that we can go ‘beyond ouch and oops’ and extend this notion of expressive meaning to further phenomena such as modal particles in German (see our discussion of modal particles in noncanonical questions in Chapter 6). It is important to note that, in this type of work, expressive meaning is thus equal to use-conditional meaning, and since the beginning of the twenty-first century a whole strand of research on the ‘expressive dimension’ of meaning has developed semantic analyses for empirical phenomena that require a use-conditional level of analysis (Gutzmann 2015; Potts 2005a, 2007a,b). In what follows, I will call this reading of expressivity ‘expressivity—broad sense (EB)’. This final chapter of the book is not the place to discuss those analyses in depth—not least because this chapter aims at a conceptual framework for the wealth of empirical data and analyses proposed in this book. However, let me briefly illustrate what is meant by saying that elements like modal particles

10.1 T WO NOTIONS OF E XPRESSIVIT Y

165

are use-conditional items and thus belong to EB.¹ Building on Kratzer’s work (1999), I use the famous example of the German particle ja (lit. ‘yes’) here because the meaning contribution of this particular particle will also play a role further below when we turn to the connection between EB and information structure. Look at the following example: (1)

A asks B: Which race did Eliud win? B: #Eliud hat ja den olympischen Marathon gewonnen. Eliud has PART the Olympic marathon won ‘Eliud has won the Olympic marathon.’

The ‘#’ symbol indicates that B’s utterance is infelicitous in the context given in (1). Crucially, this infelicity is only due to the use of the particle ja; the utterance would be perfectly felicitous without the particle. This is because the version of the utterance containing the particle features an expressive meaning that the utterance without the particle lacks. Accordingly, the utterance contains two levels of meaning, and one of them—the expressive level—does not match the context given in (1). The two levels of B’s utterance in (1) can be illustrated as follows, slightly modifying a representation used by Gutzmann (2019: 25): (2)

You and I already know that Eliud has won expressive meaning the Olympic marathon. Eliud has won the Olympic marathon. descriptive meaning

On the other hand, the use of the particle ja would be felicitous in a context where it can be assumed that both speaker and addressee already know that Eliud has won the Olympic marathon. A felicitous scenario would be the following: (3)

A says to B: Have you heard that Eliud will take some time off in the next few weeks? B: Eliud hat ja den olympischen Marathon gewonnen. Eliud has PART the Olympic marathon won ‘Eliud has won the Olympic marathon. (So it’s clear that he will need a break.)’

This brief illustration should suffice to see that the concept of expressivity I refer to as ‘EB’ distinguishes between truth-conditional contributions ¹ An alternative theoretical approach would be to analyze modal particles as truth-conditionally vacuous presupposition triggers (e.g., Grosz 2014). There is a more general debate in the literature as to whether expressive meanings would be better analyzed as presuppositions (Schlenker 2007), but this debate would go far beyond the scope of this final chapter.

166

THE MODEL OF COMMUNICATION

(aka descriptive meaning) and use-conditional contributions (aka expressive meaning). Crucially, the notion of expressive meaning, according to EB, does not only capture non-truth-conditional items such as interjections (ouch and oops; see above); rather, this concept of expressivity accounts for all items whose meaning must be analyzed by referring to their use conditions (a level of meaning distinguished from the truth-conditional level). After having clarified the broad notion of EB, let me now turn to the second concept of expressivity that can be found in the literature. When we look outside the formal semantics–pragmatics literature, the term expressivity is often used to refer to aspects of language that interface with emotions. That also corresponds to how we have used the term in the context of emphatic and ‘emotive’ questions, but it also corresponds to our use of language in everyday life: situations where someone uses ‘expressive’ language, for instance, are quite likely to be contexts where emotions play a key role. Although the formal literature on EB cited above often uses emotive phenomena to illustrate expressivity (like interjections, expressions such as damn and bastard, and so on), the expression of emotions via language, according to EB, is just a prime example of non-truthconditional (here: use-conditional) meaning; the class of expressive items, however, is much bigger. I have illustrated that point above in the context of the German modal particle ja, which does not convey any emotion whatsoever (in contrast to some of the particles in questions we have seen in Chapter 6, which at least yield an intensity reading of the question utterance). Crucially, the reading of the term expressivity is strikingly different in the non-formal literature, as indicated by the following programmatic statement from a recent handbook article: Human beings often have feelings about what they say, about what others say, or about what happens in the here-and-now environment. These feelings are typically expressed in non-verbal ways (facial expression, gesture, posture, etc.), but, as it turns out, language itself also provides means for expressing such feelings. Language not only has a referential, but also an expressive function. (Foolen 2016: 473)

Expressivity in this reading is thus about the expression of feelings, and it is distinguished from the referential function of language. In what follows, I will call this concept of expressivity ‘expressivity—narrow sense (EN)’. What this

10.1 T WO NOTIONS OF E XPRESSIVIT Y

167

concept shares with the broader concept of EB is the performative character of expressivity. To see this, let us look at Foolen’s useful distinction (2012, 2016) between the conceptualization of emotions and the expression of emotions. While the former focuses on the referential function of emotion words (joy, enjoy, happy) and more complex expressions (figurative speech and so on), the latter explores which linguistic elements convey emotional meaning by their linguistic form. In other words (and as already mentioned in Chapter 3), we can of course speak about emotions such as happiness and anger in descriptive terms (I’m enjoying my time with you versus I’m angry at you). However, the expression of emotions, according to Foolen’s distinction (2012, 2016), is not about what we say, but rather about how we say things. Expressivity, in this interpretation, thus refers to linguistic elements and strategies that convey emotions by their very linguistic form and not by their referential content. It is worth noting that this narrow concept of expressivity can also be found in classical speech act theory, where the speech act class of expressives is associated with emotive meanings; this is indicated by the following quote (see also Searle 1976): The expressive point is to express feelings and attitudes. In utterances with the expressive point the speaker expresses some psychological attitude about the state of affairs represented by the propositional content. (Searle and Vanderveken 1985: 38; emphasis in the original)

Accordingly, for many scholars working in non-formalist frameworks, expressivity refers to meanings that are somewhat emotive, and expressive items thus do not form a class with other use-conditional phenomena that have nothing to do with emotivity. Crucially, this notion of EN can be found in the more recent formal literature too. For instance, Rett (2021) proposes that expressives like damn form one class with emotive markers such as English alas, to the exclusion of other notat-issue elements (e.g., appositives). To my mind, this indicates that using the term ‘expressivity’ requires a clarification, not only across different linguistic frameworks, but also more and more so within formal linguistics. Since the beginning of the twenty-first century, the terms expressivity and ‘the expressive dimension’ have become catch-all phrases in theoretical linguistics, and this is why we need to distinguish between the different readings before we start to explore to what extent expressivity connects to specific aspects and domains of human language such as non-canonical questions. With all this in mind, let

168

THE MODEL OF COMMUNICATION

us now first turn to the connection between information structure and EB in the next section.

10.2 Information structure and EB Similarly to what we have seen in the context of expressivity in the previous paragraphs, the notion of ‘information structure’ also heavily depends on empirical orientation and theoretical goals (see Erteschik-Shir 2007; Féry and Ishihara 2016; and Leino 2013 for overviews of different fields and research traditions). Given this situation, I will not attempt to survey the various approaches to this complex issue and mention all relevant empirical observations (some of which I have already introduced in Chapters 1 and 4 of this book). Nevertheless, we have to introduce some core notions that can serve as a basis for our discussion about the connections between information structure and EB. Consider the following sentences, where capital letters indicate main stress: (4)

a [The RUNNER]F won the marathon. b The runner won [the MARATHON]F . c The runner [WON]F the marathon.

As (4) shows, in addition to the level of propositional structure, which, roughly speaking, encodes the information of ‘who did what to whom’, there is also the level of information structure, encoding the fact that, as Halliday (1967: 242) defines it, the speaker maps on to the clause, as defined in sentence structure, a structure of a different kind in terms of information units, by which he organizes the discourse into message blocks and specifies the status of the components of the message as new information or otherwise.

At the level of meaning, while the propositional structures in (4a)–(4c) are identical, the element interpreted as providing new information (the ‘focus’/ ‘F’) is different: in (4a), the element interpreted in this way is the runner, in (4b) it is the marathon, and in (4c) it is the verb won. The distinction between these two meaning components is well established in linguistics, regardless of theoretical orientation and framework. In particular, the grammar of natural languages expresses two domains of semantic interpretation: the broad component of argumental and event-related semantics and a variety of phenomena that can be subsumed under the notion of

10.2 INFORMATION STRUCTURE AND EB

169

scope-discourse semantics. The notion of information structure and interpretations like ‘focus’ clearly belong to the domain of scope-discourse semantics. The question addressed by this final chapter of the book now is what this domain of meaning has to do with expressivity. Intuitively, it is clear that discourse meanings such as old/given or new information yield matches in a concrete context and thus have a use-conditional component (see our remarks above). This has been spelled out most notably by Kratzer (2004) and, more recently, by Kratzer and Selkirk (2020). The general idea of this line of research is that core notions of information structure (such as ‘focusing/backgrounding’ and ‘givenness’) are in fact expressive meanings in the sense of EB. Let us look at the details of this approach. In a short response paper to Geurts and van der Sandt’s programmatic article (2004) on focus interpretation, Kratzer (2004) has suggested that basic information-structural features resemble the behavior of modal particles like German ja (see examples (1) and (3) above). Kratzer’s paper (2004) is often cited in the context of how to model the readings of (German) modal particles, but less so in the context of the main claim of her piece, which is: when we look at information structure, we are essentially looking at expressive meanings. Observe the following example (see Kratzer 2004: 124–6): (5)

[ELIUD]F won the Olympic marathon.

The utterance in (5) would be odd in a context such as (6a), where an utterance containing a focused constituent [ELIUD]F would not match the context. However, the same utterance with the same focusing of Eliud would be perfectly fine if the context is slightly changed as in (6b): (6)

a A: Which race did Eliud win? B: #[ELIUD]F won the Olympic marathon. b A: When did Mo win the Olympic marathon? B: [ELIUD]F won the Olympic marathon!

Kratzer’s point (2004) is that the discourse (mis)matches we observe in (6) are very similar to what we see in the context of modal particles like German ja, doch, and so on. Observe the following patterns: (7)

a A: Which race did Eliud win? B: #Eliud hat ja den olympischen Marathon gewonnen. Eliud has PART the Olympic marathon won ‘Eliud has won the Olympic marathon.’

170

THE MODEL OF COMMUNICATION

b A: Which race did Eliud win? B: Eliud hat doch den olympischen Marathon gewonnen! Eliud has PART the Olympic marathon won ‘Eliud has won the Olympic marathon.’ In (7a), B’s utterance is infelicitous because ja in that utterance indicates that both A and B should already be aware of the fact that Eliud has won the Olympic marathon—but this does not match the current context in which A is expressing their ignorance about this fact by explicitly asking about which race Eliud has won. The situation is different in (7b). Here, B uses the modal particle doch, which has a different meaning than ja insofar as this particle indicates that A should already know that Eliud has won the Olympic marathon, and Speaker B is reminding A of this fact. In such a context, B’s reaction to A’s question matches the current context and is a felicitous use of a German modal particle. We thus see that the use of particular modal particles has to match the respective contexts, and this is exactly what we also observe for different focusing/backgrounding configurations in (6). If we take the parallel between focusing/backgrounding in the domain of information structure and the use of different modal particles seriously, so Kratzer’s argument (2004) goes, common notions of information structure such as focus and givenness are in fact use-conditional items and thus, according to the notion of EB, expressive meanings. We can illustrate this by using our two-level representation again and postulating that the use-conditional meaning of the focus in (5) and (6) can be paraphrased as a corrective discourse relation (see Repp 2016: 278–9 for a detailed account and Kratzer and Selkirk 2020: 33–4 for a different proposal): (8)

You wrongly presuppose that someone other expressive meaning than Eliud won the Olympic marathon. Eliud has won the Olympic marathon. descriptive meaning

The claim that information structure and use-conditional items such as modal particles operate at the same level of meaning (namely, the expressive level) is further supported by cross-linguistic considerations. For instance, it has been argued that Romance languages such as Italian use marked word order to express the same meanings that modal particles convey in languages like German (Cardinaletti 2015b; Trotzke et al. 2020). Look at the following examples from Cardinaletti (2015b: 17):

10.2 INFORMATION STRUCTURE AND EB

(9)

Context: speaker and hearer are both well aware that the hearer has been to Paris before, and the speaker wants to make this fact salient in order to follow up on it: Du warst ja/doch schon in Paris. you were PART/PART already in Paris ‘You’ve (ja/doch) already been to Paris.’

(9′)

a

A Parigi, ci sei in Paris, there you.have ‘You’ve already been to Paris.’

b

Ci sei there you.have

c Sei d #Sei

già stato. already been

171

Left Dislocation

già stato, a Parigi. Right Dislocation already been, in Paris

già STA::to, a Parigi. già stato a Parigi.

Marginalization SVO

Cardinaletti (2015b) argues that, in a context like (9), the Italian counterparts of the German particles ja and doch would probably involve marked word orders as in (9′a,b), or the locative constituent [a Parigi] would at least be ‘marginalized’ by emphatically stressing the preceding element of the clause. Crucially, a completely unmarked construction such as (9′d) would not match the given context in (9). Cardinaletti’s data (2015b) contain two interesting facts for our discussion. First, the relevant use-conditional meanings in some languages are conveyed by lexical means (German), while in other languages the very same interpretations can be expressed via syntactic and/or phonological marking (Italian). This supports the general claim by Kratzer (2004) that not only items like modal particles, but also core informationstructural notions, are expressive meanings in the sense of EB. Second, the syntactic strategies in Italian exemplified above seem to be optional in the context of yielding a reading that matches the context; the same reading can also be achieved via phonology only (9′c). This latter point will be of interest in the next section, where we will ask whether there are also cases where marked word order is not really optional, but actually strongly preferred for conveying particular meaning effects. We have seen many of those cases in non-canonical questions, but the question is whether the relevant emphatic effects in questions instantiate syntactic mechanisms that hold across speech acts. We certainly do not want a functional hierarchy of the clause that holds only for individual speech acts (e.g., questions)—the goal is to show how the specific patterns within a class of speech acts (within specific languages) illustrate some general syntactic principles at the syntax–pragmatics interface.

172

THE MODEL OF COMMUNICATION

10.3 Information structure and EN In the previous section, we have seen that, if we adopt the notion of EB, then the most common concepts of information structure (i.e., focusing/backgrounding and givenness) can in fact be considered expressive meanings. The question addressed by the following discussion now is whether we also find phenomena of EN among those more general types of useconditional meaning. In what follows, I would like to focus on the most prominent and obvious example in this domain: so-called mirative or emphatic focus. We have already highlighted this concept in the context of non-canonical frontings in questions in Italian and German (Chapter 4), but here I would like to broaden the perspective and illustrate to what extent this might not be a language-specific phenomenon. Krifka (1995: 227) has argued that “[t]he function of emphatic focus is to indicate that the proposition that is actually asserted is prima facie a particularly unlikely one with respect to the alternatives.” Put succinctly, he argues that this meaning effect represents a general type of focus that is made explicit with particles like even or idiomatic constructions like of all persons (examples by Krifka 1995: 227): (10)

a Mary knows every place on earth. She has (even) been to BORneo! b People expected that John would win the election, followed by Bill, with Mary as a distant third. But then the election was won by MARY (of all persons)!

