Negative Inversion, Social Meaning, and Gricean Implicature: A Study Across Three Texas Ethnolects 9781501519277, 9781501512346, 9781501512360, 2020933711

Citation preview

William Salmon Negative Inversion, Social Meaning, and Gricean Implicature

Mouton Series in Pragmatics

Editor Istvan Kecskes Editorial Board Reinhard Blutner (Universiteit van Amsterdam) N.J. Enfield (Max-Planck-Institute for Psycholinguistics) Raymond W. Gibbs (University of California, Santa Cruz) Laurence R. Horn (Yale University) Boaz Keysar (University of Chicago) Ferenc Kiefer (Hungarian Academy of Sciences) Lluís Payrató (University of Barcelona) François Recanati (Institut Jean-Nicod) John Searle (University of California, Berkeley) Deirdre Wilson (University College London)

Volume 24

William Salmon

Negative Inversion, Social Meaning, and Gricean Implicature A Study Across Three Texas Ethnolects

ISBN 978-1-5015-1927-7 e-ISBN (PDF) 978-1-5015-1234-6 e-ISBN (EPUB) 978-1-5015-1236-0 ISSN 1864-6409 Library of Congress Control Number: 2020933711 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2020 Walter de Gruyter, Inc., Boston/Berlin Typesetting: Integra Software Services Pvt. Ltd. Printing and binding: CPI books GmbH, Leck www.degruyter.com

Contents List of abbreviations and symbols List of Tables

XI

Acknowledgments Introduction

IX

XIII

1

Chapter 1 Negative inversion in Texas, in three varieties of English 13 1 Negative inversion and varieties of Texas English 13 1.1 African American English 15 1.2 Anglo English 16 1.3 Chicano English 18 1.4 Shared features of African American, Anglo, and Chicano Englishes 19 1.5 Contact between the three varieties of English 25 1.6 Methods and data 27 1.7 Survey participants and socioeconomic class 30 1.8 Conclusion 31 Chapter 2 Negative inversion and its contents 32 2 A new empirical account of negative inversion 32 2.1 The present empirical account of negative inversion 33 2.1.1 The initial auxiliary 34 2.1.2 The subject noun phrase 42 2.1.3 The negative inversion main verb and copula 49 2.1.4 Intonation and accent in negative inversion 49 2.1.5 Negative concord and negative inversion 50 2.1.6 Is negative inversion emphatic? 51 2.1.7 Negative inversion allows negative polarity items in embedded clauses 52 2.1.8 Embedding of entire negative inversion sentences 53 2.1.9 Negative inversion and not-initial sentences 55 2.1.10 Negative inversion and its non-inverted counterpart(s) 56 2.1.11 Negative inversion, truth conditions, and camouflage 60 2.1.12 Negative inversion and modal existential constructions 61

VI

Contents

2.1.13 2.2

There-deletion existential constructions Conclusion 65

63

Chapter 3 Negative inversion, existential sentences, and definite subjects 67 3 Introduction 67 3.1 Definiteness effects and negative inversion 68 3.2 Subjects in negative inversion and there-existentials: Not so different after all 80 3.3 The modal existential and negative inversion constructions 87 3.3.1 Generational data for modal existential and negative inversion constructions 91 3.3.2 Texas Anglo speakers 91 3.3.3 Texas Chicano speakers 94 3.3.4 Texas African American Vernacular English speakers 94 3.4 Generational results and apparent-time analysis 95 3.5 Status of expletive subjects and negative inversion 103 3.6 Negative inversion, negative existentials, and conversational deletion 107 3.7 Conclusion 107 Chapter 4 Negative inversion and its discontents 109 4 Negative inversion is stigmatized 109 4.1 Overt and covert prestige 113 4.2 Negative inversion attitudes and meanings 115 4.3 Negative inversion, social meanings, and meaning diagnostics 120 4.3.1 Social meaning and ineffability 120 4.3.2 Negative inversion meaning and presupposition 123 4.3.3 Negative inversion meaning and conversational implicature 129 4.3.4 Negative inversion meaning and conventional implicature 133 4.4 Social meaning, social marking, and social stereotypes 135 4.5 Conclusion 136 Chapter 5 Negative inversion and emphatic meanings 138 5 History of emphatic claims 138 5.1 Definitions of emphatic in the negative inversion literature 5.1.1 Labov-emphatic and affective triggers 139

139

Contents

5.1.2 5.1.3 5.1.4 5.2 5.3 5.4

Green-emphatic, semantic widening, and absolutely no exceptions 141 Labov- and Green-emphatics in context 142 Israel-emphatic, pragmatics, and interaction 146 Emphatic is conventional, but emphasis is pragmatic NI as linguistic variable? 159 Conclusion 161

150

Chapter 6 Negative inversion is a formal idiom 162 6 Introduction 162 6.1 The nature of previous negative inversion analyses 163 6.2 The Labovian two-solution account 164 6.2.1 The Labovian expletive deletion account 164 6.2.2 The Labovian inversion movement account 166 6.3 Non-movement accounts after Labov et al. (1968) 167 6.4 Movement accounts of negative inversion motivated by disambiguation 170 6.5 Movement accounts of negative inversion with other motivations 174 6.6 Summary of previous accounts 175 6.7 Negative inversion as a formal idiom 176 6.8 Conventional requirements of negative inversion 179 6.8.1 Style and social meaning 179 6.8.2 Intonational contour 181 6.8.3 The initial verb must be an auxiliary 182 6.8.4 The auxiliary verb must be negativized 183 6.8.5 Post-verbal negative inversion subjects 185 6.8.6 Summary 190 6.9 Conclusion 190 Conclusion

192

References

201

Index

213

VII

List of abbreviations and symbols AA AAVE AE AUX CE CI CNVSI CD DE FC FI ME NC NI NP NPI PS SAI SE SM TE YGDP * # ? γ % “*, #, ?, etc.” ☞

African American African American Vernacular English Anglo English Auxiliary Chicano English Conventional Implicature Conversational Implicature Conversational Deletion Definiteness Effects Free Choice Formal Idiom Modal Existential Negative Concord Negative Inversion Noun Phrase Negative Polarity Item Presupposition Subject-Auxiliary Inversion USA Standard English Social Meaning There-existential Yale Grammatical Diversity Project Sentence is syntactically ungrammatical Sentence is infelicitous in context Sentence form is questionable Sentence is attested via Google Search Grammatical form is possible in a particular non-standard variety Acceptability judgment is reported from an earlier study but not accepted in the present study A demonstration gesture accompanying the use of a demonstrative

https://doi.org/10.1515/9781501512346-203

List of Tables Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7

Matyiku’s distribution of subjects in TEs and NI constructions 81 Number and average age of Anglo participants 92 Number and average age of Chicano participants 94 Number and average age of African American participants 94 Anglo familiarity with MEs and NI by age group 95 Chicano familiarity with MEs and NI by age group 96 African American familiarity with MEs and NI by age group 97

https://doi.org/10.1515/9781501512346-204

Acknowledgments This book is dedicated to: My mother: Patricia Salmon, who is the best there is. She worked magic for me many times, lining up survey participants young and old, from high school classrooms to retirement homes across the state of Texas. This book simply would not have happened without her. My father: Noel Salmon, an old-time Texan, who gave me lots of great Negative Inversion data and always had a ready answer for all my questions on Texas grammar. Profesora Jennifer Gómez Menjívar: Who has been an awesome linguistics spouse all these years, and who still answers my questions on how to say this and whether you can say that, and who was the best co-author ever on all the Kriol work we did in those wild and free pre-tenure years. My teacher friends: Professor Jaime Mejía at Texas State University, Professor John Poch at Texas Tech University, and Professor James T. Jones III at Prairie View A&M University, for graciously letting me test my surveys on their students. Mr. Charlie Nicholls of east Dallas: Who came correct with the derelict dialect. Professor Sue Maher: Who as my former dean was very supportive of this project and very generous in funding my Texas fieldwork trips. Professor Chongwon Park: Who has patiently tutored me and tried to keep me current on the intricacies of inversion syntax in generative and construction grammars, and who has supported my career in general in far too many ways to count. My kids: Joe, Noel, and Willie, who have grown up far far from Texas or the South, but who I hope will one day acquire Negative Inversion grammar and pragmatics anyway.

https://doi.org/10.1515/9781501512346-205

XIV

Acknowledgments

Sabina Matyiku: Who introduced me to Negative Inversion and without whose magnificent dissertation this book would surely not exist. My teachers: Háj Ross and Larry Horn, without whom wouldn’t nothing.

Introduction Can’t nobody teach you to be a cowboy. Can’t no college teach you. – Tom Scott, Fort Worth Stockyards (2003) Can’t nobody teach ya how to bang. Can’t nobody teach ya proper slang. – Chinx, ‘How to Get Rich’ (2015)

In his classic little book entitled The Hedgehog and the Fox (Berlin 1953: 2), Oxford philosopher Isaiah Berlin, drawing inspiration from the ancient Greek poet Archilochus, writes of two kinds of thinkers. The first of these, the hedgehog, will “relate everything to a single central vision [. . .] in terms of which alone all they are and say has significance.” The second of Archilochus’s and Berlin’s thinkers is the fox, which pursues “many ends, often unrelated and even contradictory, connected, if at all, only in some de facto way.” In this present book, it will seem that we start off as a fox, pursuing a range of seemingly disparate ideas with respect to Negative Inversion in three varieties of Texas English. By the end of the book, however, it will be clear that it has been a hedgehog type of process all along, with the facts of pragmatics, sociolinguistics, and grammar coming together to make a wholistic, all-embracing, hedgehog type of sense of the construction. Negative Inversion will ultimately be shown to be a very complex construction, but one that can be understood nonetheless with only a few simple sociolinguistic facts and a pragmatic system in which to interpret them.

Introducing negative inversion Negative Inversion (NI) constructions such as (1) have been the subject of a large body of literature in linguistics since their first mention by Labov et al. in the late 1960s, with respect to African American Vernacular English (AAVE): (1)

Can’t nobody beat ‘em. ‘Nobody can beat them.’

The sentences have been investigated in AAVE and various Southern US Englishes almost exclusively in the framework of generative grammar and with most attention paid to the syntactic “inversion” of the subject and auxiliary https://doi.org/10.1515/9781501512346-001

2

Introduction

verb.1 Such syntactic analyses have been used to make arguments about a range of puzzles, including: NI’s relationship to its non-inverted counterpart (Nobody can’t beat ‘em), to negative there-existential sentences (There is nobody who can beat ‘em), to Modal Existential sentences (There can’t nobody beat ‘em), definiteness effects (*Can’t the man beat ‘em), emphatic meaning (Can’t nobody beat ‘em [with no exceptions!]), and more. The present monograph takes up all of these questions and others, though from a very different perspective than any that has preceded it. Instead of the movement-based syntax in which most previous studies of NI are couched, this monograph draws on thought from Gricean pragmatics and sociolinguistics, offering a multi-dimensional account of NI sentences across African American, Anglo, and Chicano Englishes in contemporary Texas. The book offers a new empirical account of the construction and provides tidy solutions to the extant puzzles mentioned above, arguing that solutions are to be found not in syntactic movement but rather in Gricean reasoning, demographics of use, and a simple truth conditional semantics. First, based on an apparenttime analysis across generations of speakers, the book argues that NI is a diachronic descendant of the Modal Existential construction (There can’t nobody beat ‘em) that has lost the expletive subject over time, and that despite the name of the construction, the “inversion” itself is actually just a myth and theoretical artifact. This existential origin then motivates a straightforward account of NI’s poorly understood definiteness effects, a characteristic originally identified in there-existential sentences in Milsark (1974), but which has also been assumed to be true of Negative Inversion sentences as well since the earliest mentions of the construction. The received understanding here is that definite subjects are conventionally prohibited from appearing in NI; I show that definite subjects are in fact

1 NI has gone by various other names in previous literature, including Non-canonical Negative Inversion in White-Sustaita (2010), Verb-initial Negative Inversion, aka “V1NI” in Horn (2015, 2018), Popular Negative Inversion in Blanchette (2015), Negative Auxiliary Inversion in Matyiku (2017) and Blanchette and Collins (2018), Declarative Negative Auxiliary Inversion in L. Green (2014) and L. Green and Sistrunk (2015), and Negative Auxiliary First in Salmon (2017). The Negative Inversion of the present work also appears in Labov et al. (1968), Labov (1972a, 1972b), G. Green (1985), Martin (1992), Weldon (1994), Sells, Rickford, and Wasow (1996), Foreman (1999, 2015), Parrott (2000), Fought (2003), Salmon (2018a, 2018b), and elsewhere. Other constructions in the linguistics literature have also been referred to as “Negative Inversion” over the years, including (i), which appears in Rochemont (1978), Haegeman (2000) and Sobin (2003). Constructions such as (i) play no role in the present book. (i) Never again will I eat raw spaghetti.

Pragmatics and sociolinguistics in the present work

3

allowed in NI but that the referents must be pragmatically construable as hearer-new in the sense of Prince (1992). Not so coincidentally, Ward and Birner (1995) have argued that this same pragmatic requirement constrains the appearance of definite subjects in there-existentials. Following this, the book takes up the question of NI’s emphatic qualities. Prior debates on this topic have taken semantic or syntactic approaches; this book argues that the emphatic quality arises as an interaction of Gricean reasoning and NI’s status as a Labovian (1972b) social marker. In this way the book engages the problem of pragmatic reasoning and style-shifting, an area which is surprisingly underexplored in both the pragmatics and the sociolinguistics literature. The book also considers the ways in which Labovian social meaning relates to other dimensions of meaning from the semantics and pragmatics literature. Thus, well-known diagnostics for presupposition, conversational implicature, and conventional implicature align the social content of NI with the lattermost dimension, i.e. the conventional implicature of Grice (1975). As I discuss below, this ultimately provides a secure footing for the conversational implicature analysis given in Chapter 5 in a fully pragmatic account of NI’s emphatic effects. The final score here is that most of the interesting problems thus far puzzled over in the NI literature can be understood in terms of pragmatics and demographics, with no requirements whatsoever put on a movement-based syntax or much beyond a simple semantics. The book thus provides fresh perspectives on old NI problems and makes a strong contribution to the understudied intersection of sociolinguistics and the semantics-pragmatics interface. In the sections that follow, I discuss in more detail the manners in which pragmatics and sociolinguistics intermingle in the present work. I also provide short chapter overviews and a discussion of terminology.

Pragmatics and sociolinguistics in the present work The chapters of the present work draw in diverse ways on Gricean pragmatics as well as sociolinguistic thought from a variety of approaches. Chapter 3, for example, makes use of the apparent-time construct to argue for the diachronic relation between Negative Inversion and Modal Existential sentences.2 Apparent-time analyses have been a staple means of studying language change in variationist

2 The apparent-time construct involves surveys of language use across multiple generations of speakers and is believed to show real-time change in a speech community. See Bowie (2005) and sources therein for a detailed history of the methodology.

4

Introduction

sociolinguistics since Labov (1963, 1966). This generational approach to data collection plays an important role in Chapter 3, as it allows us to see very clearly that the younger generations surveyed herein accept only Negative Inversion, while the older generations accept both Negative Inversion and Modal Existentials, supporting the hypothesis that the former is essentially a descendant of the latter. This link between Negative Inversion and the modalized existential sentences then suggests a clear direction for an account of the definiteness effects that have been hinted at for the construction since Labov et al. (1968) and that have been directly associated with it since at least Martin (1992). As is well known, definiteness effects have been associated with existential sentences since at least Milsark (1974).3 As I go on to argue in Chapter 3, the definiteness effects of NI are very similar to those of existential sentences, and they can be given a pragmatic analysis in terms of hearer-new subject referents in the same way that Prince (1992) and Ward and Birner (1995) do for existential sentences. What we see here, then, is not so much an investigation of pragmatic issues within a Labovian variationist framework, but rather, variationist tools employed in support of a pragmatic analysis. Chapter 4 takes a close look at the dimensions of social meanings associated with NI. My survey participants view NI as highly stigmatized, and they associate it with uneducated or informal speech but also speech with a certain “cool” factor. This chapter then considers these aspects of social meaning against several meaning diagnostics in the presupposition, conversational implicature, and conventional implicature literature, arguing ultimately that the social meaning aligns very closely with the latter, which segues neatly into Chapter 5. Chapter 5, which considers the emphatic nature of Negative Inversion, is situated directly in the crossroads of variationist sociolinguistics and Gricean pragmatics. Based on the findings of Chapter 4, NI is argued to be a Labovian (1972b) social marker: i.e. it is conventionally associated with social content, and speakers are aware of this association. This conventional social content then provides the basis for deriving emphasis as a Gricean conversational implicature. The argument

3 Early mention of the seeming incompatibility of existential there and definite subjects in the generative literature can be seen in Bresnan (1970). Bolinger (1977: 115) notes also: It is only recently that grammarians seem to have been talking about the unacceptability of there with grammatically definite subjects. No such restriction is noted in any of the older comprehensive handbooks. In fact, all have a scattering of definites, on the order of five per cent or so of the examples cited. Bolinger cites Jespersen (1909–49: 3.17.72), Poutsma (1928: 401), Curme (1931: 10), and Kruisinga (1932: Section 2176).

Pragmatics and sociolinguistics in the present work

5

is essentially that NIs such as (2) and their non-inverted counterparts such as (3) are equivalent in terms of truth conditions in the Texas varieties under consideration.4 However, in order to derive the pragmatic account of emphasis, there must be something conventional to set the two constructions apart as alternates. (2)

Can’t many people lift that rock. ‘Many people can’t lift that rock.’

(3)

Many people can’t lift that rock.

I argue that NI is imbued with a particular social content, which is semantic but non-truth conditional, and this content is what sets it apart from the non-inverted form. From here, there is a clear path to a standard conversational implicature derivation of the emphasis via choice and Gricean reasoning. The pragmatic account is thus argued to have access to the sociolinguistic content. In practice, this is no different than arguing that conversational implicatures can be derived based off of conventional implicature content, which is also a type of meaning that is semantic, but non-truth conditional. For example, consider Horn and Abbott’s (2012: 345) account of definite and indefinite articles such as (4) and (5). (4)

The dog jumped the fence. Conventional Implicature: uniqueness

(5)

A dog jumped the fence.

Horn and Abbott assume that in terms of truth conditional content, definite and indefinite articles are equivalent, and thus (4) and (5) are as well. The definite article, though, has an additional conventional implicature meaning of uniqueness associated with it, which the indefinite article lacks. In Horn and Abbott’s account, it is this additional layer of meaning conventionally associated with the definite article that allows defeasible conversational implicatures, such as familiarity, to be derived via Gricean reasoning. On the other hand, if the two articles

4 Green (2014) suggests that in the AAVE variety she considers, NI and its non-inverted counterpart would actually have different truth conditions. Green’s position seems largely to be an outlier, though. In this book and many other works, NI and its non-inverted counterpart(s) are assumed to be truth conditionally equivalent. See Chapters 2, 4, and 5, below for much more discussion.

6

Introduction

were completely equivalent in all aspects of semantic meaning, there would be no way to derive pragmatic implicatures from use of one or the other. My argument that NI can be used to convey emphasis via conversational implicature runs in parallel fashion. The primary difference is that in the case of NI, the additional non-truth conditional content is social rather than information structural or identificational. However, as I show in Chapter 4, the social content of NI patterns in exactly the same way that prototypical conventional implicatures do in the conventional implicature diagnostics typically used in the literature. Thus, I claim that NI is conventionally loaded with a social meaning that is also nontruth conditional semantic content. From this extra bit of lexical content – which is not associated with NI’s non-inverted counterpart – speakers can convey conversational implicatures of emphasis. Chapter 5 also sees a coordination of Gricean pragmatics and the interactional sociolinguistics developed in the work of Gumperz (1970, 1982) and underlying work on politeness by Brown and Levinson (1978). In Gumperz’s work on code-switching, he shows that speakers can convey emphasis by switching from one language or dialect to another. This is another plausible means of accounting for the pragmatic emphasis effects seen with NI. Relying on the same semantic account of NI as mentioned above, we can posit a Gumperz-style switching as a means of conveying pragmatic emphasis. That is, if NI is socially marked as belonging to a familiar style or genres, which it is, then a speaker who uses NI to switch from one style or genre to another might invite inferences of emphasis. This is very similar to what was described above, with some difference, though. In the first example, the Gricean conversational implicature was derived due to speaker choice of using one form versus another (i.e. NI or noninverted counterpart). In the Gumperz-style interactional approach, the emphasis implicature of NI is invited by the speaker’s using NI to switch between styles or genres. In both cases, the pragmatic implicature is invited by the speaker’s choice of sentence form. In the first case, though, the inference process is assumed to depend on the fact that one form was used as opposed to another; in the second case, the implicature is believed to arise as a result of using the form to switch codes, styles, or genres, and it is in essence the speaker’s switch that invites the inference. As I argue in Chapter 5, both of these means of inference invitation are possible means of using NI to generate emphasis implicatures on different occasions of use. And, both of them depend on pragmatic reasoning processes having access to sociolinguistic, social meaning. The important thing, as we will see, is that the emphasis in question here is pragmatic and defeasible, and as such it can be generated by a host of different strategies, as is expected on a pragmatic account.

Vernacular language varieties of the present work

7

The present book is, in my mind, very clearly a book of pragmatics. In one sense, it is a fairly conservative, Anglo-American, Gricean approach to the meaning and use of Negative Inversion in three contemporary ethnolects. On the other hand, though, in both its methods and assumptions about meaning, it probes boundaries between sociolinguistics and pragmatics, relying on facets of both to solve several puzzles that have long been associated with the NI construction. I do not want to say, then, that this book is a work of sociopragmatics, as that term has been defined. For example, for Leech (1983: 10), whom Kecskes (2012) credits with the creation of the term, sociopragmatics is concerned with “specific ‘local’ conditions on language use.” As Leech (1983: 10) goes on to write, “it is clear that the Cooperative Principle and the Politeness Principle operate variably in different cultures or language communities, in different social situations, among different social classes, etc.” This language or culturespecific operation of pragmatic principles has been the purview of sociopragmatics. As we move through this present book, it will be clear to the reader that this is not the goal of this study, as I posit no special or culture-specific instantiation of any pragmatic principles. Nor does the present work fall into the domain of what has recently been termed variational pragmatics, as in the work of Barron and Schneider (2009). As Cameron and Schwenter (2013: 5) write of this field of pursuit, “variational pragmatics investigates how particular speech acts, routines, or even broader notions such as politeness, are realized across varieties of the same language. This is similar to quantitative research into cross-cultural Speech Act realizations in Applied Linguistics.” Essentially, then, work in variational pragmatics might look at how a given speech act – i.e. an apology, a promise, or a bet – is realized in differing varieties of the same language. Important as that work is, it too is not the goal of this present book. This present book relies upon general Gricean thought and sociolinguistic data to relieve much of the burden formerly placed upon a generative syntax with respect to one construction in three ethnolect varieties of English.

Vernacular language varieties of the present work In this book I use terms like “African American English,” “Anglo English,” and “Chicano English” to refer to the language varieties of interest in the present study. What I am referring to is the vernacular language varieties associated with each of those ethnic groups: i.e., the informal and unguarded, perhaps uneducated or working-class, in-group, natural, way of speaking controlled by many members of those groups. All of these varieties are ethnolects, in that

8

Introduction

they are ways of speaking associated with specific ethnic groups, though it is not necessary to be a member of a given ethnicity to speak that ethnicity’s language variety. Vernaculars are languages of the home, of Friday-night parties in the field or the bar, of the front porch, of the hunting lease, the high school football game, the barbershop, and of the street. They are not ways of speaking one learns in college; if anything, they are ways of speaking one un-learns in college. Be that as it may, they also provide the objects of consideration in this book.5 The Negative Inversion of this book doesn’t appear in standard varieties of English, and it doesn’t appear in the proper, monitored speech of Caucasian Southerners or Texans, or Chicanos and African Americans of any region. NI is at once improper but immensely useful. It is marked speech; but, it is also speech that marks a speaker’s relation to her addressee. It is a construction that raises questions of grammar, social relations, and pragmatics, all of which are addressed in this book.

The place of the author In Renaissance Self-Fashioning: From More to Shakespeare, Stephen Greenblatt (1980: 5) writes, “It is everywhere evident in this book that the questions I ask of my material and indeed the very nature of this material are shaped by the questions I ask of myself.” Greenblatt was writing with respect to historicist literary criticism, or cultural poetics, as he terms it, in which the historical particulars of the critic necessarily play a role in the interpretation that critic gives of a text. This present book is about the pragmatic and sociolinguistic particulars of Negative Inversion constructions in three varieties of Texas English. The arguments herein rely on a wide range of empirical data and interpretations of that data. As I am the one doing the interpretation, it is thus important for the reader to have a glimpse into my personal history as it relates to NI. I am a native speaker of Texas Anglo English, born in the 1970s and raised in Texas, in the middle of NI territory, and I have strong native-speaker intuitions about all of the data presented in this work. My roots are lower-middle class Anglo, and I grew up on the outskirts of a sizable city in south Texas in the middle of what has traditionally been cotton and ranching country. This city has a large working-class economy derived mostly from agriculture and 5 Labov (2006: 86) defines vernacular as “the language first acquired by the language learner, controlled perfectly, and used primarily among intimate friends and family members. Thus every speaker has a vernacular, some quite close to the network standard, some quite remote from it.”

Chapter summaries

9

petroleum, and my high school football team routinely goes to the state playoffs, occasionally winning the state championship. A recent study on literacy performed by researchers at Central Connecticut State University presented my hometown as the second least literate city in the United States, edged out of the number one spot only by Bakersfield, California.6 It is also the hometown of celebrities such as Farrah Fawcett, Eva Longoria, and Selena Quintanilla. Though I identify as Anglo, throughout my years in Texas I was in constant contact with Spanish and Chicano English as well as African American Vernacular English. My hometown is approximately 60 percent Hispanic, with the district where I attended primary and secondary school approximately 50 percent Hispanic. African American, Anglo, and Chicano Englishes clearly constitute distinct varieties of the language there, but contact is and has been so great between the three that it would be difficult to determine how exactly and with respect to what features the distinctions should be drawn. The fact that I am a native speaker of one of the NI varieties described in this book, as are most of my family members, lifelong friends, and enemies, has necessarily influenced the interpretations of the data and the research directions the book has taken.

Chapter summaries Chapter 1 of this book locates Negative Inversion constructions in the primarily vernacular varieties of African American, Anglo, and Chicano speakers of Texas English, including discussion of what is meant by these various designations. The chapter also gives an overview of the kinds of non-standard grammatical features shared by these varieties, likely as a result of the extended language contact mentioned above. There is also an extended discussion of the methodologies employed in data collection. The book makes a wide range of claims across pragmatics, semantics, and sociolinguistics, and a wide range of data gathered with various methods was required. Chapter 1 gives an overview of the kinds of data that were used and points to further specific discussions in the local chapters in which the data appears. Finally, the chapter includes a discussion on social class of the many survey participants, as well as a brief section on the prognosis of the construction, given its stigmatization, diachronic history, and the changing demographics of Texas.

6 See “The five least literate cities in America.” (2014, February 24). Time.

10

Introduction

Chapter 2 provides an in-depth empirical account of Negative Inversion, which dramatically expands the received empirical understanding of the construction. There is a significant body of preceding work on NI in AAVE and Southern Anglo Englishes, dating to Labov et al. (1968), and most previous work assumes a uniform empirical description for the construction regardless of racial, regional, or generational dialect concerns. In this chapter, however, I show that NI in the three Texas vernaculars, while quite similar to each other, in fact differ widely in their syntax, semantics, and pragmatics from all previous empirical accounts of NI that have been advanced elsewhere. This chapter then discusses the possible implications of these differences in terms of regional and generational variation as well as Spears’ (1982) notion of “linguistic camouflage.” This chapter provides a thumbnail sketch of what is to come in the rest of the book; it also provides a strong empirical portrait for those working in other linguistic fields who might not be interested in the pragmatics and sociolinguistics but who wish to explore the further possibilities of the construction in their own ways. Chapter 3 shows, contrary to what would be expected based on all previous literature, that definite and referential subjects do appear in Negative Inversion sentences. All previous literature assumes that such subjects are syntactically ungrammatical in the construction. This chapter shows that definite subjects are possible in the Texas NI but that they are subject to pragmatic constraints. The chapter argues further that the pragmatic constraints in question are the same ones that govern the acceptability of definite subjects in existential theresentences, and that this can be taken as strong evidence of a diachronic relation between Modal Existential sentences and Negative Inversion. This latter claim is supported by a cross-generational survey and apparent-time analysis. I show here, for example, that the older participants (61+) in the surveys were mostly familiar with negative sentences such as There can’t nobody lift that rock while younger participants were for the most part unfamiliar with it, accepting only the canonical Negative Inversion such as Can’t nobody lift that rock. Based on this finding, I argue that NI is essentially a negative existential sentence which has lost the expletive subject over time. This data and reasoning are completely novel to the literature and suggest that most previous accounts of NI are ultimately based upon insufficient empirical characterizations of the construction. Chapter 4 is a discussion of language attitudes and social meanings associated with NI. From the survey data a clear picture emerges of NI as a type of social marker, what Labov (1972b) would describe as a linguistic item that “may lie below the level of conscious awareness, [but] will produce regular responses on subjective reaction tests” (Labov 1972b: 314). Essentially, NI marks a dialect style and at the same time is imbued with various dimensions of social meaning: in this case, speakers know NI as informal and familiar, yet stigmatized. Or, as one

Chapter summaries

11

high school survey participant comments, as “how rednecks talk.” This chapter discusses the wide range of social meanings that accompany NI, and it explores the different dimensions of conventional and non-conventional meanings available in which to locate them. It also provides a natural segue into the following chapter, which, in terms of Gricean (1975) reasoning, gives an account of the emphatic meaning frequently associated with NI. Chapter 5 considers the question of emphaticness and Negative Inversion. Labov et al. (1968: 285) note that the construction carries an emphatic signal, or is “excited, and strongly affective.” In later literature, this observation has taken on a number of forms, including: being repeated verbatim, being denied completely, being located as a semantic feature, or as the result of a syntactic movement operation. It has also been reinterpreted, as in Green (2014) and Green and Sistrunk (2015), as making a semantically wider statement: i.e. Can’t nobody lift that rock makes a wider claim than Nobody can lift that rock. This chapter argues that the emphatic affect that is felt with the Texas NI is actually a pragmatic issue, and that it arises as an interaction of Gricean pragmatics and social marking. As shown in Chapter 4, NI is a Labovian social marker in the sense that its use clearly marks the social class and style of the speaker as well as her beliefs about her audience. It is argued here that the emphatic qualities of the construction can be derived from the resulting social dynamics of its use. This strategy yields important dividends, as it respects the intuitions of those scholars who have argued that NI is not emphatic; at the same time, it also accounts for the emphatic affect that very frequently does accompany the construction. It also provides a means of pragmatically deriving the reinterpretation of Green (2014) and Green and Sistrunk (2015), in which the construction is argued to be semantically wider. Further, as an investigation of Gricean pragmatics and style-switching, this chapter is a significant contribution to an underexplored area of sociolinguistics and pragmatics in general. Based on the accumulation of claims made in Chapters 1–5, Chapter 6 argues that the structure of NI is best understood as a formal idiom, of the kind described in Fillmore, Kay, and O’Connor (1988). NI has an idiomatic syntactic frame and encodes a specific semantic meaning: i.e. that of the social marker described above. At the same time, the rigid syntactic frame contains variable slots that are constrained to allow only particular forms of subject, auxiliary verb, and negation. Taken as a whole, the formal idiom account provided here allows us to account for the structure, meaning, and usage requirements that make up the entirety of the Negative Inversion construction. Along the way in this chapter, strong critiques are given of previous syntactic accounts. The construction has received significant attention since Labov et al. (1968) and continues to be discussed in numerous articles, conference proceedings, and very

12

Introduction

recent dissertations, such as Blanchette (2015, CUNY) and Matyiku (2017, Yale). All previous accounts have relied upon generative grammar frameworks, assuming generally the same empirical description of the construction, and running up against essentially the same puzzles. This chapter shows that the movement approaches to this particular construction are ultimately unmotivated and so clears a path forward for the new way of thinking about it. Essentially, the NI is a very complex construction, with a range of different conventional properties. It might be possible to account for some of these properties in a generative framework, but, taken as a whole, Occam’s Razor supports an idiom account for this particular construction. The book ultimately makes contributions in a variety of areas. It makes an important empirical contribution to the study of NI, broadening the received empirical account of the construction significantly to where it can be viewed as closely related to the much better-studied there-existential construction. The book is also the first attempt at studying NI across multiple language varieties at the same time, and the first to consider NI outside of Anglo and African American English contexts at all. In the sparse existing literature on Chicano English in general, NI is only mentioned once, by Fought (2003), to say that it does not occur in the California variety she considers. Finally, the book makes a larger theoretical contribution as it explores the Labovian social meaning associated with NI in terms of Gricean semantics and pragmatics. I am not aware of literature in sociolinguistics or semantics elsewhere in which a synthesis of this kind is undertaken.

Chapter 1 Negative inversion in Texas, in three varieties of English When Smitherman lists “patterns of grammar and pronunciation in Ebonics,” I recognize many of them in my own familiar [Anglo] speech. Still can’t nobody say for sure who learned what from who. – James Sledd, in ‘Race, Class, and Talking Proper’

1 Negative inversion and varieties of Texas English As suggested by the title of this book, Negative Inversion is investigated in three different varieties of English as spoken in Texas: AAVE, AE, and CE. These three varieties of English are also spoken elsewhere in the US, but the focus here is strictly Texan, with no strong claims made about speech outside of this context. AAVE, for example, is spoken throughout the American South as well as larger cities around the US. The AE under consideration is very close to Southern Anglo Englishes across the American South. The CE observed here is close to that spoken across the American Southwest as well as various larger cities elsewhere, such as Chicago, and is increasingly found in the American Southeast and Midwest as well. Thus, the varieties under consideration here are each and together part of a much larger linguistic picture. For the most part, I do not make claims about these larger linguistic contexts. I do raise questions, though, in places where the Texas NI varies with claims made about NI elsewhere. For example, the African American NI I tested in Texas appears to differ markedly with claims made about NI in all previous AAVE literature. The same is true of Texas Anglo NI: it differs dramatically from all previous work on Anglo and AAVE NI across the American South. With respect to Chicano NI, there is very little previous work on NI with which to contrast. This is likely due to the fact that there is comparatively little work on CE in general. The only case I am aware of in which NI has been mentioned in CE at all is in Fought (2003: 143), in which it is related briefly that NI is not found in her corpus or interviews from California CE.7 As we will see below, this is

7 NI does occur in some CE contexts in California. For example, the hip hop artist, Lil Rob, is a Mexican-American native of California and performs in a mix of CE and Spanish, and NI can be found in his corpus of songs. Here is a verse from Rob’s song “I Remember,” (2002): https://doi.org/10.1515/9781501512346-002

14

Chapter 1 Negative inversion in Texas, in three varieties of English

clearly not the case in Texas. NI among Chicano speakers in Texas is clearly alive, if not completely well. As this book progresses, I note these differences in empirical NI coverage, but I do not dwell upon them or make strong claims. At this point, I do not know if they are actual empirical differences, or if the differing empirical descriptions are products of the times, places, or manners of the previous research undertaken on the sentence type. For example, a trailblazing study on AAVE undertaken in the 1960s in New York City describes the speech of African Americans in a very different time and place than the AAVE of contemporary Texas – or even of contemporary New York City. Do those 1960s empirical observations hold in the New York City of 2019? Or, almost 50 years later, perhaps the empirical situation on the ground in New York City is actually closer to the data of the present work. More research is needed in order to know. Similarly, a central claim of Chapter 3 below is that definite NP subjects are definitely allowed in Texas Negative Inversion. This is a finding that contradicts all previous descriptions of NI in other times and places.8 As I show in Chapter 3, however, there is a pragmatic requirement on the construction that limits the nature of the definite subjects allowed. Perhaps there was a similar pragmatic restriction in place for earlier accounts of NI as well, but it was not known by the researchers, and in effect concealed the acceptability of this type of subject. Again, more research is needed in order to know, and in the cases of the much earlier studies, we will never know. Thus, this book does not make claims across generations or regions of NI users. The focus in this book is on the speech of contemporary Texans. To the extent that the results differ from research done elsewhere and before – whether One of the things we use to do Can’t nobody take it away from us We used to live life dangerous, time flew away like angel dust Never known as presentado levas Stay away from metiches, and chepas Talking mentiras, cause they can’t beat us Mira Lil Rob on his lowrider bicicleta Looking for muñecas, beautiful like aztecas. Of course, it is possible that the NI in Rob’s work is a product of contact with the AAVE hip hop community in California as well, for whom NI is well attested. Or, perhaps, Rob’s NIs are instances of dialectal crossing. As Fought (2006: 75) notes, “AAVE can also have an influence on the structure of Latino English varieties.” More research is definitely needed though into the existence and range of NI in CE in California. 8 The only exceptions to this claim are Salmon (2017, 2018a), upon which the present work builds.

1 Negative inversion and varieties of Texas English

15

this is due to actual regional or generational differences, or basic research methods – is ultimately beyond the scope of this work, though it is certainly fodder for subsequent research. In the next sections, I give a general background for each of the language varieties under investigation in the present work, beginning alphabetically with the African American speech of east, south, and north Texas.