But why should such a scalar focus type be relevant to EN (i.e., to the useconditional expression of emotions)? The reason for why the scalar reading of unlikelihood in (10) could be relevant in the context of expressing emotions is extensively addressed in the literature on so-called mirative focus (see our discussion in Chapter 4 and, again, Cruschina 2012, 2019; Cruschina and Bianchi 2021). Building on typological work by DeLancey (1997), mirativity is often defined in the more recent literature by referring to the emotion surprise. Krifka’s examples above (1995) express the surprise component via phonological means only (see also Asu et al. 2022; Rett and Sturman 2020; on mirative prosody). However, the discussion about mirative focus in the literature centers on word-order changes that involve the left periphery of the clause (so-called mirative fronting). We have already seen many examples in

10.3 INFORMATION STRUCTURE AND EN

173

Chapter 4; for convenience, observe the following Italian, French, and German examples²: (11)

[CONTEXT: Anna tells about a customer who complained for nothing] a Pensa te! Col direttore voleva parlare! think you with-the manager wanted speak.INF b Pensa te! Voleva parlare col direttore! think you wanted speak.INF with-the manager ‘Guess what! He wanted to speak to the manager!’ (Bianchi et al. 2015: 6)

(12)

a Même ses caleçons on lui a volé. even his underwear they him have stolen ‘Even his underwear they stole from him.’ b Trois heures (qu’)elle a passé chez son avocat. three hours (that)-she had spent at her laywer ‘Three hours she spent at her lawyer’s office.’³ (Authier and Haegeman 2019: 7–9)

(13)

A: How did Eliud do in the race? B: a ABGEBROCHEN hat er das Rennen! dropped-out has he the race b Er hat das Rennen ABGEBROCHEN! he has the race dropped-out ‘He has dropped out of the race.’

Let us briefly look at a German example such as (13) again. Both the marked (13a) and unmarked word order (13b) can emphasize the fact that Eliud dropped out of the race as remarkable and unexpected, provided speaker and hearer have the common ground that Eliud Kipchoge, at the time of utterance, is the most steady and confident marathon runner of all time. However, only the syntactic configuration in (13a) obligatorily features the relevant phonology. In other words, it is not possible in German to front ² Other examples are Spanish (Cruschina 2019) and, using a different terminology, Bangla (Bayer and Dasgupta 2016), Bavarian (Bayer 2001), Hausa (Hartmann and Zimmermann 2007), and Latin (Danckaert 2012). Let me hasten to add that Destruel et al. (2019), based on their experimental work, have recently speculated that English clefts are preferably used in mirative contexts too. ³ This French example additionally contains a special occurrence of the complementizer element que, and this strategy is known to occur in exclamative contexts in Romance (Trotzke and Villalba 2021).

174

THE MODEL OF COMMUNICATION

the predicate abgebrochen without putting non-structural stress on that constituent (see our discussion in Chapter 4). By contrast, the syntax given in (13b) could also appear with structural accent on the predicate, in which case it would ‘only’ be a regular answer to the question posed by A (i.e., without conveying any emotion such as surprise). In a broader cognitive perspective, one could say that the marked word order plus the obligatory marked phonology in (13a) indicates to the addressee that a non-default interpretation of the utterance is needed (Rosemeyer et al. 2022; Skopeteas and Fanselow 2011). More specifically, since a neutral answer in the context given in (13) could also be realized by the in-situ word order (13b), the addressee expects that there is an additional non-default meaning conveyed by (13a). Crucially, we have already seen that in some fronting patterns in noncanonical questions this non-default meaning does not have to be mirativity, and this is why ‘mirative fronting’ might be a misnomer. That is, the effect of those frontings has to be conceptualized much more broadly. We can also easily illustrate this for declarative cases like (13) above. Look at the following pattern: (14)

[Everyone knows that Eliud is by far the world’s most successful marathon runner.] A: How did Eliud do in the race?’ B: a GEWONNEN hat er! won has he b Er hat GEWONNEN! he has won ‘He has won. (What else?)’

Example (14) illustrates that a similar fronting of a predicate as in (13) could also occur in a context where the proposition is not the least likely alternative and thus surprising, but rather the most likely alternative. Specifically, the marked word order in (14a) could fit smoothly into a context where the speaker continues by saying: ‘What else?’ or ‘Why are you asking so stupidly?’ In this use, the marked word order would be chosen to express that the answer could be expected (that Eliud won the race is what should be most expected). Since in this case A’s question forces B to state the obvious, this use often has a reprimand character. Again, this interpretative effect could also be achieved by placing heavy and non-structural stress on the predicate in the unmarked word order in (14b). However, as soon as the predicate is fronted, this phonological cue for a non-standard interpretation obligatorily follows, while it is not obligatory in the in-situ version in (14b).

10.3 INFORMATION STRUCTURE AND EN

175

Given the examples above, we can conclude that not only surprise, but also negative emotions such as disapproval and reprimand—similar to what we have seen in the domain of non-canonical questions—can be conveyed by marked word-order options.⁴ Given this more general character of those marked configurations, I have already pointed out throughout the book that in my previous work I have referred to such meaning effects as emphasis for intensity (Trotzke 2017a), where intensification refers to the component that alternatives in the readings above are always ordered on a propositional-level, attitudinal scale in a way that the intensified element points to an (extreme) end of that scale (see Chapter 9). As in many other linguistic domains, intensification goes along with adding emotivity to the relevant linguistic configuration (see Beltrama and Trotzke 2019). Crucially now, all the interpretative effects of marked word order indicated above can be classified as EB, because they can all be analyzed as operating at a use-conditional level of meaning. There are many ways to show this (see Potts 2007b for some general tests and Frey 2010; Trotzke 2017a for the application to emphatic frontings). In the context of this chapter, let me point out only that in those configurations the descriptive and the expressive level of the utterance can be targeted separately in a discourse, and also that the expressive content, in contrast to the descriptive content, cannot be denied: (13′)

A: How did Eliud do in the race? B: ABGEBROCHEN hat er das Rennen! dropped-out has he the race ‘He has dropped out of the race.’ A: No, he finished the race first! A′: Ok, but why are you surprised?/ #No, you are not surprised!

descriptive meaning expressive meaning

The discussion above indicates that we also find phenomena of EN (surprise, disapproval, reprimand, and so on) among the more general connections between information structure and EB sketched in Section 10.2 above. One possible hypothesis would now be that all kinds of information-structural meanings (including emotive ones such as mirative focus) can be triggered by marked phonology only, but, if a marked word order is chosen, then the meaning effect is certainly one that interfaces with emotions. ⁴ This is also supported by Krifka’s notion (2008: 259) of emphatic focus, which covers not only cases where marked word order and/or marked phonology conveys an ordering on a scale of likelihood, but also cases such as strong polarity items (his example: [Wild HORses]F wouldn’t drag me there), which are certainly not typical cases of mirativity either.

176

THE MODEL OF COMMUNICATION

Such a hypothesis suggests itself in the context of German. Experimental work by Trotzke (2017b) shows that German native speakers clearly prefer the in-situ word-order options in a variety of different information-structural contexts—except for mirative contexts. Those emotional contexts are the only ones where the fronting is judged as good as the in-situ versions. The reason for this might be one that has already been indicated in Section 10.2 above: German has lexical means (i.e., modal particles) that can express a variety of information-structural EB meanings that would involve marked word-order choices in other languages (like Italian, see above). In German, marked word order is thus not really needed for expressing most of the common information-structural meanings, and, if fronting is used at all, it is preferably used for EN interpretations like mirativity. This seems to hold for further Germanic languages like English (modulo syntactic differences in this particular language of course). On the basis of their experiments, Destruel et al. (2019) have recently hypothesized that English clefts are preferably used in EN contexts (following the literature, they restrict their discussion to mirativity). Needless to say, the hypothesis that fronting patterns always involve EN readings faces some problems when we look beyond Germanic languages. Let us stick with Italian, because this language is famous for its various fronting strategies, and we have already introduced some basics in Section 10.2 above. Look at the fronting pattern by Bianchi et al. (2015: 6) again, repeated here for convenience: (15)

[CONTEXT: Anna tells about a customer who complained for nothing] a Pensa te! Col direttore voleva parlare! think you with-the manager wanted speak.INF b Pensa te! Voleva parlare col direttore! think you wanted speak.INF with-the manager ‘Guess what! He wanted to speak to the manager!’

While the context in this example makes it clear that a mirative interpretation is intended, Bianchi et al. (2016) have elaborated on this in more recent work and speak of ‘evaluative implicatures’, which are basically conventional implicatures (in the sense of Potts 2007b) conveying a broader interpretation. Crucially, however, part of their research agenda has also been experimental work in which they were able to show that, in addition to mirative contexts, corrective focus contexts are the preferred contexts that such focus frontings can occur in. Coming back to the hypothesis that frontings always involve EN readings, it is hard to see why a corrective focus such as (16) should involve emotivity (example from Bianchi et al. 2015: 5):

10.3 INFORMATION STRUCTURE AND EN

(16)

177

A: Hanno invitato Marina. have.3PL invited Marina ‘They invited Marina.’ B: Giulia hanno invitato (, non Marina). Julie have.3PL invited not Marina ‘They invited Julie (, not Marina).’ B′: Hanno invitato Giulia (, non Marina). have.3PL invited Julie not Marina

While correction could also be expressed by the in-situ option (B′), the empirical work by Bianchi et al. (2015) shows that the fronting is clearly preferred in such contexts (just as it is preferred in mirative contexts)—and this differs from the results obtained by Trotzke (2017b) in a similar study on German. Let us thus assume that, in Italian (and possibly further Romance languages), mirative and corrective contexts prefer focus fronting over the in-situ realization by mirative or corrective prosody only. Although the prosody (and further properties) of mirative and corrective focus may be quite different (see Bianchi et al. 2016 for experimental evidence), the two categories seem to form one class when we consider only their potential to trigger focus fronting. Why is that so? Let me try to address this question by starting with a very intuitive and quite superficial observation. In many overview articles of different types of information-structural meanings, the examples illustrating corrective focus are presented as exclamation statements (see, e.g., Krifka 2008; Repp 2016; and our detailed discussion of the exclamation speech act in Chapter 5 of the book). Look at the two following examples by Repp (2016), illustrating two different discourse types with corrective focus: (17)

a A: B: b A: B:

When did Pete clean up? [JOHN]F cleaned up! It’s raining outside. [The sun is shining!]F Look out of the window!

If we take exclamations to be expressions of an attitude of emotion (see Chapter 9), or at the very least expressive speech acts in the sense of Searle (1976), then we can hypothesize that corrective statements like the ones given in (17) are not just about substituting a discourse segment with an alternative and conveying that a proposition expressed in the preceding discourse is false. Rather, the speaker typically also expresses that they find this falsehood shocking, or at least that they cannot understand and are puzzled by the background

178

THE MODEL OF COMMUNICATION

assumptions of the preceding statement—and indeed many of the examples of corrective focus in the literature feature an emotive flavor that could be characterized as indignation on the part of the speaker (see also examples in (17) above). In this context, it is interesting that Repp (2016) uses the category of ‘degree of contrastiveness’ for defining a variety of information-structural notions, and she claims that corrective focus features the highest degree of contrast. It might be this intensification component that corrective statements share with the scalar component of mirativity readings and that often produces emotivity of corrective statements. One hypothesis would be that, in languages where corrective focus can as well be expressed without fronting (as in German, but also in Italian, with different preferences as discussed above), the fronting is not due to the mere substitution of an alternative, but rather is motivated by EN and thus the speaker’s expression of an emotion towards the current discourse and the relevant background assumptions. If this is on the right track, then we would also have to rethink prominent cases like Hungarian in this context (e.g., Horvath 2010). Repp (2016) has argued that, in Hungarian too, the discourse relation of corrective focus may be relevant, because it often goes along with exhaustiveness. To my mind, rethinking the role that EN might play for syntactic processes that up to now have been looked at only from a truthconditional and non-emotive perspective is a promising new path for future research interested in the interface between language and emotion.

10.4 Questions within a model of communication Let us take stock. In the previous sections, I have discussed the various connections that exist between expressivity and a set of discourse meanings commonly summarized in the literature as the domain of information structure. Since both empirical domains are rather broad, I have first tried to organize the discussion by proposing a distinction between ‘expressivity—broad sense (EB)’ and ‘expressivity—narrow sense (EN)’. According to this distinction, EN covers all use-conditional phenomena that also convey emotive meanings. EB, on the other hand, includes EN plus use-conditional meanings in general, probably the more common reading of ‘expressivity’ in the formal semantics–pragmatics literature. Given the distinction between EB and EN, I have then discussed to what extent general information-structural notions (such as focusing/backgrounding and givenness) can be accounted for in terms of EB. In this context, I have sketched a proposal that postulates that

10.4 QUESTIONS WITHIN A MODEL OF COMMUNICATION

179

there is no such thing as information structure proper, and the most common information-structural concepts are in fact expressive meanings. I then turned to the concept of EN and how it connects to what has traditionally been accounted for in information-structural terms. In this domain, I have focused on marked word order involving the clausal periphery, and I have argued that EN, in contrast to EB, is conveyed not merely via prosodic means, but typically via marked word-order choices. My discussion suggested that EN (and thus emotivity) in fact motivates many of the marked fronting constructions that are commonly accounted for in information-structural terms, including prima facie emotionally unsuspicious cases like corrective focus. Since emotive concepts such as mirativity pattern together with corrective focus in many respects (at least for the limited set of languages sketched above), this expressive perspective on marked word order seems worth exploring further. Recall now from previous chapters that I have proposed the following functional architecture of the clause to account for the syntax of a variety of non-canonical questions involving scalar readings at the propositional level of attitude (e.g., why-like-what questions; pseudo-questions such as How cool is that?; how come questions; can’t-find-the-value questions with fronted [wh+Prt]-constituents; and non-canonical frontings in polar questions). (18)

[XP[Force; Emp]] Force0

[XP[Force; Emp]]

Emp0[intensity] [... TP ...]

This analysis could also be applied to the cases of corrective focus above, where we see that some reading of EN is involved. Overall, our discussion above has shown that large parts of what is commonly known as the field of information structure can actually be reconceptualized as the study of expressivity across grammar. However, this also means that there is no categorial distinction between expressivity (both EB and EN) and information structure. Both are types of use-conditional meaning and can thus be considered the same. Interestingly, then, this would mean that our analysis of non-canonical questions involving expressive readings can be simplified dramatically: we do not need any special projection to account for expressivity in questions, because all relevant phenomena can be analyzed by using a functional domain below Force that is traditionally available in most of the cartographic approaches to

180

THE MODEL OF COMMUNICATION

information-structural phenomena. The only difference to those traditional approaches in my account is that I postulate that this domain inherently includes an intensification component (emphasis) that takes scope at the propositional level. Up to this point, we have proposed an account for the various phenomena that we find in the domain of the expressivity of questions. I would like to refer to the famous model of communication by Bu¨hler (1934 [2011]: 28) in this context and claim that this dimension of questions concerns an aspect of linguistic signs that can clearly be separated from the illocutionary force. To see that, take a look at the modified version of Bu¨hler’s model (1934 [2011]) in Figure 10.1. OBJECTS/STATE OF AFFAIRS representation Expressivity ( formal semantic concept; ‘expressive dimension’) expression

SENDER

Force (deontic concept: S wants from A...) S

appeal

RECEIVER

Figure 10.1 A modified form of Bu¨hler’s model of communication (1934), incorporating Expressivity and Force; English translations according to Bu¨hler (1934 [2011]).