1.1 African American English Since the 1960s, AAVE has received more attention in the linguistics literature than any other domestic variety of English. Much of the early work on AAVE was conducted in northern US cities such as New York City and Detroit, but it has since spread to investigations around the US, including urban and, to a much lesser extent, rural contexts as well, including Texas, with the rural work of Cukor-Avila and Bailey (1996, 2001, inter alia). Over the last two decades, it has become clear that there is in fact substantial regional variation in AAVE as it is spoken around the US, especially in terms of phonology and copula issues.9 There have been no claims of regional variation in AAVE with respect to Negative Inversion, though further research is certainly needed on this question.10 With respect to the AAVE of the present book, I am assuming that the variety in question here belongs to the dialect region encompassing east-central Texas and areas around the Dallas-Fort Worth Metroplex. The majority of my survey participants are from the greater Houston area, including the urban Third Ward area, the suburban areas to the south and to the west of Houston, as well as the south sides of Fort Worth and Dallas, and the east side of Austin.11

9 See the several papers included in Lanehart (2015: Part II “Lects and Variation”). Evidence to the contrary notwithstanding, however, the idea that AAVE is uniform across regions – what Wolfram (2007) refers to as the “supra-regional myth” – is still alive and well for many. 10 A further relevant point here is that many discussions of AAVE Negative Inversion in the linguistics literature do not include regional information at all. That is, NI is identified as being a feature of AAVE in general, but not specifically located to a particular region, as for example, the work of Martin and Wolfram (1998), Parrott (2000), White-Sustaita (2010), Green (2011b, 2014), Green and Sistrunk (2015), and so on, which are framed in terms of a general AAVE usage rather than a specific region. Labov et al. (1968) and Sells, Rickford, and Wasow (1999) are notable exceptions here, identifying the sources of their NI data as AAVE speakers from New York City and East Palo Alto, California, respectively. 11 The African American population of Texas was historically found in the north and eastern parts of the state, and along the Gulf coast, as those areas were home to cotton and other plantation farming in the 18th and 19th centuries. These areas are still home to a significant percentage of AAVE speakers in Texas today, as well as the major cities of Fort Worth, Dallas,

16

Chapter 1 Negative inversion in Texas, in three varieties of English

To my knowledge, there have been no detailed grammars written of AAVE anywhere in Texas; although, there has been significant work on various grammatical and phonological features, not including NI. See Cukor-Avila and Bailey (2015) and sources therein for discussion of AAVE in Texas in general. AAVE has gone by a number of labels over the years, marking a range of socio-political implications and histories: from Negro English, to Negro Dialect, Black English, Ebonics, African American English, African American Vernacular English, African American Language, African American Vernacular Language, and more. In the present work, I use African American English to refer to the complete range of styles, registers, genres, and so on, which compose the variety as a whole.12 I use African American Vernacular English (AAVE) to refer to the “vernacular” of African American English. I assume, in the manner of Spears (2015), that the AAVE variety in question here is truly a vernacular, as opposed to what Spears refers to as African American Standard English, which is “a type of [African American English]: it has [distinctively black grammatical features], but none that are stigmatized or considered nonstandard” (786). African American Standard English then, by Spears’ definition, is less likely to include NI as part of its grammar than are the vernacular forms, and so it does not play a role in the rest of this book. My participants and various survey respondents referred to the language they speak in a variety of ways, including “African American English, Black English, ghetto English, slang, broken English, Ebonics” and so on. No distinction was made by any of them between standard or vernacular means of speaking.

1.2 Anglo English AE in Texas is very similar to other Midland and Southern US Englishes. As the term “Anglo” suggests, it is the language most associated with non-Hispanic whites, but it is spoken by many others in the state as well.13 Like the description Houston, and San Antonio. AAVE has been spoken in Texas since at least the early 19th century, and so has been in contact to some extent with the ancestor varieties of Anglo and CE in the state since that time. 12 In this book, I do not hyphenate “African American.” There are differing ideas in the literature about whether or not to do so. Some authors I cite in this book do hyphenate, and I have attempted to maintain their usage preference wherever they are quoted or cited. 13 For example, a video hosted on the Texas English Project Web site at the University of Texas at Austin has an interview with a Latina university student who has a clear Anglo Texas accent. The video can be viewed at the link below, with the relevant interview occurring between 6:06–7:36 on the video:

1 Negative inversion and varieties of Texas English

17

of AAVE above, AAE in Texas consists of a wide range of styles, registers, and micro-variations across different social classes and sub-regions of the state.14 Following the general trend in linguistics literature, I refer to the language variety of interest here as Anglo English, more specifically, as Texas Anglo English. It should be noted, though, that this is not a term used by the speakers themselves. Many of my survey participants had in fact never heard the term “Anglo English” before, and they were in any case at a loss for what to call the language they speak: “Texas English, regular ol’ English, American, country English, redneck English, and Texas drawl” were some suggestions they provided for the name of the variety.15 Over the years there have been numerous attempts at dividing the state into dialect areas; none are universally accepted, though. Extensive contact with Spanish, African American English, German, and the various other European languages of its settlers has had an impact on the English of Texans over its history. Thomas (1997: 309) argues as well that due to large amounts of in-migration from outside the state, there is an urban/rural divide, which has resulted in “dialect islands” in the urban centers. [T]here are starker differences between Anglos from large metropolitan centers and those from rural areas in the production of certain vowels, particularly /ai/, as in night, and /e/, as in day. The result is the creation of dialect islands in Texas where the large metropolitan centers lie. (Thomas 1997: 309)

The Anglo data for the present book was collected in both rural and urban environments alike, in south, east, central, and west-central Texas. I did not find grammatical or structural variation with NI among these different groups. Certainly, though, there is attitudinal variation, in which the urban Anglo speakers appear to see NI as more stigmatized than their rural counterparts, though all participants found it to be stigmatized to various degrees. It is likely that there is usage variation as well, meaning NI occurs with different frequencies in perhaps different genres of speech. I did not test this aspect of NI

. Additionally, one of the Youtube comments on the video, left by a poster named Grass Snake, corroborates, with the statement: “I’m Latino from Texas and I can turn the twang up or down as needed. haha.” 14 As Cukor-Avila et al. (2012: 18) note in their perceptual dialectology treatment of Texas English, Texans “do not view themselves as a homogenous speech community [. . .].” 15 See Cukor-Avila et al. (2012) for many more terms used by Texans to refer to their speech variety. See Foley (1999: 6) for discussion of the extremely problematic application of the term Anglo in Texas in general.

18

Chapter 1 Negative inversion in Texas, in three varieties of English

variation, though, and it is not ultimately important to the claims being made in this book.

1.3 Chicano English The third language variety of interest in this study is what I refer to as “Chicano English.” Like the other two varieties under consideration in this work, it is an ethnolect, in that it is associated most directly, though not exclusively, with an ethnic group: in this case, non-White Hispanic speakers. In what appears to be the most in-depth discussion of CE in general, Fought (2003), with respect to her Los Angeles study, defines CE with two quotes: the first from Metcalf (1974), the second from Santa Ana (1993). The resulting description here is one that seems to work well for the Texas context as well, even as there appears to be regional grammatical variation distinguishing the Texas and California Chicano Englishes in question. [CE is] a variety of English that is obviously influenced by Spanish and that has low prestige in most circles, but that nevertheless is independent of Spanish and is the first, and often only, language of many hundreds of thousands of residents of California. (Metcalf 1974: 53) Chicano English is an ethnic dialect that children acquire as they acquire English in the barrio or other ethnic social setting during their language acquisition period. Chicano English is to be distinguished from the English of second-language learners [. . .]. Thus defined, Chicano English is spoken only by native English speakers. (Santa Ana 1993: 15)

As Fought emphasizes, following these two quoted works, CE is not the product of Spanish interference in a situation in which English is being learned as a second language. Rather, CE is simply a variety of English that has been influenced by Spanish over the generations. Fought also maintains that CE is not “Spanglish:” i.e. it is not a mixture of English and Spanish, nor is it simply a code-switching situation among bilinguals; it is a variety in its own right. In fact, many speakers of CE do not speak Spanish at all, though as Bayley and Santa Ana (2004: 168) note, they may have passive knowledge of Spanish. As these authors write: Since there are Chicano English speakers who do not speak any Spanish, Spanish cannot be the proximate source of their native English dialect. Nevertheless, because Chicano English speakers are often in daily contact with fluent speakers of Spanish and because many Chicano English speakers live in communities where they have only minimal contact with speakers of Anglo varieties, we acknowledge the possible influence of the Spanish substrate on features of Chicano English grammar.

1 Negative inversion and varieties of Texas English

19

Finally, I note that I use the term “Chicano” in this work, as it seems to be the term that has crystalized in the linguistics community to refer to the ethnicity and associated language variety. However, as Fought notes, it is not necessarily the term preferred by the language community itself.16 Fought (2003: 17) writes “the vast majority of [her] US-born young adults identify themselves as simply ‘Mexican’.” She continues, “A few speakers also had mixed feelings about ‘Chicano,’ which was sometimes associated with radical politics, or surprisingly even with gang members.” Minus the “gang-member” association, this has been my experience in Texas as well. The preferred ethnic term seems to be “Mexican” or more formally, “Mexican-American,” (even by those whose ancestors have been in Texas for multiple generations), or “Latino.” And similar to Fought’s participants in California, “Chicano” seems to have strong political associations in Texas as well. To avoid confusion, though, with other linguistic studies on the variety, I adopt “Chicano English” here, and assume that it overlaps to a significant extent with the California variety described in Fought (2003), while leaving open the likelihood of regional variation.

1.4 Shared features of African American, Anglo, and Chicano Englishes The three varieties of English under investigation here are all nonstandard varieties of English, meaning that they have phonological and grammatical forms that do not occur in standard varieties of English. Similarly, all three varieties have an “insider” quality to them, and they suffer from negative stereotypes and stigmatization, which is frequently known to their speakers. Most importantly, these three varieties have been in contact with each other in Texas since the mid-to-late 19th century. In this section, we will consider nonstandard grammatical forms that are shared by each of these three ethnic varieties.17

16 Chicano comedian Cheech Marin has an entertaining and highly informative word study on “Chicano” published in 2012 in The Huffington Post. . In a separate interview, available here, , Marin describes the term “Chicano” as originally being a derogatory term used to distinguish between citizens of Mexico and Mexican-Americans, which was reclaimed and taken up as a badge of pride by the Mexican-Americans from the late 1960s onward. 17 The comparison made in the present work is strictly grammatical, with no comment on phonological issues: i.e. vowel qualities, processes such as cluster reduction, consonant devoicing, and so on. Nor will there be discussion of the possibly significant differences in the prosodies, timing, intonation, or vocal qualities, and so on. Hopefully, issues such as these will be addressed in future research, especially as regarding language contact. They are beyond the needs

20

Chapter 1 Negative inversion in Texas, in three varieties of English

What follows below is influenced heavily by Fought’s (2003: Chapter 4) comparison of African American, Anglo, and Chicano Englishes as spoken in California. There is some overlap as well with Bayley and Santa Ana’s (2004: Section 2) treatment of CE grammar. Fought’s data come from California speakers of CE, while Bayley and Santa Ana rely on data from both Texas and California speakers. For the present work, I have tested each of the claims below with five speakers each of African American, Anglo, and Chicano Englishes from Texas. This is strictly preliminary research, and needs to be investigated in much more detail, but it is useful here to see the extent to which the three nonstandard varieties overlap in form. Undoubtedly, there will be quantitative differences in rates of appearance, the genres and styles in which the various forms appear, and so forth, in the three varieties. I do not test that here.18 Nor am I working from original corpus data or interviews, as is the case with Fought and the data found in Bayley and Santa Ana. I present the data below simply to give the reader an idea how the varieties in question are similar to each other, how they differ from US standard English, and how they compare to their California cousins. Further, as there are no in-depth empirical descriptions of these Texas varieties available, this section provides a preliminary springboard from which to consider Texas Negative Inversion in more detail in subsequent chapters. The examples used in the present chapter were taken from Fought (2003) and Bayley and Santa Ana (2004), with each source identified throughout on each example. Occasionally, Fought’s or Bayley and Santa Ana’s data have been adapted to soften potentially offensive language or to clean up natural speech issues found in their corpus data – i.e. repetitions, anacolutha, and various other disfluencies – that might have distracted the Texas participants from the issues of interest. The African American participants were interviewed in Austin, the Anglo speakers were interviewed in Comanche, and the Chicano speakers were interviewed in Corpus Christi. All of the participants were longtime residents of the cities and town in which they were interviewed, and they ranged in age from 19–44. None of the participants were known to me, and I

of the present study, however. See Thomas (2007) for discussion of phonological and other means by which sound might vary in African American English and other English varieties. See Thomas (2015) and Thomas and Bailey (2015) on the phonetics and phonology of AAVE generally. See Fought (2003), Bayley and Santa Ana (2004), and Field (2011) for discussion of the phonetics and phonology of CE in California. 18 Bayley and Santa Ana (2004: 181) note that “most of the features of Chicano English morphology and syntax that diverge from prescriptive norms are shared by other vernacular English dialects. However, [. . .] very few of these features have received the kind of systematic study required to determine if they pattern in the same way they do in other English dialects.”

1 Negative inversion and varieties of Texas English

21

did not ask for any identifying information other than the participants’ age, where they were from, and how long they had lived in Texas. The data points below were presented in person via clipboard and hardcopy questionnaire to research participants in Texas in 2018. Individual participants held the clipboard as they responded to the survey, frequently making comments to me and asking clarification questions as they worked their way through the process, oftentimes developing into a discussion with me on the various questions. Ultimately, there was a little variation in the results, but not enough to be interesting in a preliminary discussion such as this one. For the most part, all participants were in agreement on the intuitions reported below. Responses were often verbal to me, and generally of the form, “yeah, I could see myself saying that” or “I’ve heard that” or “yeah, that’s okay.” The actual survey items begin in (6) below and run through (21). The first set of questions, reported below in (6–10), were all acceptable in all three varieties. That is, my participant speakers of AAVE, AE, and CE indicated that each of these nonstandard forms was possible in their speech variety. (6)

a. Variable Agreement in 3rd Singular Forms. [Fought: 94] Everybody knew the Cowboys was gonna win again. b. Regularization of Irregular Verbs. [Bayley and Santa Ana: 169] When I was in school, the teacher hit my hand with a ruler. Man, when she striked me with that thing, that just blew my mind. c. ‘Is’ with Plural Subject. [Bayley and Santa Ana: 169] And the people that live there is never home.

(7)

Ain’t. [Fought: 94] It ain’t okay, but . . .

(8)

Nonstandard Pronoun Forms. [Fought: 95] Your mom and your dad will always think, he’s a guy, he could take care of hisself.

(9)

Expletive ‘It.’ [Fought: 96] It’s four of us, there’s two of them.

(10) Perfective ‘Had.’ [Fought: 96] The cops had went to my house yesterday. Question (11), which tests use of habitual be, was not accepted by all of my participants. Fought (2003) finds that some California Chicano speakers use

22

Chapter 1 Negative inversion in Texas, in three varieties of English

habitual be in the prototypical AAVE way. Of the Texas speakers, however, this was accepted only among the African American participants. Both the Anglo and Chicano groups rejected it, while simultaneously identifying it as African American speech. (11) Habitual ‘Be.’ [Fought: 96] The news be showing it too much. It is likely that this form has become such a shibboleth for AAVE that these other groups avoid it, unless they intentionally mean to signal an association with the African American culture or group, or unless it occurs in their speech beneath their level of awareness. As the data I’m relying on in this section were gathered via direct, overt survey of linguistic intuitions, it is possible that there is a mismatch between what the speech respondents believe they produce and the speech that respondents actually produce.19 The next group is taken from Fought’s Section 4.4, entitled “Features of Ambiguous Origin.” The first is negative concord in (12), which was accepted by all three Texas groups, as was the subject-auxiliary inversion (SAI) in embedded questions, as in (13). (12) Negative Concord. [Fought: 97] I didn’t have no self-confidence. (13) Subject-Auxiliary Inversion in Embedded Questions. [Fought: 97] He went up to some kids and then asked them where did they live. In the next section are forms that Fought suggests are particular to CE, beginning with “would in if-clauses” with stative verbs. (Fought is following Wald [1996] here.) The example in (14), lightly modified from Fought’s example due to suggestive content, was accepted by all of the Texas participants of all three ethnicities.20 (14) ‘Would’ in ‘If’-Clauses. [Fought: 99] If I woulda been a football player, I woulda been too busy to work.

19 See Labov (1972b) on speaker competence and reliability of linguistic intuitions, especially with respect to nonstandard dialects. See Rickford (1987) for discussion of the need to complement observed data with intuitional data. 20 The football player in (14) was a gangster in Fought’s original sentence, which I thought might be distractive to the Texas participants.

1 Negative inversion and varieties of Texas English

23

Another modal usage Fought mentions is with could, “where the basic meaning is competence” (100). In example (15), then, speakers of standard English would more likely use can in that situation. Fought remarks that the feature has not been documented in AAVE. Bayley and Santa Ana (175) also suggest that it is possibly an innovation in CE. Each of the three Texas groups accepted this sentence, however, on the relevant meaning. (15) Competence ‘Could.’ [Fought: 100] I learned that people that are left handed could draw better than people who are right handed. In (16), the adverb barely, which means something like ‘almost not’ in standard varieties of English, means something closer to ‘just recently’ in CE, according to Fought. The Texas participants in all three groups accepted (16), which is lightly modified from Fought (104) to emphasize the temporal meaning. (16) ‘Barely,’ Meaning ‘Just Recently.’ Hold on, man. Gimme a second. I just barely got here. The next section involves the use of like, be like, and be all, which are common in California CE. Bayley and Santa Ana (173) also suggest these usages are common in Texas as well. I found all three groups of Texas speakers to accept them. Focuser like, in (17) was accepted, but it did draw some laughs and comments about being bad grammar, or the way young people talk; it seems that the prescriptions against it were well known to the participants. (17) Focuser ‘Like.’ [Fought: 108] He could talk, like, smart, y’know. Quotative like in (18) was accepted by all groups, without comment. (18) Quotative ‘Like.’ [Fought: 108] She’s like, “No, you leave the house when you get married.” The final form in this group is be all, as in (19). Again, there were some comments here about this form being casual and informal, or more likely to be said by young people, but nevertheless it was universally accepted. (19) Be all. [Fought: 108] I was all, “I’m not giving you the keys to my car.”

24

Chapter 1 Negative inversion in Texas, in three varieties of English

Bayley and Santa Ana (169) provide examples of variable absence in past-tense marking in CE. Interestingly, examples such as (20) were not accepted by any of the Texas groups. (20) Variable Absence of Past-Tense Marking I saw some girl, she look pretty.21 This needs further research, but my hunch is that some speakers in these groups actually do produce sentences such as (20), but that it occurs for them beneath the radar, so to speak. Such sentences also appear to be quite stigmatized, as I received several negative comments across the three groups about bad grammar. So, it is also possible there was resistance to the feature on those grounds. A final group I tested with the Texas speakers comes from a section in Fought dedicated to nonstandard use of prepositions in and on. According to Fought (101), the nonstandard uses of these prepositions are likely tied directly to input or contact with Spanish: essentially, the meanings of both in and on are often conveyed by the Spanish en. These sentences were not preferred by any of the groups I surveyed – even the speakers of CE. Survey participants consistently “corrected” these sentences to reflect the standard usage. (21) Nonstandard Use of ‘In’ and ‘On.’ We’re really supposed to get out of here on June. [Fought: 100] We used to go stand in the porch because they never let us in the house. [Bayley and Santa Ana: 175] We see then a linguistic situation in which three ethnic varieties of English share many of the same nonstandard forms. Now, this is not to say that all of the nonstandard forms are used in precisely the same way or with precisely the same frequencies, but the forms are at least all clearly part of the grammatical inventories of the three varieties with very similar uses, a fact which is likely rooted in contact history.

21 Bayley and Santa Ana (169) also point out that it is possible (21) is actually a product of [kt] cluster reduction rather than tense-marker reduction.

1 Negative inversion and varieties of Texas English

25

1.5 Contact between the three varieties of English An issue that must be raised here, though I do not have much to say about it at the present time, is the fact of language contact among the three varieties of English in Texas. Fought (2003) notes that Negative Inversion does not appear in her corpus of CE or in her interviews with Chicano speakers in California. As noted above, however, NI was accepted by the Chicano speakers I worked with in Texas. If the Texas and California Chicano Englishes differ in this regard, it is possible that it is because of language contact. CE in Texas has been in deep contact with African American and Anglo Englishes since at least the 19th century, over many generations. Both of these latter varieties have had NI throughout this time. For example, (22) is taken from an ex-slave narrative recorded in Texas in the 1930s. Ex-slave James Boyd, on the subject of his singing ability, relates: (22) I can’t sing, now you knows can’t no old man sing what ain’t got no teef or hair. ‘I can’t sing, now you know no old man can sing who doesn’t have any teeth or hair.’ Similarly, in (23) we see a typical NI found in a blues song title, claimed to have been played on the radio in Corpus Christi, Texas, where much of the data from this book was collected, in the 1940s. (23) I remembered a story the musicologist Mack McCormick once told about Johnson playing a song called “Can’t Nobody Hide From God” over a Corpus Christi radio station during the early forties, when people on the coast were terrified of German submarines lurking in the Gulf of Mexico.22 Consider too the following descriptions from Wilkinson’s (2008: 90) oral history project of the years between 1870–1914 in east Texas. We find NI in the speech of African American Texan, Clifton Peoples, of Hopkins County, Texas. Clifton Peoples’s family grew bountiful gardens into which they invited others to gather at will, including those from households experiencing illness or financial trouble. According to Peoples, neighbors made sure ‘didn’t nobody go hungry.’

22 .

26

Chapter 1 Negative inversion in Texas, in three varieties of English

Peoples also testifies to the proximity in which African Americans and Anglos lived in east Texas at this time (Wilkinson 2008: 98). Clifton Peoples came of age in the Hopkins County community of St. Luke, named for the black Baptist church. Whites lived close and the white and black churches were about a mile apart. ‘They’d come down to our church and we’d go to theirs.’ Queried directly about what kind of people local whites were, Peoples replied, ‘They was good people. Religious type of people.’ According to Peoples, whites also came to St. Luke’s Juneteenth celebration, ‘all of them that were raised there.’

It is inevitable that there were linguistic repercussions of such close social contact. Likewise, CE speakers have been in contact with these groups since the earliest days of the state and before – at least in the southern, central, and eastern parts.23 Consider, for example, a description of this 19th century cotton-growing region found in Foley’s (1999) The White Scourge: After the Civil War former slaves and vaqueros increasingly worked together on plantations and cotton ranches in central and south-central Texas, while rapid immigration of southern whites into central Texas made it the “white belt” as well as the leading cotton belt of the state. [. . .] Almost from its beginning, the cotton economy of central Texas emerged as a hybrid culture situated between the South, the West, and Mexico.24

We thus see a long and complex history of contact between these three ethnic groups in Texas. In California, the contact situation between African American, Anglo, and Chicano English speakers is somewhat different. For one thing, the arrival of significant numbers of African Americans in California is more recent than that of Texas. The increase in the AA population in California is usually associated with the Great Migration, which began near the end of World War I. At this time, African Americans moved in large numbers from Texas and other Southern states to the northern US cities and to California. A similar outward migration to California occurred with mostly Anglo migrants from Appalachia after World War II, in which Appalachian migrants traveling the so-called “Hillbilly Highway” also headed for northern US cities and California. Even with this migration of the relevant language varieties, however, it’s not clear if there was significant contact with speakers of CE, or if the numbers and prestige dynamics themselves would have been conducive to NI’s entry into CE in

23 See Montejano (1987) and Foley (1999) for extensive discussion of the complex and intertwined relation of African Americans, Anglos, and Mexican-Americans in 19th and 20th century Texas. 24 In many ways in this part of the country, the lines dividing ethnicities – especially in the Anglo and Chicano categories – are far less clear than those distinguishing social classes.

1 Negative inversion and varieties of Texas English

27

California.25 So, if the Chicano speech of Texans and Californians does differ in this regard, it would not ultimately be too surprising, due to the differing histories of language contact. Another question that arises here in terms of language contact, for which I have no explanation, but will only note, is bidialectalism. In all of the surveys and interviews I conducted in Texas, it is likely that many of the participants and interviewees are actually bidialectal – meaning they speak at least two of the varieties in question here: Anglo and Chicano, or Anglo and AAVE, etc. This raises the question of what variety is influencing their responses to surveys, as it seems likely that all three of the varieties include NI as part of their grammars. Are survey participants competent to distinguish which language variety is influencing their answers to questions? Many of the participants might not even be aware where one variety ends and the other begins for them. With respect to Negative Inversion, this would seem not to be a problem in other studies, such as Labov’s work with AAVE speakers in New York City. Labov’s African American speakers were not in contact with Southern Anglo English, which has NI in its own right. Instead, the English with which Labov’s AAVE speakers were in contact was that of Anglo New Yorkers, which are not typically considered NI varieties. Thus, Labov could be fairly certain that the NI data he recorded from his participants originated in their AAVE grammar. It’s not clear how this plays out in the current study, though, in which participants might be bidialectal with both varieties being NI languages. It is certainly an area of future research, though. In the next section, I change pace and provide an overview of the kinds of data used in this book as well as the methods of collection used to obtain it.

1.6 Methods and data This book makes a wide range of claims about Negative Inversion and its users: from syntactic and semantic, to pragmatic, to race and generation, to social attitudes toward NI. A wide range of data is thus needed. Further, this book was not conceived of as an overarching project all at once; rather, it was built up one problem at a time, piece by piece, over three years of thinking about the issues and returning to Texas again and again to talk to speakers and to gather

25 Certainly there were Anglo NI speakers in California in the early 20th century. John Steinbeck’s California novels such as Of Mice and Men (1937) and The Grapes of Wrath (1939), for example, both contain prototypical NI sentences used by Anglo characters.

28

Chapter 1 Negative inversion in Texas, in three varieties of English

new data as new questions arose. As such, some of the methods will seem ad hoc to the strict methodologist. All were devised though to add specific pieces to the larger question of NI and all have a specific role to play. Most of the data and collection procedures are described in the chapters locally in which they are relevant. In this section, however, I provide an overview of the diverse methods that were employed. Much of the data comes from surveys and so relies upon native speaker intuitions. The book is concerned for the most part with semantic and pragmatic aspects of Negative Inversion, and so this introspective method via survey is appropriate. Importantly, the survey data in question are often situated as part of a connected discourse, so survey participants judge both felicity of an NI as well as grammatical well-formedness.26 This can be seen most clearly in Chapter 3 in the discussion of definite subject NPs in NI, in which it is argued that the definite subject NP must introduce a referent that is hearer-new. Often, seeing the proper licensing relation requires seeing the NI in a running discourse. If the NI were given in isolation, on the other hand, its appearance with a definite subject would likely be deemed unacceptable. See, for example, the methodology discussion in Matthewson (2004: Section 3.5). Matthewson (2004: 393) writes of negation in general that “a sentence containing negation will often need to be supplemented with a discourse context to be felicitous.” It is not surprising then that the semantic and pragmatic intricacies of NI are better visible to survey participants in contextualized settings. In the sociolinguistics literature, in which variation is often studied with quantitative methods, intuitional data are less commonly used. As King (2013: 448) writes, “Linguistic introspection has played a limited role in sociolinguistics.” King goes on to write that the traditional quantitative methods of sociolinguistic research on phonetic or phonological variation – i.e. examination of large numbers of tokens and contexts – “presents a challenge for research in morphosyntax, given the fact that particular grammatical structures may be infrequently occurring [. . .].” This is true of the Texas NI: there are no corpora with significant numbers of tokens, and my early recordings of relevant speakers contained very few instances of the construction as well. As such, the data sets I rely upon most heavily were collected via a wide range of surveys that queried native speaker introspection, and which are discussed in more detail in the paragraphs below. In Chapters 2, 3, 4, and 5, I rely upon survey data collected in several Texas high schools, which were variously majority Anglo or Chicano speakers, as well

26 Cf. Matthewson (2004).

1 Negative inversion and varieties of Texas English

29

as an HBCU.27 The surveys were conducted in person in the classrooms, and queried grammar, usage, and attitudinal aspects of NI. These surveys were followed by extensive in-class discussions with the students and teachers and are described in more detail below in Chapter 4. In Section 1.4 of Chapter 1, above, a comparison is made of nonstandard grammatical features that appear in the three language varieties in question in this book. For this section, I relied upon the intuitions of five speakers per variety to rate the acceptability of the nonstandard features – a total of 15 speakers. These data were gathered on the street in Austin, Comanche, and Corpus Christi. In Chapter 3, which investigates semantic and pragmatic features of NI, I relied upon the intuitions of 20 Texas speakers, 10 of whom were speakers of AE and five each who were speakers of AAVE and CE, respectively. Many of these speakers are known by me personally, and their intuitions were gathered informally through email surveys, telephone conversations, and Google Hangouts. Several speakers from each group, however, were not known to me and were surveyed on the street in Corpus Christi and Houston, Texas, in October 2018. It is telling that the results were essentially the same whether the data came from the impromptu interviews in the latter case or from those already known to me. Chapter 3 also makes quantitative, demographic claims with respect to Negative Inversion and Modal Existential constructions. These claims are based on surveys of speakers conducted online via Survey Monkey. The survey link and a call for participants was posted in three private Facebook groups, each of which was restricted to members who had grown up in the cities of Abilene, Corpus Christi, and Odessa, Texas. Chapter 4 explores attitudes toward Negative Inversion by NI speakers. Some of the data used in this chapter comes from the high school and HBCU surveys described above. Other data were collected over several days in cities such as Dallas, Austin, and Houston, and smaller towns including Comanche, Robstown, Sinton, Hamilton, and more. In these cases, language attitude data were collected using the traditional methodology of perceptual dialectology, inspired by Preston (1989). Participants were approached on the streets of the cities and towns mentioned above and presented with a survey sheet containing a

27 Survey distribution here is not completely consistent, as the Anglo and Chicano groups were surveyed in high schools, while the African American group was surveyed in a college setting. I would have preferred all three of the survey groups to come from high schools, but I could not get timely access to an African American majority school at that level. Thus, there is a slight difference in age and a difference in the education level of the three survey groups. All of the participants, however, can be categorized as working-class or low-mid, middle class, with only a few years difference in age.

30

Chapter 1 Negative inversion in Texas, in three varieties of English

blank map of Texas on one side and several attitudinal and demographic questions on the other in the style of Cukor-Avila et al. (2012).28 Participants were asked to label the areas of the state in which they believed NI would be found. They were also asked to label the areas with whatever descriptive terms they found appropriate. When participants were finished with the map task and survey questions, I discussed their results with them and asked them further openended questions regarding their opinions about NI, the people who use them, and their place in the language variety. I do not rely on the dialectology maps themselves in the present book, but I do use the attitudinal data that were gathered during that task. In general, I found the map task a very helpful and disarming way of easing into the discussion of attitudes toward NI, which was the main purpose of the interaction. Some data used in this book is taken from previous literature. When this is the case, the original source is noted. Some data illustrating various grammatical points are taken from Internet searches. These examples are indicated with γ, a practice pioneered by Laurence Horn to indicate that the data were found using Google.29 Thus, every piece of data used in this book has been assessed through formal or informal surveys, or attested online, and has also been assessed by me, as a native speaker of one of the varieties in question.

1.7 Survey participants and socioeconomic class Almost all of the participants who have contributed to this study – whether in the African American, Anglo, or Chicano groups – can be assumed to be either working class or low-mid, middle class.30 I hedge here with “almost all” because

28 Cukor-Avila et al. (2012: 11) describe their methods in this way: Our fieldwork sites included malls, restaurants, bars, hotel lobbies, university campuses, stores, gas stations, and historical sites (e.g., the Capitol Building in Austin and the statue of Sam Houston in Huntsville). Respondents were randomly approached, given a map, and instructed to indicate places where they thought people in Texas sounded different, and next, to label those areas or write down what they would call that way of talking. In this particular stage of data collection, I used a similar approach, but instead asked participants to label where they thought speakers would use NI. Most participants indicated that NI could be found all over the state. 29 See Horn (2018a) on the origins of γ as a data marker. 30 In many ways in this part of the country, the lines dividing ethnicities – especially in the Anglo and Chicano categories – are far less clear than the lines distinguishing social classes.

1 Negative inversion and varieties of Texas English

31

to my knowledge, none of my participants can be considered as upper class or elite, regardless of how those concepts are defined. As such, this work does not make claims about the upper-class demographics. For the present purposes, it is enough to assume NI exists in the informal speech of many working- and middle-class Texans and that they likely also exist to a lesser extent in the speech of the upper classes as well – especially those who are older. Matyiku (2017: 10), for example, “found middle-to-upper class speakers whose speech exhibited the phenomenon in Austin, Texas,” though it is not clear what criteria Matyiku uses to identify upper class.31 The situation is directly comparable to Wolfram and Schilling’s (2016: 34) discussion of the distribution of regional dialect features across social classes in Appalachia. With respect to certain Appalachian English features, they write: “[W]e have to be careful to note that these features are used at different rates among different social groups in Appalachia and may not be used at all by those of higher social status.”

1.8 Conclusion In this chapter Negative Inversion has been located in the primarily informal, vernacular speech of speakers of African American, Anglo, and Chicano Englishes in Texas. A survey of nonstandard grammatical features of these three ethnolects was also undertaken, which leads to a discussion of historical contact between these three groups in Texas. These three groups have been in contact in the cotton and ranching country of Texas, the territory best represented in this book, for 150 years or more. This differs from places elsewhere in which similar groups are in contact, such as California, and in which speakers of CE have been found to not have NI. Also in the chapter is an introduction to the populations who supplied most of the data for the analyses given as well as the methodologies used to collect that data. In the next chapter, we take a very close look at the empirical account of NI assumed in this book and at some discussions of how the present empirical account differs from all those that have come before.

31 According to Feagin (1979: 235) in Appalachian English, NI is used by both the upper and working classes. Feagin’s Section 3.3 defines class and caste in the study in terms of family lineage, occupation, location of residence, income, club and church memberships, and a number of other factors.

Chapter 2 Negative inversion and its contents Love your data! – Háj Ross Sometime in the 1970s, 80s, 90s or 2000s

2 A new empirical account of negative inversion This chapter dramatically expands the received empirical understanding of the Negative Inversion construction. There is a significant body of preceding work on NI in AAVE and Southern Anglo Englishes, dating to Labov et al. (1968), most of which assumes a uniform empirical description for the construction, with little attention paid to potential racial, regional, or generational dialect variation.32 In this present chapter, I show that NI in the three Texas vernaculars, while very similar to each other, in fact differ widely from all previous empirical accounts of NI that have been advanced elsewhere in terms of their syntax, semantics, and pragmatics. In this chapter also, I discuss most empirical claims that have been made about NI in previous literature, and I point out what does and doesn’t hold with respect to the current data. I also provide pointers here to theoretical arguments that have been made in the past based on these previous empirical assumptions and discuss whether similar arguments can be advanced based on the present data. In the cases where the present empirical account conflicts with previous works, I do not claim that those previous accounts of NI are necessarily inadequate. Perhaps, as they are circumstances of other places, times, people, and vernaculars, those previous accounts simply describe similar though nonidentical linguistic objects. Without more research, we have no way of knowing. Finally, this chapter is a grammatical overview of NI in the three Texas varieties in question; it does not cover statistical appearances of NI in the different varieties or across different genres and styles. There might very well be differences in genre and style in these Texas varieties. This too though must be a subject

32 Early mention of Negative Inversion in AAVE can also be found in Putnam and O’Hern (1955: 22), in which it is noted only that “the study dialect included occasional unconventional word order. Declarative sentences sometimes began with the auxiliary verb: Didn’t none of them want to go to school this morning.” There is no further elaboration on the topic. In his 1978 dissertation, Sanders dedicates a paragraph to “negative preposings,” such as Can’t nobody do it and Didn’t nobody help. In his study of African American speech in Columbia, South Carolina, Sanders finds that Negative Inversion occurs primarily in the speech of those over seventy years old. See Sanders 1978: 56–57. https://doi.org/10.1515/9781501512346-003

2 A new empirical account of negative inversion

33

of future research. For the present, in this chapter, I provide an empirical depiction of NI that is as comprehensive as possible for the varieties in question and for the needs of this present work.

2.1 The present empirical account of negative inversion We begin with a very general overview provided by Matyiku (2011/2018), writing for the Yale Grammatical Diversity Project (YGDP): Negative inversion is a phenomenon in which a declarative sentence begins with a negated auxiliary or modal, such as can’t, ain’t, or won’t, followed by a quantificational (or indefinite) subject, such as nobody.

This description sums up the most basic characteristics of the construction, and it is generally assumed to be the starting point for most if not all grammatical analyses of NI across African American and Southern US Englishes. The YGDP provides several examples, listed below, which are taken from Appalachian English, Alabama English, and West Texas English, respectively: (24) a. Can’t nobody beat ‘em. ‘Nobody can beat them.’ b. Didn’t nobody get hurt or nothin’. ‘Nobody got hurt or nothing.’ c. Won’t anybody hit us. ‘Nobody will hit us.’ d. Cain’t all o’ ya go at once.33 ‘You can’t all go at once.’

(Labov et al., 1968) (Wolfram and Christian, 1976) (Feagin, 1979) (Foreman, 1999)

In these examples, we see a range of initial auxiliary choices, positive and negative subject NPs, as well as with and without negative concord. These topics and more will be taken up in the following sections, beginning with the initial auxiliary in Section 2.1.1.

33 The eye-dialect spelling of cain’t here in Forman (1999) calls attention to the frequent pronunciation of this negative modal in Southern and African American Englishes. In these varieties the word is pronounced with mid-front vowel [e] rather than the low-front [æ] typically found in standard English. Many of my survey participants over the course of this project commented that can’t-NI sounds much more natural when pronounced [keɪnt], as opposed to [kænt]. I do not know if this is merely a fact of stylistic coherence or if there are additional linguistic, perhaps semantic, differences distinguishing [keɪnt] and [kænt].

34

Chapter 2 Negative inversion and its contents

2.1.1 The initial auxiliary Foreman (1999: 6), writing on West Texas English, states that “essentially any Aux which can bear negation” can participate in this construction, including “ain’t (which can stand for both be+n’t as in [25a] and perfective have+n’t as in [25b]), wudn’t (SE wasn’t), don’t, didn’t, won’t, wouldn’t, cain’t (variation of can’t), couldn’t, shouldn’t, and hadn’t” as shown below.”34 Essentially, if a modal or non-modal auxiliary is used in the relevant linguistic variety, it can appear in initial position in NI.35 Below we see several examples from Foreman (1999)’s account of NI in West Texas English [Foreman’s (12a–i)]. (25) a. Ain’t nobody doin’ nothing wrong. ‘Nobody is doing anything wrong.’ b. Ain’t very many people read your book. ‘Not many people have read your book.’ c. Wudn’t no more than ten people allowed in at a time. ‘No more than ten people were allowed in at a time.’ d. Don’t nobody live there. ‘Nobody lives there.’ e. Didn’t nowhere near a thousand people go to that concert. ‘Nowhere near a thousand people went to that concert. f. Won’t none of the students go to the party. ‘None of the students will go to the party.’ g. Wouldn’t no gentleman act like that. ‘A gentleman wouldn’t act like that.’ h. Cain’t no dog but Ol’ Blue do that trick. ‘No dog but Ol’ Blue can do that trick.’ i. Couldn’t neither of them fit in the car. ‘Neither of them could fit in the car.