The separation of the two fundamental communicative functions ‘expression’ and ‘appeal’ dovetail nicely with the syntactic representation in (18): Expressivity and Force are two separate components of a syntactic utterance (i.e., a complex sign in Bu¨hler’s terms). In particular, all types of Forces—not only directives such as questions—can be analyzed within a deontic framework (i.e., all Forces are volitional from the speaker’s perspective ‘S wants from A…’; see Truckenbrodt 2006 and Zaefferer 2001). Force thus belongs to the ‘appeal’ function of human language. On the other hand, the ‘expression’ function covers all communicative phenomena where the speaker expresses any type of subjectivity that does not affect the other two functions of communication; most notably in our context, it does not have an impact on the appeal function. My syntactic analysis of non-canonical questions (and other speech acts) accounts for this dimension of human communication by postulating

10.5 QUESTIONS AND THE SYNTAX OF INDIRECT SPEECH ACTS

181

the Emphasis operator, which is distinct from Force and not able to change the respective Forces. Specifically, the underlying hypothesis is that, no matter which type of non-canonical question we are looking at, the Force remains the same, because all types discussed in this book are still directive speech acts (even when not asking for information, but for some other type of response). Given this general conceptual framework, we can explain the major phenomena discussed in the literature on the syntax of non-canonical questions. However, there is one missing component in our discussion that I have briefly mentioned in Chapter 1 of the book: Sometimes, an indirect use of interrogative syntax as a means to perform an indirect speech act is considered a non-canonical question as well; see our brief discussion and some references in Chapter 1. In the final section of the book, we will now turn to this domain.

10.5 Questions and the syntax of indirect speech acts Let me start this last section with a famous puzzle mentioned by Searle (1975). As already mentioned in Chapter 1, according to Searle’s seminal paper (1975), an indirect speech act is defined as follows: In indirect speech acts the speaker communicates to the hearer more than he actually says by way of relying on their mutually shared background information […] together with the general powers of rationality and inference on the part of the hearer. […] Unless a hearer has some inferential strategy for finding out when primary illocutionary points differ from literal illocutionary points, he has no way of understanding indirect illocutionary acts. (Searle 1975: 60–3)

In his discussion, Searle (1975) thus places a premium on the context and pragmatic inferencing based on Gricean implicatures (he speaks about ‘mutually shared background information’ and the ‘power of rationality’, which is obviously related to Grice’s cooperation principle (1975)). Crucially, classical examples for indirect speech acts in Searle’s paper are indirect requests performed by means of interrogatives such as Can you pass the salt? However, Searle himself noted a fundamental problem with this approach—and that problem touches on the interface with syntax: The biggest single problem with the foregoing analysis is this: If, as I have been arguing, the mechanisms by which indirect speech acts are meant and understood are perfectly general […] and not tied to any particular syntactical form, then why is it that some syntactical forms work better than

182

THE MODEL OF COMMUNICATION

others. Why can I ask you to do something by saying Can you hand me that book on the top shelf ? but not, or not very easily, by saying Is it the case that you at present have the ability to hand me that book on the top shelf ? (Searle 1975: 75)

Searle (1975) did not provide a detailed attempt at solving this problem. In the end, his interests were philosophical in nature, and therefore he did not investigate the syntactic differences between the different versions of indirect speech acts discussed in his paper in any detail. As already mentioned at the outset of the book in Chapter 1, one linguistic account that addresses the interface between syntax and speech acts in the context of indirect speech acts is proposed by Meibauer (2019). However, in his pragmatic approach he basically maintains Searle’s concept of indirect speech acts and, by emphasizing that sentence types are indicators of illocutionary force, defends a literalist (‘literal force’) hypothesis, which cannot solve Searle’s ‘biggest single problem’ (1975) either. Some scholars working in cognitive linguistics frameworks have also proposed relevant theories to account for the fundamental problem pointed out by Searle (1975) above. For instance, Stefanowitsch (2003) has suggested a link in terms of metonymic processes between indirect forms such as Can you pass the salt? and the speech act of a REQUEST (see Panther and Ko¨pcke 2008; Thornburg and Panther 1997 on how such metonymies work for the interface between syntax and speech acts in general). I acknowledge that such abstract cognitive modeling can indeed explain very general (and possibly universal) patterns in the domain of speech act syntax. However, from a syntactic perspective, one could also look into the various forms of indirect speech acts in more detail, with the hypothesis that generalizations relevant for the speech act interpretation are also possible by referring to the respective syntactic forms only. In other words, a generative theory about those forms could be possible. The only generative proposal I know of that has suggested a way of dealing with Searle’s challenge in the context of interrogatives and indirect speech acts—but only in passing—is the cartographic approach to illocutionary force of Coniglio and Zegrean (2012). According to them, indirect requests are a good example for illustrating that we must distinguish between clause type and the speech act performed, and so they mention the example in (19a) as further evidence for their general claim that the left periphery of the clause must contain two distinct syntactic projections (19b): one for the illocutionary force (ILL) and one for clause type (CT); see Coniglio and Zegrean (2012: 446):

10.5 QUESTIONS AND THE SYNTAX OF INDIRECT SPEECH ACTS

(19)

183

a Could you close the window, please? (#Yes, I could.) ILL = directive (requesting an action); CT = interrogative b [ILL [CT … [TP …]]]

However, this analysis cannot solve the biggest single problem of Searle (1975) either, because it is just restating the facts that there is a relevant distinction, but it cannot (and does not intend to) explain why certain interrogatives can and others cannot be used to perform a request speech act. If we take a step back and look at the relevant data, not from the perspective of Searle’s theory, but rather from the perspective of empirical cognitive science, then we could say that examples such as (19a) are not ‘indirect’ speech acts in any reasonable sense of the term ‘indirect’. In particular, as for the psychological reality of the distinction between direct and indirect speech acts, experimental studies have repeatedly shown that the processing of indirect speech acts expressing a request interpretation (like (19a) above) is not characterized by additional costs when compared to the processing of direct speech act comprehension (e.g., Ruytenbeek et al. 2017), and that the distinction is thus a theoretical (and probably artificial) one. Consequently, given the problematic status of the term ‘indirect’ speech acts, one could rather refer to utterances like (19a) as ‘natural conventions’, as recently proposed by Simons and Zollman (2019) in their extensive discussion of the nature of linguistic conventions. Look at Table 1.1 again (repeated here as Table 10.1 for convenience), which illustrates natural ways of performing a request in English (Simons and Zollman 2019: 13). As Searle (1975) has already pointed out, the main challenge is that even such a big set of possible forms is not arbitrary (e.g., alternative forms such as his Is it the case that you at present have the ability to hand me that book on the top shelf ? are not possible). Another challenge is that forms such as Can you pass me the salt? seem to be the unmarked and neutral way of performing a request. The imperative form Give me the salt, (please)!, even if containing the marker please, requires a much more specific social context than the ability question—namely, one where less politeness is required. Accordingly, the use of many (if not all) of the forms of indirect speech acts listed in Table 10.1 below affect a domain of interpretation that, according to the model of communication introduced in Section 10.4 above, can be located at the addressee-directed level of appeal. In other words, since indirect speech acts modify the way of how an utterance is directed at an addressee in a given communicative situation (e.g., polite versus non-polite), they also impact the Force of an utterance. Let us develop these lines of thought a little further. My hypothesis is that those patterns are not an idiosyncratic quirk of the grammar of requests, but

184

THE MODEL OF COMMUNICATION

Table 10.1. Natural ways of performing a REQUEST in English, including a proposal of classification means

form Asking about

Can you help

Would you be

Is it at all

Do you have a

ability

me with this?

able to help me

possible for you

moment to help

with this?

to help me with

me with this?

this? Asking about

Would you mind Do you mind

Are you willing

I’m hoping you

willingness

helping me with

helping me with

to help me with

wouldn’t mind

this?

this?

this?

helping me with this.

rather hold across speech acts. For instance, have a look at assertions like (20). Assertions are of course not in the same way addressee-directed as questions might be (being directive speech acts), but the hypothesis here is that they are also crucially determined by being directed at the addressee, given what we have said above about the appeal function more generally. In particular, the form in (20) is more likely to occur in an ongoing discourse like in (20′). In out-of-the-blue contexts, the configuration in (20′′) seems to be more adequate: (20)

Eliud has dropped out of the race.

(20′)

A: Excuse me, who has dropped out of the race? B: Eliud has dropped out of the race.

(20′′)

Did you know that Eliud has dropped out of the race?

Similarly, in a situation where you approach a stranger on the street (and hence no social relation whatsoever has been established), one would most probably use syntactic forms like in (21): (21)

a Could you tell me where the train station is? b Do you know if there is a way around the traffic?

10.5 QUESTIONS AND THE SYNTAX OF INDIRECT SPEECH ACTS

185

Note that both utterances are interpreted as wh-questions, but yet the versions Where is the train station? and Which is the way around the traffic?, respectively, do not work well when approaching a stranger (i.e., without a social precontext). I hasten to add that I consider markers like excuse me (like other verb-based markers and strategies) as separate speech acts whose very function is to establish the necessary social precontext for performing the speech act that follows them. However, even if excuse me is used, there is a difference between Excuse me, could you tell me where the train station is? and the more direct version Excuse me, where is the train station? All in all, the main idea in this coda of the book is that Searle’s ‘biggest single problem’ can find a syntactic explanation once we acknowledge that the way(s) in which the speaker directs an utterance at the addressee plays a major role in defining the speech act itself. This is why Bu¨hler’s model (1934 [2011]) in Section 10.4 is a good way to illustrate that connection: Force, as being part of the appeal function of human communication, is essentially characterized by how a sentence-level utterance is directed at the addressee. As a syntactic consequence, I hypothesize that we do not need a distinction between the encoding of illocutionary force in the C-system and the encoding of social roles in an even higher domain, as proposed by recent work on the syntax of speech acts (e.g., Haegeman 2014; Miyagawa 2022; Wiltschko 2021; and many others). The version in (22a) is modeled according to Haegeman and Hill (2013), but it is representative of a whole strand of research postulating that there is a Speech Act Phrase (SAP) above the complementizer system (C-system); in the seminal approach by Rizzi (1997), which I have adopted throughout the book, speech act meaning was still encoded in syntax as part of the C-system (22b):

(22) a Recent speech act syntax SAP

b The traditional account ForceP speech act

speech act FinP

ForceP/C-system utterance/p (...)

utterance/p (...)

To be sure, there is a lot of empirical evidence for the structural claim in (22a), and I cannot re-evaluate all the literature on why we should have (at least) two distinct projections for speaker and hearer within such a functional domain of

186

THE MODEL OF COMMUNICATION

SAP. Let me just mention one of the main pieces of evidence for the encoding of the speaker in the left periphery of the clause. The following example from Hill (2007: 2093) contains the Romanian particle hai in a declarative and its relevant translations: (23)

Hai cã iar am gres¸it/s-a gres¸it! hai that again have-1 erred/ARB-se has.erred ‘Damn, I messed it up again.’/ ‘Right, it has been messed up again.’

The uninflected form hai can occur across speech acts, and, depending on the respective speech act, it can have several interpretations: In (23) above, it features an evaluative reading. Crucially, hai can inflect for person and number (24); Hill (2007: 2092): (24)

Haidem sã începem lucrul! hai-1PL SUBJ start work ‘Let’s start the work!’

I cannot discuss the Romanian system of sentential particles in any detail in this chapter. The only point I want to highlight here is that hai is a verb-based particle (derived from Turkish (h)ajde ‘go!, let’s go!’; see Tschizmarova 2005). In this respect, its properties can be compared to similar items in various other languages. One prominent verb-based particle in the languages of Europe is the Spanish item mira (lit. ‘look!’), which can inflect for number as well, and can occur sentence-initially or -finally; see González López and Trotzke (2021) for detailed discussion. The meaning of Spanish mira can be paraphrased as follows (González López and Trotzke 2021: 25): (25)

⟦mira⟧ = ‘S knows p and wants to draw attention of H to p.’

At the end of their paper, González López and Trotzke (2021) mention an analysis where mira performs a separate speech act, based on previous work by Portner (2004), who has analyzed Hanging Topics in English as performing separate speech acts (in the imperative mood; i.e., S wants to draw someone’s attention to p/parts of p). The point is that verb-based particles, much the same as the excuse me pattern mentioned further above, function to establish a social context for a certain speech act and could thus be characterized as separate performatives and not as part of the speech act itself. Needless to say, to make such a claim for Romanian hai would require an in-depth discussion of its properties (including some potential counter-evidence already mentioned in the literature cited above). However, I merely would like to point out that, if such a performative perspective can be maintained also for prominent cases like Romanian hai, the necessity of postulating separate projections

10.5 QUESTIONS AND THE SYNTAX OF INDIRECT SPEECH ACTS

187

for encoding of the speaker and the hearer within a SAP domain loses some of its plausibility. On the other hand, the evidence for an addressee projection in the syntax of speech acts is much stronger—not least because of the obvious empirical domain of vocatives (see Slocum 2016 for a recent study) and the numerous cases of encoding fine-grained addressee-directed interpretations in the morphological inventories of many languages. One of the prominent examples for this is Japanese politeness marking, which clearly expresses the social relation to the addressee. Look at the following examples from Miyagawa (2022: 39): (26)

a Hanako-wa piza-o tabe-ru Hanako-TOP pizza-ACC eat-PRS ‘Hanako will eatCOLLOQUIAL pizza.’ b Hanako-wa piza-o tabe-mas-u Hanako-TOP pizza-ACC eat-MAS-PRS ‘Hanako will eatFORMAL pizza.’

The choice between using the version with or without -mas- depends on how the speaker perceives the relationship to the addressee. The challenge for a syntactic theory is that -mas- occurs very low in the structure (i.e., below Tense/TP), although it should scope over the whole utterance and thus be located in the left periphery. This is the case in other languages featuring politeness marking. For instance, Korean has so-called speech style particles that occur at this very periphery of the clause (see Portner et al. 2019, 2022). Miyagawa builds his analysis of politeness marking in Japanese on the fact that this type of addressee-directed marking at a distance corresponds to the well-known phenomenon of allocutive agreement, another strategy of morphologically encoding the social relation to the addressee (see Alok and Haddican 2022 for a recent overview). We have already seen in Chapter 6 of this book that one language where this type of agreement has been investigated in great detail is Basque (e.g., Oyharçabal 1993). Consider the following example from Chapter 6 again: (27)

Hi, Alemanian egon nau-k/-n /∗ naiz you Germany.IN be AUX.1SG.ABS-ALLOC.M/-F AUX.1SG.ABS ‘Dude, I was in Germany.’

When one uses the informal pronoun hi (informal ‘you’) and not the neutral version zu, the finite verb in Basque requires an additional morpheme that corresponds to the overt addressee. The version without the allocutive morpheme (-k or -a for masculine and -n or -na for feminine) would be deviant in

188

THE MODEL OF COMMUNICATION

this context. Crucially, the allocutive morpheme occurs low in the structure, but nevertheless agrees with the addressee at the level of the whole utterance. This parallel to Japanese politeness marking is the main motivation for Miyagawa (2022: 51–7) to propose that politeness marking can be modeled in terms of allocutive agreement:

(27)

AddrP

addrˈ

Addressee CP = utterance

addr

Cˈ

Spec CΦALLOCUTIVE

TP

... -mas- ... My hypothesis in the context of indirect speech acts now is that this syntax of politeness marking instantiates a more general aspect of speech act syntax and can also be employed for the distinction between the many ways of performing a certain speech act indicated in Table 10.1 above. For example, a request of the form (29a) is usually perceived as less polite than its counterpart in (29b): (29)

a Give me the salt! b Can you pass me the salt?

[– polite] [+ polite]

In this form, those parallels are only sketchy and barely scratch the surface, but they nevertheless help to see the general direction of how a syntactic account of indirect speech acts could be feasible. In particular, looking at the various linguistic forms in Table 10.1 above, the hypothesis is that the means of directing the utterance at the addressee can also be spread across the functional hierarchy of the clause and thus be found at different syntactic levels (e.g., choice of modals/modal verb, word order, particles, inflection of sentential subject, and so on). This heterogenous distribution of addressee-directedness is also well known in the technical literature on how to model agreement operations for complex phenomena such as allocutivity, which can be spread across several

10.5 QUESTIONS AND THE SYNTAX OF INDIRECT SPEECH ACTS

189

syntactic domains as well—an aspect of great interest in most recent research on the topic (e.g., Haddican and Etxeberria 2022). If we adopt this general idea, some of the data discussed in the literature can be analyzed differently and, to my mind, on more general grounds—taking into account their most basic pragmatic functions. Let us look at one example. As we have discussed in Chapter 6 of this book, languages like German can feature modal particles to indicate a non-canonical reading of a question. These particles, as we have shown, often interact with the emphasis projection (i.e., the expressivity) of the clause. However, there are also particles that do not yield non-canonical interpretations, but rather are used in neutral information-seeking questions; the famous and highly frequent German denn (lit. ‘then’) would be such a case. Interestingly, this particle affects the social dimension of performing a question speech act (and not the expressive one, as all those particles signaling non-canonical readings). To see that, observe the following contrast: (30)

Context: You, on the phone with your dentist: a Wann kann ich vorbeikommen? when can I come.by b Wann kann ich denn vorbeikommen? when can I PART come.by ‘When can I come by?’