34 In the AAVE under investigation in Labov et al. (1968: 286), there is believed to be a tighter restriction on which modal verbs are allowed. Here is Labov et al., as quoted in Weldon (1994: 3), “The range of modals which can be used seems to be quite limited. From the emphatic and affective nature of negative inversion, we would infer that may, might, would, and should are not likely to be used.” According to Weldon, however, auxiliaries shouldn’t and wouldn’t do occur in NI constructions, as they of course do in Foreman’s West Texas English data and the data of the present work. 35 The negativized auxiliary typically does appear in first position in the clause. It is possible to find NI modified by sentence and speaker adverbs, but these tend to reside on a separate intonational tier, as in (i), though they need not do so. (i) Honestly, didn’t nobody go in there. ‘Honestly, nobody went in there.’

2 A new empirical account of negative inversion

35

j. Shouldn’t nobody be allowed to act like that. ‘Nobody should be allowed to act like that.’ k. Hadn’t nary a soul set foot in that house ‘til Dave moved in. ‘Not a soul had set foot in that house, until Dave moved in.’ It is worth noting in Foreman’s data above that there is only one form of be represented, in (25c), and that it has undergone the Southern /z/ → [d] assimilation, in which wasn’t is realized as [wʌdn̩t].36 This change frequently affects auxiliary verbs appearing in this position (and others) in the relevant dialects, but it doesn’t necessarily do so. On the other hand, Weldon (1994: 5) notes that in her AAVE NI data, fricative forms such as isn’t, hasn’t/haven’t, doesn’t, and wasn’t are rare but that these auxiliaries are all more acceptable with “the less formal pronunciations [ɪdn̩t], [hædn̩t], [dʌdn̩t], [wʌdn̩t].” Thus, she quotes an example from Labov et al. (1968), noting that it is “extremely unnatural in her judgment,” which she indicates with the double question marks below. (26) “??”Doesn’t nobody really know that it’s a God, you know. ‘Nobody really knows that it’s a God, you know.’ As with Weldon’s AAVE data, the fricative-bearing forms are also somewhat less natural in the Texas English of the present study, though they are certainly possible, as in (27) below. It should be noted, though, that all of the examples in (27) are completely natural with the [d] forms (i.e. [ɪdn̩t], [hædn̩t], etc.) as well.37 (27) a. Isn’t a one of ‘em gonna be there. ‘None of them will be there.’ b. Hasn’t nobody been up there since John left. ‘Nobody has been up there since John left.’ 36 See Reynolds (1994) for further discussion of Southern /z/ → [d]. 37 Weldon attributes the awkwardness of the fricative variants to a “formality co-occurrence restriction.” This would appear to cause the fricative variants, which are “more closely associated with standard English (SE) (and hence with more formal registers) [to be] unacceptable in NI constructions, which are associated with more vernacular registers” (5). This does sound plausible; but, at least for the Texas varieties, the fricative variants can appear straightforwardly in NI – especially downstream, as in (i), which is perfectly acceptable. (i) Cain’t nobody go inside if Johnny isn’t home. So, it would seem that the “co-occurrence restriction” would have to be limited to the initial auxiliary, which is rather unlikely.

36

Chapter 2 Negative inversion and its contents

c. Doesn’t anybody think about that stuff anymore. ‘Nobody thinks about that stuff anymore.’ d. Wasn’t any of ‘em really worth lookin’ into. ‘None of them were worth looking into.’ An important feature of NI is that most of the modal and non-modal auxiliaries must be negativized, bearing the -n’t inflection.38 Thus, non-inflected examples like (28a–c) are unacceptable. (28) a. *Cannot nobody beat ‘em. ‘Nobody can beat them.’ b. *Will not anybody hit us. ‘Nobody will hit us.’ c. *Did not anybody come in here.39 ‘Nobody came in here.’

38 See Zwicky and Pullum (1983a) on inflection vs. cliticization of n’t in English generally. 39 This example can be acceptable on an unrelated emphatic reading, with rhetorical escalation of the dialogue and heavy stress on the negation. Speaker A: I’m telling you, somebody came in here! Speaker B: No way. Nobody came in. Speaker A: Yes, did too! Speaker B: No. Did NOT anybody come in here! Relatedly, Green (2014: 126) mentions (i) in her Fn. 11, which appears to be an inversion sentence in the absence of negation. (i) Did too somebody steal the tea! ‘Indeed somebody did steal the tea.’ She notes, however, that it is not certain that this process is productive with different auxiliaries.” I believe that the process is in fact productive though, as in (ii). (ii) a. Can too he come over! b. Will too I go in there! c. Would too he kiss me! So, the construction does seem to be productive with other auxiliaries. Not mentioned by Green is that the sentences in (i) and (ii) have rigid restrictions on the kinds of rhetorical situations in which they can appear. Like the incredulity response construction described below in Chapter 6, these kind of echoic “Did too!” constructions must appear as a response to a contested, directly preceding assertion, as in (iii). (ii) Parent: You will not go in there! Child: Will too I go in there!

2 A new empirical account of negative inversion

37

Similarly, non-negative examples such as (29a–b) are unacceptable. (29) a. *Can nobody beat ‘em. ‘Nobody can beat them.’ b. *Will anybody hit us. ‘Somebody will hit us.’ c. *Should anybody have to do that. ‘Nobody should have to do that.’ d. *Is anybody home. ‘Somebody is home.’ Sentence (29a) also shows that, in Parrott’s (2000: 417) words, “The negative quantifier alone does not license NI.” Some modals, such as may and shall, do not occur frequently in negated form in the relevant dialects. Accordingly, these are also generally not acceptable in NI, as in (30a–b). Modals must and might appear to be possible in NI, as in (30c–d), though their statuses are a bit more marginal.40 Again, this is likely just a reflex of the fact that must and might don’t frequently bear negative inflection at all in the relevant dialects. (30) a. *Mayn’t nobody be in there. ‘Nobody may be in there.’ b. *Shan’t nobody pass. ‘Nobody shall pass.’ c. Mustn’t anybody go in there. ‘Nobody must go in there.’ d. Mightn’t anybody show up. ‘Nobody might show up.’

Negative Inversion does not have rhetorical restrictions of this kind, which is reason enough to keep it apart as a separate construction. As Lakoff (1987) writes, “one of the major ways we have of identifying grammatical constructions is to find cases where pragmatic conditions are associated with syntactic conditions” (470). See also the discussion in Chapter 6, below. 40 Weldon (1994: 4) marks an example with mightn’t ungrammatical, but allows that mustn’t and “certain semi-modals such as needn’t” are possible, if marginal in AAVE, though may be “suited to more formal registers than one would typically associate with Negative Inversion.”

38

Chapter 2 Negative inversion and its contents

In addition to the principal modal verbs, NI is also possible with semi-modals, such as need, ought, and dare, as in (31). Dare is a bit more marginal, as in (31c), though it is possible. (31) a. Needn’t anybody go over there anymore. ‘Nobody needs to go over there anymore.’ b. Oughtn’t nobody to be in there right now. ‘Nobody ought to be in there right now.’ c. ?Daren’t anybody tell him that to his face. ‘Nobody dares tell him to his face.’ In (32) we see the semi-modal idiom had better, which can also appear in NI, with nobody intervening between had and better in (32a), as well as with the subject nobody following the modal idiom in (32b). In (32c–d), we see examples in which the subject is the negative polarity item anybody, which is also completely acceptable. (32) a. {Hadn’t/Hain’t} nobody better call me again. ‘Nobody should call me again.’ b. Hadn’t better nobody call me again. ‘Nobody should call me again.’ c. Hadn’t better anybody call me again. ‘Nobody should call me again.’ d. Hadn’t anybody better call me again. ‘Nobody should call me again.’ Also in the semi-modal category, we find what I believe to be the only contemporary instances in which the initial auxiliary of NI does not require the negativization. This is with had-less better and best as in (33–34).41 (33) a. Better not anybody bother him before he gets his coffee. ‘Nobody should bother him prior to him drinking his coffee.’ b. *Bettern’t anybody bother him before he gets his coffee. ‘Nobody should bother him prior to him drinking his coffee.’

41 See Denison and Cort (2010) for extensive discussion of the modal and verbal properties of better and best.

2 A new empirical account of negative inversion

39

(34) a. Best not anybody go in there right now. ‘Nobody should go in there right now. b. *Bestn’t anybody go in there right now. ‘Nobody should go in there right now.’ Unlike the aforementioned auxiliaries and semi-modals, though, had-less better and best do not allow n’t in any syntactic environment. So, it’s not just an issue with NI. Consider (35–37). (35) a. You should go. b. You better go. c. You best go. (36) a. You should not go. b. You better not go. c. You best not go. (37) a. You shouldn’t go. b. *You bettern’t go. c. *You bestn’t go. Similar to the non-negativized examples of (33–34), we also find non-inflected examples with negative subjects, such as (38a–d), which are attested from various Internet searches.42 (38) a. γ That said, better not nobody ever make James Bond a black guy UNLESS it’s in the guise of a full, from-the-top reboot. ‘That said, nobody should ever make James Bond a black guy [. . .].’

42 These best not and better not examples, while attested on online sources, are also acceptable to many of my participants. Relatedly, based on the examples in (38), I constructed the examples below in (ia–b) with expletive subjects, and these were also found to be acceptable to the middle-aged and older groups I queried. (i) a. They obest not anybody bother him right now. ‘Nobody should bother him right now.’ b. There better not anybody talk bad about Ms. Jenkins. ‘Nobody should talk bad about Ms. Jenkins.’

40

Chapter 2 Negative inversion and its contents

b. γ Better not nobody talk bad about ms jenkins. . .except me. ‘Nobody should talk bad about Ms. Jenkins, except me.’ c. γ Born and raised in the great state of Kentucky and best not nobody say nothing negative about my state. ‘Nobody should say anything negative about Kentucky.’ d. γ [Comment on a Nicki Minaj “Anaconda” parody] Better not nobody diss my Nicki for real. this shyt is funny doe! ‘Nobody should diss my Nicki for real. This shit is funny though.’ Data such as those in (33), (34), and (38) are completely novel to the NI literature and contradict every previous empirical description of NI sentences. I will discuss them further in Chapter 6 below. In addition to these non-negativized examples, it is also possible to find historical data that appear to be non-negativized as well – especially if the expletive subject is present – as in (39), which are found via Google search, and so indicated with “γ.” I do not find any such examples in which the expletive is absent. Interestingly, most of my survey participants found these to be very awkward. (39) a. (California, 1889) γ As a matter of fact, there cannot anybody go in there and let horses for what they are letting them now and have to haul in their feed. ‘Nobody can go in there and let horses [. . .].’ b. (New Jersey, 1913) γ It has been said you might go out in the remotest part of the earth and hide yourself away the best you know how, get away where there could not anybody apparently find you. ‘Nobody could apparently find you.’ c. (Texas, 1916) γ There is another part of the conversation that I did not recall this morning, in regard to keeping the service in the Jewel Theater. Mr. Touchon said that there could not anybody get our program. ‘Nobody could get our program.’ d. (New York, 1916) γ There cannot anybody in our company make circles around me, and if this kind of performance has been going on such as you mentioned, I may say, and I am authorized by Mr. Shonts to say, that an example will be made of anybody who is engaged in that kind of business. ‘Nobody in our company can make circles around me [. . .].’

2 A new empirical account of negative inversion

41

e. (Michigan, 1944) γ A neighbor, also called as a witness by defendant, testified that “there cannot nobody mar her reputation.” ‘Nobody can mar her reputation.’ Most of the tokens I find are from prior to the mid-20th century. It is possible that at this time negative inflection on the auxiliary was not yet a requirement. Another possibility is that the negative inflection was a grammatical requirement and that the non-inflected forms above were actually hypercorrections in speech or in transcription. In any case, this older form is not a possibility with contemporary speakers I surveyed for this book – not even the older participants – which suggests a construction in flux in the 20th century and which fits nicely with the claim I make in Chapter 3 regarding the loss of the expletive subject itself across the last few generations. Both of these changes – i.e. the loss of the expletive subject and the requirement that the auxiliary bear negative inflection – can be used as supporting arguments that NI has become a type of idiomatic construction as well, a claim I will make below in Chapter 6. Moving on to other auxiliary configurations in NI, it is also possible to find double modal verbs. For example, in (40a–c), we see couldn’t used to, shouldn’t ought to, and best oughtn’t. (40) a. Couldn’t used to anybody get in there. ‘Nobody used to be able to get in there.’ b. Shouldn’t ought to anybody have done that. ‘Nobody should have done that.’ c. Best oughtn’t nobody bother him right now. ‘Nobody should bother him right now.’ In general, the initial auxiliaries keep their normal range of meanings in NI. Thus, the modal auxiliaries allow both epistemic and deontic readings, similar to normal modal usage. In (41a), we can coerce an epistemic reading from should, compared to a clear deontic reading of it in (41b). (41) a. [Speaker A, scanning a list of rules of acceptable behavior, running his finger down the list.] Let’s see here. . . Nope. According to the rules, shouldn’t nobody be allowed to act like that. ‘As far as I can tell, nobody should be allowed to act like that.’ b. [Later that day, Speaker A, watching a group of kids walk across his manicured lawn. He shakes his head in anger.] Shouldn’t nobody be allowed to act like that. ‘If I was king, nobody would be allowed to act like that.’

42

Chapter 2 Negative inversion and its contents

Similarly, in (42a) and (42b) there are clear examples of epistemic and deontic readings of can NI. (42) a. [Speaker B, watching the final playoff game, with one team dominating all the others.] They got a hell of a team this year. I guess can’t nobody beat ‘em. ‘Nobody is able to beat them.’ b. [Speaker B, giving instructions to other teams to let one team win.] All right now. Look everybody. We got to let them win this time around. Can’t nobody beat ‘em. Does everybody understand? ‘Nobody is allowed to beat them.’ Finally, it should be reiterated that lexical verbs are absolutely disallowed in the construction. Unlike the various marginal auxiliary and modal examples listed above, lexical verbs are incontrovertibly ungrammatical, with no hope of rescue. (43) a. *Ran nobody that race. ‘Nobody ran that race.’ b. *Slept nobody in that bed. ‘Nobody slept in that bed.’ c. *Drank nobody that last beer. ‘Nobody drank that last beer.’ Much of the attention given to the initial auxiliary in the NI literature has been dedicated to the putative requirement that it bear the negative inflection n’t. Beyond that, further attention is typically not paid to it. In this section, however, I have expanded the empirical description to include discussion of semimodals as well as examples in which negation need not be inflectional. In the next section, we turn to look at the nature of the subject NP, which has drawn the bulk of the attention in the literature in the last few decades. 2.1.2 The subject noun phrase In all previous literature except Salmon (2018a), subject noun phrases of Negative Inversion are required to be quantificational or indefinite.43 For example, here is Labov et al. (1968: 285): “The explanation [. . .] does not suggest any reason why

43 This point is also made in truncated form in Salmon (2017), which is a short paper published in conference proceedings. Salmon’s exception is noted briefly in Matyiku (2017) and Blanchette and Collins (2018), though it is not taken up in those works in detail.

2 A new empirical account of negative inversion

43

indefinites are always involved.” More recently, here is Matyiku (2011/2018), summing up the situation for the Yale Grammatical Diversity Project: Definite subjects such as pronouns, proper names, and DPs headed by definite or possessive elements are not possible, and so none of the following are allowed: – Don’t the police break up a fight. – Won’t they catch us. – Wouldn’t Sally and Jean help the poor man.

This characterization of subject NPs has stood for NI since its introduction into the linguistics literature in 1968 and is generally assumed to be identical for AAVE and other Southern US Englishes.44 Essentially, NI has been assumed to exhibit definiteness effects similar to those of Milsark (1974)’s claim regarding TE sentences.45 As can be seen in the many examples above, the YGDP’s description does provide for many very natural NIs. However, as I discuss in this section, and in much more detail in Chapter 3, this characterization does not match the present data. Definite and referential subjects are allowed in NI sentences as long as specific information structural requirements are met.46 Essentially, as I argue in Salmon (2018a) and in more detail in Chapter 3 of the present work,

44 See, for example, Fasold and Wolfram (1970), Martin (1992), Foreman (1999), Green (2014), Blanchette (2015), Matyiku (2017), Blanchette and Collins (2018), on this putative prohibition of definite subjects. Conversely, see Salmon (2018a) and Chapter 3 below for a detailed empirical illustration that definite subjects are not syntactically or semantically prohibited, but rather are only pragmatically constrained and are thus very possible in the construction. 45 McNally (2011: 1832) sums up Milsark’s effect: [T]he definiteness restriction amounts to a restriction on the acceptability of definite, demonstrative, and necessarily quantificational noun phrases, including proper names and personal pronouns, in the pivot. [. . .]. [T]here are restrictions [posited] on the indefinite noun phrases that are licensed as well, which limit the acceptability of partitives (including covert partitives, identifiable by stress on the determiner) and generically interpreted indefinites. In previous NI literature, all of these subjects that are prohibited in TEs are also believed to be prohibited in NI – except for the universally quantified subject. It is well known that NI does appear with this type of subject. See, for example, Foreman (1999) and Matyiku (2017). The rest are generally considered to not appear in NI. 46 Foreman (1999) uses the idea that definites are disallowed as NI subjects to distinguish NI from other SAI constructions, which do allow definite subjects. As Foreman (1999: 11) writes: A more substantial difference between WTE NI and questions and SE ‘negative inversion’ is which kinds of subjects the auxiliary is allowed to raise over. Both questions and SE ‘negative inversion’ allow the Aux to raise over Definite subjects [. . .]. However, this is not the case for WTE NI. In NI sentences, the negative Aux cannot precede a Definite subject.

44

Chapter 2 Negative inversion and its contents

definite subjects are allowed in NI as long as they are construable as hearernew, in the sense of Prince (1992) and Ward and Birner (1995). Two examples will suffice here. In (44), we see a case in which there is a new instantiation of a known type. Here, there is situation in which a type is familiar, and so can be marked definite, but the instantiation of it is hearer-new. See Chapter 3 for more in-depth explanation. (44) Speaker A: Hey baby, y’all have a good time tonight? How was the bar? Speaker B: Not too bad. Couldn’t the usual crowd get in there ‘cause of the cover charge, so there was plenty of room. Got the best table in the house. ‘The usual crowd couldn’t get in there.’ In (45), we see a case where a demonstrative is used in context with a pointinggesture demonstration. Here, then, the referent is construable as hearer-new, but at the same time is present in the context and so can be picked out via demonstrative. (45) [A man tries to walk into a concert without paying, and gets caught by security. He quickly points at a guy who walked in right in front of him.] Security Guard: Excuse me sir. It’s $5 to get in. Man: Didn’t that ☞ guy there pay! ‘That guy ☞ there didn’t pay.’ Definite examples such as (44) and (45) and more are discussed in Chapter 3. The acceptability of these kinds of subjects marks an abrupt shift away from every other treatment of NI in the literature, except for Salmon (2018a). Every other treatment of NI has assumed that definite subjects of all kinds are ruled out, and this assumption frequently plays an important role in the theoretical accounts that have been given.47 As is explained in Salmon (2018a) and in

The fact that NI does allow definite subjects, then, potentially undermines Foreman’s distinction between NI and other SAI constructions, though this is not pursued in the present work. 47 For example, Foreman (1999) and Matyiku (2017) provide syntactic movement analyses that prohibit definite subjects based on the landing site to which the definite NP moves. Here is Foreman on the subject, briefly: Definite subjects [. . .] cannot appear below NegP2, and thus cannot appear after not or ain’t, because the Definite DPs must appear in RefP which is above NegP2. Thus [. . .] if the

2 A new empirical account of negative inversion

45

Chapter 3, however, it is possible that the pragmatic constraints on the appearance of definite subjects has essentially concealed their acceptability over the last several decades, especially as almost all prior work considers singlesentence NI examples with no context or running discourse. Another important update to make with respect to NI subjects is with respect to weak quantifiers such as some and several.48 These have also been assumed to be ungrammatical in subject NPs of NI. Here, for example, is Green (2014), writing with respect to AAVE, “the weakest quantifiers on the positive side of the scale do not occur in [NI] constructions felicitously” [Green’s ex. 28]49: (46) “#”Didn’t some (of the students) show up. ‘Some of the students didn’t show up.’ Similar sentiment is found in Matyiku (2017) and Blanchette and Collins (2018), as is reported in more detail in Chapter 3, below. However, sentences such as (47) were overwhelmingly accepted by my test participant speakers of African American, Anglo, and Chicano Englishes. Relatedly, it is reported in Matyiku (2017) that several NP cannot be a felicitous subject of NI. This, too does not hold for the Texas data. Examples such as (47) are flawless and can be replicated straightforwardly. (47) Didn’t several of ‘em get off the boat. ‘Several of them didn’t get off the boat.’ As with the definite subject NPs described above, the claimed prohibition against these weakly quantified NP subjects has been used to provide support for various theoretical arguments. For example, Foreman (1999), Zanuttini and Bernstein (2014), Matyiku (2017), and Blanchette and Collins (2018), require the negation in NI sentences to have unambiguous wide scope over the rest of the sentence. On these syntactic movement accounts, it is precisely this disambiguation that motivates movement of the negative auxiliary to the front of the sentence in the first place. Thus, in an example like (48a), there are two scopal interpretations available, but in (48b), there is argued to be only one interpretation, a quality that Blanchette and Collins (2018) suggest is

Definite subject is required to move to RefP then [it] cannot appear below negation and the [NI word order] cannot be derived. (13) 48 See Milsark (1974) for discussion of weak and strong quantifiers in general. 49 Green does not specify which regional variety of AAVE or generational age group is under consideration in that work.

46

Chapter 2 Negative inversion and its contents

an instance of “scope freezing,” as described for a different set of phenomena in Collins (2016). On these scope-based accounts, movement from the canonical positions of (48a) occurs to resolve this putative scopal ambiguity, resulting in an NI like (48b). (48) a. Everybody didn’t eat dinner tonight. ‘Everybody didn’t eat dinner tonight.’ ‘Not everybody ate dinner tonight.’ b. Didn’t everybody eat dinner tonight. ‘Not everybody ate dinner tonight.’

(∀ > ¬) (¬ > ∀) (∀ > ¬)

However, with weak quantifier examples such as (49a–b), there is no difference in ambiguity between the NI form and the non-inverted form. As such, there is no need or way to motivate a disambiguation account with such examples. (49) a. Didn’t some of the students show up.‘Some of the students didn’t show up.’ b. Some of the students didn’t show up. NIs such as (49a) are predicted to be ruled out on syntactic accounts such as those mentioned above, which rely on disambiguation to motivate the inversion movement. However, as mentioned above, these NIs are found to be perfectly acceptable among the participants of the current study. A further point to note here, however, is that the characterization of the universal quantifier as necessarily having wide scope in (48b) is not representative either with respect to the present data. I will discuss this further below in Chapter 6. For the moment, though, I will simply point out that the NI in (48b) can in fact be ambiguous in a way similar to that of the non-inverted (48a). Recall, there should not be a reading available of sentences such as (48b) in which the universal quantifier takes scope over the negation: this means there should be no readings of (48b) where didn’t everybody X = nobody Xed; the only reading available should be didn’t everybody X = not everybody Xed. Consider (50), though, which depends on the nobody Xed reading in order for the entire structure to be felicitous. Example (50) sets up a specific contrast of strength between the paired statements. The first statement is weak, while the second is strong. If the NI with universal quantifier could not take wide scope over the negation, as Foreman and others suggest, the pairing would essentially be a tautology.

2 A new empirical account of negative inversion

47

(50) Speaker A: Hey man. How was the party last night? Speaker B: Not so good. Didn’t some people even bother comin’ out. Hell, what am I sayin’. Didn’t everybody come out. There wudn’t a soul there last night. It was a total bust. ‘Nobody came out to the party.’ Pace Foreman (1999) and the others, then, NI in which universal quantifier subjects take wide scope are possible in this Texas data. This reading might be difficult to tease out, but it is nonetheless possible. What this likely tells us is that we are looking at a product not of sentential syntax, but of a pragmatic restriction of some kind. In any case, syntactic accounts such as the aforementioned ones, which assume that inversion movement occurs to achieve unambiguous wide scope, are unmotivated for this kind of data. A related empirical point has also been made for NIs having overtly negative subjects, such as (51), below. According to Blanchette and Collins (2018: 5): Because they have two syntactic negations, they should offer two possible interpretations: one in which the syntactic negations contribute a single semantic negation, the socalled Negative Concord (NC) reading, and one in which each occurrence of negation contributes a distinct semantic negation, a true Double Negation (DN) reading.

However, as Blanchette and Collins go on to suggest, “Despite the existence of these two possible interpretations, for some speakers, [NI] with overtly negative subjects can only be NC” (emphasis is mine). With (51), below [Blanchette and Collins’ (7)], then, only the negative concord reading is believed to hold – again, for some speakers. (51) Didn’t nobody watch the game. Negative Concord: Nobody watched the game. Double Negation: “*”It is not the case that nobody watched the game. Much like we saw above with (48), however, a bit of context and intonational supplementation can license the reading that Blanchette and Collins argue does not exist. In (52), for example, the “double negation” reading is clearly possible. The capitalization on “NObody” represents a falling intonation on that syllable. (52) Speaker A:

Man, so there wudn’t anybody at the party? Johnny? Marvin? Nobody? Speaker B: Well, I mean, didn’t NObody come. Johnny an’ nem were there, and Little Mikey, if you wanna count them.

48

Chapter 2 Negative inversion and its contents

‘It’s not the case that nobody came. Johnny and them were there, and Little Mikey, if you want to count them.’ Blanchette and Collins in fact do leave an opening for such an interpretation in their Fn. 3, which they attach to example (51) above, and which is quoted here in full: One of our consultants reports that the DN interpretation may be possible under certain circumstances, and if possible it is highly marked. It may be that when [51] has a double negation reading it is a case of contrastive negation (see McCawley 1991), which is subject to distinct syntactic and pragmatic constraints. We set this aside as a matter for future work.

As with (48a–b) above, if the restriction in (51) is assumed to be a syntactic or semantic one, as Blanchette and Collins argue, we would not expect to find any exceptions at all, even if they are marked or less common: i.e. we would not expect to find examples such as (52). What is more likely occurring here is that the apparent issue of ambiguity and scope is not a syntactic one; rather, it has to do with finding the proper pragmatic licensing contexts. This is beyond the scope of the present chapter; what I wish to do here though is merely update the empirical record. The sentence reading that is marked as nonexistent in (51) actually does exist. A last quick point on the ambiguity resolution claim. Foreman (2015) provides a related example of NI as unambiguous. He first notes that the noninverted (53) [Foreman’s ex. 142] allows two readings: i.e. ‘many boys did like to hunt’ and ‘many boys did not like to hunt.’ (53) Many boys in the town didn’t like to hunt. As such, he argues there is no contradiction in a sentence like (53) [Foreman’s ex. 143]. (54) Many boys in town didn’t like to hunt, but many did. However, he goes on to claim that in a Negative Inversion such as (55), there is only one meaning: i.e. that ‘many boys did not like to hunt,’ without the possibility that also ‘many boys did like to hunt’ [Foreman’s ex. 144]. (55) Didn’t many boys in town like to hunt. ‘Many boys in town didn’t like to hunt.’

2 A new empirical account of negative inversion

49

Accordingly, he rules out (56) as ungrammatical due to contradiction. (56) “*”Didn’t many boys in town like to hunt, but many did. ‘Many boys in town didn’t like to hunt, but many did.’ Pace Foreman, this sentence is perfectly acceptable to my participants, however, and has the same meaning possibilities as that in (54). As such, I do not believe that it works as an argument in favor of NI as necessarily unambiguous. 2.1.3 The negative inversion main verb and copula The main verb phrase of NI consists of essentially any verb plus its arguments and modifiers. Thus, there is no need to exemplify this aspect of NI further here. White-Sustaita (2010: 432), citing Green (2006), notes that in AAVE, NI constructions can appear with habitual be as copula, as in (57). (57) Don’t nobody be drinking tea. ‘It is usually the case that not a single person is drinking tea.’ This form was not accepted by my Anglo or Chicano participants. It was accepted by some, though not all, of my AAVE participants. 2.1.4 Intonation and accent in negative inversion NI bears the falling intonation of a declarative sentence. Thus, even though the surface form of the sentence appears in print to resemble that of a yes/no question, in actual spoken use, the intonation clearly sets it off as a statement. See many attestations of this in the literature cited below in Chapter 6. The primary stress typically falls within the subject NP, but it can be shifted elsewhere in the sentence for pragmatic or information structural reasons, as seen in the accent variation in (58). (58) a. b. c. d. e.

DIDn’t a single man like him. Didn’t a SINGle man like him. Didn’t a single MAN like him. Didn’t a single man LIKE him. Didn’t a single man like HIM.

To the extent that a default accent is possible, it seems that (58b) is most natural, though stress does not occur in that position necessarily. This contrasts

50

Chapter 2 Negative inversion and its contents

with the situation reported for AAVE by Parrott (2000: 416), in which “[Negative Inversion] adds strong stress to the negation.”50 2.1.5 Negative concord and negative inversion NI sentences are acceptable in the relevant dialects with and without negative concord, involving a wide range of sentence negation and n-words, such as nothing and nobody.51 Thus, (59a–d) are all acceptable, and all have the same essential meaning. (59) a. Can’t anybody beat a Texan. ‘Nobody can beat a Texan.’ b. Can’t anybody beat no Texan. ‘Nobody can beat a Texan.’ c. Can’t nobody beat a Texan. ‘Nobody can beat a Texan.’ d. Can’t nobody beat no Texan. ‘Nobody can beat a Texan.’ It has been claimed that NI among AAVE speakers must exhibit negative concord, as in Labov (1968, 1972a), but that is not the case with the speakers of the present study, regardless of race or ethnicity: even the AAVE speakers surveyed in Texas did not require NI to exhibit negative concord. This could be a result of regional variation between the populations surveyed, or it could be a result of generational change. Labov’s study is almost five decades old at this point. It is possible that a requirement of negative concord existing in 1972 has weakened since that time.52 In addition, it is straightforward to find examples on the Internet of NI in AAVE that do not exhibit negative concord. For example, here is a line from C. Winn’s novel Don’t Stop Keep Goin’ (Fam First, 2011): (60) The weed game was gravy in the Yard. A lot of niggas had it but couldn’t anybody keep it. ‘[. . .] nobody could keep it.’ 50 Parrott does not identify the AAVE community for which this assumption is made. 51 Labov et al. (1968: 285) write that NI is “well known in colloquial Southern speech without negative concord: Didn’t anybody see it.” Negative concord is not required in the Alabama English NI reported in Feagin (1979) nor the West Texas English NI reported in Foreman (1999, 2015) or Matyiku (2017). 52 Negative concord is still required for the AAVE constructions in Green (2014); though, as mentioned above, it is not clear with which AA population Green is concerned.

2 A new empirical account of negative inversion

51

Here is another example, from New Yorker and heavyweight boxing champion, Mike Tyson, who is berating a reporter at a 2002 press conference for his fight with Lennox Lewis. (61) [. . .] can’t anybody in here fuck with this, this is the ultimate. ‘Nobody in here can fuck with this, this is the ultimate.’ Similar non-concord examples in AAVE contexts are easy to multiply on Twitter. In (62), for example, we see a non-concord NI from 2013 from a female speaker in a context that includes an instance of zero copula, which is a wellknown AAVE feature (i.e. My momma Ø hella mad).’ (62) R.@determinedAM_I. 28 Jul 2013; My momma hella mad cuz won’t anybody in the house speak to her ‘My momma is very mad because nobody in the house will speak to her.’ In (63), we see a female AAVE speaker using multiple instances of NI, some of which exhibit negative concord, and one of which (i.e. Can’t anybody say . . . ) does not, indicating that this speaker has both (non)concord possibilities at her disposal. (63) bri marie. @_ThaQueenB. 24 Jan 2012; Cant no nigga say he did this & that with me. Cant no female call me a hoe. & cant anybody say im any nigga’s sideline. #SelfRespect ‘No guy can say he did this and that with me. No girl can call me a hoe. And nobody can say I’m any guy’s secondary relationship.’ It is not clear, then, to what extent Labov’s characterization of AAVE NI requiring negative concord actually holds in general. Importantly, however, it doesn’t hold at all for the Texas speakers tested for the present work.

2.1.6 Is negative inversion emphatic? Beginning with Labov et al. (1968), NI was argued to be emphatic compared to its non-inverted counterpart. Thus, a sentence like (64a) would be argued to be more “emphatic, excited, and strongly affective” (Labov et al. 1968: 285) than its non-inverted counterpart in (64b).

52

Chapter 2 Negative inversion and its contents

(64) a. Wouldn’t many people do that. ‘Many people wouldn’t do that.’ b. Many people wouldn’t do that. This claim, which is investigated in detail in Chapter 5, has been understood in various ways over the years. For example, with respect to AAVE Negative Inversion, Labov et al. would understand (64a–b) to be truth conditionally equivalent, with (64a) simply bearing a further unspecified component of emphatic meaning. Parrott (2000), also with respect to AAVE, would also assume truth conditional equivalence for (64a–b), and he suggests that the emphatic quality is pragmatic, though he does not show how the pragmatics would work. In the accounts of Green (2014) and Green and Sistrunk (2015), also writing on AAVE, it appears that (64a–b) are not truth conditionally equivalent; for these authors, the Negative Inversion makes a statement that is semantically wider than its non-inverted counterpart. With respect to non-AAVE NI, there has been comparatively less written about the emphatic character. Many authors are simply not concerned with whether NI is emphatic or conveys emphasis at all. As I discuss in Chapter 5, authors such as Foreman (1999), Zanuttini and Bernstein (2014), and Matyiku (2017), writing on Texas and Appalachian NI, are primarily concerned with the internal generative syntax of the construction. The emphatic quality, to the extent that it is considered, is assumed not to trigger subject-auxiliary inversion and so ceases to be of further interest in those accounts. The emphatic question does not arise at all in Blanchette (2015) or Blanchette and Collins (2018), who are also writing on Appalachian English. I will argue in Chapter 5 that the Texas NI of the present work is emphatic, but that emphatic here should be taken to mean speaker involvement, rather than emphasis. To the extent that NI is used to convey emphasis, I will argue that this is done via Gricean reasoning. 2.1.7 Negative inversion allows negative polarity items in embedded clauses Martin (1992) shows negative polarity items anymore and a damn thing licensed in downstream embedded clauses in NI sentences. Consider (65) and (66) [Martin’s ex. 29b–c]. (65) [Don’t nobody say [that dealers sellin’ drugs in the school yard {no more / anymore}.]] ‘Nobody says that dealers are selling drugs in the school yard anymore.’

2 A new empirical account of negative inversion

53

(66) [Can’t a man in this place say [he happy ‘bout {nothin’ / a damn thing}.]] ‘No man in this place can say he is happy about anything.’ These examples are worth noting because Martin takes them as an argument against a syntactic movement account for Negative Inversion. That is, in Martin’s account, upward movement of the negative auxiliary would result in intervention effects that should block NPI licensing. Since NPIs are licit in these sentences, however, Martin concludes that the negative auxiliary must not move after all: i.e. NI sentences are created in some other way, such as expletive deletion. I will return to discussion of Martin’s ideas in Chapter 3.53 2.1.8 Embedding of entire negative inversion sentences It has been pointed out numerous times in the literature for various reasons that NI can serve as complement clauses in embedding environments.54 Foreman (1999) and others have, for example, argued that this fact distinguishes NI from yes/no questions in the generative frameworks they employ, as these other SAI constructions were formerly believed to be unembeddable.55 This embedding behavior is worth mentioning here; though, it ultimately plays little role in the account given in this book. In (67a–b), we see examples of NI embedded both with an overt that complementizer (67a) and without (67b). Example (67a) is spoken by an Anglo farmer in Alabama, while (67b) belongs to a middle-aged AAVE speaker in Cleveland. Both are acceptable to the Texas participants. (67) a. γ “I didn’t tell you that you could have it; and he said, well, you told me that I could have parts off of it and you said that couldn’t nobody get it out and you said if I wanted something off of it that I could have it.” ‘You said that nobody could get it out [. . .].’

53 Martin, writing in the early 1990s, assumes a syntactic approach to NPI licensing, which is not a universal assumption. More recently, there are a range of non-syntactic approaches to NPI licensing on the market as well. See Israel (2011) for discussion of syntactic, semantic, pragmatic, and rhetorical approaches. On a non-syntactic approach to NPI licensing, Martin’s point regarding NPIs appearing in downstream clauses in NI would be less relevant. 54 In other words, NI is not what G. Green (1976) would refer to as a Main Clause Phenomenon. 55 Newmeyer (1998: 46–49), for example, argues that SAI constructions in general (such as those in Chapter 6 below) cannot be embedded in larger sentences. Goldberg (2006: 180–181), however, shows that Newmeyer’s claim is too strong and that SAI can be embedded in some environments. See also the references in Goldberg (2006) on embedding SAI constructions. This potentially raises problems for Foreman’s (1999) account of NI, but this is beyond the scope of the present work.