[– polite] [+ polite]

There is a vast amount of literature on the topic of the exact interpretation of denn (and its cognates in other languages), some of which is summarized in Trotzke and Haegeman (2022). My goal here is merely to indicate that the use or non-use of German denn can (and not must) have a social meaning component, and this is in fact often highlighted for particles in general in the non-technical literature on learning those particles as a second-language user, which speaks of “negative consequences in social interaction” (Busse 1992: 39) in the context of either using or not using pragmatic elements like German denn. Based on this illustration, one hypothesis we could entertain is that other ways of performing information-seeking questions politely (such as Excuse me, Q or Could you tell me Q; see examples above) are instantiations of the same allocutive morpheme that can be expressed by denn. Looking at questions through the glasses of this new syntactic approach may have other important consequences. For instance, coming back to the Japanese politeness marker -mas- introduced above, it is an interesting observation that the marker -mas-

190

THE MODEL OF COMMUNICATION

is obligatory in some unmarked information-seeking wh-questions; data and judgments from Miyagawa (1987: 362): (31)

a Dare-ga ki-mas-u who-NOM come-MAS-PRS ‘Who will come?’ b ∗ Dare-ga kuru ka? who-NOM come Q ‘Who will come?’

ka? Q

This could indicate the interesting observation that information-seeking questions, without any specific precontext, must agree for a [+polite] feature (i.e., come with a social cost). On the other hand, we would predict that morphemes that signal colloquial and non-polite language should be ruled out in those neutral questions. And this is what we can observe in Basque allocutive marking: as we have discussed in detail in Chapter 6, Basque features particles that mark non-canonical questions, but some varieties also feature particles that are obligatory in unmarked questions. In eastern dialects, the particle -a is used to mark neutral polar questions. Crucially, allocutive marking signaling no social distance to the addressee is ruled out in those questions, as we have illustrated with the following example in Chapter 6: (32)

Jinen da(∗ -n)-a hire laguna afarira come.FUT AUX-ALLOC.F.PART your friend.ABS dinner.ADL ‘Dude, is your friend coming to the dinner?’

The overall idea behind this new way of analyzing questions and speech acts (be they ‘direct’ or ‘indirect’) is thus that speech act syntax (the Force/appeal component of the utterance) is fundamentally shaped by how the utterance is (or is not) directed at the addressee. As a consequence, it puts the speaker’s perceived social relation to the addressee at the center of speech act syntax. With those general concepts in mind, let us now draw some general conclusions and end this book.

10.6 Conclusions In this chapter, I have first introduced the distinction between ‘expressivity— broad sense (EB)’ and ‘expressivity—narrow sense (EN)’ in order to provide a framework for how we can explain the diverse syntactic analyses proposed in

10.6 CONCLUSIONS

191

this book for different types of non-canonical questions. Given the distinction between EB and EN, the main take-home message was that non-canonical questions encoding the Emphasis operator in the syntactic representation of the clause pattern with other speech acts in involving the same kind of operator. Moreover, the relevant functional domain of this emphasis projection (below Force, but still taking propositional scope) is not categorically different from other types of use-conditional meanings such as information structure— they all belong to the same communicative function of linguistic signs, which can be referred to as ‘expression’ and ‘expressivity’, respectively. After that, I have briefly looked at one of the other basic communicative functions: the addressee-directed function of appeal. I have indicated how a syntax of noncanonical questions in the domain of indirect speech acts could be possible, and I concluded that both the Force of an utterance and the various ways of how an utterance can be directed at the addressee (e.g., polite versus nonpolite) basically refer to the same type of communicative effect—namely, to the appeal function. At the end of this book, I would like to point out that this perspective on speech act syntax is connected to a different view of language as a whole (which may meet the requirement of a different type of ‘outside theory of language and discourse to interact with syntax’, demanded by Ga¨rtner and Steinbach 2006 to tackle the issue of mapping syntax onto Force). Specifically, this view diverges from the standard approach that both the speaker and the addressee play a fundamental role in speech act syntax. This standard account has its semantic–pragmatic background in the pervasive concept that the pragmatics of speech acts can be analyzed by modeling them as manipulations of the common ground and public belief sets for both speaker and addressee (Farkas and Bruce 2010; Stalnaker 2002; and many more). From the philosophical perspective of so-called addressee-directed intentionalism (a position very much in line with the ideas outlined in this final chapter of the book), Harris (2020) provides a critical discussion of this recent trend to view speech acts within such a discourse model. One of the arguments against such an account of speech acts is the so-called coordinated-attack problem. The scenario illustrates that, although two participants in a communication can be unable to add information to the common ground, they can still communicate with each other in every sense of the term. Harris’s conclusion (2020) is that changing the conversational context (the common ground) is a by-product of speech acts, but not their essential aim. The nature of speech acts can thus not be understood by referring to the common ground

192

THE MODEL OF COMMUNICATION

and a public set of beliefs. Rather, the nature of speech acts is fundamentally addressee-directed. The scenario is as follows (Harris 2020: 2720): two generals are perched on hills straddling a valley in which their common enemy is encamped. If they attack at the same time, they will defeat the enemy, and this is their goal. But uncoordinated attacks will result in a defeat so severe that they must avoid that outcome at all costs, and so neither is willing to attack unless they are certain that they will do so in a coordinated way. The hitch is that the only channel by which they can communicate is unreliable: they can send messengers to each other through the valley, but it is always likely that a messenger will be captured.

Given this scenario, the two generals can send the message: ‘we will attack at dawn’ to each other, but they can never be sure that the other general has received their respective messages. In particular, general A can start the conversation by sending a first message (‘we will attack at dawn’), but, even if general B’s reply ‘I got your message. We will attack at dawn’ reaches general A, then general B cannot be sure whether the message has reached general A (and general A knows that general B cannot be sure). This back-and-forth of messages can continue ad infinitum. The point is that we would still say that the two generals communicate with each other. Given this view, the essential goal of speech acts is to change the addressee’s mind, without being sure, however, that the abstract common and public context will be changed. The hypothesis put forward in this final chapter of the book is that this addresseedirected nature of performing speech acts is also reflected in the syntax of natural language. Situations of successful human communication, where the speaker cannot add information to the common ground, but succeeds in the relevant communicative aim anyway, are actually quite common. Harris (2020: 2723–4) provides an additional example that refers to email communication: If you write to a colleague whose inbox is notoriously overwhelmed that, owing to an urgent meeting, you will be late for the colloquium, you cannot be sure whether the message has really reached your addressee. Nevertheless, you have performed a communicative act with the aim of changing the belief of your addressee—and you can later pick up on that act when meeting your colleague at the department (A: Did you read my mail? B: Yes. A: The meeting has been canceled, so I won’t be late). We constantly perform speech acts, although we cannot be sure whether the addressee(s) will understand. Just think about

10.6 CONCLUSIONS

193

the digital world: the whole domain of social media is characterized by that uncertainty. The new perspectives presented in this chapter on how syntactically to account for indirect speech acts in the domain of questions indicated some first ideas about how speakers try to minimize the risk of social failure, and how this human capacity (essentially rooted in mindreading capabilities) is reflected in the linguistic choices we constantly make.

References Abraham, Werner (1991). ‘Discourse particles in German: How does their illocutive force come about?’, in Werner Abraham (ed.), Discourse Particles: Descriptive and Theoretical Investigations on the Logical, Syntactic and Pragmatic Properties of Discourse Particles in German. Amsterdam: John Benjamins, 203–52. Adli, Aria (2006). ‘French wh-in-situ questions and syntactic optionality: Evidence from three data types’, Zeitschrift fu¨r Sprachwissenschaft 25: 163–203. Aikhenvald, Alexandra Y. (2011). ‘Mood’, in Patrick Colm Hogan (ed.), The Cambridge Encyclopedia of the Language Sciences. Cambridge: Cambridge University Press, 512–13. Aikhenvald, Alexandra Y. (2012). ‘The essence of mirativity’, Linguistic Typology 16: 435–86. Alberdi, Xabier (1994). Euskararen tratamenduak: Erabilera. Vitoria-Gasteiz: UPV-EHU. Alcázar, Asier (2016). ‘Minor speech acts in a Basque mirative construction’, Florida Linguistics Papers 3: 1–12. Alcázar, Asier (2017). ‘A syntactic analysis of rhetorical questions’, in Aaron Kaplan, Abby Kaplan, Miranda K. McCarvel, and Edward J. Rubin (eds), Proceedings of the 34th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Press, 32–41. Alok, Deepak, and Bill Haddican (2022). ‘The formal heterogeneity of allocutivity’, Glossa: A Journal of General Linguistics 7: 1. Arregi, Karlos, and Andrew Nevins (2012). Morphotactics: Basque Auxiliaries and the Structure of Spellout. Dordrecht: Springer. Asher, Nicholas, and Brian Reese (2007). ‘Intonation and discourse: Biased questions’, Interdisciplinary Studies on Information Structure 8: 1–38. Asu, Eva Liina, Heete Sahkai, and Pa¨rtel Lippus (2022). ‘The prosody of surprise questions in Estonian’, Journal of Linguistics: First View: 1–21. Authier, J.-Marc, and Liliane Haegeman (2019). ‘The syntax of mirative focus fronting: Evidence from French’, in Deborah L. Arteaga (ed.), Contributions of Romance Languages to Current Linguistic Theory. Cham: Springer, 39–63. Bailyn, John F. (2012). The Syntax of Russian. Cambridge: Cambridge University Press. Baker, Carl L. (1970). ‘Notes on the description of English questions’, Foundations of Language 6: 197–219. Bayer, Josef (2001). ‘A symmetry in emphatic topicalization’, in Caroline Féry and Wolfgang Sternefeld (eds), Audiatur vox sapientiae: A Festschrift for Arnim von Stechow. Berlin: Akademie-Verlag, 15–47. Bayer, Josef, and Lisa Lai-Shen Cheng (2017). ‘Wh-in-situ’, in Martin Everaert and Henk van Riemsdijk (eds), The Blackwell Companion to Syntax. Oxford: Blackwell, 4979–5022. Bayer, Josef, and Probal Dasgupta (2016). ‘Emphatic topicalization and the structure of the left periphery: Evidence from German and Bangla’, Syntax 19: 309–53. Bayer, Josef, Jana Ha¨ussler, and Markus Bader (2016). ‘A new diagnostic for cyclic whmovement: Discourse particles in German questions’, Linguistic Inquiry 47: 591–629.

REFERENCES

195

Bayer, Josef, and Hans-Georg Obenauer (2011). ‘Discourse particles, clause structure, and question types’, Linguistic Review 28: 449–91. Bayer, Josef, and Andreas Trotzke (2015). ‘The derivation and interpretation of left peripheral discourse particles’, in Josef Bayer, Roland Hinterho¨lzl, and Andreas Trotzke (eds), Discourse-Oriented Syntax. Amsterdam: John Benjamins, 13–40. Beijering, Karin, Gunther Kaltenbo¨ck, and Marı´a Sol Sansiñena (2019). ‘Insubordination: Central issues and open questions’, in Karin Beijering, Gunther Kaltenbo¨ck, and Marı´a Sol Sansiñena (eds), Insubordination: Theoretical and Empirical Issues. Berlin: Mouton de Gruyter, 1–28. Beltrama, Andrea (2018). ‘Totally between subjectivity and discourse: Exploring the pragmatic side of intensification’, Journal of Semantics 35: 219–61. Beltrama, Andrea, and Emily A. Hanink (2019). ‘Marking imprecision, conveying surprise: Like between hedging and mirativity’, Journal of Linguistics 55: 1–34. Beltrama, Andrea, and Laura Staum Casasanto (2017). ‘Totally tall sounds totally younger: Intensification at the socio-semantics interface’, Journal of Sociolinguistics 21: 154–82. Beltrama, Andrea, and Andreas Trotzke (2019). ‘Conveying emphasis for intensity: Lexical and syntactic strategies’, Language and Linguistics Compass 13: e12343. Benz, Anton, Manfred Krifka, Tue Trinh, and Kazuko Yatsushiro (forthcoming) (eds). Perspectives on Biased Questions. Berlin: Language Science Press. Beukeboom, Camiel J., and Christian Burgers (2017). ‘Linguistic bias’, Oxford Research Encyclopedia of Communication, https://doi.org/10.1093/acrefore/9780190228613.013. 439. Beun, Robbert-Jan (1990). ‘The recognition of Dutch declarative questions’, Journal of Pragmatics 14: 39–56. Bianchi, Valentina, Giuliano Bocci, and Silvio Cruschina (2015). ‘Focus fronting and its implicatures’, in Enoch O. Aboh, Jeannette C. Schaeffer, and Petra Sleeman (eds), Romance Languages and Linguistic Theory 2013: Selected Papers from ‘Going Romance’. Amsterdam: John Benjamins, 1–20. Bianchi, Valentina, Giuliano Bocci, and Silvio Cruschina (2016). ‘Focus fronting, unexpectedness, and evaluative implicatures’, Semantics & Pragmatics 9: 1–54. Bianchi, Valentina, and Silvio Cruschina (2016). ‘The derivation and interpretation of polar questions with a fronted focus’, Lingua 170: 47–68. Biezma, Maria (2011). Anchoring Pragmatics in Syntax and Semantics. U. Mass. dissertation. Biezma, Marı´a, and Kyle Rawlins (2017). ‘Rhetorical questions: Severing asking from questioning’, Semantics and Linguistic Theory 27: 302–22. Bocci, Giuliano, Valentina Bianchi, and Silvio Cruschina (2021). ‘Focus in wh-questions: Evidence from Italian’, Natural Language & Linguistic Theory 39: 405–55. Brandner, Ellen (2010). ‘On the syntax of verb-initial exclamatives’, Studia Linguistica 64: 81–115. Brunetti, Lisa, Hiyon Yoo, Lucia M. Tovena, and Rachel Albar (2021). ‘French reasoncomment “how” questions: A view from prosody’, in Andreas Trotzke and Xavier Villalba (eds), Expressive Meaning across Linguistic Levels and Frameworks. Oxford: Oxford University Press, 248–78. Bu¨hler, Karl (1934) [2011]. Theory of Language: The Representational Function of Language, trans. Donald Fraser Goodwin in collaboration with Achim Eschbach. Amsterdam: John Benjamins. Bu¨ring, Daniel, and Christine Gunlogson (2000). ‘Aren’t positive and negative polar questions the same?’, TS, https://semanticsarchive.net [last accessed April 20, 2023].

196

REFERENCES

Busse, Dietrich (1992). ‘Partikeln im Unterricht Deutsch als Fremdsprache’, Muttersprache 102: 37–59. Camino, Iñaki (2019). ‘An overview of Basque dialects’, Linguistic Minorities in Europe Online [De Gruyter database], https://doi.org/10.1515/lme.9497680. Caponigro, Ivano, and Jon Sprouse (2007). ‘Rhetorical questions as questions’, Proceedings of Sinn und Bedeutung 11: 121–33. Cardinaletti, Anna (2011). ‘German and Italian modal particles and clause structure’, Linguistic Review 28: 493–31. Cardinaletti, Anna (2015a). ‘Italian verb-based discourse particles in a comparative perspective’, in Josef Bayer, Roland Hinterho¨lzl, and Andreas Trotzke (eds), DiscourseOriented Syntax. Amsterdam: John Benjamins, 71–91. Cardinaletti, Anna (2015b). ‘What do you do if you don’t have modal particles?’, in Ellen Brandner, Anna Czypionka, Constantin Freitag, and Andreas Trotzke (eds), Webschrift for Josef Bayer. Konstanz: University of Konstanz, 16–21, https://ling.sprachwiss.unikonstanz.de/pages/WebschriftBayer/2015/contents_files/Cardinaletti.pdf [last accessed April 20, 2023]. Carlson, Matthew T., Antonio Fábregas, and Michael T. Putnam (2021). ‘How wide the divide? Theorizing “constructions” in generative and usage-based frameworks’, Frontiers in Psychology 12: 601303. Castroviejo Miró, Elena (2006). Wh-exclamatives in Catalan. Universitat de Barcelona dissertation. Celle, Agnès, Anne Jugnet, and Laure Lansari (2021). ‘Expressive questions in English and French’, in Andreas Trotzke and Xavier Villalba (eds), Expressive Meaning across Linguistic Levels and Frameworks. Oxford: Oxford University Press, 138–66. Chang, Lisa (1997). Wh-in-situ Phenomena in French. University of British Columbia dissertation. Cheng, Lisa Lai-Shen (1991). On the Typology of wh-questions. MIT dissertation. Cheng, Lisa Lai-Shen (1997). On the Typology of wh-questions. New York, NY: Garland. Cheng, Lisa Lai-Shen (2003). ‘Wh-in-situ’, Glot International 7: 103–9. Cheng, Lisa Lai-Shen (2009). ‘Wh-in-situ, from the 1980s to now’, Language and Linguistics Compass 3: 767–91. Cheng, Lisa Lai-Shen (2022). ‘Response particles hai6 in Cantonese’, Journal of Chinese Linguistics 50: 532–44. Chernilovskaya, Anna (2014). Exclamativity in Discourse: Exploring the Exclamative Speech Act from a Discourse Perspective. Utrecht University dissertation. Chernilovskaya, Anna, Cleo Condoravdi, and Sven Lauer (2012). ‘On the discourse effects of wh-exclamatives’, in Nathan Arnett and Ryan Bennett (eds), Proceedings of the 30th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla, 109–19. Chomsky, Noam (1986). Barriers. Cambridge, MA: MIT Press. Chomsky, Noam (1995). The Minimalist Program. Cambridge, MA: MIT Press. Chomsky, Noam (2000). ‘Minimalist inquiries: The framework’, in Roger Martin, David Michaels, and Juan Uriagereka (eds), Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik. Cambridge, MA: MIT Press, 89–155. Chomsky, Noam (2001). ‘Derivation by phase’, in Michael Kenstowicz (ed.), Ken Hale: A Life in Language. Cambridge, MA: MIT Press, 1–52. Cinque, Guglielmo (1999). Adverbs and Functional Heads: A Cross-Linguistic Perspective. Oxford: Oxford University Press. Clark, Billy (2021). Pragmatics: The Basics. London: Routledge. Cohen, Ariel (2007). ‘Incredulity questions’, Decalog 2007: 133.

REFERENCES

197

Collins, Chris (1991). ‘Why and how come’, MIT Working Papers in Linguistics 15: 31–45. Collins, Peter (2020). ‘Clause types’, in Bas Aarts, April McMahon, and Lars Hinrichs (eds), The Handbook of English Linguistics. Second edition. New York: Wiley, 131–44. Coniglio, Marco (2011). Die Syntax der deutschen Modalpartikeln: Ihre Distribution und Lizenzierung in Haupt- und Nebensa¨tzen. Berlin: Akademie Verlag. Coniglio, Marco, and Iulia Zegrean (2012). ‘Splitting up force: Evidence from discourse particles’, in Lobke Aelbrecht, Liliana Haegeman, and Rachel Nye (eds), Main Clause Phenomena. Amsterdam: John Benjamins, 229–55. Craenenbroeck, Jeroen van (2005). ‘Adverbial modification under sluicing’, in Kiyong Choi and Changguk Yim (eds), Ellipsis in Minimalism: Proceedings of the Seventh Seoul International Conference on Generative Grammar. Seoul: Hankook, 77–102. Croft, William (1994). ‘Speech act classification, language typology, and cognition’, in Savas L. Tsohatzidis (ed.), Foundations of Speech Act Theory. London: Routledge, 468–85. Cruschina, Silvio (2012). Discourse-Related Features and Functional Projections. Oxford: Oxford University Press. Cruschina, Silvio (2019). ‘Focus fronting in Spanish: Mirative implicature and information structure’, Probus 31: 119–46. Cruschina, Silvio, and Valentina Bianchi (2021). ‘Mirative implicatures at the syntax– semantics interface: A surprising association and an unexpected move’, in Andreas Trotzke and Xavier Villalba (eds), Expressive Meaning across Linguistic Levels and Frameworks. Oxford: Oxford University Press, 86–107. Cutting, Joan, and Kenneth Fordyce (2020). Pragmatics: A Resource Book for Students. London: Routledge. d’Avis, Franz (2016). ‘Different languages—different sentence types? On exclamative sentences’, Linguistics and Language Compass 10: 159–75. Danckaert, Lieven (2012). Latin Embedded Clauses: The Left Periphery. Amsterdam: John Benjamins. Dayal, Veneeta (2016). Questions. Oxford: Oxford University Press. De Clercq, Karen (2017). ‘Prosody as an argument for a layered left periphery’, Nederlandse taalkunde 22: 31–9. De Haan, Ferdinand (1999). ‘Evidentiality and epistemic modality: Setting boundaries’, Southwest Journal of Linguistics 18: 83–101. de Marneffe, Marie-Catherine, and Judith Tonhauser (2019). ‘Inferring meaning from indirect answers to polar questions: The contribution of the rise-fall-rise contour’, in Malte Zimmermann, Klaus von Heusinger, and Edgar Onea (eds), Questions in Discourse. Leiden: Brill, 132–63. de Reuse, Willem J. (2003). ‘Evidentiality in Western Apache’, in Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds), Studies in Evidentiality. Amsterdam: John Benjamins, 79–100. Degen, Judith, Andreas Trotzke, Gregory Scontras, Eva Wittenberg, and Noah D. Goodman (2019). ‘Definitely, maybe: A new experimental paradigm for investigating the pragmatics of evidential devices across languages’, Journal of Pragmatics 140: 33–48. Dehé, Nicole, and Bettina Braun (2020). ‘The prosody of rhetorical questions in English’, English Language & Linguistics 24: 607–35. Del Gobbo, Francesca, Nicola Munaro, and Cecilia Poletto (2015). ‘On sentential particles: A cross-linguistic study’, in Sylvie Hancil, Alexander Haselow, Margje Post (eds), Final Particles. Berlin: Mouton de Gruyter, 359–86. DeLancey, Scott (1997). ‘Mirativity: The grammatical marking of unexpected information’, Linguistic Typology 1: 33–52.

198

REFERENCES

den Dikken, Marcel, and Anastasia Giannakidou (2002). ‘From hell to polarity: “Aggressively non-d-linked” wh-phrases as polarity items’, Linguistic Inquiry 33: 31–61. Desmets, Marianne, and Antoine Gautier (2009). ‘Comment n’y ai-je pas songé plus tôt? Questions rhétoriques en comment’, Travaux de linguistique 1: 107–25. Destruel, Emilie, David I. Beaver, and Elizabeth Coppock (2019). ‘It’s not what you expected! The surprising nature of cleft alternatives in French and English’, Frontiers in Psychology 10: 1400. Do¨rre, Laura (2018). Primary and Secondary Meaning in the Processing of German Particles. University of Konstanz dissertation. Do¨rre, Laura, and Andreas Trotzke (2019). ‘The processing of secondary meaning: An experimental comparison of focus and modal particles in wh-questions’, in Daniel Gutzmann and Katharina Turgay (eds), Secondary Content: The Semantics and Pragmatics of Side Issues. Leiden: Brill, 143–67. Duryagin, Pavel (2020). ‘On some factors affecting the choice of tune in Russian whquestions’, Proceedings of Speech Prosody 2020: 235–9. Eckardt, Regine (2007). ‘Inherent focus on wh-phrases’, Proceedings of Sinn und Bedeutung 11: 209–28. Egg, Markus, and Malte Zimmermann (2012). ‘Stressed out! Accented discourse particles: The case of doch’, Proceedings of Sinn und Bedeutung 16: 225–38. Elliott, Dale E. (1974). ‘Toward a grammar of exclamation’, Foundations of Language 11: 231–46. Elordieta, Arantzazu (2001). Verb Movement and Constituent Permutation in Basque. University of Utrecht dissertation. Elordieta, Gorka (1997). Morphosyntactic Feature Chains and Phonological Domains. University of Southern California dissertation. Endo, Yoshio (2018). ‘Variation in wh-expressions asking for a reason’, Linguistic Variation 18: 299–314. Endo, Yoshio (2022). ‘Non-standard questions in English, German, and Japanese’, Linguistics Vanguard 8: 251–60. Erteschik-Shir, Nomi (2007). Information Structure: The Syntax–Discourse Interface. Oxford: Oxford University Press. Espinal, M. Teresa, and Susagna Tubau (2019). ‘Response systems: The syntax and semantics of fragment answers and response particles’, Annual Review of Linguistics 5: 261–87. Etxepare, Ricardo, and Myriam Uribe-Etxebarria (2008). ‘On negation and focus in Spanish and Basque’, Anuario del Seminario de Filologı´a Vasca ‘Julio de Urquijo’, https://ojs.ehu. eus/index.php/ASJU/article/download/9791/9221 [last accessed April 20, 2023]. Euskaltzaindia [The Royal Academy of the Basque Language] (1987). Euskal gramatika: Lehen urratsak II. Bilbao: Euskaltzaindia. Evans, Nicholas (2007). ‘Insubordination and its uses’, in Irina Nikolaeva (ed.), Finiteness: Theoretical and Empirical Foundations. Oxford: Oxford University Press, 366–431. Fanselow, Gisbert (2002). ‘Quirky subjects and other specifiers’, in Ingrid Kaufmann and Barbara Stiebels (eds), More than Words. Berlin: Akademie Verlag, 227–50. Farkas, Donka (2020). ‘Canonical and non-canonical questions’, TS, UC Santa Cruz/Princeton University, https://semanticsarchive.net [last accessed April 20, 2023]. Farkas, Donka (2022). ‘Non-intrusive questions as a special type of non-canonical questions’, Journal of Semantics 39: 295–337. Farkas, Donka, and Kim Bruce (2010). ‘On reacting to assertions and polar questions’, Journal of Semantics 27: 81–118.

REFERENCES

199

Fernandez, Beatriz, and Milan Rezac (2016). ‘Differential object marking in Basque varieties’, in Beatriz Fernandez and Jon Ortiz de Urbina (eds), Microparameters in the Grammar of Basque. Amsterdam: John Benjamins, 93–138. Féry, Caroline, and Shinichiro Ishihara (2016) (eds). The Oxford Handbook of Information Structure. Oxford: Oxford University Press. Foolen, Ad (2012). ‘The relevance of emotion for language and linguistics’, in Ad Foolen, Ulrike M. Lu¨dtke, Timothy P. Racine, and Jordan Zlatev (eds), Moving Ourselves, Moving Others: Motion and Emotion in Intersubjectivity, Consciousness and Language. Amsterdam: John Benjamins, 349–68. Foolen, Ad (2016). ‘Expressives’, in Nick Riemer (ed.), The Routledge Handbook of Semantics. London: Routledge, 473–90. Frank, Jane (1990). ‘You call that a rhetorical question? Forms and functions of rhetorical questions in conversation’, Journal of Pragmatics 14: 723–38. Frey, Werner (2004a). ‘The grammar–pragmatics interface and the German prefield’, Sprache und Pragmatik 52: 1–39. Frey, Werner (2004b). ‘A medial topic position for German’, Linguistische Berichte 198: 153–90. Frey, Werner (2010). ‘Ā-movement and conventional implicatures: About the grammatical encoding of emphasis in German’, Lingua 120: 1416–35. Garcı´a-Carpintero, Manuel (2013). ‘Explicit performatives revisited’, Journal of Pragmatics 49: 1–17. Garmendia, Joana (2014). ‘Ote: hiztunaren ziurtasunik eza’, Gogoa 12: 7–26. Ga¨rtner, Hans-Martin, and Markus Steinbach (2006). ‘A skeptical note on the syntax of speech acts and point of view’, in Patrick Brandt and Eric Fuß (eds), Form, Structure, Grammar. Berlin: Akademie Verlag, 213–22. Geurts, Bart, and Rob van der Sandt (2004). ‘Interpreting focus’, Theoretical Linguistics 30: 1–44. Ghomeshi, Jila, Ray Jackendoff, Nicole Rosen, and Kevin Russell (2004). ‘Contrastive focus reduplication in English (the salad–salad paper)’, Natural Language & Linguistic Theory 22: 307–57. Giannakidou, Anastasia, and Alda Mari (2021). Truth and Veridicality in Grammar and Thought: Mood, Modality, and Propositional Attitudes. Chicago, IL: University of Chicago Press. Giannakidou, Anastasia, and Alda Mari (2022). ‘Bias and modality in questions’, TS, University of Chicago and Institut Jean Nicod, École Normale Superieure Paris. Giannakidou, Anastasia, and Suwon Yoon (2011). ‘The subjective mode of comparison: Metalinguistic comparatives in Greek and Korean’, Natural Language & Linguistic Theory 29: 621–55. Ginzburg, Jonathan, and Raquel Fernández (2010). ‘Computational models of dialogue’, in Alexander Clark, Chris Fox, and Shalom Lappin (eds), The Handbook of Computational Linguistics and Natural Language Processing. New York: Wiley, 429–81. Ginzburg, Jonathan, and Ivan A. Sag (2000). Interrogative Investigations: The Form, Meaning, and Use of English Interrogatives. Stanford, CA: CSLI Publications. Giorgi, Alessandra (2016). ‘On the temporal interpretation of certain surprise questions’, SpringerPlus 5: 1390. Giorgi, Alessandra (2018). ‘Ma non era rosso? (but wasn’t it red?): On counterexpectational questions in Italian’, in Lori Repetti and Francisco Ordóñez (eds), Selected Papers from the 46th Linguistic Symposium on Romance Languages (LSRL). Amsterdam: John Benjamins, 69–84.