54

Chapter 2 Negative inversion and its contents

b. γ “We had got into it with a group of guys, we’re young, and the whole bar came out,” says Knuckles. “They were a group of Caucasian people, and they came out and said, ‘Excuse me, you derogatory-word-forblack-people, go back to where you came from.’ We got to fighting – there were seven of us and about 30 of them – and I hit this guy in the head with a bat. The fight ended, but days later, the police come to my door to arrest me. [. . .] At that moment, I said couldn’t no one have delivered me from that but the Lord.” ‘Nobody could have delivered me from that but the Lord.’ This differs somewhat from what has been reported in past studies. For example, In Sells et al. (1996), AAVE participants from East Palo Alto, California, rated as ungrammatical embedded NI in which the complementizer that is present. Thus, sentences such as (68) were ruled out in that study. [Sells et al. ex. 50]. Interestingly, Labov et al.’s (1968) participants marked this same sentence as acceptable, if unusual. (68) “*”I know a way that can’t nobody start a fight. ‘I know a way that nobody can start a fight.’ According to Sells et al., the East Palo Alto participants preferred sentences such as (69), in which the complementizer is absent [Sells et al. ex. 51b]. (69) It’s a reason didn’t nobody help him.56 ‘There’s a reason that nobody helped him.’ Regardless of these earlier studies, in the present study, NI with and without the various complementizers were accepted by the participants of all three Texas vernaculars under consideration. We also see imbedding of NI under if, in (70), which is taken from Matyiku (2017: 60), and which is acceptable to the Texas participants. (70) I don’t particularly care if didn’t anybody like your idea. It ain’t my concern. ‘I don’t particularly care if nobody liked your idea.’

56 Cf. That way you have something can’t nobody take away from you, spoken by Troy Maxson, in August Wilson’s (1986) drama Fences.

2 A new empirical account of negative inversion

55

Relatedly, it is also possible for NI to appear in the protasis (71a) and apodosis (71b) of conditional sentences. These examples are attested online and are also easily acceptable by my test consultants. (71) a. γ If don’t nobody else have my back, I Know God does. ‘If nobody else has my back, I know God does.’ b. γ If I can’t play then can’t nobody play. . . Lights out! ‘If I can’t play then nobody can play. . .Lights out.’

2.1.9 Negative inversion and not-initial sentences Another empirical claim made with respect to NI, which was initially made in Foreman (1999) and carried through Matyiku (2017) and Blanchette and Collins (2018), is that the class of subjects allowed in NI is the same class allowed in sentences with initial not, as in (72a–b) and (73a–b) [Foreman’s ex. 29/32; Blanchette and Collins’ ex. 5/6]. (72) a. Didn’t many people go to the party. ‘Many people didn’t go the the party.’ b. Not many people went to the party. (73) a. *Ain’t Jack seen the baby yet. ‘Jack hasn’t seen the baby yet.’ b. *Not Jack has seen the baby yet. ‘Jack hasn’t seen the baby yet.’ While this parallelism does seem to hold at first blush, it can be shown not to hold if additional context is added. As I discuss above, and in much more detail below in Chapter 3, definite subjects, proper names, and more are allowed in NI if they are construable as hearer-new, following a principle set out in Ward and Birner (1995) with respect to felicitous subjects of there-existential sentences. This is not possible with the not-initial sentences though. So, consider now (74), which is hearer-new but also uniquely identifiable, via the modifying relative clause. (74) Ain’t the Jack that I know been to see the baby yet. ‘The Jack that I know has not been to see the baby yet.’ The same strategy, however, does not rescue the initial-not sentence, as in (75), which is still ungrammatical even with the modifying relative.

56

Chapter 2 Negative inversion and its contents

(75) *Not the Jack that I know has been to see the baby yet. ‘The Jack that I know has not been to see the baby yet.’ Blanchette and Collins (2018) point out a different asymmetry in the putative NI/not subject comparison. Thus, they note the pair in (76a–b), which do not pattern as would be expected if NI and not-sentences were indeed to share the same subject constraints [Blanchette and Collins’ (Fn. 5)]. (76) a. Didn’t but John show up. ‘Nobody but John showed up.’ b. *Not but John showed up. ‘Nobody but John showed up.’ Finally, Foreman (1999, 2015) notes that the same quantifiers that are ruled out of NI are also ruled out in not-initial sentences. One of his examples involves some NP, as in (77): (77) a. *Ain’t some student finished the exam yet. ‘Some student hasn’t finished the exam yet.’ b. *Not some student finished the exam. ‘Some student hasn’t finished the exam yet.’ As we saw above, though, NI does allow some NP in subject position. However, Foreman is correct that some NP is ruled out of subject position in the notinitial sentences. The same is true for the other quantifiers listed in Foreman (2015: 144). In addition to ruling out some NP from NI and not-initials, Foreman also rules out several, each, most, and a few from both sentence types. As shown in Chapter 3, each of these latter quantifiers patterns the same way as some: i.e. they are allowed in NI in richer contexts, but they are not allowed in the not-sentences. As such, the subject comparison between NI and not-initial sentences cannot be maintained. Although there does seem to be significant overlap in allowable subjects of NI and not-initial sentences, they are clearly not the same. 2.1.10 Negative inversion and its non-inverted counterpart(s) White-Sustaita (2010: 430) writes that “in varieties where we find NI, we also find non-inverted constructions with a negative indefinite subject in the canonical clause-initial position.” What this means is that if a language variety

2 A new empirical account of negative inversion

57

contains NI such as (78a), it should also contain the inverted counterpart in (78b).57 (78) a. Don’t nothing come to a sleeper but a dream. ‘Nothing comes to a sleeper but a dream.’ b. Nothing don’t come to a sleeper but a dream. ‘Nothing comes to a sleeper but a dream.’ White-Sustaita indicates that this co-distribution is found in the AAVE described in Green (2002) as well as the Alabama English of Feagin (1979). Things are more complicated in the Texas English of Foreman (1999), Matyiku (2017), and the present work, though, in which negative concord is dispreferred in the non-inverted forms like (78b), rendering (78b) actually rather unacceptable for Foreman’s speakers, Matyiku’s, and those of the present work. Here is Foreman (1999) on the matter: It is not true that all of the non-inverted forms of NI sentences are acceptable in WTE. Generally, if only the subject and the Aux bear negative morphology then speakers prefer the inverted word order. [. . .] If further material bearing negative morphology is permitted in the clause, then its presence after the Aux can license the pre-verbal negative subject, making the non-inverted word order acceptable. [. . .]. Compared with examples from Labov et al. (1968) and based on the discussion in Sells et al. (1996), this restriction on non-inverted sentences does not hold in AAVE.

Matyiku (2017: 136) finds similar results in her study of Texas NI, writing: the non-inverted counterpart of a negative auxiliary inversion construction exhibiting negative concord is judged as ill-formed or marginally acceptable [. . .]. Foreman points out that in West Texas English, [79a] is either completely ill-formed or only very marginally acceptable while its inverted counterpart in [79b] is strongly preferred instead.

(79) a. ?* Nobody don’t like you.58 ‘Nobody likes you.’ b. Don’t nobody like you. ‘Nobody likes you.’ c. Nobody likes you. Foreman’s and Matyiku’s findings with respect to the non-inverted counterparts here are very much in line with those of the present work. The result is that in

57 Sentences (78a–b) are taken from Green 2002 (2c) and (7c), respectively. 58 The judgement on (2.57a) is found in Foreman (2000: ex. 53), and reported in Matyiku.

58

Chapter 2 Negative inversion and its contents

the present work, many perfectly acceptable NIs do not have a direct, morphemefor-morpheme, non-inverted counterpart. The preferred non-inverted counterpart of (79b) for the present speakers is not (79a); rather, it is (79c). Thus, it is not (79a) that competes with the inverted (79b); it is (79c) that does so. Relatedly, the NI of the present work sometimes also has no direct, morpheme-for-morpheme, non-inverted counterpart when the subject is positive. As such, there is no non-inverted form (80b) to compete with (80a), nor does (81b) compete with (81a). The competing forms are found in (80c) and (81c), respectively. (80) a. Don’t anybody like him. b. *Anybody don’t like him. c. Nobody likes him. (81) a. Don’t a damn thing phase him. b. *A damn thing don’t phase him. c. Nothing phases him. It is possible that the unacceptability of (80b) and (81b) stems from the negative polarity status of anybody and a damn thing. In the inverted examples, these NPI terms are easily in the scope of the negation, and can be licensed in that way.59 If a positive subject such as many is used, which is not a negative polarity item, then NI and its non-inverted counterpart can be an exact match, modulo the inversion, as in (82a) and (82b). (82) a. Don’t many people like him. ‘Many people don’t like him.’ b. Many people don’t like him. In addition to NPI subjects, we also find free choice (FC) any in subject position, especially with what Horn (2000) refers to as an anti-indiscriminative reading, when reinforced by just.60 This fact is also noted in Foreman (1999) and

59 Ladusaw’s (1979) early treatment of NPI licensing argued that NPIs were licensed if they appeared in the scope of negation. Over the intervening decades, however, this scope argument has been recognized to be insufficient for capturing the licensing requirements on all NPIs. See Israel (2011) for discussion of various other means of NPI licensing that have been proposed over the years. 60 See Israel (2011) for extensive discussion, references, and intellectual history of NPI and FC any.

2 A new empirical account of negative inversion

59

Matyiku (2017). Consider (83), which is a lyric from Texan James McMurtry’s song entitled “The Governor” (2008).61 (83) Them cigarette boats, they cost money Can’t just anybody buy a boat like that ‘Not just anybody can buy a boat like that.’ This reading is easiest to obtain with just preceding any, as in (83), though it is also possible with just following the subject, as in the AAVE example in (84) from Twitter. Note the possessive determiner, they ‘their,’ which is a well-known AAVE feature. (84) Nicki ❄□ @__crysten 8 Sep 2018; Okay so like I really want my hair done. . . but can’t anybody just have they hands in my head. ‘Not just anybody can have their hands on my head.’ Similarly, as Matyiku (2017: 236) notes, with the right intonation or the right context, the anti-indiscriminative reading can be had without just at all. It is worth pointing out here that the anti-indiscriminative NI examples such as (83–84) have clear non-inverted counterparts. Thus, the NI in (85a) has the direct counterpart of (85b). (85) a. Can’t just anybody buy a boat like that. ‘Not just anybody can buy a boat like that.’ b. Anybody can’t just buy a boat like that. Similar sentences are easily attested on Google as well. For example, here is Texas fiddler Wilmer “Cowboy” Little in (85a), speaking in 2009. Like (85b), sentence (86a) has a perfectly acceptable NI counterpart in (86b). (86) a. γ “Anybody can’t just get up there and keep a crowd alive all day long,” he says. “If you let a crowd slump off, first thing you know they’re going to be half-asleep.” ‘Not just anybody can get up there . . .’ b. Can’t just anybody get up there . . .

61 Thanks to Laurence Horn (p.c.) for pointing out this example to me.

60

Chapter 2 Negative inversion and its contents

As we have seen above, though, with NPI any in subject position in (80), there is no direct non-inverted counterpart for Negative Inversion. 2.1.11 Negative inversion, truth conditions, and camouflage It has been suggested for AAVE by Green (2014) that NI and its non-inverted counterparts can appear in different semantic contexts, which would mean that sentences such as (87) and (88) [adapted from Green] would have different truth conditions in Green’s account. (87) Don’t nobody ride bus #201. ‘Absolutely nobody rides bus #201.’ (88) Nobody rides bus #201. For Green, although (87) and (88) would overlap in the contexts in which they can be used, the contextual sets are not equal. Here is Green on the matter: [T]here is some overlap, such that the two can be used in absolute negative contexts; however, the non-inverted constructions can also be used in contexts that refer to a very small number on a scale. (129)

This is not the case in any of the vernaculars of the present book. As a native speaker of one of the relevant varieties, and a linguist with training in semantics and pragmatics, it is difficult for me to imagine a situation in which (87) would hold, but (88) would not. It is worth mentioning here that Green does not identify the population of AAVE speakers within which this description holds. To complicate matters further, Parrott (2000), also writing on an unidentified AAVE population, claims NI is in fact truth conditionally equal to its non-inverted counterpart. What are we to make, then, of Green’s description of the absolute negation as compared to the descriptions of the Texas vernaculars, which clearly do not have this meaning? Salmon (2018b) suggests that this difference in NI might be an instance of linguistic camouflage. This sense of camouflage dates at least to Spears (1982), who credits influence (though not use of the term camouflage) to Bickerton (1975). A clear and recent definition of the concept is given in Wolfram (2004: 114): One type of sociolinguistic process associated with urban AAVE is linguistic camouflaging, in which a vernacular form resembles a standard or different vernacular form so closely that it is simply assumed to be identical to its apparent structural counterpart. However, this similarity may disguise the fact that the form carries a distinctive semanticpragmatic meaning or is constructed in a subtly different way.

2 A new empirical account of negative inversion

61

A classic case of camouflage in AAVE is distributive be, as in I be sleeping when they call, which means ‘I am usually sleeping when they call.’ Remote-time been is another example, in which Bruce been playing the guitar glosses as ‘Bruce has been playing the guitar for a long time.’62 Indignant-come (Spears 1982) is another case, in which come, in the sentence They come in here talking that trash, expresses speaker indignation as well as its normal verbal properties. Similarly, the adverb steady (Baugh 1984) expresses progressive aspect in sentences like Them fools steady hustlin’ everybody they see, which translates to ‘Them fools are currently hustling everybody they see.’ What is clear with all of these examples is that while they look identical to forms in standard English or other vernacular Englishes, they have rather different and unexpected meanings. Thus, they are camouflaged in a fashion perhaps similar to what seems to be the case with Green’s NI and the NI of the present study. 2.1.12 Negative inversion and modal existential constructions It is frequently mentioned in studies of NI that many of them can take an expletive subject. White-Sustaita (2010: 431), writes for example, “In some dialects that have NI, such as Alabama English, Appalachian English and Smoky Mountain English, we find constructions that are identical to NI save for a clause-initial expletive.” Foreman (2000: 51) also assumes that they are “very similar to any other [NI] sentence.” White-Sustaita provides the following examples, taken from Feagin (1979) [Alabama English] and Montgomery and Hall (2004) [Smoky Mountain English], respectively. (89) There wouldn’ nothin’ go down through there. ‘There wasn’t anything that would go down through there.’ (90) They didn’t none of us ever get snakebit, but their work animal did. ‘There wasn’t any of us that ever got snakebitten, but their animal did.’ Similar sentences are referred to as Modal Existential sentences in Ball (1991) and von Fintel (1992), and as Transitive Expletives in Zanuttini and Bernstein (2014). For all practical purposes, these different labels seem to refer to the same object. See also the Yale Grammatical Diversity Project’s entry on Split Subjects, which appear to overlap to a large extent with the expletive constructions. It is also frequently pointed out that expletive sentences such as (89) and (90) do not occur in contemporary AAVE, but that they did occur in earlier

62 Examples of distributive be and remote-time been taken from Green (1998).

62

Chapter 2 Negative inversion and its contents

AAVE.63 See, for example, Weldon (1994), Sells et al. (1996), Martin and Wolfram (1998), and Green (2014) on the claim that contemporary AAVE does not allow the expletive construction. Pace all of these works, however, it is possible to find such sentences in song lyrics by contemporary African American artists. For example, here is West Coast hip hop artist Brotha Lynch Hung, from Sacramento, from the 1993 song “Back Fade”: I’m the so-called Devil that Brotha Lynch was talking bout you know they can’t nobody fade this this nigga packs his own people. you know what I’m saying. ‘Nobody can question this’

Twitter also offers tokens of this type in AAVE, in which the Tweeter in (91) uses an NI in the first conjoined sentence and an ME in the second conjunct.64 (91) nlf @__Blueflowers Jan 27; I take it that don’t nobody wanna talk to me today since they can’t nobody answer the phone ‘There is nobody that will answer their phone.’ In the present study, expletive sentences of this kind were almost uniformly rejected by the young participants in my surveys of African American, Anglo, and Chicano speech. The sentences also seem to be highly stigmatized among these groups. One survey participant, for example, commented that he hoped he would “never meet anyone that could use those kinds of sentences” and then hummed part of the “Dueling Banjos” tune from the 1972 movie Deliverance.65 As we will see below in Chapter 3, these expletive sentences, which I will refer to as Modal Existentials, were in fact accepted by older Anglo and many older Chicano speakers in the present study, and as White-Sustaita and others have shown, were acceptable in older AAVE varieties as well. This leads me to argue

63 White-Sustaita, for example, notes the following token found in the ex-slave narratives found in Bailey, Maynor & Cukor-Avila (1991). (i) But they’d give me a note so there wouldn’ nobody interfere with me. ‘But they’d give me a note so there wasn’t anyone who would interfere with me.’ 64 Note that (91) is another naturally attested example of NI embedding with overt complementizer in AAVE, as described in Section 2.1.8 above. 65 The movie Deliverance and especially the iconic dueling banjo song has become a wellknown symbol, evoking stereotypes of rural people in general as inbred, scary, and dangerous. The scene in question can be viewed here: .

2 A new empirical account of negative inversion

63

in Chapter 3 that the Modal Existential is actually the diachronic source of Negative Inversion, with the expletive having undergone deletion across generations in the 20th century. 2.1.13 There-deletion existential constructions In many varieties of English, there are structures that appear on the surface quite similar to Negative Inversion but which differ in important ways. Sentences such as (92a–b) differ from NI primarily in the fact that the auxiliary in question is what White-Sustaita (2010: 432) refers to as existential be in the former sentences, while the Negative Inversion of the present work has other modals and non-be auxiliaries in this position. Thus, consider (92a–b), which are White-Sustaita’s example (18) and Foreman’s (1999: Fn. 3), respectively.66 (92) a. Ain’t nothin’ happenin’. . . ‘There is nothing happening.’ b. Ain’t no black Santa Claus. ‘There is no black Santa Claus.’ While forms such as (92a–b) look similar to Negative Inversion, they have been argued to have quite different structures from NI, and they have played a variety of roles in previous NI literature.67 White-Sustaita (2010: 432) sums up the relation between the existential be sentences and NI: These constructions are the equivalent of there-insertion existential constructions; the clause-initial auxiliary is the result of phonological deletion of the expletive or syntactically motivated expletive absence (see also Green 2001; Martin 1992; Labov et al. 1968; Sells, Rickford & Wasow 1996). These constructions differ from NI insofar as they permit expletives across all dialects, and they can occur within positive constructions provided that they have an expletive. Moreover, these types of constructions are available in any number of non-canonical varieties of English that do not allow NI, as well as colloquial fast speech canonical English, which does not allow NI.

66 These examples originate as Labov et al. (1968: ex. 350) and Sells et al. (1996: ex. 36). 67 See Horn (2015, 2018) on what he terms nexistentials. Here, also, is Matyiku (2017: 27): Negative existentials are sometimes confounded with negative auxiliary inversion constructions or receive a parallel analysis in the literature (Martin 1992, 1993; Sells et al. 1996; Parrott 2000). I distinguish between the two types of constructions, in line with most of the literature on negative auxiliary inversion, including Labov et al. (1968), Labov (1972), Foreman (1999), Green (2008, 2011, 2014) and White-Sustaita (2010).

64

Chapter 2 Negative inversion and its contents

White-Sustaita goes on to write that she does not consider the existential be sentences to be NI at all, and so excludes them from her analysis. In this book, I agree with White-Sustaita’s characterization and do not consider the existential be sentences to be NI either, considering them instead to be straightforward instances of conversational deletion or allegro speech, as shown in Chapter 3. In most of the literature it is normally assumed that no inversion takes place in these existential be sentences, and that they are a product of expletive deletion of one sort or another. On the other hand, the true Negative Inversion sentences are typically argued to be derived via some type of syntactic auxiliary movement, in which the auxiliary moves to a position higher than the subject. This will be discussed in much more detail in Chapter 6. Additionally, it has been pointed out over the years that Negative Inversion sentences typically have a well-formed non-inverted counterpart, as in (93a–b), but that existential be sentences do not, as in (94a–b), which are taken from Martin (1992) [cited in Matyiku (2011/2018)]. (93) a. Can’t nobody lift that rock. ‘Nobody can lift that rock.’ b. Nobody can lift that rock. (94) a. Ain’t no trouble to make another trip. ‘It ain’t no trouble to make another trip.’ b. *No trouble ain’t to make another trip. It ain’t no trouble to make another trip.’ Likewise, as Green (2001) points out [cited in Matyiku (2011/2018) and alluded to in the White-Sustaita quote above] the existential be construction can appear in non-negative contexts, as in (95a), which further distinguishes it from Negative Inversion, as in (95b), which is unacceptable in the absence of negation. (95) a. Should be some candy in the dish. ‘There should be some candy in the dish.’ b. *Should nobody eat that candy. ‘Nobody should eat that candy.’ I assume in the present work that the there-deletion existential sentences are basically the product of synchronic conversational deletion, as in Thrasher (1974), Napoli (1982), Weir (2012), etc., in which sentence-initial material can be phonologically deleted in casual speech from a wide variety of sentence types.

2 A new empirical account of negative inversion

65

Thus, (96a) is derived from (96b) in casual speech in much the same way (97a) is derived from (97b). (96) a. Isn’t anybody here. b. There isn’t anybody here. (97) a. You seen Bill? b. Have you seen Bill? Example (97a) is possible in many varieties of English, and it differs from (97b) along dimensions of casualness or rate of speech, rather than being a regional dialectal indicator as is the case with NI.68 These there-be sentences and their relation to NI will be discussed in more detail below in Chapter 3. I argue in this work that the be-existentials are ultimately not the same construction as NI, though it is clear that the two are closely related in their internal grammar and their discourse function. As such, in Chapters 3 and 6, I rely heavily on what is known about these existential sentences to illuminate the grammar and function of NI.

2.2 Conclusion In this chapter I have provided a detailed empirical account of NI in three Texas English ethnolects. Often, this empirical account overlaps with what has been presented elsewhere in the NI literature with respect to other language varieties. I have also pointed to significant empirical differences here in the Texas varieties as well, which sets the NI described herein apart from all other

68 Another clear example of this sort of conversational deletion can be seen in the lyrics of the song “Way Over Yonder in the Minor Key,” written by Woody Guthrie, and put into music in 1998 by Billy Bragg and Wilco. The final two lines below are a recurring couplet in the song. The first line of the couplet is a clear negative existential, which is echoed in the second line without the expletive. She said it’s hard for me to see How one little boy got so ugly Yes my little girly that might be But there ain’t nobody that can sing like me Ain’t nobody that can sing like me The song appears as Track 3 on the album Mermaid Avenue (Elektra Records), by Billy Bragg and Wilco. The stanza here was transcribed by me.

66

Chapter 2 Negative inversion and its contents

accounts of NI given elsewhere. Whether this is ultimately due to microdialectal variation – i.e. there really is variation across AAVE, Southern, and Texas NI – or is due to the nature of the empirical observation made here in general, I cannot say. I suspect it is the latter and that a closer look at NI in those other varieties will show that it is closer to the Texas NI than appears to be the case here. That is not a task to be taken up in this book, though. As has been mentioned only briefly in this chapter but which will be taken up in more detail below, the aspects of the empirical account given here cause significant problems for almost all other syntactic accounts of NI on the market. For example, the fact that weak quantifiers like some and several are shown to be acceptable raises problems for scope-based accounts like that of Foreman (1999, 2015), Zanuttini and Bernstein (2014), Matyiku (2017), and Blanchette and Collins (2018) as shown in more detail in Chapter 3. Likewise, the acceptance of definite subject NPs in the present book raises problems for these same accounts, as their syntactic apparatuses depend on the presence of definite subject NPs to predict where various NI constituents reside or land in the generative syntax. Another problem for these same accounts is also given. It has long been assumed that in NI with universal quantifiers in the subject NP, the universal quantifier must take scope over the negation. This is also shown not to be true in this chapter, and it is discussed in more detail in Chapter 6. Finally, the observation that NI conveys emphasis only pragmatically, and not conventionally, is problematic for accounts like that of Green (2014), which provides an extended-CP syntax specifically to account for negative focus and so conventional emphasis, as discussed in Chapter 5. None of these concerns, however, are problematic for the existential account of NI developed in the present book. To conclude, the seed for much of the rest of this book is planted in this chapter, with enough detail provided here to give the reader a whiff of what is to come. This chapter is also meant to serve as an empirical reference guide for future work on any other aspects of NI grammar and usage.

Chapter 3 Negative inversion, existential sentences, and definite subjects One of them bunkers, they used it for a long time to store it full of whiskey, and they couldn’t nobody steal it. – Dave Conder, Killeen, Texas69

3 Introduction As we have seen in previous chapters, with the exception of Salmon (2017, 2018a), it is universally assumed in previous literature that Negative Inversion constructions prohibit the appearance of definite subjects. Matyiku (2011/2018), writing for the Yale Grammatical Diversity Project (YGDP), sums this up succinctly: Negative inversion is a phenomenon in which a declarative sentence begins with a negated auxiliary or modal, such as can’t, ain’t, or won’t, followed by a quantificational (or indefinite) subject, such as nobody.

As I show in the present chapter, however, this description of NI does not hold for the language varieties of the present work.70 Definite subjects do appear in the construction; yet, they are subject to pragmatic constraints. Importantly, these constraints are shown to be the same ones that restrict the appearance of definite subjects in there-existential sentences, which provides a strong argument for a connection between NI and existential sentences. Building upon this connection to existentials, I then argue that the NI actually is a type of modalized TE sentence that has lost its expletive subject over time. As such, what I refer to in this book as a Modal Existential sentence is really the same sentence as a Negative Inversion, except that in the former, the expletive subject is still intact. In the sections that follow, I first illustrate the behavior of definiteness effects in Negative Inversion constructions and show that they are governed by the same pragmatic constraints governing definite subjects in TEs. In the

69 Quoted in Dase and Katauskas (2011). 70 An earlier version of this chapter, cited here as Salmon (2018a), appears in Journal of Pragmatics. That article is concerned only with AE and AAVE data and reaches a narrower set of conclusions. The JoP article was preceded by Salmon (2017), which is an even shorter paper that was published in Proceedings of the Berkeley Linguistics Society. https://doi.org/10.1515/9781501512346-004

68

Chapter 3 Negative inversion, existential sentences, and definite subjects

following section, relying on synchronic generational differences, I argue that the Modal Existential is the constructional precursor – i.e. syntactic etymon – to the Negative Inversion. Finally, I consider the relation of Negative Inversion and Modal Existentials among older speakers, who accept both as grammatical. I interpret this latter fact as simply the interplay of synchronic conversational deletion and Jespersen’s (1917) need to place negation as near as possible to the beginning of a sentence.

3.1 Definiteness effects and negative inversion Since Milsark (1974), there has been a general assumption in the syntax and semantics literature that an assortment of NP types, including those with definite marking, are prohibited in the post-expletive subject position of TE sentences.71 According to Milsark, the following definite NPs are excluded [Milsark’s (64)]: (98) *There is {the dog, John’s dog, that dog, John, him/he} in the room. Over the last four decades, this data has been treated to an immense range of syntactic and semantic approaches.72 Since at least Martin (1992), it has been believed that NI patterns very closely to Milsark’s existential sentences. 71 This NP has gone by various names in the literature, including pivot, focus NP, and anchor subject. It should be noted that many researchers no longer accept the definite restrictions as originally arranged in Milsark (1974). For example, Ward and Birner (1995) consider Milsark’s definite restrictions as being ultimately “epiphenomenal,” while McNally (2011) writes: In fact, as noted in section 1.2, there is good evidence that true definite noun phrases are indeed acceptable (see, in addition to the references cited elsewhere in this article, Ziv 1982 and Lumsden 1988). But let us maintain for a moment the view that they are not [. . .]” (1841). Relatedly, based on quantitative corpus work, Beaver, Francez, and Levinson (2005) argue that definiteness effects are not categorical, as is frequently assumed in the semantics and syntax literature, but is instead a gradient phenomenon, with some NPs more likely to occur in subject position than other NPs. These authors note also that the gradient settings differ across languages: i.e. an NP type that occurs as subject frequently in one language may occur less frequently in another language. Lambrecht (1994: 179) notes too that definite descriptions and proper names are possible in French existential sentences. Unfortunately, however, many researchers continue to assume Milsark’s definiteness restriction for there-existentials tout court, to the point that it has reached a kind of mantra status in the literature. 72 See Francez (2007) and McNally (2011) for detailed discussions of previous literature on the syntactic, semantic, and pragmatic issues surrounding existential sentences.

3 Introduction

69

Here again, in stronger language than the quote above, is the Yale Grammatical Diversity Project in 2018: Definite subjects such as pronouns, proper names, and DPs headed by definite or possessive elements are not possible, so none of the following are allowed: – Don’t the police break up a fight. – Won’t they catch us. – Wouldn’t Sally and Jean help the poor man.

The examples cited here from the YGDP are African American Vernacular English, but the YGDP intends them to exemplify the grammaticality pattern of Southern Englishes in general. To be more specific to the non-AAVE varieties in question in the present work, we can turn to Martin (1992: 66) on non-AAVE Southern Englishes, who writes that “the Definiteness Condition ([. . .] which prohibits the use of definite NPs after pleonastic subject constructions) operates in the case of negative inversion constructions.” Similarly, Foreman (1999: 10), with respect to West Texas English, writes that “the restriction against Definite subjects in NI sentences still needs to be explained.” In discussion of Appalachian English NI, Blanchette and Collins (2018: 10) write “proper names, definite descriptions, pronouns and demonstrative phrases, all of which have been observed in the previous literature to be unacceptable as [NI] subjects [. . .] are ungrammatical in [NI] subject position.” This grammaticality pattern has been assumed by every researcher up to and including the most recent work on the subject, which appears to be Matyiku (2017), and Blanchette and Collins (2018).73 Salmon (2018a) and the present work are the first to argue that definite subjects are not obligatory in this construction and that there are systematic, pragmatic reasons for why this is the case. A quick scan of previous literature reveals that almost without exception, example sentences of NI have been given in isolation, with very little supporting context and no indication of intonation, stress, and the like. Just simple

73 Foreman (1999), Matyiku (2017), and Blanchette and Collins (2018) are syntactic accounts focusing directly on the status of subjects in NIs. These accounts attempt to explain the putative prohibition against definite subjects as a result of various scopal interactions posited between the subjects and the negation. For example, Matyiku (1) writes: The types of subjects that are possible in [NIs] behave uniformly in their scopal interaction with negation. This discovery furthers our understanding of why subject restrictions arise. The details of the aforementioned syntactic, scopal accounts need not concern us here, however, as the more fundamental claim is being advanced that NI actually does allow definite subjects. As such, NI is actually a different empirical object from what the syntactic accounts mentioned above sought to describe.

70

Chapter 3 Negative inversion, existential sentences, and definite subjects

sentences, bearing an ungrammatical marking, much like those in the YGDP quote above. As is argued in the present work, though, the constraints at work are pragmatic, meaning that their application can be supported or defeated in different contexts. Given this, it is unsurprising that the acceptability of definite subjects in the construction would be missed over the years. The only exception of which I am aware, in which it is suggested that the prohibition against definite subjects might be more complicated than it seems, is found in Green (2014) with respect to AAVE. In the context of a larger discussion, Green provides one example, given in (99) below, and suggests that the subject requirements of NI in AAVE are more complicated than we have been led to believe over the years. As this is not Green’s focus in the article, however, the subject is then left with little further comment. (99) Speaker A: Many old fraternity guys showed up for homecoming. I think even Vince Jackson was there. Speaker B: No, didn’t no Vince Jackson show up! Green (2014: 131) describes the example like this: Although B’s sentence can also have a nonreferential reading of no Vince Jackson, such that no one showed up by that name (reading 1), the [NI] construction can indeed refer to the nonappearance of a specific old fraternity guy Vince Jackson (reading 2). There are obviously different flavors of [NI], and I do not want to confound the issues here, where the focus is on negative concord (i.e., NegAux. . .Neg Quantifier) or what I have been referring to as canonical [NI]. However, it is worth noting that the subject position in [NI] is not limited to negative quantifiers or non-referential DPs.

To my knowledge this is the only other time that the definite subject prohibition has been breached. Green is a native speaker of AAVE, and it is noteworthy that the example provided contains sufficient context to license what would likely otherwise have received an ungrammatical rating in the literature. In the presentation of examples to come, I follow the work of Ward and Birner (1995) on there-existential sentences.74 In this work, Ward and Birner show that contrary to most of the literature on TE sentences, definite subject

74 The insights of Ward and Birner (1995) are also included in the discussion of thereexistentials in Huddleston and Pullum’s (2002) The Cambridge Grammar of the English Language. Chapter 16 of that work, which is co-authored by Ward, Birner, and Huddleston, takes the same line on the acceptability of definite subjects in there-existentials. The present work follows the 1995 paper, though, as the arguments are supported in much more detail there.

3 Introduction

71

NPs are in fact allowed. They argue further that the definiteness effects for TEs claimed in Milsark (1974) is actually not a real effect, and is instead epiphenomenal. Thus, in Ward and Birner’s account, subjects – or post-verbal NPs – in existential sentences are allowed to be grammatically definite as long as they are pragmatically construable as being hearer-new.75 I argue that the definite subjects allowed in NI are constrained by precisely this same pragmatic principle. Here are Ward and Birner (1995: 728) on the pragmatic and cognitive constraints on definites in existential sentences: [For] an entity to be hearer-old entails that the speaker believe or assume it to be present within the hearer’s knowledge store, while for an entity to be hearer-new entails that it not be assumed to be present within the hearer’s knowledge store. As we will show, however, it is possible for an entity to be treated by the speaker as being hearer-old in one respect yet hearer-new in another. [. . .] We will discuss a number of circumstances in which an entity may be both hearer-new and uniquely identifiable, supporting our claim that definiteness and postverbal position in there-sentences are subject to distinct discourse constraints. [. . .] Specifically, we have identified five distinct cases in which formally definite yet hearer-new NPs may felicitously occur in there-sentences. In each case, the definiteness of the NP was found to be licensed by the unique identifiability of the referent, while the there-construction is licensed by the hearer-new status of the referent.

The five distinct cases of definite but hearer-new NPs mentioned above are the following: (i) Hearer-old entities treated as hearer-new (ii) Hearer-new tokens of hearer-old types (iii) Hearer-old entities newly instantiating a variable (iv) Hearer-new entities with uniquely identifying descriptions (v) False definites As with Ward and Birner’s definite subjects in TEs, the definite subjects found in NI can also be shown to fit naturally into each of these classes. Let’s consider them one at a time, with first an existential example from Ward and Birner followed by an NI.76 Group (i), hearer-old entities treated as hearer-new, includes what are commonly referred to as “reminder there-sentences,” with their example given below in (100) [Ward and Birner’s (14a)]. The idea here is that the entity

75 Ward and Birner attribute the concept hearer-new in general to Prince (1992). 76 Ward and Birner’s data is taken from a corpus of 1.3 million words of “transcribed oral data drawn from transcripts of The Presidential Commission on the Space Shuttle Challenger Accident” (1995: 723).

72

Chapter 3 Negative inversion, existential sentences, and definite subjects

in question has been previously evoked, but “there are sufficient grounds for the speaker to believe that the entity has been momentarily forgotten” (Ward and Birner 1995: 730). (100) Mr. Rummel: Well, didn’t the designer of the orbiter, the manufacturer, develop maintenance requirements and documentation as part of the design obligation? Mr. Collins: Yes, sir. And that is what we showed in the very first part, before the Pan Am study. There were those other orbiter maintenance and requirement specifications, which not only did processing of the vehicle, but in flow testing, pad testing, and what have you, but also accomplished or was in lieu of an inspection plan. It is straightforward to compose NIs with similar definite subjects and analysis, as in (101).77 (101) Speaker A: Man, of everything that coulda gone wrong today. Tire blows out. Phone’s dead. Cold as hell outside. Speaker B: Oh yeah. Didn’t that oil heater end up comin’ on either. ‘Oh yeah. That oil heater didn’t end up coming on, either.’ As Ward and Birner explain, “The use of the definite in conjunction with the thereconstruction reflects the treatment of the referent as simultaneously hearer-new and uniquely identifiable. Indeed, it is precisely this mixed marking that leads the hearer to infer that the utterance is a reminder” (Ward and Birner 1995: 731). The same intuition clearly holds for the reminder NI in (101). In addition, it is straightforward to find NI cases where there are no grounds for the “speaker to believe that the entity has been momentarily forgotten.” In such cases, the definite subject is ruled out. This is what we see in (102), in which the subject referent in question was evoked in the directly preceding utterance and so the “reminder” function is misplaced, rendering the definite subject infelicitous.

77 Some of the results based on examples (101–118) were reported in Salmon (2018a). In that paper, 10 native speakers were queried. Six of those participants were speakers of AE, and two each were speakers of African American and Chicano Englishes, all of whom are known to me personally. For the present publication, I queried an additional 10 speakers: four of whom were Anglo, and six each who are speakers of AAVE and CE. These latter 10 participants were not known to me personally and were surveyed on the street in Corpus Christi and Houston, Texas, in October 2018.

3 Introduction

73

(102) Speaker A: Man, of everything that coulda gone wrong today. Tire blows out. Phone’s dead. Cold as hell outside. Speaker B: Oh yeah. #Didn’t that phone work either. ‘Oh yeah. That phone didn’t work either.’ In (102), it is unlikely that the phone has been momentarily forgotten in such a short span. As such, it is difficult to treat the referent as hearer-new, which gives rise to either infelicity or to the possibility of the speaker’s intention to remind the hearer of a different malfunctioning phone. Group (ii), hearer-new tokens of hearer-old types, includes NPs with a series of adjectives indicating that a type is known (such as, same, usual, regular, traditional, obligatory, expected, etc). As Ward and Birner note, the NPs in question have dual-reference: one to a type and one to a token. If the type is old or uniquely identifiable, this allows for the definite use, while at the same time, the new instantiation or token of the type, allows for the existential use in an example like (103) [Ward and Birner’s (21)]: (103) There was the usual crowd at the beach today. It is clear that “the usual crowd” could refer to the exact same group of people that is always at the beach; or, it could refer to the fact there is always a crowd of people, whatever their identity, at the beach. It is this dual reference, and dual dependence on the new and old/uniquely identifiable, that licenses the definite use. It is straightforward to find similar examples with the NI in (104). (104) Speaker A: Hey baby, y’all have a good time tonight? How was the bar? Speaker B: Not too bad. Couldn’t the usual crowd get in ‘cause of the cover charge, so there was a lot of room. Got the best table in the house. ‘The usual crowd couldn’t get in . . . .’ As Ward and Birner write of TEs, “The definite is licensed by the identifiability of the (hearer-old) type, while the there-construction is licensed by the hearernew instantiation of that type” (Ward and Birner 1995: 732). The same is true in the NI in (104). We can alter (104) a bit, however, so that the NI in question bears a subject NP that is both identifiable and hearer-old, and the results are infelicitous, as in (105). (105) Speaker A: Hey baby, y’all have a good time tonight? How was the usual bar?

74

Chapter 3 Negative inversion, existential sentences, and definite subjects

Speaker B: Not too bad. #Didn’t the usual bar fill up ‘cause of the cover charge, so there was a lot of room. Got the best table in the house. ‘The usual bar didn’t fill up . . . .’ Here, the pragmatic requirements are not met, and felicity suffers predictably. Group (iii), hearer-old entities newly instantiating a variable, is home to the most frequently discussed definite subject in existential sentences: i.e. the “list reading,” as in (106) [Ward and Birner’s (30b)]. The idea here is that the rhetorical structure of the list renders entities that might be hearer-old as new; i.e., they are new in the manner in which they are enumerated on the list.78 (106) Speaker A: Who was at the party last night? Speaker B: There was John, Mary, Fred, Susan, Hilda, Xavier, and Ethel. Here we see that John, Mary, and so on are known to Speaker A, and so hearerold. However, their arrangement on the list puts them into a relation that is new to Speaker A, and so they qualify as hearer-new in this individual situation. Specifically, as Ward and Birner write, “the individuals constitute hearer-new instantiations of the variable in some salient open proposition” (Ward and Birner 1995: 734). It is this fact of instantiating the variable slot of the open proposition that licenses the definite NP in that position. Comparable list-reading examples with NI are straightforward and are acceptable for the reason just mentioned. (107) Speaker A: Hey man, how was y’all’s meetin’ last night? Y’all have a good turnout? Speaker B: Yeah, it was pretty good. Bob and Mary, Jenny and Doug, was all there. Couldn’t Tommy, Darin, or little Mikey come out, but we still had a pretty good time. ‘Tommy, Darin, and little Mikey couldn’t come out [. . .].’

78 Cf. Gundel (1999: 4) on two types of givenness and newness: One type [of givenness/newness] is referential; it involves a relation between a linguistic expression and a corresponding non-linguistic entity in the speaker/hearer’s mind, the discourse, or some real or possible world, depending on where the referents or corresponding meanings of these linguistic expressions are assumed to reside. [. . .] The second type of givenness/newness is relational. It involves two complementary parts, X and Y, of a linguistic or conceptual representation, where X is given in relation to Y, and Y is new in relation to X.

3 Introduction

75

Relatedly, we can find less felicitous NIs of this type when the necessary context of the open proposition is missing, as in (108). In (107) above, A and B set up the expectation of a list and a salient open proposition. There is no similar expectation in (108), which renders the definite subject NP infelicitous. (108) I had kind of a boring time last night. #Couldn’t Tommy, Darin, or little Mikey come out to the party I went to. What this suggests is that it is not the “list form” per se that licenses the definite NP, but is rather the manner in which the sentence fills the open proposition. It is worth mentioning here as well that Green’s (2014) “no Vince Jackson” example from above in (99) also fits in this category. Ward and Birner describe the list example this way: “the individuals listed are uniquely identifiable; however, their membership in the set being enumerated is new to the hearer” (Ward and Birner 1995: 734). This describes the Vince Jackson example as well. In that example, the first speaker identifies Vince Jackson and locates him within a particular set (i.e. the set of those who showed up for homecoming). The second speaker then adds Vince Jackson to a different set (i.e. the set of those who did not show up for homecoming). As such, Vince Jackson’s “membership in the set being enumerated is new to the hearer.”79

79 There is another type of example that I believe fits in this category as well. Foreman (2015: 145) notes in a non-related context that NP subjects such as John can appear in subject position if preceded by even, in examples such as (i) [Foreman’s ex. 154]. i Can’t even John do that. ‘Not even John can do that.’ For Foreman, (i) is acceptable, but (ii) would not be acceptable. ii *Can’t John do that. Foreman does not elaborate as to what makes the even example in (i) acceptable, and apparently does not consider (i) to be a counterexample to the general prohibition against definite or proper noun subjects. Standard understandings of even in the semantics and pragmatics literature tell us that use of even evokes a set of alternatives, in which the focus of even is the least likely member of the set to do whatever is being predicated. For example, here is Potts (2007: ex. 1c), lightly adapted for clarity: iii Even Bart passed the test. Descriptive Meaning = Bart passed the test Conventional Implicature ≈ Bart was among the least likely to pass the test

76

Chapter 3 Negative inversion, existential sentences, and definite subjects

Group (iv), hearer-new entities with uniquely identifying descriptions, does not “depend on the prior context for felicity” (Ward and Birner 1995: 735). Rather, the NPs in question contain enough descriptive content to render them felicitous without a co-text or situation. Among other forms, this group includes formal superlatives (109a) and deictics (109b) [Ward and Birner’s (38a–b)]. (109) a. There was the tallest boy in my history class at the party last night. b. You can see the runway and the HUD that overlays the Edwards runway, and then there is this line which comes out to the outer glide slope aim point. Deictic examples, such as demonstratives, are actually some of the most facile definite subjects in NI, with examples such as (110). These in-situation demonstratives clearly pick out hearer-new entities, but entities that are at the same time uniquely identifiable due to the demonstration in the context. (110) [Looking at a photograph of contestants. The speaker points at each one as he speaks.] Won’t this ☞ one or that ☞ one end up finishin’ the race. Superlatives are similarly straightforward in NI, as in (111). (111) Speaker A: Hey man, how’s that work comin’ on your back porch? You get them cracks all sealed up? Speaker B: Yeah, it’s sealed up tight. Cain’t the littlest mouse get hisself through there now. ‘Even the littlest mouse can’t get himself through there now.’ For this group, Ward and Birner note also that NPs with modifying relative clauses can be felicitous in existentials, in which “the relative clause serves to fully specify the referent [. . .]” resulting in unique identifiability (Ward and Birner 1995: 737), as in (112) [Ward and Birner’s (39)].

Even places Bart at the least likely end of a list of those able to pass the test. Thus, like Vince Jackson, Foreman’s John, above in (i), who is likely known information to the hearer, becomes hearer-new by virtue of being ordered on a list as least likely to pass the test.

3 Introduction

77

(112) There are those who would claim that computers will take over the earth within the next decade. There are similarly straightforward NP-relative clause examples in NI. (113) [Two friends talking about their dates from the previous night.] Speaker A: What a sweetheart he was. He wouldn’t let me pay for anything. Dinner. Movie. He covered it all. I tried to pay my part, but he wouldn’t allow it. Speaker B: Well, I don’t know where you found that one. Couldn’t the guy I was with be bothered to pay for ANYthing. ‘The guy I was with couldn’t be bothered to pay for anything.’ Finally for this group, Ward and Birner note that cataphoric references are allowed, as in (114). (114) There are the following reasons for this bizarre effect. . .. [Ward and Birner’s (38c)] Cataphoric references are similarly allowed in NI, as in (115). (115) [In a military boot camp setting. The sergeant comes into the barracks to share the results of a field test. Everyone snaps to attention. The sergeant barks:] Okay, listen up, you maggots! Didn’t the following losers pass the exam! Park, Swanberg, Turner! Get your butts in my office. Now! ‘The following losers didn’t pass the exam.’ For all of the different cases in this group, the key fact is, as Ward and Birner point out, “the unique identification justifies the definite, while the hearer-new status of the referent of the postverbal NP licenses the there-sentence” (Ward and Birner 1995: 737). With the NI, we can test this by observing cases where the forms in question don’t uniquely identify a referent. For example, in (113), above, the post-modifying relative uniquely identifies the date in question. A sentence of the same form can be rendered much less felicitous by weakening the identificational content, as in (116).

78

Chapter 3 Negative inversion, existential sentences, and definite subjects

(116) [Two friends talking about their dates from the previous night.] Speaker A: What a sweetheart he was. He wouldn’t let me pay for anything. Dinner. Movie. He covered it all. I tried to pay my part, but he wouldn’t allow it. Speaker B: Well, I don’t know where you found that one. #Couldn’t the guy I know be bothered to pay for ANYthing. ‘The guy I know couldn’t be bothered to pay for anything.’ This fact illustrates once again that it is not the form of the definite NPs at issue here, but rather its relation to the discourse context and the common ground assumptions of the speaker and addressee. The last group Ward and Birner identify are the false definites. These are forms which are grammatically definite but are cognitively indefinite, and introduce or “refer to a specific entity that the speaker believes is unknown to the hearer.”80 They provide examples such as those in (117a–d) [Ward and Birner’s (43) and (44)]. (117) a. One day last year on a cold, clear crisp afternoon, there was this huge sheet of ice in the street. b. There are all sorts of other false definites. c. There is the most curious discussion of them in our paper. d. There is every reason to study them. There is no expectation that the hearer will be aware of the huge sheet of ice in (117a), nor would anyone take the speaker to believe the discussion of false definites in the paper is the most curious of a set, as a literal superlative. As such, these forms are only grammatically definite, but they are cognitively indefinite. It is possible to find false definites of this sort in NI as well. For example, in (118) we see a superlative form that is not used as a literal superlative, similar to (117c), but instead has an intensification function. (118) [Police officer A is watching an interrogation through the one-way interrogation mirror at the police station. The person being interrogated refuses to talk, and the interrogating officer stands up quickly, spilling his cup of coffee on his lap in the process. Officer A turns to Officer B and says:] Officer A: Well, cain’t the best discussion be goin’ on in there. ‘The best discussion can’t be going on in there.’ 80 See Prince (1992).

3 Introduction

79

It is thus evident that a wide range of definite subjects are allowed in the NI constructions, contra all the literature on the subject since Labov et al. (1968). There are many different questions to ask in light of this observation. I will ask three of them here. The first is, is this a new development; i.e. has the prohibition against definite subjects gradually bleached out over the years? The second question is, is there a rhyme or reason for which definites are allowed in subject position? The third question regards the relation between the “copula” thereexistentials that Ward and Birner discuss and the Modal Existentials, such as there can’t nobody lift that rock, which is argued here to be the syntactic etymon of the NI. There is insufficient historical data to determine when definite subjects became possible in NI, so there is no conclusive answer to the first question in that respect. It is possible that the construction’s prohibition against definite subjects has weakened over the years as a result of some kind of discourse pressure – perhaps in favor of becoming a general negation strategy. On the other hand, the feeling that there is still a strong prohibition against definite subjects is very real. Not just any definite in any context is allowed in the construction. This supports the belief that it’s not a general weakening of the prohibition that is being observed, but that instead the distribution is systematic. The second question, which asks if there is a rhyme or reason to which definite subjects are allowed, can receive a positive answer – the NI allows definite subjects when they do not clash with the pragmatic and cognitive constraints set out above in Ward and Birner’s discussion of TEs. This positive answer also goes some distance toward answering the first question raised above, as to whether the acceptance of definite subjects is a new phenomenon. I believe it is not. As I will argue below, the NI is essentially a Modal Existential sentence that has lost its expletive over time. The modified definiteness effect that we see with the NI today is very likely just the same modified definiteness effect that has been associated with its ME etymon all along. The third question regards the relation between the copula-there-existential and the Modal Existential. Why did the expletive delete from the latter, rendering NIs such as (119a), but not the former, that would have resulted in an expletiveless existential, such as (119b)?81

81 It is worth pointing out that there can appear to be expletiveless existentials in English, such as: (i) ‘z two cats on the counter. (Cf. There’s two cats on the counter) (ii) A: There is no cat on the counter. B: Is too!

80

Chapter 3 Negative inversion, existential sentences, and definite subjects

(119) a. There Can’t nobody lift that rock. b. There *Is a man in the house. I argue later in this chapter that the difference between (119a) and the impossible (119b) is in the negation, and in what Jespersen (1917: 5) laid out as an organizing principle of language: namely, that “there is a natural tendency in language [. . .] to place the negative first.” Before doing so, however, it is necessary to address an objection to the relation between NI and existential sentences based on subject distribution. This is done in the next section.

3.2 Subjects in negative inversion and there-existentials: Not so different after all The discussion in the preceding section argues that the pragmatic subject restrictions on NI are the same as those that restrict definite subjects in TEs. Matyiku (2017: 262–264), however, responding to Salmon (2017), argues that this parallelism can only be a partial story, as NI and TEs do not completely overlap in the kinds of subjects they allow. Based on this apparent discrepancy of permitted subjects, Matyiku reasons that the distance between NI and TEs must be wider than assumed in Salmon (2017). Table 1 lays out Matyiku’s distribution of subjects for the two constructions [Matyiku’s Figure 2.1, see Matyiku 2017: 45]. We see that the second half of the chart lists the subjects where the distributions are believed to differ in Matyiku’s system. As I show below, however, the differences turn out to be only apparent, and that in fact there is very little distributional difference of subjects between the two constructions. This observation undermines the distinction drawn between NI and TEs in Matyiku (2017) and other works.82 It also sets the stage for the existential treatment given to NI in this chapter and Chapter 6 of the present work.

These examples do not strike me as the same thing described above with respect to (119). For example, in (i), the reduced clitic form of the auxiliary is similar to what it would be if the expletive were phonologically present. It is likely that (i) is simply an instance of allegro speech deletion. Note that a fully deleted expletive as in (119b) is clearly unacceptable. Regarding (ii), it seems likely that it this is also an instance of interactional conversational deletion and so not a true expletiveless existential, as we might find in Hebrew, for example, as described in Francez (2007). 82 Foreman (1999: 7–8) argues also that NI and there-existentials allow different subjects and thus must not be related. The present section thus undermines Foreman’s claim as well.

3 Introduction

81

Table 1: Matyiku’s distribution of subjects in TEs and NI constructions.

Uniform distribution

Different distribution

Subject

Existential

NAI construction

Jamie (proper names)

×

×

he (pronouns)

×

×

the NP

×

×

their NP

×

×

any NP

√

√

no NP

√

√

many NP

√

√

more than

√

√

# NP

√

√

a NP

√

√

every NP

×

√

half the NP

×

√

many of (the) NP

×

√

none of (the) NP

×

√

some NP

√

×

few NP

√

×

several NP

√

×

Let’s begin with every NP and half the NP. These strong subjects are historically believed to not occur in TEs, yet they are well attested in NI.83 This former claim is of course part of Milsark’s (1974) original conception of the definiteness effect in TEs, as shown in (121a–b) [judgments on (121a–b) for standard English are Matyiku’s]. (120) a. Ain’t every student here yet. b. Didn’t half the students do their homework.

83 Examples (120a–b) appear as (5.9a–b) in Matyiku (2017) and originate in Foreman (1999). Examples (121a–b) appear as (5.9a–b) in Matyiku (2017).

82

Chapter 3 Negative inversion, existential sentences, and definite subjects

(121) a. “*”There’s every student here yet.84 b. “*”There are half the students outside. Pace Milsark (1974), however, and a vast literature since then, strong quantificational subjects are in fact allowed in TEs. Consider the exchange below in (122a), which is taken from a “College Confidential” chatroom discussion. Similarly, the example in (122b), which is taken from an academic essay on academic networks, is equally acceptable.85 (122) a. γ ace550: Has the teacher posted the practice problems yet? If yes, are they the EXCEL files posted in the Calendar section inside OnCourse? hkem123: Yeah, there’s every midterm from the past 4 years on the OnCourse calendar. Plus a problem that is due Wed/Thurs.86 b. γ Thus, even the students who have not filled in the survey are all embedded in our sample with at least three social ties. This means that there is every single student accounted for in mapping nonsymmetric networks.87 Examples such as these can be multiplied via Google or the freely available Corpus of Contemporary American English.88

84 Undoubtedly part of the awkwardness of this sentence stems from the unlicensed NPI yet. See Levinson (2008) on licensing of this particular NPI. 85 Many thanks to Itamar Francez for discussion of these examples with me (2018, p.c.). 86 . 87 . 88 Further, in a brief section in Chapter 16 of The Cambridge Grammar of the English Language (2002: 1401), Ward, Birner and Huddleston write “As with definite NPs, however, there is no categorical ban on displaced subject NPs determined by most, all, each, and every. Under conditions that are not yet well understood, such NPs can also appear in existential clauses, as in [ia–d].” i a. Among our dresses there were most kinds of shabby and greasy wear and much fusion and corduroy that was neither sound nor fragrant. b. There are all kinds of insurance policies that can meet your needs. c. There’s still each student in Group C to be interviewed. d. I think that’s probably still a NASA job because of the number of contractors involved. In firing room two, there’s every contractor we’ve got, just about, over there. Thus, pace most of the literature on the subject, including Matyiku (2017), strong-quantificational subjects are allowed in there-existentials and are most likely pragmatically restricted similar to what we have seen with definite subjects. The point for the present discussion is that these kinds of subjects cannot be used to distinguish the existential and NI sentence types.

3 Introduction

83

The same sources also provide ample tokens of strong subjects such as half, which are again historically assumed to be absent from TEs yet permitted in NI. Thus, in a 2017 Variety interview with Lester Holt and David Muir regarding the 2016 US presidential election, Muir speaks the perfectly natural sounding (123).89 (123) γ I think as long as we’re asking what people at home are asking, from all ends of the spectrum, we recognize that particularly after this election, we’re broadcasting to a divided America. You know, there’s half of the country that feels as though finally they’ve been heard. That they’ve been left behind by all this talk of a recovering economy that they didn’t have access to. Similarly, on SmokingMeat.com, poster “Hillbilly Jim” writes the perfectly natural sounding (124).90 (124) γ I put two butter stick size pieces of sharp cheddar and about a half cup of coarse sea salt in the smoker about 5:30 PM. I got the pellets lit, checked on it an hour later and they were still producing smoke. It’s now 7 AM, the pellets are still smoking and there’s half of the last run in the smoker still left. Let’s look next at the partitive subjects none of (the) NP and many of (the) NP. These too are believed to be prohibited in TEs yet allowed in NI, as in (125) and (126), respectively [Matyiku’s (2.20a–b) & (2.21a–b)].91 Personally, I find the sentences in (125) to be perfectly acceptable. And, it is worth noting that Hoeksema (1989), cited in McNally (2011), also argues that partitive subjects can occur in TEs. So, there are problems with the empirical assumptions regarding TEs from the beginning. (125) a. *There are none of the students present. b. *There are many of the students present. (126) a. Won’t none of the students go to the party. b. Didn’t many of them live aroud here.

89 . 90 . 91 Matyiku cites Foreman (1999: 6) and Labov (1972a: 812) for (126a–b), respectively.

84

Chapter 3 Negative inversion, existential sentences, and definite subjects

Not surprisingly, it is straightforward to find examples of TEs exhibiting these types of subjects. For instance, consider (127), which is taken from a response to a grammar question on Quora.com (which is unrelated to whether or not partitive subjects can occur in TEs).92 (127) γ Use of the future tense in your second sentence would indicate that you are aware, as you speak, that there are none of the things you request right now in the present moment. In the first sentence, there may, or may not, be any of what you request, depending upon the answer you get. Otherwise, I would like to call your attention to your syntax – how you are asking permission to ask a question, but you ask it anyway. Changing registers, we find a perfectly natural sounding TE with subject many of (the) NP, this time moving from the realm of pretentious grammar advice to a farming feed and supply store in Columbia, Missouri, advertising baby chickens for sale.93 (128) γ We have chicks. . . there are many of the ‘calm’ chicks as seen in the one pic. . . and we have the ‘attack’ chicks as seen in the very close up pic. . . they want to go home with you. And we will have FISH April 1st. . . call in your orders for fish by the 25th. . . The last putative distributional mismatch between the two constructions involves some NP, few NP, and several NP, all of which are believed to occur in TEs but not in NI. Here are Matyiku’s examples [(2.22a–c) & (2.23a–c)].94 (129) a. There are few students outside. b. There’s somebody in the hall. c. There are several students present. (130) a. *Won’t few boys go to the party. b. *Can’t somebody get in the hall. c. *Cain’t several students read that book.

92 . 93 . 94 Matyiku cites Foreman (1999) for the NI examples in (36).

3 Introduction

85

As we saw in Chapter 2, sentences such as (130b) are indeed possible, as in (131a), below. Similarly, though there might be a slight register clash in (131b), the NI there is certainly possible with several NP as subject. In the last case, however, I cannot hear a well-formed NI example with few NP as subject. Thus, (131c) appears to be ill-formed and the only apparent point in which NI and TEs differ in subject permission. This is not completely surprising, however, as few is a well-known positive polarity item, which eschews appearing under negation in any case. See, for example, Israel (2011: 87) on the positive polarity status of few NP. (131) a. Didn’t some (of the students) show up. b. Didn’t several of ‘em get off the boat. c. *Didn’t few students show up. Matyiku also reports that few, some, and several do not appear as part of the subject in negative TEs outside of metalinguistic negation or contrastive focus. Consider (132a–c) [Matyiku’s Fn. 4, see Matyiku 2017: 44, with Matyiku’s judgments]. (132) a. “*”There aren’t few students outside. b. “*”There isn’t somebody in the hall. c. “*”There aren’t several students present. With respect to (132b–c), I disagree with Matyiku’s judgments. These sound perfectly possible to me even outside of metalinguistic negation or contrastive focus contexts, and it is easy to find attested examples.95 Thus, in (133), we see a very natural sounding someone as subject of a negative TE in a TripAdvisor. com review of an Extended Stay America hotel in Dallas, Texas. I don’t see this example as requiring contrastive focus or metalinguistic negation to be licit. (133) γ Has potential, but there isn’t someone willing to go the extra mile. If you ask for something that is offered by Extended stay America you receive the bare minimum if you are lucky. Welcome is not exactly the feeling you get when you are there. Unfortunate because it could really be a great place to stay.

95 Consider also (i), taken from McNally (1992: ex. 129), which is perfectly acceptable. (i)

The food critic was annoyed because there wasn’t some type of wine on the list.

86

Chapter 3 Negative inversion, existential sentences, and definite subjects

Similarly, in the introduction of a web list of “50 best smartphone apps” on askmen.com, we find a natural token of there aren’t several NP. Again, there is no metalinguistic quality here, and no sort of contrastive focus stress or intonation is necessary. (134) γ Best-app lists are a dime a dozen, but we assure you this list is the absolute best, as it’s been carefully curated by someone who had nothing better to do with his week than experiment with hundreds of apps to select the 50 best. And while it’s understood that 50 apps may very well consume all of the storage on your precious phone, you can pick and choose from the selection based on your own preferences. So go on, now, open up that App Store. Disclaimer: We assure you there aren’t several iterations of the same app and we’ve left out the more obvious selections like Facebook and Twitter. Because we’re good like that. There are two things to note here. First, the last set of subjects believed to have a different distribution in the two constructions (i.e. some NP, few NP, and several NP) are not all that different in their distributions. And, to the extent that few NP is ruled out in the NI, it’s not clear that this actually makes a difference, as few NP is also ruled out in negative TEs due to its strong positive polarity properties. If anything, this range of subject distributions actually supports the connection between NI and existentials rather than distinguishing them. Thus, the objection to the relation of NI and TE sentences based on subject distribution largely falls apart. Beyond Matyiku’s claims, Horn (2015, 2018b) argues that Negative Inversion (Horn’s “V1NI”) and negative TEs (Horn’s “nexistentials”) overlap almost completely in subject distribution, with the exception of universals such as every. Thus, Horn writes “But V1NI doesn’t follow [the nexistential] pattern in every respect,” and he provides the following pair to exemplify the exception [Horn’s 2018b: ex. 46/47]. (135) #There is(n’t) every unicorn in the garden. (136) Corn. Lettuce. Easy cheese. Medium sauce. And the special dressing they got in the back that don’t everybody know about. [= that not everybody knows about] This would be a sticking point, then, in which the negative existential rules out the universal subject, while NI allows it. However, it seems that, pace Horn, negative existentials are in fact possible with the strongly quantified every NP as subject. Consider this interview fragment from Washington Post reporter

3 Introduction

87

David Fahrentold. Here, Fahrentold, who is from Houston, Texas, talks about his philosophy of covering news stories over time and of keeping his readers informed as a story develops: (137) γ I wanted to let people – give people a way to follow what I was doing everyday, even if there was not a story that I had written that day or you know a web story, like, let’s get people – let them know what I’m doing every single day and let them sort of see the punt. I mean people like – they make movies about reporters. There’s not every job that they make movies about. And then most of the movies are totally wrong. But they make movies about us, because there’s a chase, there’s a hunt.96 The strongly quantified subject is perfectly acceptable here, and readers can listen to Fahrentold produce it on the video linked in the footnote below. The case then for a strong relation between NI and existential sentences is quite compelling, pace much previous literature.97 I will return to this relation between the two sentence types below in Chapter 6. In the next section, we turn to the Modal Existential construction, which I argue is the source and syntactic etymon for the contemporary Negative Inversion.

3.3 The modal existential and negative inversion constructions In this section I argue that the expletive subject of the Modal Existential – i.e. there or they – is no longer available to most younger speakers of the language varieties under consideration here, though it is still available to many older speakers.98 This hypothesis provides a straightforward account of the definiteness

96 . The nexistential of interest occurs at 10:58 in the video. 97 G. Green (1985) also suggests an existential analysis of NI, though she does not fully provide it. Along the same lines, White-Sustaita (2010) argues that all NIs receive an existential reading. 98 Montgomery (2006) writes that expletive they in general “can be traced to Ulster and ultimately to Scotland in the seventeenth century and has been in variation with expletive there for 400 years.” He hypothesizes that they’s presence in the American South is due to immigration from Ulster and Scotland, as opposed to resulting from a post-vocalic deletion process of [ɹ] in Southern and Appalachian dialects. For present purposes, I note only that my older participants in Texas state that expletive they is common and that there is no meaning difference between it and expletive there. My suspicions are that there is a stylistic difference, perhaps of formality. This will require further research to determine though.

88

Chapter 3 Negative inversion, existential sentences, and definite subjects

effects described in Section 3.1 above, and it is supported by the quantitative results of cross-generational surveys, reported on below. Here are the basic empirical facts of the ME construction. Ball (1991) and von Fintel (1992) trace the roots of the construction through Middle English (138a) and suggest it ultimately descends from the Old English in (138b) [von Fintel’s (1) and (12), respectively]. (138) a. But kynge Arthur was so currageous that there might no maner of knight lette hym to lande. ‘But King Arthur was so courageous that there could no manner of knight prevent him from landing.’ b. Ne mai no man Dese word seggen Danne he godes milce bisec∂. ‘There may no man these words say when he beseeches god’s mercy.’ In the US, currently, MEs are believed to exist only in Southern or Appalachian varieties; although, it is likely that these sentences had a wider distribution in the past. For example, (139) is found in the novel Madelon, by Mary Wilkins Freeman of Randall, Massachusetts, published in 1896. Freeman has been described as “a local-color realist [. . .] in the tradition of New England women’s writing” Reichardt (1997).99 (139) [. . .] but I guess there didn’t nobody have any knife, and I guess he’ll git out of prison pretty soon.

99 It was also possible, for example, for Senator Thompson, of the New York State Senate in the Minutes and Testimony of the Joint Legislative Committee in 1916 to utter the Modal Existential sentence in (i). (i)

γ If you make a contract down here like a letter now, it’s perfectly possible for President Shouts of the Interborough and President Williams of the Brooklyn Rapid Transit to sit down and write a letter involving a transaction of millions of dollars and they will put it in typewritten sheets of ordinary letter paper and it will be perfectly understandable, but just the minute those people turn it over to the lawyers and engineers it covers two or three printed pages and there can’t anybody understand it, isn’t that right? (Laughter.) (1916: 1017)

Similarly, across the continent in California, only a few years earlier in 1912, it was possible for J. W. Jeffrey, State Commissioner of Horticulture, addressing the Convention of the California State Growers, to utter (ii). (ii)

γ The citrus trees of Butte county – that is so far away that there can’t anybody jump on to us – just raised three quarters of a box to the tree last year; that is about the average, and yet they have some very profitable orchards there. (40)

3 Introduction

89

It is also believed that MEs do not occur in contemporary AAVE (see e.g. Martin and Wolfram (1998), Green (2014) and references therein), but that NI does occur in AAVE. MEs are known, however, to have existed in older varieties of AAVE, as tokens are found in ex-slave narratives reported in Bailey, Maynor, and Cukor-Avila (1991). See, for example, White-Sustaita (2010: 432). MEs have been discussed only occasionally over the years in the sociolinguistics and dialectology literature, such as Feagin (1979) and Montgomery and Hall (2004). The semantic and syntactic description following below, however, comes primarily from Zanuttini and Bernstein (2014), who refer to the sentences as “Transitive Expletive Constructions,” though I will use the terminology “Modal Existential,” which comes from Ball (1991) and von Fintel (1992). Consider the following examples, from Zanuttini and Bernstein.100 (140) a. There can’t nobody ride him. b. There wouldn’ nothin’ go down through there. c. [. . .] but there didn’t anyone want to leave their church. Zanuttini and Bernstein (147–148) include the following descriptive properties, which are summarized below: (i) co-occurrence of an expletive subject with quantificational associates like nobody, nothing, anyone, many people, no girl, etc. (ii) only an expletive element can occur in the position preceding the modal; a lexical noun phrase cannot; (iii) the associate typically co-occurs with only one of two expletive elements, namely they or there; (iv) the subjects always occur in the presence of a modal (like can’t, couldn’t, won’t) or finite auxiliary (like didn’t, ain’t); (v) the subjects under investigation occur in sentences in which the modal or finite auxiliary carries the negative morpheme n’t. Von Fintel (1992) is concerned with similar data and provides the syntactic description in (141), describing the sentences as “very uniform and idiosyncratic.” (141) there + modal auxiliary + negative subject + verb + object

100 Zanuttini and Bernstein cite Feagin (1979) and Montgomery and Hall (2004) as sources for these examples.

90

Chapter 3 Negative inversion, existential sentences, and definite subjects

As can be seen from the previous data and empirical descriptions of the Modal Existential, it bears a striking resemblance to the Negative Inversion. So similar are the two, that my older test participants – as well as those of Zanuttini and Bernstein (2014) – could detect no difference in meaning or use of the MEs and NI. There is, however, a crucial demographic difference in the results of my survey participants: namely, older participants accepted the ME sentences, while younger participants rejected them as ungrammatical. As I discuss below, this finding is one of the motivating factors for the diachronic account being pursued in this book. The claim that NI is actually a truncated Modal Existential should not be controversial. Martin and Wolfram (1998) assert of AAVE NI that the most straightforward account of these types of sentences would be to say that “inversion does not occur. Rather, negative inversion sentences are negative existential sentences” (Martin and Wolfram 1998: 27).101 However, Martin and Wolfram cannot ultimately take this line of argumentation, as they believe that AAVE inversion sentences do not allow an expletive subject to be expressed, as in (142) [Martin and Wolfram’s (32a–b)].102 The fact that expletive sentences such as (142) don’t seem to surface in the AAVE with which Martin and Wolfram are concerned, leads them away from a deletion account. Crucially, however, they are concerned with synchronic deletion, and do not consider the possibility that the deletion might have been diachronic, as argued in the present work. (142) a. *There/*It didn’t nobody laugh. b. *There/*It can’t no man round here get enough money to buy they own farm. Sentences such as (143) are in fact acceptable to many of the older and middleaged speakers in the Texas English varieties of the present work, however. (143) a. There didn’t nobody laugh. b. They cain’t no man round here get enough money to buy his own farm. As such, Martin and Wolfram’s reasons for objecting to the expletive hypothesis for AAVE NI do not hold for the Texas English sentences. I argue that the NI of the 101 Montgomery and Hall (2004: lxiv), writing on Smoky Mountain English, suggest that expletive sentences such as (i) may be the basis of clauses with negative inversion,” though these authors do not follow up on the idea. (i)

They didn’t none of us ever get snakebit, but their work animal did.

102 This inability of NI in AAVE to take expletive subjects is also noted in White-Sustaita (2010) and Green (2014).

3 Introduction

91

present work is essentially an ME sentence, or what is left of one, after the expletive has deleted over time. In the sections that follow, I consider the generational data brought to bear on this question across the three dialect groups in question. 3.3.1 Generational data for modal existential and negative inversion constructions The data in this section show relative levels of familiarity with MEs and NI among speakers of African American, Anglo, and Chicano Englishes in Texas. I have a much richer data set for Anglo speakers than for the other two varieties, so the strongest claims below will be made for that group based on that data set. I have generational data for African American and Chicano speakers as well, and this trends very closely to that of the Anglo speakers. The primary differences in the data sets are that there is much more of the Anglo data and it thus supports a four-way age split, while the African American and Chicano data sets are smaller and support only two-way age groupings. As such, the trends over time for the Anglo data can be shown more clearly. Another difference in the data, which will be discussed in more detail below, is that it was gathered using a variety of methods, with some data being gathered through online surveys, other data gathered in in-person classroom visits, and still other data gathered in one-on-one interviews. The sections below give brief descriptions of the source and methods from which the three data sets were taken, beginning with that of the Anglo speakers. 3.3.2 Texas Anglo speakers The sociolinguistic, demographic claims made with respect to NI and the MEs in this section come from surveys of 121 native speakers of AE in Texas. Ninetytwo surveys were administered online via Survey Monkey. The survey link and a call for participants was posted in three private Facebook groups, each of which was restricted to members who had grown up in the cities of Abilene, Corpus Christi, and Odessa, Texas.103 The results of these surveys have been sorted by age, resulting in three groups, with roughly equal numbers of men

103 Abilene (pop. 117,000) and Odessa (pop. 118,000) are two of the primary cities of west Texas, which both have significant agricultural and working-class populations. Corpus Christi (pop. 326,000) is located in south Texas. This city was chosen for similar reasons. Furthermore, it is also the author’s home town, and I wanted to ascertain that my linguistic intuitions were in line with members of my home speech community.

92

Chapter 3 Negative inversion, existential sentences, and definite subjects

and women. This is given in Table 2. The majority of respondents identified as Caucasian; those who identified otherwise were excluded from this data. Table 2: Number and average age of Anglo participants. Age Group

Average Age

Number in Group

–

.



–

.



+

.



–

.



In the final row of Table 2, we see the fourth and youngest group, which is composed of participants aged 16–18, who were surveyed in three high school classes in a rural west-central Texas town in December 2016. For these surveys, the author attended each class, introduced himself and the bare essentials of the research questions to each group of students, and then administered the surveys in hard copy. As with the online surveys above, the majority of these participants were Anglo. A small number of participants identified as Latino or Mexican-American here, but these were excluded.104 There were no African American participants in these groups. All of the surveys queried meaning, use, and attitudes toward NI and ME constructions, and contained questions such as the following: Consider the two statements (a) and (b). Who (in terms of age, gender, socio-economic class, race) do you think is most likely to say something like (b)? (a) Nobody came to the party. (b) Didn’t nobody come to the party. When someone says (b), what do they mean by it? Is there a difference in meaning between (a) and (b)? Does (c) differ in meaning from (d)? If so, how does it differ? Who do you think is most likely to say a sentence like (c)? (c) They can’t nobody lift that thing. (d) Can’t nobody lift that thing.

104 I do not mean to imply that identifying with a particular ethnonym entails that one does or doesn’t speak a given variety of English. Thus, someone who identifies as Latino could be a native speaker of Anglo or African American English or some other variety. In this particular case, it is likely that the Latino participants actually are native speakers of Texas AE, as the Latino community in this area is quite small. Just to be safe, though, I excluded all of those participants from this data set in order to create a sample that was as homogenous as possible.

3 Introduction

93

The survey also gathered demographic information of participants, such as age, race, and socioeconomic class. Please tell us your age, race, gender, where in Texas you are from, and what your parents do for a living.

Results were compiled by age group into a spreadsheet. Rather than querying participant intuitions directly, the surveys do so indirectly for a variety of reasons. First, as a native speaker of one of the relevant varieties, I know that NI and MEs are stigmatized as “uneducated” workingclass speech; and, pilot surveys at two state universities confirmed this.105 To a participant who is unaware of the goals of linguistic research, a direct query about such linguistic features can be tantamount to asking if the participant herself is uneducated or working-class, and so might cause offense, undermining the intent of the survey. The indirect nature of the instructions – i.e. “who do you think is most likely to say this” – was meant to mitigate this possibility. I further wished to avoid survey responses that attempted to “correct” test stimuli to what would be prescriptively acceptable: i.e. ‘the proper way to say this is Nobody can lift that thing.’ By posing the stimuli as sentences associated with a third party rather than directly with the participants themselves, I hoped to receive a more objective consideration of the grammatical realities.106 Instructions also included questions regarding differences in meaning between the NI/ME stimuli and the canonically ordered sentences: “When someone says (b), what do they mean by it? Is there a difference in meaning between (a) and (b)?” These latter questions were crucial to inferring familiarity with and understanding of the NI and ME test stimuli. For example, most participants provided clear descriptions of differences between the NI and canonically ordered sentences. Or, they responded that there were no differences in meaning. On the other hand, many younger participants indicated that they had no idea what the ME sentences meant and so didn’t know what the difference in meaning would be. From these types of responses it is straightforward to infer that MEs are likely not part of the younger participants’ linguistic experience, while NI is much more likely to be so.

105 Pilot surveys were administered at Texas State University in San Marcos and Texas Tech University in Lubbock. 106 See also Labov’s (1972c: 111) Principle of Subordinate Shift, as cited in Sells et al. (1996: 593). As Labov writes, “When speakers of a subordinate dialect are asked direct questions about their language, their answers will shift in an irregular manner toward [or away from] the superordinate dialect.”

94

Chapter 3 Negative inversion, existential sentences, and definite subjects

3.3.3 Texas Chicano speakers For the Chicano speakers, I gathered data in a range of ways. For the younger speakers, I visited a majority Chicano high school in south Texas. There, I conducted surveys in much the same manner as described above for the Anglo high school students. For the older Chicano speakers, I visited retirement communities in the same community as the aforementioned high school. I also conducted individual surveys in person on the street to bring the participant numbers up to a sufficient level. While this method was ultimately fruitful, I was unable to obtain the spread of age ranges that I did with the Anglo community. Thus, rather than the four-way grouping seen above in Table 2, the Chicano data is divided into only young and middle-aged groups, as seen in Table 3. Table 3: Number and average age of Chicano participants. Average Age

Number in Group

.



.



3.3.4 Texas African American Vernacular English speakers For the Texas AAVE speakers, I also relied upon a range of collection methods. For the younger speakers, I visited a classroom at an HBCU in southeast Texas and conducted surveys comparable to those classroom surveys mentioned above with the Anglo and Chicano groups. For the older speakers, I conducted individual surveys on the street and in people’s homes in the greater Houston area. Similar to the description of the Chicano groups above, I was unable to obtain the age range I did with the Anglo community. As such, the African American data is divided into two groups, including younger and middle-aged participants, as seen in Table 4. Table 4: Number and average age of African American participants. Average Age

Number in Group

.