200

REFERENCES

Glasbergen, Aliza (2021). Questions in Context: The Case of French wh-in-situ. Leiden University dissertation. Goddard, Cliff (2013). ‘Interjections and emotion (with special reference to “surprise” and “disgust”)’, Emotion Review 6: 53–63. González López, Laura, and Andreas Trotzke (2021). ‘¡Mira! The grammar–attention interface in the Spanish left periphery’, Linguistic Review 38: 5–31. Goodhue, Daniel (2018). On Asking and Answering Biased Polar Questions. McGill University dissertation. Goodhue, Daniel (2022). ‘Isn’t there more than one way to bias a polar question?’, Natural Language Semantics, OnlineFirst: 1–35. Grice, H. Paul (1975). ‘Logic and conversation’, in Peter Cole and Jerry L. Morgan (eds), Syntax and Semantics, vol. iii: Speech Acts. New York: Academic Press, 41–58. Grimshaw, Jane (1979). ‘Complement selection and the lexicon’, Linguistic Inquiry 10: 279–326. Groenendijk, Jeroen, and Martin Stokhof (1991). ‘Dynamic predicate logic’, Linguistics and Philosophy 14: 39–100. Grosz, Patrick (2012). On the Grammar of Optative Constructions. Amsterdam: John Benjamins. Grosz, Patrick G. (2014). ‘Optative markers as communicative cues’, Natural Language Semantics 22: 89–115. Grosz, Patrick G. (2016). ‘Information structure and discourse particles’, in Caroline Féry and Shinichiro Ishihara (eds), The Oxford Handbook of Information Structure. Oxford: Oxford University Press, 336–58. Grosz, Patrick G. (2020). ‘Discourse particles’, in Daniel Gutzmann, Lisa Matthewson, Cécile Meier, Hotze Rullmann, and Thomas Ede Zimmermann (eds), The Wiley Blackwell Companion to Semantics. New York: Wiley. https://doi.org/10.1002/ 9781118788516.sem047. Gunlogson, Christine (2003). True to Form: Rising and Falling Declaratives as Questions in English. London: Routledge. Gutzmann, Daniel (2015). Use-Conditional Meaning: Studies in Multidimensional Semantics. Oxford: Oxford University Press. Gutzmann, Daniel (2017). ‘Modal particles ≠ modal particles (= modal particles)’, in Josef Bayer and Volker Struckmeier (eds), Discourse Particles: Formal Approaches to their Syntax and Semantics. Berlin: Mouton de Gruyter, 144–72. Gutzmann, Daniel (2019). The Grammar of Expressivity. Oxford: Oxford University Press. Gutzmann, Daniel, and Katharina Turgay (2014). ‘Expressive intensifiers and external degree modification’, Journal of Comparative Germanic Linguistics 17: 185–228. Haddican, Bill (2001). ‘Basque functional heads’, TS, NYU, https:// www.semanticscholar.org/paper/Basque-Functional-Heads-Haddican/ b879bbe6a77f9a10a21e2cdc293fbe206c42554a [last accessed April 20, 2023]. Haddican, Bill (2008). ‘Euskararen espez-buru-osagarri hurrenkeraren aldeko argudio batzuk’, in Iñigo Arteatx, Xabier Artiagoitia, and Arantzazu Elordieta (eds), Antisimetriaren Hipotesia vs Buru-parametroa: euskararen oinarrizko hurrenkera ezbaian. Bilbao: UPV-EHU, 87–124. Haddican, Bill (2018). ‘The syntax of Basque allocutive clitics’, Glossa: A Journal of General Linguistics 3: 1. Haddican, Bill, and Urtzi Etxeberria (2022). ‘Embedded allocutivity in Basque’, Syntax 25: 545–66.

REFERENCES

201

Haegeman, Liliane (2002). ‘Anchoring to speaker, adverbial clauses, and the structure of CP’, Georgetown University Working Papers in Linguistics 17: 109–41. Haegeman, Liliane (2014). ‘West Flemish verb-based discourse markers and the articulation of the speech act layer’, Studia Linguistica 68: 116–39. Haegeman, Liliane, and Virginia Hill (2013). ‘The syntacticization of discourse’, in Raffaella Folli, Christina Sevdali, and Robert Truswell (eds), Syntax and its Limits. Oxford: Oxford University Press, 370–90. Hall, Joan Kelly (2019). ‘The contributions of conversation analysis and interactional linguistics to a usage-based understanding of language: Expanding the transdisciplinary framework’, Modern Language Journal 103: 80–94. Halliday, Michael A. K. (1967). ‘Notes on transitivity and theme in English: Part 2’, Journal of Linguistics 3: 199–244. Hamblin, Charles L. (1976). ‘Questions in Montague English’, in Barbara H. Partee (ed.), Montague Grammar. New York: Academic Press, 247–59. Han, Chung-hye (2002). ‘Interpreting interrogatives as rhetorical questions’, Lingua 112. 201–29. Harris, Daniel W. (2020). ‘We talk to people, not contexts’, Philosophical Studies 177: 2713–33. Hartmann, Katharina, and Malte Zimmermann (2007). ‘In place—out of place? Focus in Hausa’, in Kerstin Schwabe and Susanne Winkler (eds), On Information Structure, Meaning and Form: Generalizing across Languages. Amsterdam: John Benjamins, 365–403. Hedberg, Nancy, and Juan M. Sosa (2002). ‘The prosody of questions in natural discourse’, Proceedings of Speech Prosody 2002: 375–78. Hedberg, Nancy, Juan M. Sosa, and Emrah Go¨rgu¨lu¨ (2017). ‘The meaning of intonation in yes–no questions in American English: A corpus study’, Corpus Linguistics and Linguistic Theory 13: 321–68. Heim, Irene (1982). The Semantics of Definite and Indefinite Noun Phrases. U. Mass. dissertation. Henriksen, Nicholas C. (2012). ‘The intonation and signaling of declarative questions in Manchego Peninsular Spanish’, Language and Speech 55: 543–76. Heycock, Caroline (2017). ‘Embedded root phenomena’, in Martin Everaert and Henk van Riemsdijk (eds), The Blackwell Companion to Syntax. Oxford: Blackwell, 174–209. Hill, Virginia (2002). ‘Complementizer phrases (CP) in Romanian’, Rivista di Linguistica 14: 223–48. Hill, Virginia (2007). ‘Vocatives and the pragmatics–syntax interface’, Lingua 117: 2077– 105. Hinterho¨lzl, Roland, and Nicola Munaro (2015). ‘On the interpretation of modal particles in non-assertive speech acts in German and Bellunese’, in Josef Bayer, Roland Hinterho¨lzl, and Andreas Trotzke (eds), Discourse-Oriented Syntax. Amsterdam: John Benjamins, 41–70. Hinterho¨lzl, Roland, and Nicola Munaro (2021). ‘On the illocutionary force of exclamatives and non-canonical questions’, in Andreas Trotzke and Xavier Villalba (eds), Expressive Meaning across Linguistic Levels and Frameworks. Oxford: Oxford University Press, 43–65. Hohaus, Vera, and M. Ryan Bochnak (2020). ‘The grammar of degree: Gradability across languages’, Annual Review of Linguistics 6: 235–59. Holmberg, Anders (2013). ‘The syntax of answers to polar questions in English and Swedish’, Lingua 128: 31–50. Holmberg, Anders (2015). The Syntax of Yes and No. Oxford: Oxford University Press.

202

REFERENCES

Hooper, Joan, and Sandra A. Thompson (1973). ‘On the applicability of root transformations’, Linguistic Inquiry 4: 465–97. Holtgraves, Thomas (1998). ‘Interpreting indirect replies’, Cognitive Psychology 37: 1–27. Hornstein, Norbert, and Jairo Nunes (2014). ‘Minimalism and control’, in Andrew Carnie, Yosuke Sato, and Daniel Siddiqi (eds), The Routledge Handbook of Syntax. London: Routledge, 239–63. Horvath, Julia (2010). ‘Discourse features, syntactic displacement and the status of contrast’, Lingua 120: 1346–69. Irwin, Patricia (2014). ‘So [totally] speaker-oriented: An analysis of Drama SO’, in Raffaella Zanuttini and Laurence R. Horn (eds), Microsyntactic Variation in North American English. Oxford: Oxford University Press, 29–70. Jackson, Joshua Conrad, et al. (2019). ‘Emotion semantics show both cultural variation and universal structure’, Science 366: 1517–22. Jakobson, Roman (1960). ‘Closing statements: Linguistics and poetics’, in Thomas A. Sebeok (ed.), Style in Language. Cambridge, MA: MIT Press, 350–77. Julien, Marit (2015). ‘The force of V2 revisited’, Journal of Comparative Germanic Linguistics 18: 139–81. Jurafsky, Daniel (2004). ‘Pragmatics and computational linguistics’, in Laurence Horn and Gregory Ward (eds), Handbook of Pragmatics. Oxford: Blackwell, 578–604. Kadmon, Nirit, and Fred Landman (1993). ‘Any’, Linguistics and Philosophy 16: 353–422. Kaplan, David (1999). ‘The meaning of ouch and oops: Explorations in the theory of meaning as use’, TS, UCLA, https://eecoppock.info/PragmaticsSoSe2012/kaplan.pdf [last accessed April 20, 2023]. Karttunen, Lauri (1977). ‘Syntax and semantics of questions’, Linguistics and Philosophy 1: 3–44. Kasper, Gabriele (2004). ‘Speech acts in (inter)action: Repeated questions’, Intercultural Pragmatics 1: 125–33. Kennedy, Christopher, and Louise McNally (2005). ‘Scale structure, degree modification, and the semantics of gradable predicates’, Language 81: 345–81. Kim, Okgi, and Jong-Bok Kim (2017). ‘English how come constructions: A diachronic perspective’, Language and Information 21: 121–34. Kiss, Katalin É. (1999). ‘The English cleft construction as a focus phrase’, Amsterdam Studies in the Theory and History of Linguistic Science Series 4: 217–30. Kratzer, Angelika (1999). ‘Beyond Ouch and Oops: How descriptive and expressive meanings interact’, paper presented at the Cornell Conference on Context Dependency, Cornell University, https://www.semanticsarchive.net [last accessed April 20, 2023]. Kratzer, Angelika (2004). ‘Interpreting focus: Presupposed or expressive meanings? A comment on Geurts and van der Sandt’, Theoretical Linguistics 30: 123–36. Kratzer, Angelika, and Elisabeth Selkirk (2020). ‘Deconstructing information structure’, Glossa: A Journal of General Linguistics 5: 1. Krifka, Manfred (1985). ‘Word order and word order change in Swahili: A categorial grammatical study’, Kiswhahili. Journal of the Institute of Kiswahili Research (Dar es Salaam) 52: 13–72. Krifka, Manfred (1995). ‘The semantics and pragmatics of polarity items’, Linguistic Analysis 25: 209–57. Krifka, Manfred (2001). ‘Quantifying into question acts’, Natural Language Semantics 9: 1–40.

REFERENCES

203

Krifka, Manfred (2008). ‘Basic notions of information structure’, Acta Linguistica Hungarica 55: 243–76. Krifka, Manfred (2011). ‘Questions’, in Klaus von Heusinger, Claudia Maienborn, and Paul Portner (eds), Semantics: An International Handbook of Natural Language Meaning, vol. ii. Berlin: Mouton de Gruyter, 1742–85. Krifka, Manfred (2014). ‘Embedding illocutionary acts’, in Tom Roeper and Margaret Speas (eds), Recursion: Complexity in Cognition. Dordrecht: Springer, 125–55. Krifka, Manfred (2015). ‘Bias in commitment space semantics: Declarative questions, negated questions, and question tags’, Semantics and Linguistic Theory 25: 328–45. Krifka, Manfred (2023). ‘Layers of assertive clauses: Propositions, judgements, commitments, acts’, in Jutta Hartmann and Angelika Wo¨llstein (eds), Propositionale Argumente im Sprachvergleich / Propositional Arguments in Cross-Linguistic Research: Theorie und Empirie / Theoretical and Empirical Issues. Tu¨bingen: Narr, 116–83. Kuong, Io-Kei Joaquim (2008). ‘Yes/no question particles revisited: The grammatical functions of mo4, me1, and maa3’, Proceedings of the North American Conference on Chinese Linguistics 20: 715–33. Kuwabara, Kazuki (2013). ‘Peripheral effects in Japanese questions and the fine structure of CP’, Lingua 126: 92–119. Laka, Itziar (1990). Negation in Syntax: On the Nature of Functional Categories and Projections. MIT dissertation. Laka, Itziar (2019). On the Syntax of Negation. London: Routledge. Lasersohn, Peter (1999). ‘Pragmatic halos’, Language 75: 522–51. Lasnik, Howard, and Mamuro Saito (1984). ‘On the nature of proper government’, Linguistic Inquiry 15: 235–89. Lee, Seung-Hee (2012). ‘Response design in conversation’, in Tanya Stivers and Jack Sidnell (eds), The Handbook of Conversation Analysis. New York: Wiley, 415–32. Leino, Jaakko (2013). ‘Information structure’, in Graeme Trousdale and Thomas Hoffmann (eds), The Oxford Handbook of Construction Grammar. Oxford: Oxford University Press, 329–44. Levinson, Stephen C. (1983). Pragmatics. Cambridge: Cambridge University Press. Li, Boya (2006). Chinese Final Particles and the Syntax of the Periphery. University of Utrecht dissertation. Liddicoat, Anthony J. (2021). An Introduction to Conversation Analysis. Third edition. London: Bloomsbury. Macagno, Fabrizio, and Sarah Bigi (2017). ‘Analyzing the pragmatic structure of dialogues’, Discourse Studies 19: 148–68. McCready, Elin, and Norry Ogata (2007). ‘Evidentiality, modality and probability’, Linguistics and Philosophy 30: 147–206. McTear, Michael (2020). ‘Conversational AI: Dialogue systems, conversational agents, and chatbots’, Synthesis Lectures on Human Language Technologies 13: 1–251. Majid, Asifa (2019). ‘Mapping words reveals emotional diversity’, Science 366: 1444–45. Marocchini, Eleonora, and Filippo Domaneschi (2022). ‘Can you read my mind? Conventionalized indirect requests and Theory of Mind abilities’, Journal of Pragmatics 193: 201–21. Meibauer, Jo¨rg (1986). Rhetorische Fragen. Berlin: Mouton de Gruyter. Meibauer, Jo¨rg (1991). ‘Existenzimplikaturen bei rhetorischen W-Fragen’, in Marga Reis and Inger Rosengren (eds), Fragen und Fragesa¨tze. Tu¨bingen: Niemeyer, 223–42. Meibauer, Jo¨rg (1994). Modaler Kontrast und konzeptuelle Verschiebung: Studien zur Syntax und Semantik deutscher Modalpartikeln. Tu¨bingen: Niemeyer.

204

REFERENCES

Meibauer, Jo¨rg (2019). ‘What is an indirect speech act? Reconsidering the literal force hypothesis’, Pragmatics & Cognition 26: 61–84. Meisnitzer, Benjamin (2012). ‘Modality in the Romance languages: Modal verbs and modal particles’, in Werner Abraham and Elisabeth Leiss (eds), Modality and Theory of Mind Elements across Languages. Berlin: Mouton de Gruyter, 335–60. Miyagawa, Shigeru (1987). ‘LF affix raising in Japanese’, Linguistic Inquiry 18: 362–7. Miyagawa, Shigeru (2022). Syntax in the Treetops. Cambridge, MA: MIT Press. Monforte, Sergio (2018). ‘Question particles in Basque’. Isogloss 4: 29–53. Monforte, Sergio (2020a). ‘Syntactic analyses of discourse particles through the microvariation of Basque ote’, Glossa: A Journal of General Linguistics 5: 1. Monforte, Sergio (2020b). Galderetako-a, al eta ote partikulak euskaraz: Sintaxia, mikroaldakortasuna eta interpretazioa. University of the Basque Country dissertation. Monforte, Sergio (2021). ‘On the pragmatic development of modal particles in Navarrese– Lapurdian Basque’, in Alexander Haselow and Sylvie Hancil (eds), Studies at the Grammar-Discourse Interface: Discourse Markers and Discourse-Related Grammatical Phenomena. Amsterdam: John Benjamins, 57–76. Montague, Richard (1970). ‘Universal grammar’, Theoria 36: 373–98. Morzycki, Marcin (2009). ‘Degree modification of gradable nouns: Size adjectives and adnominal degree morphemes’, Natural Language Semantics 17: 175–203. Morzycki, Marcin (2012). ‘Adjectival extremeness: Degree modification and contextually restricted scales’, Natural Language & Linguistic Theory 30: 567–609. Morzycki, Marcin (2016). Modification. Cambridge: Cambridge University Press. Munaro, Nicola, and Hans-Georg Obenauer (1999). ‘On underspecified wh-elements in pseudo-interrogatives’, University of Venice Working Papers in Linguistics 9: 181–253. Munaro, Nicola, and Cecilia Poletto (2002). ‘Ways of clausal typing’, Rivista di grammaticagenerativa 27: 87–105. Munaro, Nicola, and Cecilia Poletto (2009). ‘Sentential particles and clausal typing in Venetan dialects’, in Benjamin Shaer, Philippa Cook, Werner Frey, and Claudia Maienborn (eds), Dislocated Elements in Discourse: Syntactic, Semantic and Pragmatic Perspectives. London: Routledge, 173–99. Nasu, Norio (2012). ‘Topic particle stranding and the structure of CP’, in Lobke Aelbrecht, Liliane Haegeman, and Rachel Nye (eds), Main Clause Phenomena. Amsterdam: John Benjamins, 205–28. Neeleman, Ad, and Reiko Vermeulen (2012). ‘The syntactic expression of information structure’, in Ad Neeleman and Reiko Vermeulen (eds), The Syntax of Topic, Focus, and Contrast: An Interface-Based Approach. Berlin: Mouton de Gruyter, 1–38. Newmeyer, Frederick J. (2009). ‘On split CPs and the “perfectness” of language’, in Benjamin Shaer, Philippa Cook, Werner Frey, and Claudia Maienborn (eds), Dislocated Elements in Discourse: Syntactic, Semantic, and Pragmatic Perspectives. London: Routledge, 114–40. Nicolle, Steve (2000). ‘Markers of general interpretive use in Amharic and Swahili’, in Thorstein Fretheim and Gisle Andersen (eds), Pragmatic Markers and Propositional Attitude. Amsterdam: John Benjamins, 173–88. Niebuhr, Oliver (2010). ‘On the phonetics of intensifying emphasis in German’, Phonetica 67: 170–98. Norrick, Neal R. (2009). ‘Interjections as pragmatic markers’, Lingua 41: 866–91. Noveck, Ira A., and Dan Sperber (2004) (eds). Experimental Pragmatics. Basingstoke: Palgrave Macmillan. Nunes, Jairo (2012). ‘Sideward movement: Triggers, timing, and outputs’, in Myriam Uribe-Etxebarria and Vidal Valmala (eds), Ways of Structure Building. Oxford: Oxford University Press, 114–42.