.



3 Introduction

95

As mentioned above, the African American and Chicano data, which are each divided into two age groups, do not allow for as fine a picture as seen with the Anglo data, which is divided into four groups. Nonetheless, the results from the AA and Chicano data to differing degrees do support the overall apparent-time hypothesis of expletive deletion across generations, which we turn to now, beginning with the Anglo speakers.

3.4 Generational results and apparent-time analysis The manner of diachronic hypothesis formation employed in this section, based on age-stratified variation in real time, is referred to as “apparent-time” analysis.107 It is a commonly used method to assess language change when it is unrealistic or impossible to gather longitudinal data over an extended period of time, or when there are no corpora available for the necessary range of times. This is precisely the situation with the ME and NI data. In this section, I show substantial decreases in familiarity with MEs across generations: i.e. older generations are more familiar with MEs than are younger generations. This suggests that a change has occurred across the relevant time span. The change in question here is deletion of the ME expletive, resulting in the extant NI construction. We look first at the Texas Anglo data, in Table 5. Table 5: Anglo familiarity with MEs and NI by age group. Age Group

Average Age

Number in Group

Familiar with ME

Familiar with NI

–

.





%

–

.



%

%

–

.



%

%

+

.



%

%

Table 5 presents the relative familiarity of NI and MEs across generations of Anglo speakers. As we can see, the youngest survey participants were uniformly unaware

107 The apparent-time construct has been used in many studies over the years, including Labov’s (1963) Martha’s Vineyard research. See Bailey (2002), Bowie (2005) and sources therein for history and criticism of the apparent-time analysis. Boberg (2010: 188) notes that apparent-time analysis is “largely valid for phonology and syntax” but that it can be problematic in consideration of the lexicon.

96

Chapter 3 Negative inversion, existential sentences, and definite subjects

of the ME sentences, though they were quite aware of the NI. As the age of the groups increases, we see a parallel increase in familiarity with the ME constructions, with the oldest group being 83% familiar with them. Meanwhile, there is broad familiarity with NI across age groups, ranging from 80–100%, suggesting that NI is still very much in the realm of experience for these participants – whether or not they admit to using it themselves. The variation in familiarity with the two constructions suggests that a change has occurred, or is occurring, with the ME over the second half of the 20th century. The most straightforward way to describe this change is the loss of the expletive subject in the ME. We find similar patterns in the generational data of the Chicano participants, which is shown in Table 6. As mentioned above, the Chicano data supports only a two-way split, between younger and middle-aged groups; however, the trend showing the decrease in familiarity with the ME across the generations is clear, with 41% of the middle group being familiar with MEs, and 6% of the younger group being familiar with them. At the same time, familiarity with NI is high for both age groups. Table 6: Chicano familiarity with MEs and NI by age group. Average Age

Number in Group

Percent Familiar with ME

Percent Familiar with NI

.



%

%

.



%

%

Finally, we see the African American generational data in Table 7, below. Here, we see much less difference across the generations. The younger group shows 14% familiarity with MEs, while the middle-aged group shows 23% familiarity, though both groups are 100% familiar with NI.108 As such, this data does not strongly support a claim that the construction has changed – i.e. the expletive 108 A further comment on inferring familiarity: as mentioned above, most of the young survey participants from all of the ethnolects responded to the ME questions with comments indicating they had no idea what the MEs meant, or that the MEs were garbled nonsense. One young AA participant responded, for example, “That don’t make no sense.” Several of the middle-aged Chicano and AA participants, however, responded in ways that indicated they understood MEs, whether or not they used them themselves. For example, one test sentence on the survey from Houston in October 2018 was “There didn’t nobody like him.” This time was in the middle of a highly controversial midterm election in Texas, and several respondents asked questions along the lines of the following: “You talkin’ about Trump, right?” or, “Is this about Ted Cruz?” An AA man told me that the only person for whom the sentence could actually be

3 Introduction

97

has deleted – across these generations, as can be argued much more clearly for the Anglo and Chicano groups. This data does partially support the general claim made by Martin and Wolfram (1998), Green (2014), and others that MEs do not occur in AAVE. It does not completely support that claim, though, because MEs are possible for some AAVE speakers, and as I show above in Chapter 2, MEs are attested for AAVE speakers in various lyrics, Twitter posts, etc. Table 7: African American familiarity with MEs and NI by age group. Average Age

Number in Group

Percent Familiar with ME

Percent Familiar with NI

.



%

%

.



%

%

What we do know about AAVE in general is that MEs are found in early instantiations of the language variety, such as the ex-slave recordings mentioned above. MEs also are found sporadically in contemporary AAVE varieties, as shown above in Chapter 2, though not as commonly as in the Anglo variety. It is worth pondering whether the expletive deletion process began earlier in the relevant AAVE varieties than it did in the Anglo and Chicano varieties under consideration. The addition of data from older generations of AAVE speakers, comparable to the Anglo data, could possibly illuminate the situation further. At the present time, however, I do not have access to this group of speakers. In the discussion that follows, I will assume that the claim of diachronic expletive deletion holds clearly for the Anglo and Chicano data, while holding out the likelihood that it also might for AAVE, though more data is needed to show this. Reconstructed modal existentials In addition to the survey data just considered, I worked closely with four older participants (two of whom were Anglo and two of whom were Chicano) and tested many reconstructed ME sentences with them, including the following examples from the Yale Grammatical Diversity Page for Negative Inversion. Thus, attested NI from the YGDP page in (144), were reconstructed into the MEs in (145). These older consultants were adamant that there was no difference in the meanings of NI and MEs.

true is the devil. These kinds of responses indicate grammatical comprehension and so allow the inference of familiarity with the construction. This contrasts with the young survey participants who frequently did not know what the ME meant at all.

98

Chapter 3 Negative inversion, existential sentences, and definite subjects

(144) a. Can’t nobody beat ‘em. b. Didn’t nobody get hurt or nothin’. c. Won’t anybody hit us. (145) a. They can’t nobody beat ‘em. b. They didn’t nobody get hurt or nothin’. c. They won’t anybody hit us. I provide only a few examples above; however, I and the older consultants went through many similar NI/ME pairs with the same result. Similarly, as a middle-aged native speaker of Texas AE, I have not come across any examples in any of the literature that refuse to take an expletive subject in this way. These findings parallel those of Zanuttini and Bernstein (2014: 163), who note that their Appalachian English consultants had minimal ability to point out any difference between NI and MEs. As these authors write: When asked about the difference between a sentence with [negative inversion] and one with an expletive, our consultants say that the one with the expletive allows for prosodic emphasis on the quantificational associate. In the example below, we represent such emphasis with capital letters: (i) Wouldn’t nobody go to see her. (ii) They wouldn’t NOBODY go to see her.

The difference in prosodic potential is the only difference Zanuttini and Bernstein’s consultants could provide. This is very little difference indeed, however, as NI itself seems to bear primary stress on the quantificational associate by default.109 It seems then that there is a strong case to be made for the diachronic deletion account. An important question that must be addressed, though, is why deletion occurred in the first place. While it cannot be known for sure, it is likely that deletion of the expletive is motivated by the need to be negative, following a pathway set out for language in general in Jespersen (1917).110 As Jespersen (5) writes: [T]here is a natural tendency, also for the sake of clearness, to place the negative first, or at any rate as soon as possible, very often immediately before the particular word to be negatived (generally the verb, see below). At the very beginning of the sentence it is found comparatively often in the early stages of some languages [. . .].

109 See Section 2.1.4 above. 110 Horn (2015) notes that NI follows the “Neg-First” principle, which is “the tendency for negative force to be marked as early as possible within a sentence.” See Horn and Wansing (2017) for further discussion of the Neg-First principle and its roots in Jespersen (1917).

3 Introduction

99

A very obvious way of following Jespersen’s tendency is to simply delete the expletive subject, as in There can’t anybody lift that rock. This is especially plausible given the semantic and phonological weakness of the expletive subject. As is noted repeatedly in the linguistics literature, expletive subject there is a semantically and phonologically bleached derivative of locative there (Huddleston and Pullum 2002; etc.). Thus, it should not be surprising for expletive there (or its dialectal cousin they) to drop off of the ME sentence completely over time. This idea is very much in accord with work in synchronic conversational deletion, in which sentence-initial material can be deleted in spoken English from a wide variety of sentence types. Napoli (1982), for example, following Dwight Bolinger, argues for a phonological rule that deletes the first part of a sentence “preceding the first main accent,” resulting in constructions like those in (146) [Napoli (85)]111: (146) a. Wish Tom were here. (I wish Tom were here) b. You seen Tom? (Have you seen Tom?) c. Good thing you decided to come along. (It’s a good thing you decided to come along) This pre-stress deletion would be precisely the type to result in an NI. Consider a pre-deletion ME such as (147a). The natural sentence stress falls on “nobody.” Thus, a rule such as the Bolinger/Napoli/Weir rule would delete content to the left of that. It makes sense too that the negativized modal would be spared the deletion, as emphasis of the negative is the most likely motivator for the deletion in the first place, resulting in the NI in (147b).112

111 See, for example, Thrasher (1974), Napoli (1982), Zwicky and Pullum (1983b) and more recently Weir (2012: 110) on conversational deletion in general. Weir (110) argues, following Napoli, and Zwicky and Pullum, that conversational deletion is post-syntactic. This means that though the phonological form in question has deleted, the syntactic content of it is still in residence. This would seem to be a natural step for the ME-to-NI story told here, in which the ME expletive drops off and the definiteness effects, such as they are, continue to be associated with the construction. 112 White-Sustaita (2010: 439) argues that expletive deletion cannot be motivated by stress in AAVE. As she writes: We also find evidence that the ban on expletives is not prosodic (contra Martin 1992), based on the ungrammaticality of expletives in embedded NI in AAVE. If expletive deletion were a phonotactic phenomenon associated with the left periphery, then an embedded NI clause with an expletive should be able to survive the deletion. Yet, as the following

100

Chapter 3 Negative inversion, existential sentences, and definite subjects

(147) a. They cain’t nobody win there. b. Cain’t nobody win there. Importantly, Negative Inversion is not alone in this behavior. We find similar deletion behavior in standard negative TEs as well. Consider (148a–b). It is perfectly possible for (148a) to undergo conversational deletion resulting in (148b) without altering the meaning of the sentence or speech act performed. (148) a. There isn’t a dog in the yard. b. Idn’ a dog in the yard.113 This same manner of conversational deletion is not available for positive TEs, as in (149a–b). This is a strong argument for the role negation likely played in the shift from Modal Existential to Negative Inversion: i.e. the absence of negation diminishes motivation for subject deletion and maximization of negative force.114

example demonstrates, an expletive in an embedded clause is still ungrammatical in [African American English]. (i)

They told me that *{it/dey}didn’t none of the children see anything . . . ‘They told me that there weren’t any of the children who saw anything . . . ’

The likeliest explanation for White-Sustaita’s pattern in (i) is that the diachronic process of expletive deletion is already completed in the variety of White-Sustaita’s informant(s), so the expletive is unacceptable in any context, whether embedded or not. On the other hand, it doesn’t hold of AAVE universally. See Chapter 2, Section 2.1.12 above for an AAVE counterexample and discussion. Similarly, as we have seen in Chapter 3, White-Sustaita’s claim doesn’t hold universally for the older African American, Anglo, or Chicano speakers in the present work. 113 An additional argument here in favor of the preservation of the negative can be seen in the difference between (i) and (ii) below. In (ii), which is a more natural locution in this variety than (i) would be, the fricative [z] has assimilated to the [-continuant] feature of the [n] of the immediately adjacent negative, further accentuating Jespersen’s early negation. (i) Isn’t anybody there. (ii) Idn’ anybody there. See Reynolds (1994) for further discussion of the Southern /z/ → [d] rule, in which it is suggested that this alternation is far more likely to occur in contracted negative auxiliary forms such as [ɪzn̩t] → [ɪdn̩] than in non-auxiliary items with similar phonological environments, such as [bɪznəs] → [bɪdnəs], which do occur but much less frequently. 114 Another clear example of this negative-existential deletion is seen in the lyrics of the Woody Guthrie song “Way Over Yonder in the Minor Key” mentioned in Chapter 2.

3 Introduction

101

(149) a. There is a dog in the yard. b. *Is a dog in the yard. The same deletion behavior is also seen in other types of negative sentences containing expletive subjects. Consider the extraposition sentences in (150). In (150b) the negative extraposition sentence can survive without the expletive; however, the positive sentence in (151b) is ungrammatical without the expletive. (150) a. It isn’t likely that Tony left. b. Idn’ likely that Tony left. (151) a. It is likely that Tony left. b. *Is likely that Tony left. The same is true for the negative weather it sentences in (152), but not for the positive weather sentence in (153b). (152) [Looking out the window] a. It isn’t raining now. b. Idn’ raining now. (153) [Looking out the window]115 a. It is raining now. b. *Is raining now. c. Raining now. The same is true of negative it-cleft sentences. (154) a. It wasn’t John that ate the pizza. b. Wadn’t John that ate the pizza.

There ain’t nobody that can sing like me Ain’t nobody that can sing like me Were the polarity of the sentences reversed, Guthrie’s lyrics would have been ungrammatical: There is somebody that can sing like me *Is somebody that can sing like me. 115 Note that it is possible to delete the expletive and auxiliary together, as in (153c), but it is not possible to delete just the expletive, as in (153b).

102

Chapter 3 Negative inversion, existential sentences, and definite subjects

(155) a. It was John that ate the pizza. b. *Was John that ate the pizza. In each case, the expletive subjects can only be deleted when the following auxiliary is negatively inflected. To push the issue further, however, we find this same correlation of negation and subject-deletion in many other sentences, whether the subject is an expletive or not. In the examples, below, the negated sentences are unobjectionable, while the positive sentences are unacceptable (or, they are only acceptable in a humorous telegraphic speech style). (156) Can y’all come over later? a. Nope. Can’t make it. (We can’t make it) b. Sure. *Can make it. (157) Is that how you’d do it? a. Nope. Wouldn’t do it that way. (I wouldn’t do it that way) b. Yep. *Would do it that way. (158) Is John around? a. Nope. Idn’ here right now. (He isn’t here right now) b. Yep. *Is here right now. In all of these examples and more there appears to be a type of synchronic conversational deletion, working in accord with the maximizing negation. It is straightforward, then, to suppose that a comparable deletion process has been occurring with the Modal Existential in the last few generations, with the difference that the reduced form – i.e. the NI – has become the main form for the younger speakers, while the non-reduced form – i.e. the ME – is no longer available to that group. For the diachronic purposes of this book, it is not necessary to fully understand or argue for the particulars of the deletion; it is enough to see that it is a very common phenomenon and that, when coupled with the negative tendencies of Jespersen (1917), can provide a straightforward account of how we got to where we are with the NI data.116 Combined with the

116 A question one might ask here is why would the reduced form (i.e. the NI) become the conventional form, but the reduced forms of the other deletion examples illustrated above, such as the cleft, the extraposition, etc., did not. One big difference is that the ME and NI are both non-standard, socially marked forms, while the cleft, extraposition, and others, are not socially marked and so are found across a wider range of spoken and written genres. This would very likely reinforce their status with the non-reduced form.

3 Introduction

103

generational data in which only the elder participants were familiar with the MEs, it is very likely that the ME construction has undergone systematic expletive deletion, resulting in the present-day NI. Importantly, this hypothesis accounts for the extant definite effects still seen in the NI. Thus, the definiteness effect in the syntax remains in residence with the NI, even as the expletive subject itself has been deleted, in effect concealing an ME behind the missing expletive subject. A useful historical analogy can be seen here with nasal vowels in contemporary French, in which certain vowels retain their nasal character long after the nasal consonant to which they originally assimilated has dropped off. The nasality, however, remains behind on the vowel, reminding us of what the earlier coda structure of the syllable had been.117

3.5 Status of expletive subjects and negative inversion In the diachronic account suggested in this book, I assume that the expletive subject is no longer part of the NI construction for the younger speakers – primarily because there is little availability for it to surface in ME form for that group. Others have suggested synchronic accounts in which the expletive subject is phonologically null but syntactically present. For example, Parrott (2000) argues that AAVE has a lexical item that equates to a phonologically null expletive subject: i.e. a silent there. Martin (1992) suggests for Appalachian English that an expletive “subject [could] delete at PF [. . .] because of the heavier stress which is assigned to the Neg Infl and to the following NP” (Martin 1992: 65). In these cases, the phonologically null expletive would retain its syntactic and semantic properties, leaving a sentence that appears to be an NI but that is actually an ME in covert form. I agree with the spirit of Martin’s and Parrott’s hypotheses, but disagree with how they are operationalized. If there were a phonologically null expletive subject in NI, we would expect it to surface with all ages of speakers and to participate perhaps in predictable ways in subjecthood tests such as tag questions; however, as is shown below, this is not the case. Foreman (1999: 8), for example, argues that NI is unlikely to be the result of expletive subject deletion. (His interests are synchronic, however.) His argument here relies primarily on data from tag questions. As he writes,

117 See, for example, Rochet (1976) and the plentiful references and discussion included in that work.

104

Chapter 3 Negative inversion, existential sentences, and definite subjects

[T]ag questions provide evidence that underlying subjects of NIs are in fact in the normal syntactic subject position [. . .] If [NI] sentences are derived by eliminating expletive pronouns from [subject position] then we would expect the expletive to resurface in tag questions. However, tags of NI sentences show an agreement with the true subject of the NI sentence, instead of introducing a dummy pronoun.

Consider (159a–b) [Foreman’s (17c–d), with Foreman’s judgments]: (159) a. Ain’t no man gonna cheat on a woman like that, is he? b. “*”Ain’t no man gonna cheat on a woman like that, is there? According to Foreman, the subject of the tag question reflects the subject no man, rather than an unarticulated expletive subject. As such, Foreman marks (159b) ungrammatical and concludes that NI is not simply a synchronically truncated form of the ME construction. I agree with Foreman’s judgments in (159a), but do not agree with his judgment on (159b). In fact, I find both examples in (159) to be perfectly acceptable and can straightforwardly provide other pairs indicating the same judgment patterns, as in (160a–b). (160) a. Didn’t no man lift that rock, did he? ‘No man lifted that rock, did he?’ b. Didn’t no man lift that rock, did there/they? ‘No man lifted that rock, did there/they?’ So, it seems that tag questions targeting either “subject” are possible in NIs such as (159) and (160). Further, the same is true of the ME constructions, in which the expletive subject is fully articulated, as in (161) and (162). (161) a. Couldn’t no man lift that rock, could he/they? ‘No man could lift that rock, could he/they?’ b. They couldn’t no man lift that rock, could he/they? ‘There is no man who could lift that rock, could he/they?’ (162) a. Ain’t no man would cheat on a woman like that, would he/there? ‘No man would cheat on a woman like that, would he/there?’ b. They ain’t no man would cheat on a woman like that, would he/they? ‘There is no who would cheat on a woman like that, would he/they?’ The indeterminate results of the tag question in these examples raise a couple of questions. On the one hand, the results do not provide evidence for a covert,

3 Introduction

105

yet syntactically present, expletive. If there were a covert expletive, we would expect it to surface as the subject of the tag question on every occasion – if indeed tag questions work as Foreman assumes. The fact that the expletive is present as the tag subject at the same time that the anchor subject is present in the tag subject, however, suggests that there must be something else going on. One possibility is that the construction is in transition with the loss of the expletive. As such, “both” subjects might be equally available for some speakers, or available for some and not for others. Another possibility is that the tag question subjects of these constructions are created and linked pragmatically to some type of discourse referent, depending on the exigencies of the discourse. For example, in Huddleston and Pullum (2002: 893), the authors note that some tag questions are sensitive to communicative – i.e. pragmatic – meaning rather than syntactic identity. In their description of (163), below, they write, “Grammatically, it’s legal is subordinate to the think clause, but communicatively it is the subordinate clause that is primary [. . .] the form of the tag reflects the communicative meaning rather than the grammatical structure [. . .].” (163) I think it’s legal, isn’t it? In a bit more detail, Sailor (2009: Section 3.2) provides a range of tag question data in which there is no syntactic parallelism or identity with the main clause subject, arguing instead that “tag questions are simply adjoined yes/no questions that undergo [verb phrase ellipsis] by way of their close semantic and pragmatic relationship to the clauses that host them” (Sailor 2009: 8). In other words, as Sailor goes on to write, “[T]here is no independent requirement on ‘tag question identity’ that needs to be satisfied. Instead, the syntactic form of a tag question follows straightforwardly from pragmatics and ellipsis.” This way of understanding tag question formation, in which there is no syntactic identity expected, would account for the varying intuitions shown above with NI subjects. One last possibility to mention here with respect to NI subjects and varying tag question intuitions comes at the problem from yet another direction. Siemund (2013), writing on a range of vernacular Englishes, provides the data below in (164), and suggests that “the agreement properties found in Standard English tag questions are often relaxed or simply non-existent in vernacular varieties.” (164) a. We saw some the other day, isn’t it? [Welsh English] b. You said you’ll do the job, isn’t it? [Indian English] c. Yall didn’ buy no clothes from town, inni? [Gullah]

106

Chapter 3 Negative inversion, existential sentences, and definite subjects

Accordingly, if the Texas vernaculars under consideration in the present book do not require strict identity in tag question formation, it would appear that they are not alone in doing so. Ultimately, I view tag question diagnostics as not informative for the purposes of the present work, which assumes that the expletive subject has been lost over time. Tag questions do not provide an argument in favor of the anchor subject as the main subject, pace Foreman (1999) and Matyiku (2017); neither do they provide evidence for a covert expletive subject, pace Martin (1992) and Parrott (2000).118 Older speakers accept both modal existential and negative inversion constructions A further question that arises is that of the older speakers who accept both MEs and NI. What is happening with these speakers? Recall that the youngest speakers rejected MEs almost completely, while familiarity with the constructions increased parallel to an increase in age. Thus, the oldest speakers were very familiar with both constructions. As a middle-aged native speaker of Texas AE, I too have intuitions about MEs and NI. I assume that what is really occurring here with the older speakers (myself included) is that there is actually one construction: i.e. the ME. However, the presence of Jespersen’s need to be negative and the application of synchronic conversational deletion simply makes it appear as if there are two constructions: i.e. the ME and the NI. For the younger speakers, however, for whom there is only NI and no ME, the diachronic process of expletive-deletion is complete.119 It thus seems unlikely for these younger speakers that NI is simply the result of a de-stressing and phonological deletion rule, as suggested in Martin (1992) – although this could very well be what is happening with the older speakers.

118 The results of this section do raise questions regarding the adequacy of the diagnostic for subjecthood in this language variety, though, and perhaps others. See Croft (1991a) for lengthy discussion of the necessity for different syntactic tests across different languages: i.e. the idea that diagnostics in one language don’t necessarily apply to or pick out the same categories in another language. 119 It is well known that language shift can take place over three generations. Cheshire (2005) gives the example of Generation 1 arriving in London from Turkey and beginning to acquire English as a second language. Generation 2 might then be bilingual in Turkish and English, while Generation 3 might speak only English. The constructional change we see here with Modal Existentials and Negative Inversion seems to follow a similar generational timeline.

3 Introduction

107

3.6 Negative inversion, negative existentials, and conversational deletion It is worth mentioning briefly that an advantage of the diachronic account given in this chapter is that only one mechanism is needed to account for both Negative Inversion (165) and what are frequently referred to as negative existentials, as in (166): namely, the independently attested mechanism of conversational deletion. (165) Can’t nobody beat ‘em. (166) Ain’t no black Santa Claus. [Sells et al. (1996: ex. 36)] The question of a relation between these two sentence types goes back to Labov et al. (1968), who ultimately suggest that two different accounts are necessary: i.e. an inversion account for sentences such as (165) and an expletive deletion account for existentials such as (166). Some researchers, such as Sells et al. (1996), have attempted to formulate a single solution for both sentence types. On the other hand, some accounts of Negative Inversion simply do not address existential sentences such as (166), with the account limited to only true NI. These accounts simply observe that the (166) existential sentences are different sentence types and so are not part of the account in question. Here is WhiteSustaita (2010: 432), for example: These constructions differ from [NI] insofar as they permit expletives across all dialects [. . .] and they are available in any number of non-canonical varieties of English, which do not allow [NI], as well as colloquial fast speech canonical English, which does not allow [NI]. I will argue that true NI results from a separate syntactic operation from these existential constructions.

As White-Sustaita mentions, it is common to hear sentences such as (166) in casual spoken English of many varieties that do not also allow NI. This is a strong argument that the two are different constructions. For the present purposes, I assume that NI and negative existentials are closely related constructions, and that one can tell us much about the other as they share a strong family resemblance. Beyond this, it is only necessary to note that both are, or have been, subject to the independent process of conversational deletion, which accounts for their similar surface forms and absent expletives.

3.7 Conclusion I have argued that Negative Inversion in at least AE and CE is actually a diachronic descendant of the Modal Existential construction: essentially, the expletive subject

108

Chapter 3 Negative inversion, existential sentences, and definite subjects

of the former has undergone deletion over time. This argument is supported by quantitative data in an apparent-time analysis, and it provides a compelling account of the origins and distribution of the so-called “definiteness effect” in the NI. I believe this account also holds for the African American constructions, though my data for this variety is not as rich as that of the AE and CE. Most sources agree that AAVE did host Modal Existential constructions in the past, as the various ex-slave narratives such as Bailey et al. (1991) clearly show. I think it is likely, though, that the ME faded earlier from AAVE than it did from AE. I predict that a survey of older AAVE participants than I have provided here might show even greater familiarity with the ME, and so provide further support to the general claims of this chapter. I also argue above that the definiteness effect in question with NI is quite different from what has been maintained in the literature for the last five decades. It has been assumed that definite subjects are prohibited in NI; I have shown that definite subjects are in fact allowed but that they are subject to pragmatic constraints, which are the same as those that constrain definite subject NPs in there-existential sentences. It is possible that the definite subject NPs shown in the Texas NI constructions are a point of micro-variation. It is also possible that the pragmatic constraints described herein have contributed to the illusion of strict prohibition over the last five decades.

Chapter 4 Negative inversion and its discontents Whoever uses sentences like those is ignorant and lazy. Probably voting for Trump. – Respondent 52 Absolutely uneducated. Typical Clinton voter. – Respondent 67

4 Negative inversion is stigmatized A sentiment that has been made very clear in just about every survey or interview I have conducted for this work is a perception of NI as inappropriate, incorrect, bad grammar, and just plain bad. Descriptions like the following were ubiquitous. NIs are: the way rednecks talk; bad grammar; uneducated; the speech of gang members; speech of old hillbillies; low class; rural; inner city; lazy; bastardized; ignorant; and more. Essentially, my participants viewed NI as bad English grammar from English dialects that are themselves bad. Very few of my survey respondents were neutral on the construction. Here, for example, is a longer quote taken from the survey of a 35-year-old, lower middle class, Anglo male. [NI is used by] redneck, male or female, lower class and i think rednecks are only white. although the guy i work with is 35 and he would definitely say something like this. He is as texan as it gets, he ranches, farms, bails hay and his family has farmed ranched and bailed hay for generations. 7 out of 10 people in the county are related to him. he has managed to knock himself unconscious with a hammer while repairing a fence, gotten stuck in a tree upside down and unable to extract himself without the aid of his wife and a tractor, damaged all the nerves in his left forearm from a tractor repair accident and numerous other things i cant remember at the moment.

Even so, when pushed a little further, participants often mentioned other values they associated with the construction. NIs are: just an informal way of talking; my uneducated family members talk like this and I have no trouble understanding them; I use this kind of stuff when I’m comfortable; just for use with friends; this stuff is okay but not by the book; people just use them out of habit; and so on. Here is a survey quote from a lower-middle class African American female in her mid-40s. [The NI] is okay for television, or on the street, if you’re listening to ebonics. It’s improper though and bad grammar, but it’s normal talk for a lot of people.

https://doi.org/10.1515/9781501512346-005

110

Chapter 4 Negative inversion and its discontents

So, NI is seen to be improper grammar, but acknowledged to exist and to be used in specific times and places. This was a common sentiment among participants across the three varieties. Less often, but still more than occasionally, upon deeper reflection, NI sentences were afforded a bit of positivity, described as: what kids say when they wanna sound bad; language of young people trying to be cool; people trying to fit in like tough guys; a young lady tryna be sassy; no formal education but that doesn’t mean they’re not smart!, etc. Here is a longer quote from a lower middleclass Anglo female in her mid-40s: [NI speakers] in Texas are 65–90 years of age, male, working class as a rule, or educated working class for funsies. Caucasian. Possibly females of the same class 80+ years of age. Sometimes young male rednecks, working class or no class. In other states this particular way of speaking is affected by inner city hip-hop punks, i.e. young male African Americans, or Caucasians in rap outfits who wish they were. Any socio-economic class at all. But it isn’t real.120

And so appears a familiar picture with NI. They are highly stigmatized but acknowledged to be used in non-formal situations, and there is also a certain cool factor associated with them. This is by now a well-established pattern in the sociolinguistic and language-attitude literature in general, especially in the literatures on AAVE, creoles, and vernacular languages, which are often associated with both negative and positive qualities. See Wolfram and Schilling (2016) as well as Section 4.1, below.

120 This survey participant suggests that for some of the young folks who use NI, “it isn’t real.” This participant believes NI isn’t part of these speakers’ native varieties but that they are borrowing as an affectation, or, to be cool. This motivation for borrowing is well known in general. See, for example, Eckert’s (2004: 48) discussion of final [t] release among American speakers, which she argues is borrowed from British varieties of English to invoke social prestige. British speakers of English regularly release their [t]s, and Americans adopt this feature when imitating British English. The age-old stereotype of the British, and British English, as superior, intelligent and educated, is no doubt at work here. A distinction made at the international level, then, seems to be providing American speakers with a resource for signaling superiority of a variety of sorts, but it seems to be primarily limited to intelligence, education, and articulateness. Interestingly, we see this same kind of borrowing with NI, which is a testament to its social marking. For example, a recent song by 2NE1, which is a Korean pop group, has the lyric “Can’t nobody hold us down.” Undoubtedly, there is an intention here on the part of the Korean artists to borrow on the hipness factor of AAVE and American hip hop. The point is that the NI is recognized enough as a marker that the K-pop artists deem it useful to borrow in the first place.

4 Negative inversion is stigmatized

111

Let’s look at a few more quotes from survey participants, which build in all of the aforementioned qualities. The following comes from a Latino police officer in his early 40s. When I read [the NI], I immediately thought of some dumb backwoods redneck. The movie Deliverance came to mind.121 I was brought up in a middle-class family and honestly can say that I haven’t heard anyone speak like that in the circles I grew up with. As a police officer, I’ve worked all areas of town. I would say that phrases similar to that can be heard among gang members (Hispanics and African Americans) and their associates. Their socio-economic status is typically considered lower class with minimal education. I would go so far as to say that they know how to speak formally but choose to speak that way as it seems to be gang culture.

In this quote we see both the negative stigmatization – attributed to both rednecks and gang members – as well as the acknowledgment of NI as dangerous and cool. Speaking that way is a choice. Below we see additional comments that were left on one of the surveys I posted in a Facebook group, in which the same sorts of themes are developed.122 The Facebook group in question is dedicated to people who were born and raised in Corpus Christi, Texas, which is also the hometown of the author of this book, a fact which was made known to the participants in the posting of the survey. (167) Carrie: Unfortunately, I have heard this type of speaking. Laziness. Melissa: I completed the survey. I have heard basically all of those. . .. grates on my nerves. Jane: [To the researcher], I was actually shocked by your survey. I grew up in Corpus Christi. The only place I ever heard any sentence construction that was that thoroughly colloquial was in DEEP East Texas where my mother was raised. Her significantly less intelligent sister and her husband regularly bastardized the English language. They were both white. What is your research trying to identify or correlate?123

121 Note that the 1972 movie Deliverance was invoked by several other participants as well with respect to NI. See, for example, the discussion in Chapter 2, Section 2.1.12. 122 Names have been changed to hide commenter identities. Gender is consistent with the actual names, and the posts themselves are verbatim quotes of the originals. All of these posters appear to be Anglo, except for Susan, who is Latino. 123 “Deep East Texas” is typically believed to include Houston and areas north and east of there toward Louisiana – a region located approximately 3–5 hours away by car from Corpus Christi.

112

Chapter 4 Negative inversion and its discontents

Mary:

Susan:

My mother grew up in truly deep east Texas. Not an overly educated spot. I remain completely offended that [the researcher] felt compelled to posting this survey on our site. You’re dead on, Jane. LMAO. . ..don’t be offended Jane. [The researcher] is offended because he moved up North. They probably made fun of his Native Language. East Texas does have that deeper draw. Them Yankees wish they could talk like us. Texas is its own lil world. That survey is funny.

As a result of this series of comments, I apologized for any offense that was caused, and pulled the survey from the group immediately, as I believed the negativity of these comments had compromised the survey for any subsequent participants who might have wished to take it.124 In this exchange, though, we see outright hostility toward NI, and, worse, toward the posters’ assumptions that I might think they use NI themselves.125 In this hostility, though, is a tacit admission of familiarity with NI, and an exhibition of a linguistic self-loathing that is a testament to just how stigmatized NI can be for some speakers.126 We also see a glimpse at a kind of covert prestige in the last quote: “Them Yankees wish they could talk like us. Texas is its own lil world.” This idea of covert prestige will be important to arguments made in Chapter 5, and it is thus worth exploring in more detail in the following section.

124 When I posted the survey to the other groups (i.e. Abilene, Odessa, and a different one from Corpus Christi), I was careful to disable comments on the Facebook posting to which the survey was linked to prevent this type of occurrence in those groups. 125 Fortunately, this particular online exchange is the only time I faced such hostility. All of the in-person surveys, classroom surveys, and other online exchanges I have conducted over the years of this project have been quite pleasant. 126 This is not unlike Labov’s (1966/2006) findings of some New Yorkers’ perceptions of their own speech. Labov (329–330) writes: New Yorkers show a general hostility towards New York City speech which emerges in countless ways. The term ‘linguistic self-hatred’ is not too extreme to apply to the situation which emerges from the interviews. [. . .] The terms which New Yorkers apply to the speech of the city give some indication of the violence of their reactions. It’s terrible. Distorted. Terribly careless. Sloppy. It’s horrible. Lou-zay.

4 Negative inversion is stigmatized

113

4.1 Overt and covert prestige In the contemporary sociolinguistics literature, the notion of covert prestige is usually assumed to begin with observations made in Labov (1966), and has been the subject of an immense literature since that time. Covert prestige contrasts with overt prestige. Linguistic forms which are overtly prestigious tend to be those associated with high status groups and are frequently those “perpetuated by the agents of standardization in our society – teachers, the media, and other authorities who hold themselves responsible for setting standards of linguistic behavior” (Wolfram and Schilling 2016: 175). On the other hand, it is often the case that speech styles or grammatical forms that are stigmatized by standard speakers will carry a type of covert prestige, among the speakers who use the forms or speech style. As Wolfram and Schilling continue: The notion of covert prestige is important in understanding why vernacular speakers do not rush to become [Modern American English] speakers [. . .]. Features that are widely socially stigmatized, such as negative concord, subject-verb agreement, and different irregular verb paradigms, may function at the same time as positively valued, covertly prestigious features in terms of local interactional norms. Furthermore, covert prestige can actually become quite widespread and well recognized running as an undercurrent beneath the norms of mainstream societal institutions, as in the case of widespread use of features of African American Language by young people of various ethnicities and social status affiliations who may use the features to index lack of concern with, or even defiance of, mainstream norms.127

In this long quote, we see answers to an important question that arises immediately with respect to prestige and Negative Inversion. That is, if NI is so stigmatized, why would anyone use it in the first place? The answer, of course, is that NI indexes a more local kind of prestige and use of the constructions indicates group membership, belonging, and solidarity, as well as possibly rejection or defiance of standard norms.128 We can see all this in the specific construction of NI,

127 An early description of this last fact is found in Sapir (1931) who writes: A word may be added in regard to the social psychology of dialectic forms of speech. In the main, markedly dialectic peculiarities have been looked upon as symbols of inferiority of status, but if local sentiment is strongly marked and if the significance of the local group for the larger life of the nation as a whole allows, a local dialect may become the symbol of a kind of inverted pride. [Reprinted in Sapir (1949: 88)] 128 Here is Sapir (1931) again with a very early take on motivations for the survival of stigmatized speech: Obviously the question of the conservatism of dialect is not altogether a negative matter of the inertia of speech and of the failure of overriding cultural influences to permeate

114

Chapter 4 Negative inversion and its discontents

but it is really part of a much larger pattern, as a very common situation among vernacular and creole languages of all stripes. Smitherman (2006: 6), for example, uses the term “linguistic push-pull” with respect to AAVE, to describe “Black folk loving, embracing, using Black Talk, while simultaneously rejecting and hatin on it.” Rickford and Traugott (1985: 252) describe the “paradoxical combination of negative and positive attitudes which are found in communities where pidgin and creole varieties of English are spoken.” This situation of paradoxical attitudes is a general tendency across vernaculars and creoles, but it is reflective of what we see in the specific instance of NI across the Anglo, African American, and Chicano speakers of the present study as well. The source of the negativity associated with NI is clear. It is clearly identified with a way of speaking and with specific groups of people, which are themselves stigmatized. As is well known in the language attitudes literature, negative perception of a grammatical form or speech variety is very likely rooted in a negative perception of the speakers themselves.129 Let’s consider, then, who the users of NI are. First, perhaps, NI is a marker of stigmatized regional speech in general. As Cukor-Avila et al. (2012: 17) note, “Texas speech is rarely (if ever) perceived as “standard” or “correct” by non-Texans (cf. Preston 1996; Fought 2003).”130 Within this regional stigma, though, NI further marks lower social class – a quality which is itself frequently stigmatized.131 Continuing on, NI also marks an informal style within a stigmatized regional dialect and social class. As is well known in the sociolinguistics literature, informal styles are frequently highly stigmatized. So, NI, as a clear indicator of these multiple stigmatizing factors, has triple the chance of bearing a negative stigma itself – and this is only for NI of the Anglo

into all corners of a given territory. It is, to a very significant degree, a positive matter of the resistance of the local dialects to something which is vaguely felt as hostile. This is easily understood if we look upon languages and dialects not as intrinsically good or bad forms of speech but as symbols of social attitudes. [. . .]. The home speech was associated with kinship ties and with the earliest emotional experiences of the individual. [Reprinted in Sapir (1949: 87)] 129 As Preston (2006: 40) writes, “It is perhaps the least surprising thing imaginable to find that attitudes towards languages and their varieties seem to be tied to attitudes towards groups of people.” 130 See Preston (2015, 2018) and the many sources therein on the nature of stigmatization of Southern US English in general. 131 The stigma associated with working-class speech has been well known since at least Goffman (1963). See also the linguistic inferiority principle of Wolfram and Schilling-Estes (1998: 6) in which “the speech of a socially subordinate group will be interpreted as inadequate in comparison to that of a socially dominant group.” See also Fought (2006: 54–55) and (Lippi‐Green 2012: 70) on this idea.