REFERENCES

205

Nuyts, Jan (1999). Epistemic Modality, Language, and Conceptualization. Amsterdam: John Benjamins. Nye, Rachel (2009). How Pseudo-Questions and the Interpretation of wh-clauses in English. University of Essex MA dissertation. Obenauer, Hans-Georg (2004). ‘Non-standard wh-questions and alternative checkers in Pagotto’, in Horst Lohnstein and Susanne Trissler (eds), The Syntax and Semantics of the Left Periphery. Berlin: Mouton de Gruyter, 343–84. Obenauer, Hans-Georg (2006). ‘Special interrogatives: Left periphery, wh-doubling, and (apparently) optional elements’, in Jenny Doetjes and Paz González (eds), Romance Languages and Linguistic Theory 2004, Amsterdam: John Benjamins, 247–73. Odria, Ane (2014). ‘Differential object marking and the nature of dative case in Basque varieties’, Linguistic Variation 14: 289–317. Oraby, Shereen, Vrindavan Harrison, Amita Misra, Ellen Riloff, and Marilyn Walker (2017). ‘Are you serious? Rhetorical questions and sarcasm in social media dialog’, arXiv preprint arXiv:1709.05305. Ortiz de Urbina, Jon (1989). ‘Dislocaciones verbales en estructuras de polaridad’, ASJU 23: 393–410. Ortiz de Urbina, Jon (1993). ‘Checking domains in Basque and Breton’, ASJU 27: 751–75. Ortiz de Urbina, Jon (1994). ‘Verb-initial patterns in Basque and Breton’, Lingua 94: 125–53. Oyharçabal, Bernard (1993). ‘Verb agreement with nonarguments: On allocutive agreement’, in Jóse Ignacio Hualde and Jon Ortiz de Urbina (eds), Generative Studies in Basque Linguistics. Amsterdam: John Benjamins, 89–114. Panther, Klaus-Uwe, and Klaus-Michael Ko¨pcke (2008). ‘A prototype approach to sentences and sentence types’, Annual Review of Cognitive Linguistics 6: 83–112. Park, Jinho (2020). ‘Evidential and/or mirative components of some verbal endings and periphrastic constructions in Korean’, in Chungmin Lee and Jinho Park (eds), Evidentials and Modals. Leiden: Brill, 399–411. Partee, Barbara H. (1984). ‘Compositionality’, in Fred Landman and Frank Veltman (eds), Varieties of Formal Semantics: Proceedings of the 4th Amsterdam Colloquium, Dordrecht: Foris, 281–311. Paul, Waltraud, and Victor Junnan Pan (2017). ‘What you see is what you get: Chinese sentence-final particles as head-final complementizers’, in Josef Bayer and Volker Struckmeier (eds), Discourse Particles: Formal Approaches to their Syntax and Semantics. Berlin: Mouton de Gruyter, 49–77. Pesetsky, David, and Esther Torrego (2007). ‘The syntax of valuation and the interpretability of features’, in Simin Karimi, Vida Samiian, and Wendy K. Wilkins (eds), Phrasal and Clausal Architecture: Syntactic Derivation and Interpretation. Amsterdam: John Benjamins, 262–94. Peterson, Tyler (2010). ‘Examining the mirative and nonliteral uses of evidentials’, University of British Columbia Working Papers in Linguistics 28: 129–59. Peterson, Tyler (2013). ‘Rethinking mirativity: The expression and implication of surprise’, TS, University of Toronto. Petrone, Caterina, and Oliver Niebuhr (2014). ‘On the intonation of German intonation questions: The role of the prenuclear region’, Language and Speech 57: 108–46. Portner, Paul (2004). ‘Vocatives, topics, and imperatives’, TS, Georgetown University, https://pdfs.semanticscholar.org/7977/af790507b218c682b9a500b512cdaf5be1b5. pdf [last accessed April 20, 2023]. Portner, Paul (2018). Mood. Oxford: Oxford University Press. Portner, Paul, Miok Pak, and Raffaella Zanuttini (2019). ‘The speaker-addressee relation at the syntax-semantics interface’, Language 95: 1–36.

206

REFERENCES

Portner, Paul, Miok Pak, and Raffaella Zanuttini (2022). ‘Dimensions of honorific meaning in Korean speech style particles’, Glossa: A Journal of General Linguistics 7: 1. Potts, Christopher (2005a). The Logic of Conventional Implicatures. Oxford: Oxford University Press. Potts, Christopher (2005b). ‘Lexicalized intonational meaning’, University of Massachusetts Occasional Papers 30: 129–46. Potts, Christopher (2007a). ‘The expressive dimension’, Theoretical Linguistics 33: 165–98. Potts, Christopher (2007b). ‘Conventional implicatures: A distinguished class of meanings’, in Gillian Ramchand and Charles Reiss (eds), The Oxford Handbook of Linguistic Interfaces. Oxford: Oxford University Press, 475–501. Rampton, Ben (2019). ‘Interactional sociolinguistics’, in Karin Tusting (ed.), The Routledge Handbook of Linguistic Ethnography. London: Routledge, 13–27. Rathcke, Tamara (2006). ‘A perceptual study on Russian questions and statements’, Arbeitsberichte des Instituts fu¨r Phonetik und digitale Sprachverarbeitung der Universita¨t Kiel (AIPUK) 37: 51–62. Recanati, François (1987). Meaning and Force: The Pragmatics of Performative Utterances. Cambridge: Cambridge University Press. Recanati, François (2005). ‘Literalism and contextualism: Some varieties’, in Gerhard Preyer and Georg Peter (eds), Contextualism in Philosophy: Knowledge, Meaning, and Truth. Oxford: Oxford University Press, 171–96. Reese, Brian (2006). ‘The meaning and use of negative polar interrogatives’, Empirical Issues in Syntax and Semantics 6: 331–54. Reese, Brian (2007). Bias in Questions. The University of Texas at Austin dissertation. Reese, Brian, and Nicholas Asher (2007). ‘Prosody and the interpretation of tag questions’, Proceedings of Sinn und Bedeutung 11: 448–62. Reimer, Laura, and Andreas Trotzke (2022). ‘The landscape of non-canonical “only” in German’, Linguistics Vanguard 8: 129–38. Repp, Sophie (2010). ‘Defining “contrast” as an information-structural notion in grammar’, Lingua 120: 1333–45. Repp, Sophie (2013). ‘D-linking vs. degrees: Inflected and uninflected ‘welch’ in exclamatives and rhetorical questions’, in Holden Ha¨rtl (ed.), Interfaces of Morphology. Berlin: Akademie Verlag, 59–90. Repp, Sophie (2016). ‘Contrast: Dissecting an elusive information-structural notion and its role in grammar’, in Caroline Féry and Shinichiro Ishihara (eds), The Oxford Handbook of Information Structure. Oxford: Oxford University Press, 270–89. Repp, Sophie, and Ljudmila Geist (forthcoming). ‘Responding to negative biased questions in Russian’, in Anton Benz, Manfred Krifka, Tue Trinh, and Kazuko Yatsushiro (eds), Perspectives on Biased Questions. Berlin: Language Science Press. Repp, Sophie, and Lena Rosin (2015). ‘The intonation of echo wh-questions’, Proceedings of Interspeech 16: 938–42. Rett, Jessica (2008). Degree Modification in Natural Language. Rutgers University dissertation. Rett, Jessica (2011). ‘Exclamatives, degrees and speech acts’, Linguistics and Philosophy 34: 411–42. Rett, Jessica (2021). ‘A comparison of expressives and miratives’, in Andreas Trotzke and Xavier Villalba (eds), Expressive Meaning across Linguistic Levels and Frameworks. Oxford: Oxford University Press, 191–215. Rett, Jessica, and Sarah Murray (2013). ‘A semantic account of mirative evidentials’, Semantics and Linguistic Theory 23: 453–72.

REFERENCES

207

Rett, Jessica, and Beth Sturman (2020). ‘Prosodically marked mirativity’, Proceedings of WCCFL 37. de Rijk, Rudolf (1969). ‘Is Basque an S.O.V. language?’, FLV 3: 319–51. Rizzi, Luigi (1990). ‘Speculations on verb-second’, in Joan Mascaró and Marina Nespor (eds), Grammar in Progress: GLOW Essays for Henk van Riemsdijk. Dordrecht: Foris, 137–52. Rizzi, Luigi (1997). ‘The fine structure of the left periphery’, in Liliane Haegeman (ed.), Elements of Grammar: Handbook in Generative Syntax, Dordrecht: Kluwer, 281–337. Rizzi, Luigi (2001). ‘On the position of “Int(errogative)” in the left periphery of the clause’, in Guglielmo Cinque and Giampaolo Salvi (eds), Current Studies in Italian Syntax. Amsterdam: Elsevier, 267–96. Rizzi, Luigi (2014). ‘Syntactic cartography and the syntacticisation of scope-discourse semantics’, in Anne Reboul (ed.), Mind, Values, and Metaphysics. Dordrecht: Springer, 517–33. Roberts, Craige (1996). ‘Information structure in discourse: Toward a unified theory of formal pragmatics’, Ohio State University Working Papers in Linguistics 49: 91–136. Roberts, Craige (2012). ‘Information structure in discourse: Towards an integrated formal theory of pragmatics’, Semantics and Pragmatics 5: 1–69. Rojas-Esponda, Tania (2014). ‘A discourse model for u¨berhaupt’, Semantics and Pragmatics 7: 1–45. Rosemeyer, Malte, Daniel Jacob, and Lars Konieczny (2022). ‘How alternatives are created: Specialized background knowledge affects the interpretation of clefts in discourse’, in Davide Garassino and Daniel Jacob (eds), When Data Challenges Theory. Amsterdam: John Benjamins, 116–46. Ross, John R. (1970). ‘On declarative sentences’, in Roderick A. Jacobs and Peter S. Rosenbaum (eds), Readings in English Transformational Grammar. Waltham, MA: Ginn & Co, 222–72. Ruytenbeek, Nicolas (2021). Indirect Speech Acts. Cambridge: Cambridge University Press. Ruytenbeek, Nicolas, Ekaterina Ostashchenko, and Mikhail Kissine (2017). ‘Indirect request processing, sentence types and illocutionary forces’, Journal of Pragmatics 119: 46–62. Sadock, Jerrold (1971). ‘Queclaratives’, Proceedings of the Annual Meeting of the Chicago Linguistic Society 7: 223–31. Sadock, Jerrold (1974). Towards a Linguistic Theory of Speech Acts. New York: Academic Press. Sadock, Jerrold, and Arnold Zwicky (1985). ‘Speech act distinctions in syntax’, in Timothy Shopen (ed.), Language Typology and Syntactic Description: Clause Structure, vol. i. Cambridge: Cambridge University Press, 155–96. Safárová, Marie, and Marc Swerts (2004). ‘On recognition of declarative questions in English’, Proceedings of Speech Prosody 2004, https://www.isca-speech.org/archive_ open/sp2004/sp04_313.pdf [last accessed April 20, 2023]. Sauerland, Uli, and Kazuko Yatsushiro (2017). ‘Remind-me presuppositions and speech-act decomposition: Evidence from particles in questions’, Linguistic Inquiry 48: 651–77. Schlenker, Philippe (2007). ‘Expressive presuppositions’, Theoretical Linguistics 33: 237–45. Searle, John R. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, John R. (1975). ‘Indirect speech acts’, in Peter Cole and Jerry L. Morgan (eds), Syntax and Semantics. Speech Acts, vol. iii. New York: Academic Press, 59–82.

208

REFERENCES

Searle, John R. (1976). ‘A classification of illocutionary acts’, Language in Society 5: 1–23. Searle, John R., and Daniel Vanderveken (1985). Foundations of Illocutionary Logic. Cambridge: Cambridge University Press. Seeliger, Heiko, and Sophie Repp (2018). ‘Biased declarative questions in Swedish and German’, in Christine Dimroth and Stefan Sudhoff (eds), The Grammatical Realization of Polarity Contrast: Theoretical, Empirical, and Typological Approaches. Amsterdam: John Benjamins, 129–72. Servidio, Emilio, Giuliano Bocci, and Valentina Bianchi (2018). ‘(Dis)agreement, polarity, and focus: Answering negative polar questions in Italian’, Glossa: A Journal of General Linguistics 3: 1. Shlonsky, Ur (2010). ‘The cartographic enterprise in syntax’, Language and Linguistics Compass 4: 417–29. Shlonsky, Ur (2012). ‘Notes on wh in situ in French’, in Laura Brugè, Anna Cardinaletti, Giuliana Giusti, Nicola Munaro, and Cecilia Poletto (eds), Functional Heads: The Cartography of Syntactic Structures 7. Oxford: Oxford University Press, 242–52. Shlonsky, Ur, and Gabriela Soare (2011). ‘Where’s why?’, Linguistic Inquiry 42: 651–69. Si, Fuzhen, and Luigi Rizzi (2021) (eds). Current Issues in Syntactic Cartography: A Crosslinguistic Perspective. Amsterdam: John Benjamins. Siemund, Peter (2015). ‘Exclamative clauses in English and their relevance for theories of clause types’, Studies in Language 39: 698–728. Siemund, Peter (2018). Speech Acts and Clause Types: English in a Cross-Linguistic Context. Oxford: Oxford University Press. Simons, Mandy, and Kevin J. Zollman (2019). ‘Natural conventions and indirect speech acts’, Philosophers’ Imprint 19: 1–26. Skopeteas, Stavros, and Gisbert Fanselow (2011). ‘Focus and the exclusion of alternatives: On the interaction of syntactic structure with pragmatic inference’, Lingua 121: 1693– 1706. Slocum, Poppy (2016). The Syntax of Address. State University of New York at Stony Brook dissertation. Stalnaker, Robert (2002). ‘Common ground’, Linguistics and Philosophy 25: 701–21. Stefanowitsch, Anatol (2003). ‘A construction-based approach to indirect speech acts’, in Klaus-Uwe Panther and Linda L. Thornburg (eds), Metonymy and Pragmatic Inferencing. Amsterdam: John Benjamins, 105–26. Stefanowitsch, Anatol (2011). ‘Keine Grammatik ohne Konstruktionen: ein logischo¨konomisches Argument fu¨r die Konstruktionsgrammatik’, in Stefan Engelberg, Anke Holler, and Kristel Proost (eds), Sprachliches Wissen zwischen Lexikon und Grammatik. Berlin: Mouton de Gruyter, 181–210. Stivers, Tanya (2010). ‘An overview of the question–response system in American English conversation’, Journal of Pragmatics 42: 2772–81. Stivers, Tanya, N. J. Enfield, Penelope Brown, and Stephen C. Levinson (2009). ‘Universals and cultural variation in turn-taking in conversation’, Proceedings of the National Academy of Sciences 106: 10587–92. Svetozarova, Natalija (1998). ‘Intonation in Russian’, Intonation systems 261–74. Takano, Yuji (2020). ‘Exploring Merge: A new form of sideward movement’, Linguistic Review 37: 7–45. Taniguchi, Ai (2017). The Formal Pragmatics of Non-At-Issue Intensification in English and Japanese. Michigan State University dissertation.

REFERENCES

209

Thornburg, Linda L., and Karl-Uwe Panther (1997). ‘Speech act metonymies’, in WolfAndreas Liebert, Linda R. Waugh, and Gisela Redeker (eds), Discourse and Perspective in Cognitive Linguistics. Amsterdam: John Benjamins, 205–22. Thurmair, Maria (1989). Modalpartikeln und ihre Kombinationen. Tu¨bingen: Niemeyer. Tovena, Lucia M. (2022). ‘A comparative corpus study on a case of non-canonical question’, Linguistics Vanguard 8: 209–17. Trotzke, Andreas (2017a). The Grammar of Emphasis: From Information Structure to the Expressive Dimension. Berlin: Mouton de Gruyter. Trotzke, Andreas (2017b). ‘German grammar and mirativity: Experimental evidence’, Review of Cognitive Linguistics 15: 460–88. Trotzke, Andreas (2018). ‘DP-internal modal particles: A case study of German ja’, Studia Linguistica 72: 322–39. Trotzke, Andreas (2019). ‘Approaching the pragmatics of exclamations experimentally’, in Eszter Ronai et al. (eds), Proceedings of the 54th Annual Meeting of the Chicago Linguistic Society (CLS 54). Chicago: Chicago Linguistic Society, 527–40. Trotzke, Andreas (2020a). ‘How cool is that! A new “construction” and its theoretical challenges’, The Journal of Comparative Germanic Linguistics 23: 327–65. Trotzke, Andreas (2020b). ‘Constructions in minimalism: A functional perspective on cyclicity’, Frontiers in Psychology 11: 2152. Trotzke, Andreas, Ermenegildo Bidese, and Manuela Moroni (2020). ‘German discourse particles in the second language classroom: Teasing apart learning problems at the syntax-pragmatics interface’, Pedagogical Linguistics 1: 184–210. Trotzke, Andreas, and Anna Czypionka (2022). ‘The pragmatics of surprise–disapproval questions: An empirical study’, Linguistics Vanguard 8: 239–49. Trotzke, Andreas, and Liliane Haegeman (2022). ‘PP-internal particles in Dutch as evidence for PP-internal discourse structure’, in Remus Gergel, Ingo Reich, and Augustin Speyer (eds), Particles in German, English and Beyond. Amsterdam: John Benjamins, 297–322. Trotzke, Andreas, and Sergio Monforte (2019). ‘Basque question particles: Implications for a syntax of discourse particles’, Linguistic Variation 19: 352–85. Trotzke, Andreas, and Giuseppina Turco (2015). ‘The grammatical reflexes of emphasis: Evidence from German wh-questions’, Lingua 168: 37–56. Trotzke, Andreas, and Xavier Villalba (2020). ‘Exclamatives as responses at the syntax– pragmatics interface’, Journal of Pragmatics 168: 139–71. Trotzke, Andreas, and Xavier Villalba (2021). ‘Expressive insubordination: A crosslinguistic study on that-exclamatives’, in Andreas Trotzke and Xavier Villalba (eds), Expressive Meaning across Linguistic Levels and Frameworks. Oxford: Oxford University Press, 108–20. Trotzke, Andreas, and Jan-Wouter Zwart (2014). ‘The complexity of narrow syntax: Minimalism, representational economy, and simplest Merge’, in Frederick Newmeyer and Laurel Preston (eds), Measuring Grammatical Complexity. Oxford: Oxford University Press, 128–47. Truckenbrodt, Hubert (2006). ‘On the semantic motivation of syntactic verb movement to C in German’, Theoretical Linguistics 32: 257–306. Truckenbrodt, Hubert (2013). ‘Selbsta¨ndige Verb-Letzt-Sa¨tze’, in Jo¨rg Meibauer, Markus Steinbach, and Hans Altmann (eds), Satztypen des Deutschen. Berlin: Mouton de Gruyter, 232–46. Tschizmarova, Ivelina K. (2005). ‘Hedging functions of the Bulgarian discourse marker hajde’, Journal of Pragmatics 37: 1143–63.

210

REFERENCES

van der Wouden, Ton, and Ad Foolen (2017). ‘A most serious and extraordinary problem. Intensification of adjectives in Dutch, German, and English’, Leuvense Bijdragen 101: 82–100. Villalba, Xavier (2003). ‘An exceptional exclamative sentence type in Romance’, Lingua 113: 713–45. Villalba, Xavier (2008). ‘Exclamatives: A thematic guide with many questions and few answers’, Catalan Journal of Linguistics 7: 9–40. Villalba, Xavier (2017a). ‘Exclamatives, imperatives, optatives’, in Andreas Dufter and Elisabeth Stark (eds), Manual of Romance Morphosyntax and Syntax. Berlin: Mouton de Gruyter, 603–46. Villalba, Xavier (2017b). ‘Non-asserted material in Spanish degree exclamatives: An experimental study on extreme degree’, in Ignacio Bosque (ed.), Advances in the Analysis of Spanish Exclamatives. Columbus: Ohio State University Press, 139–58. Wa¨lchli, Bernhard (2023). ‘The interplay of contrast markers (“but”), selectives (“topic markers”) and word order in the fuzzy oppositive contrast domain’, Linguistic Typology, OnlineFirst. Walker, Traci, Paul Drew, and John Local (2011). ‘Responding indirectly’, Journal of Pragmatics 43: 2434–51. Waltereit, Richard (2001). ‘Modal particles and their functional equivalents: A speech-acttheoretic approach’, Journal of Pragmatics 33: 1391–417. Wiemer, Bjo¨rn, and Katerina Stathi (2010). ‘The database of evidential markers in European languages: A bird’s eye view of the conception of the database (the template and problems hidden beneath it)’, STUF-Language Typology and Universals 63: 275–89. Wierzbicka, Anna (1999). Emotions across Languages and Cultures: Diversity and Universals. Cambridge: Cambridge University Press. Wiltschko, Martina (1997). ‘D-Linking, scrambling und superiority in German’, Groninger Arbeiten zur germanistischen Linguistik 41: 107–42. Wiltschko, Martina (2021). The Grammar of Interactional Language. Cambridge: Cambridge University Press. Woods, Rebecca (2016). Investigating the Syntax of Speech Acts: Embedding Illocutionary Force. University of York dissertation. Zaefferer, Dietmar (2001). ‘Deconstructing a classical classification: A typological look at Searle’s concept of illocution type’, Revue Internationale de Philosophie 2/2001: 209–25. Zanuttini, Raffaella, and Paul Portner (2003). ‘Exclamative clauses: At the syntax–semantics interface’, Language 79: 39–81. Zeevat, Henk (2004). ‘Particles: Presupposition triggers, context markers or speech act markers’, Reinhard Blutner and Henk Zeevat (eds), Optimality Theory and Pragmatics. New York: Palgrave, 91–111. Zimmermann, Malte (2008). ‘Discourse particles in the left periphery’, in Benjamin Shaer, Philippa Cook, Werner Frey, and Claudia Maienborn (eds), Dislocated Elements in Discourse: Syntactic, Semantic and Pragmatic Perspectives. London: Routledge, 200–31. Zimmermann, Malte (2011). ‘Discourse particles’, in Klaus von Heusinger, Claudia Maienborn, and Paul Portner (eds), Semantics: An International Handbook of Natural Language Meaning. Berlin: Mouton de Gruyter, 2011–38. Zwicky, Arnold, and Ann Zwicky (1973). ‘How come and what for’, Ohio State University Working Papers in Linguistics 8: 173–85.

Index adverb 20, 33, 51, 61–2, 68, 92–5, 105, 111, 115, 150 affirmation 20, 119, 121–3, 133–6 allocutivity 97, 109–10, 113, 187–90 answer 6, 21–9, 35, 38–9, 41–2, 63, 68–9, 75, 86, 95–6, 120–3, 133–4, 139–46, 174 see also response answer particle 142–3 assertion, assertive 3, 8, 17, 37, 63, 73–7, 90, 94, 120, 125–6, 129–32, 140–3, 153, 184 see also at-issue content at-issue content 119, 127–8, 133, 136, 145, 153 attitude 9–12, 33, 41, 45, 62, 66, 86, 120, 127–30, 134, 139, 151–61, 167, 175, 177, 179 Bangla 173 Basque 20, 61, 88, 92, 96–115, 157, 187, 190 Bavarian 173 Bellunese 66 bias (speaker’s) 20, 32–8, 46, 56, 111, 119, 120, 122, 136 can’t-find-the-value question 3, 32, 38–41, 46, 62–3, 157–60, 179 common ground 27–8, 86, 130–2, 153, 173, 191–2 complementizer 82–4, 92, 109, 113, 173, 185 contrastive focus 10, 53–4, 57, 61, 150–1, 157, 178 see also focus corrective focus 170, 176–9 see also focus declarative question 3, 14, 19–20, 49–50, 58–61, 63–4, 119 degree 20–1, 62, 65, 68–81, 86, 107, 119, 126–8, 140–4, 150–4, 157, 159, 178 see also scale

denial 73–4, 129, 153–4, 175 deontic meaning 180 descriptive content 71, 73–5, 124–33, 140, 144, 160, 164–6, 170, 175 see also at-issue content discourse commitment 24–5, 95, 131 discourse particle 97 see modal particle dislocation (left and right) 91–2, 170–1 Dutch 34, 88 Dynamic Semantics 23–5, 27–8 echo question 20, 49, 54–8 emotion, emotive 12, 14–5, 20–1, 32–5, 43–6, 50, 56, 65, 69, 119–20, 126, 129, 134, 136, 149–53, 161, 166–7, 172, 174–9 emphasis 11–2, 14, 61–4, 76–7, 80–1, 84, 107–8, 135, 150–62, 175, 180–1, 189, 191 English 5–11, 14, 18–9, 29–30, 32–7, 43–4, 49–51, 56–61, 65–66, 68–83, 88, 121, 124, 129–30, 133, 137–45, 150–4, 168–9, 172, 177, 183–85 epistemic, epistemicity 3, 20, 23, 25, 32–5, 38–43, 56, 85–6, 94–5, 136, 151 evaluative meaning 62, 109, 132, 157–9, 176, 186 evidential, evidentiality 33, 94, 97, 155–6 exclamation 21, 66–81, 94, 122–34, 140, 149, 177 exclamative 4, 14–5, 34, 67, 70–81, 94, 123, 125–30, 133–4, 139–44, 156, 173 see also exclamation expectation speaker’s expectation 36–7, 66, 69, 71, 86, 122 unexpectedness 69, 71, 86 expressives 167 see also expressivity expressivity 21, 45, 128, 140, 149, 160, 163–80, 189–91

212

INDE X

French 49, 55–6, 84–6, 173 German 16–7, 34, 38–41, 51–4, 58, 61–2, 66, 74, 79, 89–96, 105, 111–2, 122–3, 135, 151, 155, 157–60, 164–5, 169–71, 173, 189 Greenlandic 31 Hausa 173 how come question 82–6 illocutionary force 4–5, 7–10, 16–7, 22, 63–4, 67–9, 76, 80–1, 86–91, 122, 124–30, 134–6, 149, 155, 162, 180–2, 185 see also speech act implicature conventional 64, 155, 159–60, 176 conversational 16, 139, 145, 156, 181 indirect speech act 6, 15–9, 181–90 see also speech act inference 16, 137, 145, 181 inflection 97–103, 109, 113, 186, 188 information-seeking question 21, 24–5, 27–9, 38, 40, 42, 44, 67, 70, 73, 75, 88–9, 119–24, 129, 133–41, 189–90 intensity, intensification 21, 33–4, 38–9, 43–4, 62–4, 69, 76, 81, 86, 107, 135, 149–61, 166, 175, 178–80 interjection 34, 126, 164, 166 intonation 12–5, 29–31, 33, 35–6, 49–51, 59–60, 69, 75, 77, 90, 124, 162 Italian 63, 71, 83, 92, 154, 156, 170–1, 173, 176–7 Japanese 49, 92, 187–90 Korean 187 Latin 173 Mandarin Chinese 13–4, 30, 55 mirative, mirativity 64, 76–7, 84, 100, 155–6, 172–7 modal particle 17, 20, 38–41, 61–2, 88–115, 125, 156–8, 165–6, 169–71, 176, 189 Montague Semantics 26–7

negative polar question 32, 35–7, 46 not-at-issue content 108–9, 119, 136, 139 optative clause 4, 14–5 Paduan 71 polar question 3, 8, 14, 18, 29–32, 35–7, 46, 49, 63–4, 68, 95, 98–100, 108–11, 115, 142–3, 154, 179, 190 see also negative polar question politeness 18, 21, 23, 59–60, 120, 162, 183, 187–91 presupposition 29, 32, 41, 43, 45–6, 71–2, 89, 120, 125, 132–3, 137, 140, 145, 165 see also not-at-issue content projection semantic 24–5, 131 syntactic 4, 9–10, 54, 57–8, 61–2, 64, 67–8, 80, 86–7, 92, 94, 103, 106, 108, 110–1, 115, 134, 179, 182, 185–7, 191 propositional content 65, 71, 76, 86, 95, 122, 127–30, 135, 143–5, 154–5, 158–62, 167–8, 175, 179–80, 191 prosody 20, 49–60, 64, 149, 163, 172, 177 quantifier, quantification 26, 29, 40 response 21–3, 29, 35–6, 39–40, 69, 71, 73, 75, 120–1, 129, 132–3, 137–46, 181 rhetorical question 3, 16–7, 32, 38–9, 41–3, 46, 68, 84–6, 88–9, 100, 112, 119–20, 129–30, 139, 140–1, 162 Romanian 91–2, 110, 186 root (syntactic) 57–8, 80, 91, 105, 107 root/non-root asymmetry 57–8, 80 Russian 30, 50 scale, scalar reading 38, 63, 71, 74, 77–86, 140, 144, 151–7, 172, 175, 178–9 see also degree scope 21, 76, 86, 93–5, 103, 105–12, 135, 154–5, 161, 169, 180, 187, 191 Sicilian 64 Spanish 74, 173, 186 stress 29–30, 36, 49–51, 53–4, 57, 94, 150, 157, 168, 171, 174 see also prosody

INDE X surprise-disapproval question 58, 67–8, 100, 122–3, 130–6, 139–40, 157–9 Swahili 13–4, 30–1 topic, topicalization 9–10, 51, 53, 91, 142, 163, 186

213

V2 property, V2 syntax 62, 98–9, 104–5, 107, 157 wh-in-situ question 20, 49, 54–8, 60, 64, 80, 119 wh-the-hell question 40–1, 43–5, 57–8, 80 why-like-what question 20, 58, 65–8, 86, 119, 122, 131, 134–6, 158–9, 179