4 Negative inversion is stigmatized

115

variety.132 Add to this the fact that many survey participants across all three ethnicities saw NI as a marker of African American speech as well – even though they themselves used it – and an additional source of NI stigma can be added to the mix. NI bears similar marking on the African American side of things as well. In this case, it is not regionally bounded, as NI is commonly found in AA speech around the US, the construction is, however, socially marked in a range of ways, none of which mark it as belonging to the African American Standard English described in Spears (2015). NI is seen as variously uneducated and working class, but also as hip. My AA participants mostly considered NI to be part of uneducated and working-class speech, but the fact remains that they are ubiquitous in mainstream hip hop and AA popular culture, which gives them a certain cool factor. It seems to be a standard sort of relationship of overt and covert prestige that is frequently seen in vernacular languages and among differing creole strata.

4.2 Negative inversion attitudes and meanings Throughout the course of the fieldwork undertaken in support of this book, I have surveyed, interviewed, and spoken both formally and informally with approximately two hundred native speakers of the Texas English varieties in question. These speakers came from a variety of backgrounds but were essentially uniform in their attitudes toward NI. Almost everyone indicated that NI is nonstandard language, indicative of low socioeconomic class, and whether rural or urban, indicated a lack of education. As described above, though, wrapped up in this characterization was an additional sentiment of familiarity and an often grudging admission that NI was believed to be cool by some speakers. In this section I present the results of one particular survey question across the three speech groups to illustrate these facts. The surveys targeted several different semantic, syntactic, and pragmatic aspects of the NI construction; I report here, however, a question that speaks directly to the meaning of NI and its

132 For the NIs that also contain instances of negative concord, which is itself a stigmatized feature, such as (ii), the negative regard toward the NI is likely even stronger. I did not specifically test for this in any of my surveys, which were composed of equal numbers of NI with and without negative concord. Interestingly, though, I frequently had survey participants “correct” what they viewed as the improper grammar of NI examples by removing the negative concord, essentially changing sentences like (ii) to those like (i). (i) Can’t anybody lift that rock. (ii) Can’t nobody lift that rock.

116

Chapter 4 Negative inversion and its discontents

non-inverted counterpart. These particular surveys were conducted in an HBCU in east-central Texas, a rural Chicano-majority high school in south-central Texas, and a rural Anglo majority high school in central Texas. The teachers in all cases were gracious enough to allow me an entire class period in which to survey and to talk about language variation. This amounted to approximately 10 minutes of introductory discussion on language variation in general, 15–20 minutes of survey time, and then 30–50 minutes of post-survey discussion. In each case, upon entering the classroom, I was introduced by the host teacher or professor as a visiting speaker from the University of Minnesota who had come to talk to the students about linguistics. In each case I then introduced myself further, with a short discussion of my own background as a native Texan, my ties to the particular communities in which I was speaking, and my professional interests in how Texans spoke. At this point, I provided an overview of some ways in which English varies around the United States in order to pique the participants’ interest in the topic of variation. For example, I started with varying pronunciations: bag is pronounced [beg] and coffee is pronounced [ˈkæfi] in Minnesota,133 as well as lexical differences, rubber bands can be called gumbands in Pittsburgh, the regional distribution of coke/pop/soda, and so on. I also briefly mentioned some grammatical variation, such as The car needs washed and positive anymore from Western Pennsylvania and Ohio. My hope with this language-variation warm up was that it would prepare the students to talk more neutrally about variation in their own speech varieties.134 After the introduction and variation overview, I handed out a survey containing approximately three pages of questions targeting a range of syntactic, semantic, pragmatic, attitudinal, and demographic issues. There were several

133 At least in the parts of Minnesota currently experiencing the Northern Cities Vowel Shift. See Labov et al. (2006). 134 Many of these students indicated that they had never traveled outside of the state. As such, they have little exposure to regional variation in US English speech. Most of them were astounded that English varied so much in the US. Discussion of the Minnesota [ˈkæfi] pronunciation and the positive anymore of the Midwest had them laughing out loud and trying out the pronunciations and grammatical forms for themselves. At the same time, the students are also aware that the speech in which they are immersed in their daily lives is viewed as “improper.” My hope with the variation warm up was to try and soften some of this linguistic insecurity and so create a more open and welcoming environment in which to conduct the survey. See Labov (1966) for an early discussion of linguistic insecurity. See Macaulay (1975) for suggestions on how to conduct research in such an environment, an important one of which is that the researcher have open discussions of the topic with members of the community in question, as is done here.

4 Negative inversion is stigmatized

117

different types of question on the survey, which required participants to give short descriptive answers, to assess the felicity of conversational exchanges, and to rate the acceptability of sentences and contexts on Likert-style scales. Question (168), which asks for a short descriptive answer, deals specifically with the meaning differences between NI and its non-inverted counterpart. As we see below, a strong majority of all three groups said there was no basic difference in the meanings of the two constructions, but that there were clear differences in what I am calling social content. (168) When someone makes a statement like (a) below, what do they mean by it? Is there a difference in meaning between statement (a) and statement (b)? If so, what is the difference? (a) Didn’t nobody come to the birthday party. (b) Nobody came to the birthday party. The African American group consisted of 21 university students, with an equal gender distribution and an average age of 20.2 years. All participants were from central and east Texas. Of these 21 students, 18 (86%) indicated that the sentences mean essentially the same thing, but that (a) would be more likely to be used by rural or uneducated folks and also in situations that were informal, comfortable, more dramatic, or with friends and family, while (b) was proper grammar, and would be used in more formal situations, etc. The Anglo participants consisted of 19 high school students with a balanced gender distribution and an average age of 16.8 years. All of these participants were from a majority-Anglo, rural west-central Texas town. Of these 19 students, 15 (79%) indicated that there was no difference in the meaning of (a) and (b). And as with the AA group, many participants in this survey also indicated differences in social class, education, formality, proper grammar, and so forth. Finally, the CE group consisted of 33 high school students in a Chicanomajority high school, with slightly more female participants, and an average age of 16.2 years. Of these 33 participants, 27 (82%) indicated that there was no difference in the meaning of (a) and (b), and they had similar comments on NI being used at home, on the street, etc., but not being proper grammar. Several of these participants also noted that they associated NI with AAVE and Anglo speakers more than with Chicano speakers.135 Thus, among all three groups we see a pair

135 One student, for example, as he was taking the survey, raised his hand for me to approach his desk. When I did so, he whispered, furtively, “Sir, is it okay if I say these sentences are like what black people use?”.

118

Chapter 4 Negative inversion and its discontents

of sentences that appears to vary little in terms of truth conditional content, but which triggers associations of style, setting, felicity, emotion, and so on. After each of the three groups above had completed their surveys and returned their forms to me, we discussed the results of the survey, with students sharing their thoughts publicly with the rest of their class. This was done to make the goals of the research clearer to the students and to the hosting teachers, but also as a way of gathering further information on my part.136 Openended survey questions such as (168) are not as precise as, say, a question of forced-choice; however, what I was looking for here was actually a wider probe of the NI’s potential meaning. In what ways does it differ from the non-inverted form? Participants’ initial impressions were recorded on the survey, and these were then followed up on in the subsequent class discussions. When the surveys had all been handed in, I would then ask if there were any follow up questions about anything from the survey. There were usually several questions, which would often lead to lengthier discussions in general. During this time, I would also write pairs such as (169a) and (169b) on the chalkboard or marker board at the front of the room so I could point to them and we could talk about them as a class . (169) a. Didn’t nobody come to the party. b. Nobody came to the party. As these discussions heated up, I would throw out scenarios to the class and ask if the NI on the board and its non-inverted counterpart could both be used to describe them. For example: “So, what if I have a birthday party, and not a single person shows up. I mean, it’s absolutely empty in there. Not a soul. Can I describe that with a Negative Inversion sentence like (169a)? What about (169b), which is not inverted? Can I use that?” The discussion continues: “What about if there was just one person that showed up for the party? Then could I use (169a)? What about (169b)? What would it have to be like for you to use (169a)? What if (169a) is true, then could (169b) be false at the same time?” I also pressed further on social issues during this time. “Who would use (169a)? Who would you say it to? When would you say it? Could you say it to your grandma? Could your grandma say it?” And on and on. These bull sessions were lively, with questions and answers flying, side arguments raging, one student saying he would never use the improper (169a),

136 An additional goal here was to spark an interest in linguistics, perhaps, among the students, as these discussions can be quite interesting and entertaining.

4 Negative inversion is stigmatized

119

another student calling him out as stuck up, and so on. At the end of the day, however, there was no occasion in any of the classrooms in which we could come up with a scenario requiring differing truth conditions of the NI and its non-inverted counterpart – i.e. a case where one would hold and the other would not. This is telling, because it often became somewhat of a game for the students to be able to do so. The differences between the two constructions always ended up coming down to formal and informal style, stigmatization, education level, emphasis, emotion, situational suitability, and so on, but not to anything that could point to a propositional descriptive difference.137 These results of these free-flowing discussions were more informative even then the quantitative survey results reported above – which is a testament to the importance of mixed methods in this kind of research. With the survey data, with each participant answering privately without consulting others, it is very possible for a participant’s answer to be affected by prescriptive attitudes toward the subject matter, or for the participant to simply not understand what information is actually being sought.138 However, I believe that this is less likely in the bull session scenario described above, as scenarios were created and negotiated publicly by multiple participants and “tested” in front of – and with the input of – the rest of the participants in the class. Every response or claim was negotiated with the class until a satisfactory conclusion was agreed upon. The result was that no truth conditional difference between the NI and its non-inverted counterpart could be identified by any of the survey groups, pace what is predicted in Green (2014) and Green and Sistrunk (2015) for such constructions in AAVE. Though no difference in truth conditional meaning could be determined between NI and its non-inverted counterparts, as was mentioned above, a range of differing non-truth conditional, social meanings were identified. I will argue below that these meanings are a conventional part of the NI construction. What we will see with NI then is a construction with a simple truth conditional semantics, roughly equivalent to that of its non-inverted counterpart; however, the construction is also saturated with a range of non-truth conditional semantic meanings, which speakers can lean upon to generate pragmatic implicatures in context, an operation which will be the topic of Chapter 5, below. Meanwhile,

137 As a native speaker of one of the relevant dialects who was trained in linguistic semantics and pragmatics, I concur completely with the results of the surveys and bull sessions. I cannot imagine a scenario in which (169a) would hold but (169b) would not, or vice versa. I can certainly imagine scenarios in which (169a) would be socially inappropriate, but, as Frege might say in “The Thought” (1918), this would be “of no concern to logic.” 138 Cf Schütze (1996).

120

Chapter 4 Negative inversion and its discontents

though, in the next sections we will look at the non-truth conditional meaning in a range of meaning diagnostics from the semantics and pragmatics literature.

4.3 Negative inversion, social meanings, and meaning diagnostics In the next sections, I ask what kinds of meaning dimensions are available to host the obvious social meanings associated with NI. I argue that the social meaning of the construction is conventional but that it is also non-truth conditional. In the semantics and pragmatics literature, the most likely categories of meaning available are those of Gricean (1975) conventional implicature (CI) and linguistic presupposition (PS). I explore each of these with respect to NI social meanings in the sections below; I also consider Gricean conversational implicature (CNVSI). Given the standard properties of all three of these meaning dimensions, NI social content is very easily analyzed as conventional implicature, and very easily dismissed as either CNVSI or PS. We turn to this exercise now, beginning with the question of meaning and ineffability. This will provide a natural segue into the following discussions of the various kinds of implicature and PS. 4.3.1 Social meaning and ineffability One striking aspect of NI’s social meaning is that my participants provided many different descriptions of what it actually is. As shown above, participants described NI with a wide range of negative terms centering around lack of education, rurality, and the inner city; but they also used a wide range of positive terms centering around comfortable speech, informality, and coolness. In essence, participants recognized social meaning in the construction, but were not able to put their finger on exactly what it might be in clear, consistent, neutral language. One way of describing this result in the semantics and pragmatics literature would be to say that the social meaning of NI is ineffable. This corresponds with the way that Gricean conventional implicature meaning is frequently described as well: see, for example, Blakemore (2001, 2011) on discourse particles, Potts (2007) on expressives, Horn (2013) on the personal dative construction, and more.139 Here is Potts (2007) on the ineffability of expressives, which he argues

139 Blakemore (2001, 2011) is presented in the framework of relevance theory (Sperber and Wilson, 1986/1995), which does not rely on the notion of conventional implicature. However, Blakemore’s subject matter – i.e. discourse particles – have been given conventional implicature accounts by many other researchers over the years.

4 Negative inversion is stigmatized

121

contribute conventional implicatures, and, alluding to Blakemore’s discussion on various discourse particles: Blakemore (2001) observes that speakers are generally unable to articulate meanings for a wide range of discourse particles. When pressed for definitions, they resort to illustrating where the words would be appropriately used. Expressives in general manifest this descriptive ineffability. [. . .] I’ve only once been told that an uncontroversial expressive had an accurate paraphrase in descriptive terms: bastard was claimed to mean ‘vile contemptible person.’ But this paraphrase misses its wide range of affectionate uses [. . .].

A related quote from Cruse (1986), quoted in Potts (2007), gets at the variability of expressive meaning in a slightly different way, with a metaphor of digital versus analog. Here is Cruse (1986: 272): [P]resented meaning is for the most part coded digitally – that is to say, it can vary only in discrete jumps; expressive meaning, on the other hand, at least in respect of intensity, can be varied continuously, and is therefore analogically coded.

What we see here then is a particular quality or qualities of expressive and other conventional implicature meaning that is very similar to the ineffable qualities of NI. Turning to the sociolinguistics literature, the descriptions above of ineffability are reminiscent of the “indexical field” description of social meaning described in Eckert (2008b). According to Eckert, the social meaning associated with sociophonetic variables is fluid and context dependent. In Eckert’s (2008b: 454) words: [T]he meanings of variables are not precise or fixed but rather constitute a field of potential meanings – an indexical field, or constellation of ideologically related meanings, any one of which can be activated in the situated use of the variable. The field is fluid, and each new activation has the potential to change the field by building on ideological connections.

Similarly, Horn’s (2013) account of personal dative constructions such as (i) suggests that these constructions contribute a conventional implicature (referred to as “F-implicature” in Horn’s terminology): (i)

I love me some Elvis.

The content in question, as Horn (2013: 167) argues, is one of: subject affect, imposing a use-conditional constraint on its felicitous assertion, viz. that the speaker assumes that the action expressed has or would have a positive effect on the subject, typically satisfying the subject’s perceived intention or goals. Note in particular that [Personal Datives] share with (other) F-implicatures the property of descriptive ineffability cited above: the content of such implicatures is notoriously elusive. [emphasis mine].

122

Chapter 4 Negative inversion and its discontents

This description fits nicely with the variety of social meanings we see with NI, and it provides a clear explanation for the origins of some instances of ineffability in the first place. That is, if the essential social content is enriched indexically in a given situation, it makes sense that individual participants would describe the meaning of the construction in a range of different but related ways. Horn (2013) gets at something like this in his discussion of tu/vous pronouns, suggesting that the elusive shades of meaning distinguishing these forms are underdetermined, i.e. contextually enriched, which is essentially what Eckert is getting at with the indexical fields. For comparison, here is Horn (2013: 168): If you know generally that my use of vous rather than tu signals something in the range of formal respect, distancing, and/or lack of intimacy, my precise motives can remain underdetermined, but if you don’t know whether I’m using a 2nd person or 3rd person pronoun, the indeterminacy would be more serious.

In essence, we see ineffability in both social meaning and conventional implicature meaning: NI shares in this with vous, but, and the numerous other conventional implicature devices identified in the semantics literature over the years.140 As Potts (2007) and Horn (2013) note, ineffability is a strong, if imperfect, indicator of membership in the conventional implicature category.141 The presence of this quality with NI, as well as the fact that NI’s social content appears to be conventional but non-truth conditional, provides a strong prima facie motivation for considering other ways in which this kind of social meaning compares to Gricean conventional implicature. It also provides motivation for considering this aspect of NI against other dimensions of meaning such as PS and Gricean CNVSI. In the sections below, we do exactly this, beginning with presupposition.142

140 See Horn (2013) for a lengthy list of linguistic devices that have been argued to convey Gricean conventional implicatures. 141 While ineffability is a strong indicator of conventional implicature meaning, it is not a distinguishing criterion, because, as Geurts (2007) illustrates, there are various other clearly nonCI forms that share this quality. Horn (2013) cites Geurts (2007) to the effect that “the presence of ineffability can at most be a necessary, not sufficient, condition [. . .].” 142 In Sections 4.3.2 through 4.3.4, the semantic and pragmatic intuitions relied upon initially were my own, as a native speaker of the AE in question. I have also checked each of the intuitions in these sections in person with five other participants, one of whom is a native speaker of Texas AE, and two each who are speakers of AAVE and CE, respectively, all of whom are well known to me. This work was done in person with discussion of each of the diagnostics in question, and the results were uniform.

4 Negative inversion is stigmatized

123

4.3.2 Negative inversion meaning and presupposition The category of presupposition is much less clearly defined than is the case with the other two categories of meaning discussed in this section: i.e. Gricean CI and CNVSI. Nonetheless, there are several defining criteria for PSs upon which many researchers agree.143 In this section, I will consider the behavior of NI and its social content in these more or less standard PS diagnostics, beginning with what is often referred to as the ‘family of sentences’ as termed by Chierchia and McConnell-Ginet (1990/2000), as this series of tests seems to be best known in the literature.144 Chierchia and McConnell-Ginet (1990/2000: 350) assume that the primary empirical characteristics of PS content are that it is both backgrounded and taken for granted, and as such, that PS triggers should be felicitous in sentence types which are typically non-assertive, such as negated sentences, questions, and in the protasis of a conditional. Here is Chierchia and McConnell-Ginet’s arrangement of the diagnostic, which is their (31). (170) a. b. c. d.

S. It is not the case that S. Is it the case that S? If S, then S’.

The idea is that PSs of a sentence S will survive in all of the different nonassertive contexts of the family of sentences, while asserted content will not. Consider (171), which is Chierchia and McConnell-Ginet’s (30). A clear PS of (171a) seems to be that of (171b). (171) a. Joan has stopped drinking wine for breakfast. b. Joan used to drink wine for breakfast. The question is, then, does the presupposed (171b) survive in the sentence family. Consider (172a–c), adapted from Chierchia and McConnell-Ginet’s (32). In each of these examples, there is a clear sense in which (171b) is still assumed; that is, it is not overturned or suspended in the negative, interrogative, or conditional contexts. As such, it is argued that (171b) is presuppositional.

143 See Potts (2015) for an overview of the current debates on the topic as well as an intellectual history. 144 As Potts (2015) notes, the essentials of the ‘family of sentences’ idea originate in Karttunen (1973).

124

Chapter 4 Negative inversion and its discontents

(172) a. It is not the case that Joan has stopped drinking wine for breakfast. b. Has Joan stopped drinking wine for breakfast? c. If Joan has stopped drinking wine for breakfast, she has probably begun to drink more at lunch.

Negation Interrogative Protasis

Let us see now what the same kind of reasoning brings for NI in similar contexts. In (173) there is a prototypical NI. The question we are pursuing here is whether the social meaning of NI in (173) – i.e. informality, lower socioeconomic class, etc. – survives in the family of sentences in (174a–c). (173) Didn’t nobody come in here. Social Meaning: informality, uneducated, lower class, camaraderie, etc. (174) a. It just isn’t the case that didn’t nobody come in here. Negation I can see the footprints coming right in the door. ‘It just isn’t the case that nobody came in here [. . .].’ b. So, you’re saying that didn’t nobody come in here? Interrogative ‘So, you’re saying that nobody came in here?’ c. If didn’t nobody come in here, then why are there Protasis footprints there on the floor? ‘If nobody came in here, then why are there footprints [. . .].’ In these non-assertive examples in (174), NI retains all of the meaning and social flavor that it has in (173). There really is no difference. What this tells us is that the social meaning, ineffable though it may be, survives in the family of sentences contexts and so we have an argument in favor of a PS treatment.145 The contexts above in (172) and (174) in which presuppositions are allowed to “survive” were originally described with the metaphor of “holes” that PSs are allowed to slip through, as in Karttunen (1973). In addition to the three contexts above, Karttunen also noted PSs slip through modal contexts of possibility. Thus, in (175), under the modal auxiliary might, the PS of (171b) is similarly allowed to slip through.

145 But see Tonhauser et al. (2013: 67) who show that conventional implicatures pattern the same way that presuppositions do in the family of sentences – or in what these authors refer to as “projective” contexts. This certainly reduces the effectiveness of the family of sentences as a presupposition diagnostic. It is also not surprising, then, that if NI are CI devices, then they would pattern in the same way as presuppositions in these particular contexts.

4 Negative inversion is stigmatized

125

(175) Joan might have stopped drinking wine for breakfast. Presupposition: Joan used to drink wine for breakfast. The same effect is present for the NI in (176), which continues to assume the social meaning of (173). (176) It might be that didn’t nobody come in here. ‘It might be the case that nobody came in here.’ Social Meaning: informality, uneducated, lower class, camaraderie, etc. NI appears thus far then to be indistinguishable from PS in these diagnostics. As is well known, though, these diagnostics are not conclusive in and of their own right. There are others to consider as well. Karttunen (1973) also discussed a group of contexts he referred to as “plugs,” which mirror the behavior of the PS holes we just saw. Briefly, when a PS trigger appears in the embedded sentence of verbs of saying (say, tell, etc.) or certain propositional attitude verbs (want, believe, etc.), the PSs are frequently plugged in the embedded context and so do not associate with the larger embedding sentence. Thus, in a non-plugged sentence like (177), the definite article typically carries a PS of existence. In Russell’s (1905) famous example, the definite description carries the PS that the King of France exists. (177) The King of France is bald. Presupposition: there is a King of France However, in the plugged example in (178), taken from Levinson (1983: Chapter 4), the existence PS no longer need be present. It is plugged under the propositional attitude verb believe. (178) Loony old Harry believes he’s the king of France. Let’s compare with an NI embedded under the same propositional attitude verb in (179). Unlike the plugged existence PS of (178), the social content of the NI in (179) is allowed to project out of the embedded sentence: i.e. to associate with the speaker of the higher clause. The same is true with the NI under the verb of saying in (180). The social content is still present. (179) Crazy old John believes that can’t nobody see him, that he’s invisible. ‘Crazy old John believes that nobody can see him, that he’s invisible.’ Social Meaning: informality, uneducated, lower class, camaraderie, etc.

126

Chapter 4 Negative inversion and its discontents

(180) J.W. McVeigh said that won’t nobody go home. ‘J.W. McVeigh said that nobody will go home.’ Social Meaning: informality, uneducated, lower class, camaraderie, etc. This NI behavior is evidence against a presuppositional treatment of the social content, as it is not plugged under the attitude verb. Another PS diagnostic to consider is defeasibility. It has long been known that many (if not all) PSs exhibit this property.146 According to Levinson (1983: 114), an “inference is defeasible if it is possible to cancel it by adding some additional premises to the original ones.” PSs can also be defeased in contexts that contradict them, as in (181), adapted from Levinson (1983). Here, the complement of before is said to be presupposed in (181a), and we can see that this is so as it survives in the negation of (181b). However, in (181c) it is quite clear that the complement of before is not presupposed in the face of the conflicting matrix verb.147 This is because we know that when someone dies, they do not go on to finish their thesis. (181) a. John cried before he finished his thesis. Presupposition: John finished his thesis b. John didn’t cry before he finished his thesis. c. John died before he finished his thesis. With respect to NI, the social content of NI is not cancelable or defeasible, which provides yet another argument against a presuppositional identification of NI social content. The clearest way to see this is the fact that the social content is present even when there is no context at all, as in many of my survey examples

146 This is an issue of some debate in the literature, with some researchers assuming all presuppositions are defeasible, and others assuming only some are defeasible, for a variety of different reasons. Kadmon (2001: 223) writes, for example, “I believe that even the most hardcore pss sometimes ‘disappear’ even from simple affirmative examples.” On the other hand, Beaver and Geurts (2014) argue that presuppositions are only defeasible in embedded contexts. Here also is Huang (2007: 67): [. . .] unlike semantic entailments, presuppositions are cancelable. They are nullified if they are inconsistent with (i) background assumptions, (ii) conversational implicatures, and (iii) certain discourse contexts. Furthermore, they can also drop out in certain intrasentential contexts, some of which give rise to the projection problem of presupposition. Defeasibility has in general been taken as the second most important property of presupposition. 147 This observation dates back to Heinämäki (1972).

4 Negative inversion is stigmatized

127

which were presented with no context yet still gave rise to the social meanings. Further, it is straightforward to create contexts that conflict with the social content but do not cancel it, as in (182). Speaker B’s response to Speaker A is given in crisp standard English, with no other social indicators in his speech. (182) [Middle class, middle-aged males, standing on a metro platform somewhere in Texas] Speaker A: Excuse me, sir, can you tell me how long the train has been gone? Speaker B [in crisp standard English]: No sir. Didn’t any train come by here. ‘No sir. No train came by here.’ The result is an oddity. It is clearly not ungrammatical or unresponsive, but it is incongruent. As Frege (1892) might say, it feels as if “a song with a sad subject were to be sung in a lively fashion.” Thus, cancelation attempts, whether direct or contextual, do not support the idea of NI social content as presuppositional. If the social information were presuppositional, it should have been defeated in this context in which it was not supported. There are still other diagnostics to consider, though, such as the if-clause suspension of PSs discussed in Horn (1972), Levinson (1983), and others. In (183), below, taken from Levinson (1983), again triggers the PS that John cheated in the past. (183) John didn’t cheat again. Presupposition: John cheated in the past. This PS can be suspended with an if-clause, as in (184), in which the PS does not survive the if-clause operation.148 (184) John didn’t cheat again, if indeed he ever did. Notice that the at-issue content in (185) is not suspendable in this way. The attempts here to do so are infelicitous.

148 Abrusán (2016) subsumes this type of suspension under defeasibility. The differing categorization of the phenomenon does not affect the empirical fact, though, that clear presuppositions behave one way with respect to if-clause suspension and NI behaves in a different way.

128

Chapter 4 Negative inversion and its discontents

(185) #John didn’t cheat, if indeed he ever did. With respect to NI, if the social content were presuppositional, we would expect it to also be suspendable, as in (186). However, it does not appear to be so, patterning instead as does the at-issue content above in (185). (186) #Didn’t nobody come in here, if indeed I am being informal or uneducated. ‘Nobody came in here [. . .].’ The social content of NI is not suspendable with an if-clause, and so there is another argument against a PS analysis. Finally, we will consider the intuition that PSs convey backgrounded information. Consider example (187) and discussion here from Potts (2015: 9). (187) Arguably the defining feature of presuppositions, at least in the pre-theoretical sense, is that the speaker acts as if he presumes them already to be in the common ground. As a result, it is generally possible for the speaker to explicitly articulate the presupposed content before relying on it presuppositionally: a. Sam once smoked, but she quit smoking. b. Sam has a dog, and her dog is sick. Of course, such examples might seem plodding, given the ease of accommodation [. . .] but they are not perceived to be redundant. One of the things Potts is getting at here is the fact that PSs are generally taken to be a result of relations between propositions in a common ground: i.e. assumptions of information in a discourse to which new information is added. This is decidedly different from the situation with the social content of NI. With NI, the social content does not serve as a foundation to which new information is added, as in (188). (188) #I am being informal and uneducated, but didn’t nobody come in here.149 ‘Nobody came in here [. . .].’

149 There is an unrelated reading here which should not be confused with the relevant one. Example (188) is acceptable, for example, as a kind of trigger warning. Cf. I’m going to be blunt here, but . . . This is unlike the presupposition cases given in the Potts quote in (187). There, the presupposed content is explicitly articulated, and then built upon in the following clauses.

4 Negative inversion is stigmatized

129

In using NI, a speaker does not put propositions of social content into a common ground and then build upon them with new information; rather, the speaker establishes or maintains a relationship with an addressee, and on some occasions, leans on this relationship maintenance pragmatically to convey further emphasis, as will be discussed in Chapter 5. Typologically, then, the social content of NI is quite distinct from what is normally considered presuppositional content. Let’s look now at another possibility, which is Gricean conversational implicature. 4.3.3 Negative inversion meaning and conversational implicature Horn (2004: 3) defines CNVSI as “an aspect of what is meant in a speaker’s utterance without being part of what is said.” Spelled out in more formal terms, Potts (2015) defines CNVSI, as in (189) [Potts, ex. 9]. (189) Proposition q is a conversational implicature of utterance U by agent A in context C if, and only if: a. it is mutual, public knowledge of all the discourse participants in C that A is obeying the Cooperative Principle; b. in order to maintain (189a), it must be assumed that A believes q; and c. A believes that it is mutual, public knowledge of all the discourse participants that (189b) holds. Essentially, if the social content of NI is a result of CNVSI, it must be inferred from a speaker’s use of NI, based on an assumption of the speaker’s intentions in doing so. As Levinson (1983: 114) notes, Grice (1975) isolates five characteristic properties of CNVSI. Over the years, researchers have modified or adapted these for different purposes. I will consider the five properties described in Levinson (1983), as they are quite loyal to Grice’s original arrangement and are laid out in a very accessible fashion.150 The clear result will be that NI social content does not behave as a conversational implicature with respect to these diagnostics. The first property is cancelability or defeasibility. Like PS discussed above, and even more so, CNVSIs are believed to be cancelable by most researchers.

150 See also Levinson (2000), Horn (2004), Huang (2007), and Potts (2015) for discussion of these properties and others.

130

Chapter 4 Negative inversion and its discontents

As Levinson (1983: 114) writes, “An inference is defeasible if it is possible to cancel it by adding some additional premises to the original ones.”151 In (190), adapted from Levinson’s (51)–(52), we see that (190) can be taken to convey the CNVSI that John has only three cows. (190) John has three cows. Conversational Implicature: John has only three cows and no more. However, if we add additional premises which contradict the CNVSI, the conventional implicature is seen to be canceled, as in (191) [Levinson’s (57)]. (191) John has three cows and maybe more. Levinson also shows that entailed descriptive meaning cannot be canceled in this way. In (192), then, when we try to add additional premises which contradict the entailed meaning, it is not canceled, and the resulting sentence is contradictory. (192) *John has three cows and maybe none. Now let’s consider NI cancelability, with the NI in (193). (193) Didn’t nobody go to the party. ‘Nobody went to the party.’ Social Content: uneducated, rural, inner city, comfortable, etc. What is needed here to show that the social content is cancelable is to add additional premises which contradict the social content, as in (194). (194) #Didn’t nobody go to the party, and I am an educated urbanite. ‘Nobody went to the party [. . .].’ While the attempt at cancelation in (194) doesn’t result in a contradictory or ungrammatical sentence, it does result in one that is very odd and very unlike the spirit of the legitimate cancelation of the CNVSI shown above in (191). Thus, direct cancelation does not seem to be an option for the social content of NI.

151 See Haugh (2014: 63) for significant discussion of this diagnostic as well as an overview of debates focusing on what it accomplishes and how it operates.

4 Negative inversion is stigmatized

131

As is made clear in in Grice (1978: 115), however, there are other ways to cancel CNVSIs as well. These implied meanings can also be canceled contextually “if one can find situations in which the utterance of the form of words would simply not carry the implicature.” What is needed here then is to find a situation in which NI is produced but does not carry any of the social content in question. As we saw above in (182) in the discussion of PS, the social content of NI is simply not cancelable.152 Thus, an assessment of NI meaning in terms of CNVSI is already on unsure footing. Non-detachability is the second property listed in Levinson (1983: 116). As Levinson writes, “the implicature is not attached to the semantic content of ‘what is said’.” The most straightforward way of testing detachability is to look for another way of saying the same thing, but which does not convey the CNVSI; in other words, paraphrasing the sentence in question but maintaining propositional equivalence. Huang (2007: 34) provides the following example in (195), which shows that “the use of any linguistic expression that is synonymous with almost will trigger the same conversational implicature.”153 (195) The film almost/nearly won/came close to winning an Oscar. Conversational Implicature: The film did not quite win an Oscar. The CNVSI, then, is said to be non-detachable. On the other hand, as we will see below in example (202), the social content of NI is clearly detachable via paraphrase: i.e. the non-inverted counterpart is truth-conditionally equivalent to NI but it does not convey the social meaning. So, once again, NI patterns unlike conversational implicature. The third property of CNVSI is calculability. Conversational implicature content is not considered to be part of ‘what is said’; rather, this content must be inferred based off of ‘what is said’ in a particular context. Levinson (1983: 117) writes: (196) for every putative [conversational] implicature it should be possible to construct an argument [. . .] showing how from the literal meaning or the sense of the utterance on one hand, and the co-operative principle and

152 Levinson (1983: 93) makes a related argument with respect to the conventionalization of social meaning of various honorifics cross-linguistically, noting that ironic use of honorifics would be impossible if social meaning and rules of use for them were not conventional. See also Bender (2007) on the conventional association of social meaning with specific instances of a linguistic variable. 153 Implicatures arising via the Maxim of Manner, which are dependent on the particular form of the utterance, behave somewhat differently.

132

Chapter 4 Negative inversion and its discontents

the maxims on the other, it follows that an addressee would make the inference in question [. . .]. S’s saying that p conversationally implicates q iff154: (i) S has said that p (ii) there’s no reason to think S is not observing the maxims, or at least the co-operative principle (iii) in order for S to say that p and be indeed observing the maxims or the co-operative principle, S must think that q (iv) S must know that it is mutual knowledge that q must be supposed if S is to be taken to be co-operating (v) S has done nothing to stop me, the addressee, thinking that q (vi) therefore S intends me to think that q, and in saying that p has implicated q So, CNVSIs are not part of the semantic content of a sentence but rather require a reasoning process that takes into account ‘what is said’ and the speaker’s likely intention in saying it, and the co-operative principle and maxims. It is very straightforward to show that the social content of NI is not calculable. For one thing, all of my survey participants looked at context-free instances of NI and still came to the conclusion that they carried the relevant negative or informal social content. NI social content thus does not pattern as a CNVSI with respect to this property either. The fourth property Levinson mentions of CNVSI is that they are nonconventional; i.e. they are “not part of the conventional meaning of linguistic expressions” (Levinson 1983: 117). It should be clear by now from the previous discussion that the social content of NI is conventional; thus, no further comment is required here. The fifth property Levinson discusses is indeterminacy. By this Levinson means that “an expression with a single meaning can give rise to different implicatures on different occasions” (Levinson 1983 118). Consider (197), which is Levinson’s (69). (197) John’s a machine. Of this sentence, Levinson writes “This could convey that John is cold, or efficient, or never stops working, or puffs and blows, or has little in the way of grey matter, or indeed any and all of these.” In an infinite number of contexts, (197) can express

154 The configuration in (196) is Levinson’s (1983: ex. 45).

4 Negative inversion is stigmatized

133

an infinite number of implicatures. This is very different from what we see with the social content of NI, which is always identified as the same related, if ineffable, meanings described above. So again, we see the relevant NI meaning as something quite apart from conversational implicature. Levinson (2000: 15) and Potts (2015) discuss reinforceability as another property of CNVSI.155 Levinson writes “it is often possible to add explicitly what is anyway implicated with less sense of redundancy than would be the case if one repeated the coded content.” We see in (198a) that almost X conversationally implicates not quite X. We see that this CNVSI can be reinforced in (198b). (198) a. John almost won the game. Conversational Implicature: John didn’t quite win the game b. John almost, but not quite, won the game. However, non-CNVSI content is more difficult to reinforce in this way, as in (199), in which it is attempted to reinforce the content of ‘what is said.’ (199) #John won the game, but John won the game. Let’s try to reinforce the social content of an NI now in (200). Here too we see that the reinforcement attempt is very awkward on the relevant reading. (200) #Didn’t nobody win the game, but I’m trying to be informal. ‘Nobody won the game [. . .].’ It is apparent, then, that the social content of NI has little in common with CNVSI content or PS content. In the next section, I consider the social content of NI alongside standard defining criteria of Gricean conventional implicature, and I show that NI meaning in question is conventional yet non-truth conditional, much as prototypical Gricean conventional implicatures are assumed to be. 4.3.4 Negative inversion meaning and conventional implicature Gricean (1975) conventional implicatures are generally taken to be non-calculable, non-cancelable, and detachable.156 They are part of the conventional meaning of a

155 For early discussion of reinforeability, see Sadock (1978) and Horn (1991). 156 Note that there have been different formal arrangements and understandings of the category conventional implicature over the years. Most significantly, Karttunen and Peters (1979)

134

Chapter 4 Negative inversion and its discontents

lexical item, and are not truth conditional, nor are they derivable as a conversational inference. The following descriptions of conventional implicature from Levinson (1983), Horn (2004), and Huang (2007) are clear in this: [C]onventional implicatures will be non-cancelable because they do not rely on defeasible assumptions about the nature of the context; they will be detachable because they depend on the particular linguistic items used (e.g. if you substitute and for but you lose the conventional implicature but retain the same truth conditions); they will not be calculated using pragmatic principles and contextual knowledge, but rather given by convention [. . .]. (Levinson, 1983: 128) Such detachable but non-cancelable aspects of meaning that are neither part of what is said nor calculable from what is said are conventional implicatures. (Horn, 2004: 2) A conventional implicature is a non-truth-conditional inference which is not deductive in any general, natural way from the saying of what is said, but arises solely because of the conventional features attached to particular lexical items and/or linguistic constructions. (Huang, 2007: 54)

The most commonly cited example of a Gricean conventional implicature device is the conjunction but, as in (201a), which entails logical conjunction and carries the additional conventional implicature of contrast between the two conjuncts. (201) a. He is poor but honest. Conventional Implicature: there is some contrast between being poor and being honest. b. He is poor and honest. c. He is poor but honest – #not that there is any contrast between being poor and honest. As we see in the descriptions above from Levinson, Horn, and Huang, the proper Gricean CI is considered to be detachable, non-cancelable, and non-calculable. We see in (201b) that it is possible to detach the conventional implicature of contrast, where but is absent but where the truth conditions are the same as that of

aligned conventional implicature very closely with presupposition, while Bach (1999) argued that conventional implicature is a myth and that what is really occurring is foregrounding and background of propositions. Finally, Potts (2005) argued that conventional implicatures necessarily project out of embedding contexts. All of these accounts have their advantages, but they do alter what we find in Grice in various ways. In this book, I assume a very traditional Gricean understanding of conventional implicature, with no comment on the aforementioned adaptations of Grice’s category.

4 Negative inversion is stigmatized

135

(201a). In (201c), we see that it is impossibly awkward to try and cancel the CI directly, illustrating its non-cancelability. Lastly, per Grice’s original description, the CI is non-calculable. This means that the CI of contrast is part of the conventional meaning of but and no inference process is required in order to recover it. The diagnostics for detachability can also illustrate this point. Let us observe, then, how the social meaning of NI stacks up in these diagnostics. There is a prototypical NI in (202a), and its non-inverted counterpart, which is truth-conditionally equivalent, in (202b). It is clear that the social content of (202a), however exactly it is articulated, is not found in (202b) and is thus detachable. (202) a. Don’t many boys in town hunt. ‘Many boys in town don’t hunt.’ Social Meaning: informality, uneducated, lower class, camaraderie, etc. b. Many boys in town don’t hunt c. Don’t many boys in town hunt – #but I don’t mean to sound informal . . . – #but I’m not working class or from a rural area . . . In the impossibly awkward (202c), we see that it is not clear how the social content could be directly canceled, similar to the situation of (201c) above. As with (201), the difference between (202a–b) can serve to illustrate that the meaning in question is associated with the construction rather than a product of calculation. What we see here, then, is that the social content of NI patterns in the conventional implicature diagnostics in the same way that prototypical conventional implicature devices like but do.

4.4 Social meaning, social marking, and social stereotypes I have so far discussed NI social meaning with respect to how it can be categorized in terms of semantics and pragmatics. As we have seen, there is a clear space for it in the category of conventional implicature. This alone, though, tells us only some things about the social meaning: essentially, that it is conventional and non-truth conditional. We know in addition that the social meaning is ineffable and that it is likely enriched through contextual indexation. As we saw above in this chapter, other conventional implicature meanings have also been described in this way. In this section I want also to consider briefly two labels from the sociolinguistics literature that apply to NI as a consequence of its social meaning: these

136

Chapter 4 Negative inversion and its discontents

are social marker and social stereotype, both of which originate in Labov (1972b: 314). Consider the passage below from Labov (2001: 196), in which these categories are elaborated upon. Changes from below begin as indicators, stratified by age group, region, and social class. At this stage, they show zero degrees of social awareness, and are difficult to detect for both linguists and native speakers. As they proceed to completion, such changes usually acquire social recognition as linguistic markers, usually in the form of social stigma, which is reflected in sharp social stratification of speech production, a steep slope of style shifting, and negative responses on subjective reaction tests. Ultimately, they may become stereotypes, the subject of overt comment [. . .].

Clearly, NI is well beyond Labov’s description of indicators. As we have seen, speakers and community members are very aware of and capable of detecting NI in speech. Labov’s markers, however, sound more familiar with respect to NI, with the social recognition, stigmatization, stratification of speech production, and so forth. Unlike some other features of the varieties in question that still fly beneath the radar, speakers and hearers are clearly aware of NI, and most have strong ideas about them. The Southern pin-pen merger, for example, in which /ɪ/ and /ɛ/ merge before nasal consonants, is a dialectal feature found in Texas, though most speakers are likely not aware of it and do not have strong feelings about it. Returning to NI, for many speakers, especially the younger ones, I suspect that the construction has reached the point of being a stereotype: hence the many survey comments about NI being the speech of rednecks, gang members, and hip hoppers. This awareness of NI among its users and larger community, whether as a social marker for some folks or a stereotype for others, is ultimately important for justifying the pragmatic accounts of NI emphasis developed below in Chapter 5.

4.5 Conclusion We saw in this chapter that NI is associated with a wide range of social meanings. The form is stigmatized in a variety of ways, but it also carries an association of coolness and that of an in-group marker. This seemingly unexpected gathering of meanings is actually seen to occur quite frequently in studies of vernacular and non-standard language varieties, in which stigmatization and in-group expectation often go hand in hand. It is no surprise to find the same here with NI. Also in this chapter I asked whence NI’s social meaning. Where does it come from, and where does it reside? NI was tested in a wide range of well-known meaning diagnostics, including those for PS, CNVSI, CI, and descriptive propositional meaning. A very clear fit for NI social meaning was

4 Negative inversion is stigmatized

137

found in the realm of Gricean conventional implicature. This means that NI social meaning is conventionally associated with some aspect of the form of NI: i.e. it is lexical meaning. The conventional implicature association also means that NI social meaning is not truth conditional, propositional meaning. This finding fits well with empirical observations of NI, and it is very convenient for deriving further pragmatic effects, such as emphasis, which have often been assumed to be conventionally associated with NI. That is, if NI social meaning patterns as a conventional implicature, then there is an uncontroversial path forward for deriving further conversational meaning from the use of NI via Gricean reasoning, as is done in Chapter 5.

Chapter 5 Negative inversion and emphatic meanings The speaker’s very selection of a particular grammatical form is a stylistic act. – Mikhail Bakhtin, ‘The Problem of Speech Genres’

5 History of emphatic claims In the earliest contemporary discussion of Negative Inversion, which was Labov et al. (1968), it was claimed for AAVE that NI constructions are emphatic compared to their non-inverted counterparts: i.e. that a sentence such as (203) should be more emphatic than its non-inverted counterpart in (204). (203) Wouldn’t many people buy that meat. ‘Many people wouldn’t buy that meat.’ (204) Many people wouldn’t buy that meat. This is an idea that is still very much with us today, though it persists somewhat uneasily. The Yale Grammatical Diversity Project reports, for example, that “[NIs] are believed to have an ‘affective’ or ‘emphatic’ interpretation in Labov et al. (1968) and Green (2002, 2011a, 2011b).” The claim is also assumed in Martin (1992), Weldon (1994), Sells et al. (1996) and continues to be maintained for AAVE in Green (2014) and Green and Sistrunk (2015). It is also mentioned as a potential empirical characteristic of NI in almost all other works that treat the construction, whether or not those works ultimately focus on that specific aspect of the construction in their account. The emphatic signal has been operationalized in a range of syntactic and semantic ways over the years. It has also been argued to not always be present, as in Foreman’s (1999) treatment of NI in West Texas English. This raises a few important questions. For example, where does the emphatic signal reside? Is it a consequence of syntax and semantics? Is it conventional or a product of pragmatic usage? Further, can the same account be given for NI across varieties of English? Or even across micro-varieties of AAVE? I argue in the present chapter for a new understanding of emphatic with respect to NIs. Instead of simply being an emphasis marker, I argue that NI is a marker of dialectal belonging and speaker involvement, and that the emphasis that often accompanies the construction can be derived pragmatically based on these facts. As such, it is not necessary to posit a conventional meaning of https://doi.org/10.1515/9781501512346-006

5 History of emphatic claims

139

emphasis as has been done in the past. But this particular process is even more interesting, as it shows that the social meaning of Labov (1963; 1972b; inter alia) itself can be conventional and thus available for pragmatic inference processes. Finally, I show that the empirical description of Texas AAVE NI given in the present book clearly conflicts with recent descriptions of the emphatic signal of AAVE NI made elsewhere (i.e. Green 2014; Green and Sistrunk 2015) and so point to a potential instance of micro-variation within the grammar of AAVE in general.

5.1 Definitions of emphatic in the negative inversion literature For many researchers of this construction, the question of whether NI is emphatic or not is taken up only fleetingly. It is mentioned, and then it is put aside as the researcher moves on to other aspects of the construction, as in Weldon (1994), Sells et al. (1996), Foreman (1999), Matyiku (2017), and others. Labov et al. (1968) and Labov (1972a) are exceptions, as are Green (2014) and Green and Sistrunk (2015), all of which will be discussed below. For the most part, the literature on this construction is situated in the generative syntax tradition, and the emphatic qualities are only interesting there to the extent that they might or might not play a role in triggering or motivating movement of the auxiliary. In this chapter, however, it is the emphatic signal itself that is of interest, and it is argued to be a consequence of socio-pragmatic interaction, quite independent of any syntactic triggering mechanism. In what follows, I first consider the different conceptions of emphatic that have been made by the different researchers, which I will refer to as Labov-emphatic and Greenemphatic. I then reject both of these notions for a third understanding of emphatic, which grows out of Israel’s (1996, 2011) work on negative polarity items (NPI) and which I will refer to as Israel-emphatic. 5.1.1 Labov-emphatic and affective triggers The earliest mention of emphatic with respect to this construction is Labov et al. (1968), which is concerned with AAVE constructions in New York City. Labov et al. describe the construction impressionistically in terms such as these:

140

Chapter 5 Negative inversion and emphatic meanings

“the strong affective character of this construction” (284) “it is emphatic, excited, and strongly affective” (285) “From the emphatic and affective nature of negative inversion, we would infer that may, might, would, and should are not likely to be used” (286)157 “It has a strongly affective character wherever it occurs” (288)

The emphatic quality itself is posited as a result of syntactic inversion. Labov et al. argue that inversion is triggered in such sentences by the presence of an affect feature – essentially the affect-attraction of Klima (1964). For Klima, affect is a “grammatico-semantic feature” common to negative terms and various syntactic configurations, which among other things, was thought to trigger subject-auxiliary inversion and license “indefinite quantifiers like ever and any.” (Klima 1964: 313). Labov et al. look to this affect feature as motivation for the inversion movement, and posit inversion as an optional rule: Negative inversion with affective value. This is an optional process which gives additional prominence to the negative, and takes different forms in different dialects. It has a strongly affective character wherever it occurs. (Labov et al. 1968: 288)

Labov et al. assume here that “options in transformational rules represent alternative ways of saying the ‘same thing’” (1968: 78), and that the motivation for inversion movement is to accentuate the negative, with no further comment on the mechanism by which the emphatic qualities arise. This suggests that the inversion sentences were assumed to mean the “same thing” as their non-inverted counterparts, differing in the prominence of the negative but not affecting truth conditional meaning.158 For Labov et al. (1968), then, it seems likely that a Gricean pragmatic account of the emphatic quality would also have been a possibility: i.e. if the inversion is optional, then in so doing the speaker might invite the hearer to infer the emphatic meaning. However, as Grice’s William James lectures and introduction to the implicature-reasoning process had only just been presented at Harvard in 1967 and not published until 1975, it’s very likely this manner of thinking wasn’t yet well known when Labov et al. appeared in 1968. In any case, what Labov-emphatic seems to mean with respect to NI is simply a more prominent negative character of the sentence that is used in more excited contexts.

157 As an aside, Weldon (1994: 3) argues that contra Labov’s inference here, auxiliaries shouldn’t and wouldn’t actually do occur in AAVE NI constructions. 158 Sells et al.’s (1996: 623) discussion of NI is in accord with that of Labov et al. As those authors write, “since ‘affect’ or ‘emphasis’ does not affect truth conditions, [there is no need to] modify the criterion of referential equivalence which Weiner and Labov (1983) establish as prerequisite to the analysis of syntactic variation.”

5 History of emphatic claims

141

5.1.2 Green-emphatic, semantic widening, and absolutely no exceptions Green (2014), also writing on AAVE, provides a semantic characterization of emphatic that is different from Labov’s, and more fleshed out as well. Green’s emphatic quality is linked to syntactic movement, specifically to a particular landing site of the negated auxiliary: i.e. the negated auxiliary moves higher than the subject NP, landing in a CP node that Green labels “negative focus.” The details of the syntax need not concern us here.159 What this translates to in terms of meaning, however, in Green (2014: 124) is: (205) a type of absolute negation or strong domain reading. As such [205] would be interpreted along the following lines, as represented by the paraphrases in double quotes: Don’t nobody want no tea. “Not even one person wants tea” “Absolutely nobody wants tea” Green goes on to write that the proper interpretation of NI is that in which there can be no exceptions at all. Thus, Don’t nobody want no tea must mean that absolutely nobody wants any tea. As such, sentences like (206) are argued to be “strange or contradictory” [Green’s (20)]: (206) #Don’t nobody ride bus number 201 – just the three people who live in the country. Most of the students in this class ride bus number 99.160

159 Green’s basic derivation is seen below in (i) [Green’s (16)]. As she writes, “One of the advantages of this analysis is that the structural positions of the moved elements correlate with the absolute negation/strong domain interpretation” (125). (i)

Don’t nobody want no tea. [CPNEGFOC[C’[NegFoc] DON’Tj]][TPNOBODYi [T’ do+n’tj]][NegP [Neg’ do+n’tj]] [VP nobodyi [V’ want no tea]].

Thus, Green links the domain widening of NI to syntactic position. The domain-widening hypothesis in general, which originates in Kadmon and Landman (1993) with respect to the negative polarity item any, and which influences Green’s argument presented here, posits that the domain-widening property of any is actually part of that term’s lexical meaning. See Chierchia (2006) for extensive discussion of different approaches to domain widening in general. 160 As shown in Chapter 2, however, this sentence was acceptable to most of my African American, Anglo, and Chicano participants, suggesting regional or perhaps generational variation between the AAVE of Green’s informants and those of the current work.

142

Chapter 5 Negative inversion and emphatic meanings

As Green describes (206), “If the absolute negation reading is associated with the [NI], then the sentence should not be used in the context in which some students, or any students at all, actually do ride bus number 201” (2014: 127). This is in contrast to NI’s non-inverted counterparts, which Green argues “can have the absolute negation reading but do not necessarily have to have that meaning.” Thus, for Green, NI has a non-ambiguous, strong, absolute negation reading, which brooks absolutely no exceptions. As such, on Green’s account, NI should have different truth conditions than its non-inverted counterpart, meaning that an NI can be judged false in a context in which a comparable non-inverted example would be true. This means that the two sentence types are not linguistic variables in Green’s account in AAVE: in fact, they mean different things and can be used in different contexts and situations. The noninverted form allows for a small number of riders, while NI allows for no riders whatsoever. As Green writes: The inverted and non-inverted constructions differ in the position of the negated auxiliary, and this seems to be directly related to the possible meanings in the two types. [. . .]. That is, there is some overlap, such that the two can be used in absolute negative contexts; however, the non-inverted constructions can also be used in contexts that refer to a very small number on a scale.

5.1.3 Labov- and Green-emphatics in context Over the last several decades, the question of whether NI is emphatic or not has taken a backseat to questions of syntax for two primary reasons. First and foremost, there is strong agreement among many researchers that the emphatic/affective quality is not responsible for triggering the inversion of the negative auxiliary and the subject. As most of the literature on this construction has been conducted in the framework of generative syntax, there would seem to be little reason to pursue the question of emphatic meaning further for those researchers.161 Sells, Rickford, and Wasow (1996: 595) write: (207) While many researchers feel that the NI construction has some important functional and pragmatic motivations, and there seems to be evidence that this is so, we do not think that such considerations are sufficient by themselves to explain the necessity of the presence of negation. [. . .] If the

161 Weldon (1994) is a notable exception, which pursues a syntactic account in the headdriven phrase structure grammar of Pollard and Sag (1994), though Weldon does not offer an account of the emphatic nature of NI.

5 History of emphatic claims

143

inversion were driven solely by some emphatic or affective pragmatic effect, it would be expected that [207] should be acceptable, but it is not. *Is somethin’ went down. ‘Something happened’ [. . .] Thus, the notion of affective-meaning does not seem to us to be a promising direction for the solution to the two problems of why [the NI] crucially involves negation and inversion. The second reason the emphatic quality has not been given much attention is that it has been claimed to not exist in every NI instance. Foreman (1999) for example, writing on West Texas English, claims “In WTE, not all NI sentences are terribly emphatic in nature.” As a native speaker of Texas AE, I absolutely agree with Foreman here. But like Sells et al.’s motivation in the quote above, Foreman is interested in the emphatic component only to the extent that it does or does not trigger inversion. Having concluded that NI can be found which isn’t “terribly emphatic,” Foreman then moves on to other potential motivators for the inversion. Moving beyond Sells et al. and Foreman, we find a further complication for the idea of NI as necessarily emphatic: i.e. it can combine freely with various compromisers, downtoners, and hedges in general, as in (208a–e).162 (208) a. Couldn’t pretty much anybody make me understand back then. ‘Almost nobody could make me understand back then.’ b. I don’t know. I guess, couldn’t nobody get in there. ‘I don’t know. I guess nobody could get in there.’ c. Can’t anybody really do much, if I had to say. ‘Nobody can really do much, if I had to say.’ d. I guess so, yeah, don’t anybody really seem to be around there these days. ‘I guess so, yeah, nobody really seems to be around there these days.’ e. Well, don’t nobody appear to be there yet except for a few early birds. ‘Well, nobody seems to be there yet except for a few early birds.’ Similarly, NI can appear in a range of non-assertive sentence contexts. In (209a) below, we see NI in the protasis of a conditional sentence, which is generally assumed to be non-assertive. See, for example, Dancygier (1998: 110),

162 Examples (208a–e) were constructed by the author.

144

Chapter 5 Negative inversion and emphatic meanings

who writes that the “conditional protasis instructs the hearer to treat the assumption in its scope as not assertable in the usual way.” We also find NI in an interrogative context in (209b), a negated context in (209c), and in the scope of weak possibility operator in (209d). (209) a. If don’t nobody else have my back, I know God does.163 ‘If nobody else has my back, I know God does.’ b. So, you’re saying that can’t nobody get in there anymore? ‘So, you’re saying that nobody can get in there anymore?’ c. It just ain’t the case that didn’t nobody come in here. ‘It just isn’t the case that nobody came in here.’ d. Cold as it is outside, it might be that couldn’t nobody get their car started this morning. ‘Cold as it is outside, it might be that nobody could get their car started this morning.’ The contexts in (209) are often referred to in the semantics and pragmatics literature as the ‘family of sentences test,’ and are well known diagnostics for presuppositions. See Chapter 4 above, Karttunen (1973), and Chierchia and McConnellGinet (1990/2000), with whom the term family of sentences originates. The idea is that backgrounded or non-asserted content survives in these contexts, while primary content, or what Grice (1975) refers to as ‘what is said,’ does not. Survival in these contexts is an argument that the content is presupposed, though it is not a conclusive argument.164 I am not suggesting here that the social meaning of NI is presupposed; I am simply illustrating that NI can occur in nonassertive sentential contexts and that this provides compelling reasons to be suspicious of a claim that NI is a conventional emphasis marker. As was shown above in Chapter 4, the social content of NI ultimately patterns more like a Gricean (1975) conventional implicature than a traditional presupposition. Let’s return now to other claims about NI and emphasis. What about the situation in AAVE beyond Texas? Here too we see differing empirical claims. Drawing on ex-slave texts, Schneider (1989: 195) writes that “The most common way of achieving emphasis in Early Black English is to invert subject and predicate, as in the following example:

163 . 164 See, for example, discussion in Potts (2015), who writes that these sentence contexts “do not themselves provide a sufficient condition for classifying a meaning as presupposed, since non-presupposed material can also project past them [. . .].”

5 History of emphatic claims

145

(210) Couldn’t nobody dat lived in Mer-ree-dian right after de Surrender ever forgit Seventh Street an’ where it head to. ‘Nobody that lived at Meridian right after the Surrender will ever forget Seventh Street and where it heads to.’ For Green (2014), all NI is strongly affective and emphatic, driven by the wide semantic interpretation that results in a statement that is more exhaustive than the non-inverted counterpart. Parrott (2000) is less sure about the status of the emphatic character and argues that the AAVE NI and its non-inverted counterpart have the same truth conditions. As Parrott (2000: 416) writes: [NI] seems to have more emphasis than a non-inverted negative sentence. [. . .]. However, this is a purely pragmatic difference. An [NI] sentence and its non-inverted counterpart are identical in their truth conditional meaning.

A further complication for the emphatic character of the AAVE sentences is pointed out in White-Sustaita (2010). In her discussion of Sells et al. (1996) – who note that NI occurs more frequently in their corpus than do the non-inverted examples – White-Sustaita writes: if the [NI] version were the marked form (i.e. marked by affective emphasis or negative focus), it would be curious that the speakers feature a greater majority of emphatic forms/affective forms than unmarked forms.165

This is an important point. For Labov et al. (1968), NI in AAVE was taken to be strongly emphatic, yet truth conditionally equivalent to the non-inverted counterpart, similar to Parrott (2000). Almost 30 years after Labov, however, in Sells et al. (1996), this is less clear. This difference in result could be due to stylistic particularities of Labov’s and Sells et al.’s differing corpora. Or, it could be an indication that the emphatic qualities of NI have eroded over the years, as often happens with negation, and thus we see a generational difference.166 For example, Labov (1972a: 815) suggests that the impetus for NIs’ existence in the first place has been that general negative concord “has lost its emphatic character in [AAVE]” and so the language has devised other ways to reinforce negation,

165 This would be curious indeed if NI were “emphatic-WIDE” à la Green or “emphaticAFFECTIVE” à la Labov. However, it is not surprising at all if NI is “ISRAEL-emphatic” as on the present story. 166 See Jespersen (1917) and the oft-discussed phenomenon referred to in Dahl (1979) as “Jespersen’s Cycle.” See also Croft (1991b: 14), who writes “The evolution of unemphatic grammatical constructions of various types from the corresponding emphatic ones is also quite well-known [. . .].”

146

Chapter 5 Negative inversion and emphatic meanings

such as negative inversion. Another possibility here is that there are simply regional differences in the emphatic quality of NI across all various studies, which sample different regional populations, or which do not identify the population they sample. Thus, Labov et al.’s corpus was created in the late 1960s in New York City; Sells et al.’s corpus reflects mid-1990s Palo Alto; Green’s and Parrott’s various works do not identify region; and so on. In any case, there are a range of different emphatic claims for NI on the market. In the present chapter, it is acknowledged that NI can certainly be used to convey emphasis; however, I argue that this is done pragmatically via invited inferences. It should be noted here too, however, that the pragmatic approach argued for in this book offers the strong benefit of explaining the varying intuitions across the literature with respect to NIs’ emphatic nature – that is, why some researchers say that NI is necessarily emphatic and others say it is not. In the pragmatic account offered here, both groups are allowed to be correct. In the next section, I discuss the preferred understanding of emphatic content for this chapter, which is adapted from Israel’s (1996, 2011) work on negative polarity items. 5.1.4 Israel-emphatic, pragmatics, and interaction In Israel’s understanding of emphatic, he is not concerned with NI constructions; his focus is on the licensing conditions and lexical semantics of negative polarity items such as any, ever, at all, drink a drop, and so forth. However, the conception of emphatic he uses in that endeavor seems exactly what is needed in the present discussion. For Israel, emphatic describes a speaker’s involvement with her propositions and/or her audience in which the speaker is highly involved. Israel contrasts this sense of emphatic with attenuating which is used to describe a speaker’s detachment from her claims and/or her audience. Thus, the notion of emphatic employed here differs substantially from those of Labov and Green as well as from what is typically assumed in the semantics and pragmatics literature, in which emphatic claims are generally assumed to be strong, forceful, direct, excited claims, which impart emphasis, etc. It is worth quoting Israel here at length to illustrate his concept. Informativity is fundamentally an expression of speaker affect. In producing an emphatic or an attenuating utterance, a speaker does more than simply assert a proposition – she expresses an attitude toward that proposition, toward its significance and its informative strength. Most importantly, she expresses an attitude toward her audience, saying “you are the sort of person with whom I would express myself in this way.” Emphasis and attenuation are, in essence, rhetorical strategies for the presentation of self in discourse. [. . .] I argue that informativity in general, and the linguistic encoding of informative value in particular,

5 History of emphatic claims

147

reflect general strategies for the negotiation of social interaction. They are, in effect, devices for the expression of speaker involvement. [. . .] Roughly, understatement and attenuation serve to give a hearer more options in responding to a speech act. As such, they may be seen as expressions of a speaker’s detachment and deference to the hearer. Emphasis, on the other hand, is a sign of intensity and speaker involvement. As such, it serves as a marker of camaraderie and solidarity with the hearer. (Israel 2011: 110–111, emphasis mine)

For Israel, NPIs lexically encode a quantitative value and an informative value. The quantitative value locates the NPI on a scale, and the informative value, which he characterizes as either emphatic or attenuating, determines the rhetorical effect of the NPI, and, crucially, determines the kinds of contexts in which NPIs are licensed to appear. As Israel writes, informative value and quantitative value “do not simply add information to a proposition; rather, they situate a proposition within a sort of informational matrix” (2011: 85). Lexical items may encode one or both of these properties, but it is only when both are encoded in specific high-low configurations that NPIs are actually licensed. Thus, minimizer NPIs such as drink a drop, budge an inch, etc., have a low quantitative value (e.g. a single drop, a single inch) but a high informative value: i.e., to say a person doesn’t drink even a drop of beer, is to make a highly informative claim. In Israel’s system, it is this combination of low and high values within a scalar model that constrains where NPIs can appear. For the present purposes, however, we are concerned only with Israel’s “informative value” and his understanding of emphatic in terms of speaker involvement. Let’s look now at an NPI sentence with this in mind. Consider (211a–b). On Israel’s account, these statements would express truth conditionally equivalent propositions. Neither technically allows for John to have drunk any beer; however, they differ in the speaker’s attitudes toward the propositions and audience.167 (211) a. John didn’t drink beer last night. b. John didn’t drink a drop of beer last night. In (211b), the speaker has a higher degree of emotional commitment to the proposition, and conveys at the same time a level of familiarity to, or solidarity

167 As Israel (2011: 9) writes regarding NPI content: Marking an expressed proposition as either emphatic or attenuating is basically just a way of calling attention to its logical status with respect to background assumptions. But the act of calling attention itself is always rhetorically loaded. An argumentative operator thus does not add to the logical content of what is said but expresses an attitude about that content and so situates it in a larger context.

148

Chapter 5 Negative inversion and emphatic meanings

with, the hearer. On the other hand, (211a) expresses the same truth conditional content without the affect or emotional commitment of (211b). A consequence of the heightened speaker-involvement here with (211b) is that it is more face-threatening to disagree with it, and less face threatening to disagree with (211a).168 A disagreement with (211a) can be a disagreement with the proposition itself without having to contend as well with the emotional involvement of the speaker. On the other hand, it is more face threatening to disagree with (211b), because in this case, the disagreement would also be threatening the speaker’s emotional commitment. This contributes to the illusion of propositional strength in sentences such as (211b). As an aside here, this way of understanding NPIs provides a fruitful explanation for the “reduced tolerance of exceptions” feeling that Kadmon and Landman (1993) describe for their account of NPI any. For Kadmon and Landman, NPIs semantically widen the domain of interpretation and thus contribute truth conditional content to the host sentence.169 Thus, for Kadmon and Landman, (211a–b) would not be truth conditionally equivalent; (211b) would express a stronger proposition that would unilaterally entail (211a) and so tolerate fewer exceptions.170 In Israel’s account of NPIs, however, (though he does not specifically describe it in this way) the “reduced tolerance of exceptions” intuition would arise as a result of the additional face threat involved in disagreeing with, or making an exception to, a speaker’s emotive statement. I believe this is exactly what we want in describing NI as well: i.e. an understanding that NI and its noninverted counterparts are truth conditionally equivalent but that they differ in terms of speaker involvement. Thus, the meaning of the NI in (212a), below, is actually closer to that of the NPI-bearing (212b) with at all than it is to the bare sentence in (212c). (212) a. Can’t nobody climb that hill. ‘Nobody can climb that hill.’ b. Nobody at all can climb that hill. c. Nobody can climb that hill.

168 See Brown and Levinson (1978) on face-threatening acts. 169 Kadmon and Landman’s conception of NPIs is cited in Green (2014) as influencing her account of semantic widening in the NI construction, which is described above in Section 5.1. 170 Kadmon and Landman’s semantic widening account of NPIs has been forcefully questioned in numerous places. See especially, however, Rohrbaugh (1994) and Krifka (1995), which argue that Kadmon and Landman’s reduced tolerance of exceptions is actually a product of focal stress. See also Rullmann (1996), Duffley and Larrivée (2010), and Israel (2011).

5 History of emphatic claims

149

Importantly, although (212a–c) are truth conditionally equivalent, they differ in speaker involvement, or emphatic qualities, in Israel’s terminology. The NPIbearing (212b) shows more speaker involvement and emotion than a default statement such as (212c). As such, we might expect (212b) to be used in contexts in which the speaker is more comfortable with her audience, in which it is safer to make a more emotionally-laden, involved statement. The same is true of the NI in (212a). Not only is (212a) dialectally marked and socially stigmatized, it also belongs to a highly informal style. Accordingly, it is clear how the use of (212a) in conversation signals something about the interpersonal relationship of the speaker and addressee. Returning to the quoted passage above, I argue that use of NI conveys to the hearer that “you are the sort of person with whom I would express myself in this way” and so serves as a marker of camaraderie and solidarity. Because it is highly stigmatized, use of NI requires significant trust in one’s speech partner. In other words, if a speaker can trust another speaker to judge her in the desired way in her use of an NI, then NI can come to indicate a kind of shared trust or intimacy. It is no surprise, then, to understand NI as a stigmatized construction, but then also to understand it as a marker of camaraderie and solidarity.171 It is this latter part of the social meaning that I am assuming is Israel-emphatic. In this way, NI conventionally encode Israel-emphatic qualities rather than Green- or Labov-emphatic qualities, and this ultimately allows us to derive the emphasis of the construction, which is the topic of this chapter, via Gricean reasoning, as discussed below.

171 We can compare this aspect of NI to the tu forms of second-person singular pronouns in some Romance languages. While tu forms are not stigmatized, use of tu does require trust in one’s speech partner. Just as use of tu in Parisian French signals speaker beliefs about the interpersonal relationship with her interlocutor, so too does use of NI. And just as with tu, successful use of NI requires interlocutors to share similar beliefs about this relationship, or else the speaker risks committing a stylistic faux pas comparable to an unreciprocated or unappreciated tutoiement. The fact that this contextual mismatch is possible with NI suggests that the social information it conveys must be conventionally encoded. If it were not conventionally encoded and was conveyed instead as a Gricean (1975) conversational implicature, for instance, we would expect the non-supporting context to cancel the implied social content, as described in Grice (1978: 115) and in Chapter 4, above. Thus, in a context where use of NI is not warranted – say, when speakers are strangers to each other and of obviously different social statuses – the social content of NI is still present, and it could very well result in conversational or social awkwardness.

150

Chapter 5 Negative inversion and emphatic meanings

5.2 Emphatic is conventional, but emphasis is pragmatic As we saw above, the idea that NI is a conventional means of conveying emphasis has been with us since the beginning, especially in the AAVE literature with Labov et al. (1968). This assumption is not universal though for all NI dialects, nor for all authors writing on AAVE.172 For example, Matyiku (2017) writes: In other varieties exhibiting negative auxiliary inversion, the phenomenon does not seem to correlate with an affective property or a widening effect. Foreman (1999) observes that it does not in West Texas English and Zanuttini and Bernstein (2014), and Blanchette (2015) also observe that it does not for their Appalachian English consultants.

This presents a question as to how emphasis arises and where it is to be found. Based on my own native speaker intuitions as well as survey results and many informal conversations, I argue that the NI of the present book does not conventionally, obligatorily, convey emphasis, though it certainly can and frequently does so, pragmatically. Two questions on the speaker surveys described above in Chapter 4, targeted NI and emphasis in particular. In (213), participants were asked to rate the emphasis of an NI sentence on a modified Likert scale of 1–5.173 The AAVE participants had a mean score of 1.5 (SD=.59), the Anglo participants had a mean score of 1.48 (SD=.5), and the Chicano participants had a mean score of 1.52 (SD=.53), suggesting that all of the groups found the NI in (213) to show relatively little emphasis.174 (213) On a scale of 1 to 5, with 5 showing the most emphasis and 1 showing the least emphasis, please rate Speaker B’s statement. [Standing outside the public swimming pool, which is closed early for the day]

172 One possible way of reconciling this seeming mismatch across dialects is to say that NI conveys emphasis conventionally in some varieties but not others. For example, perhaps emphasis is conventional in the varieties studied by Labov et al. (1968) [New York City], Sells et al. (1996) [East Palo Alto, CA], and Green (2014) [unknown], but is not conventional in the African American, Anglo, and Chicano varieties under investigation here. This would be a possible last-resort strategy, but it is not a preferable one. 173 Two different survey forms were distributed in each class. The forms contained identical questions, but the questions were scrambled in order to minimize possible order effects. 174 Given the small sample sizes and the fact that the questions are individual Likert items rather than multi-item scales, no inferential statistics were conducted. The one-sided means are quite telling in their own right, however.

5 History of emphatic claims

151

Speaker A: Hmm. Wonder why they closed the pool early? Speaker B [shrugging shoulders]: Didn’t nobody show up, I guess. Consider (214) for comparison, which asks the same question of a non-inverted negative sentence. The results of the three groups were very similar to those reported in (213). The AAVE participants had a mean score of 1.65 (SD=.58), the Anglo participants had a mean score of 1.53 (SD=.67), and the Chicano participants scored 1.52 (SD=.44). (214) On a scale of 1 to 5, with 5 showing the most emphasis and 1 showing the least emphasis, please rate Speaker B’s statement. [Standing in the gym with friend, looking for a pickup basketball game] Speaker A: I wonder where everybody is? Speaker B [shrugging shoulders]: Nobody showed up, I guess. The scores in (213) and (214) suggest a level of emphasis for NI that is very close to that of its non-inverted counterpart. This would be unexpected if NI was affective or emphatic or conveyed emphasis in the conventional ways described in Labov (1972a) or Green (2014). Under those accounts, we would expect NI to always be strong and convey emphasis on every use. For the present examples, the hedged contexts and language played a strong role in diminishing the level of emphasis, but if there were truly a conventional association of emphasis with NI, we would not expect to see this occur. However, on the account pursued in this chapter, in which emphasis is taken to be a pragmatic effect, the survey responses are not surprising. Another means of testing strength and emphasis with respect to NI can be adapted from Horn’s (1972, 1989) diagnostics for teasing apart scalar strength. Horn relies on connectives such as X or at least Y and X in fact Y to test relative strength in quantitative scales. These tools can be adapted here to attempt ordering NI sentences and their non-inverted counterparts in the same way. Thus, in (215), which takes Horn’s X or at least Y test, the X constituent should be stronger than the concessive Y constituent. Let us see how NI fares in this diagnostic when paired with its non-inverted counterpart. If NI were stronger or more emphatic or wider in some way, we would expect (215a) to be acceptable. The reversed ordering in (215b) is similarly unacceptable. This suggests there is no difference in strength in these two sentence types: especially when compared with (215c), in which the Y constituent is obviously weaker, and in which Horn’s diagnostic applies in the way in which it is expected.

152

Chapter 5 Negative inversion and emphatic meanings

(215) a. #Didn’t nobody come to the party; or, at least, nobody came. b. #Nobody came to the party; or, at least, didn’t nobody come. c. Didn’t nobody come to the party; or, at least, didn’t many people come. Now consider the second Horn diagnostic, with X in fact Y in (216) What this diagnostic is argued to show is that the Y term is stronger than the X term. The results mirror those shown in (215) and provide a compelling argument that there is no conventional difference in strength or widening between the two sentence types. (216) a. b. c. d.

#Didn’t nobody come to the party; in fact, nobody came. #Nobody came to the party; in fact, didn’t nobody come. Didn’t many people come to the party; in fact, didn’t nobody come. #Didn’t nobody come to the party; in fact, didn’t many people come.

The results of the Horn diagnostics here, paired with the survey results in (213) and (214), suggest there is little difference in conventional strength or emphasis between NI and its non-inverted counterpart. As such, conventional accounts of emphasis like those of Green (2014) seem an unlikely match for the present data. Still, the intuition that NI is somehow stronger or more likely to convey emphasis is very real. In the next sections, I provide a pragmatic account for why this is so. The pragmatics of emphasis As we have seen thus far, NI bears conventional social meanings above and beyond that of its truth conditional meaning. This fact allows for the emphasis often associated with NI to be derived pragmatically. In this section I illustrate several straightforward means in which NI can convey emphasis, each of which is dependent on NI’s Israel-emphatic nature and/or its Labovian social marking. Undoubtedly there are further ways in which NI might convey emphasis in particularized contexts; the point here, however, is to show that emphasis can be derived pragmatically from the independently attested conventional contents of NI, which saves the need of building emphasis into the construction’s conventional semantics in Chapter 6. The first pragmatic means I will mention is a variety of code-switching, or style-shifting. It is well known that all speakers control different linguistic styles. Wolfram and Schilling (2016: 281) write: Everyone, no matter who they are or where they live, uses a range of speech styles – that is, they engage in style shifting. This is because language variation is intricately tied to social meaning. On one level, social meanings and their associated linguistic usages may

5 History of emphatic claims

153

seem to be relatively fixed. [. . .] Speakers may use language based on their region of origin, their social class, ethnicity, and/or gender. On another level, however, speakers very often use language to change social meanings, including social and interpersonal relations and personal identities. A young white female in Northern California, for example, may use linguistic style shifting to show closeness to or detachment from her regional ties, affiliation with or distance from her conversational partners or the topics they’re talking about, and a range of different facets of, or even types of, personal identity – perhaps a serious university student, a flirtatious date, or a caring daughter.

With respect to the Texas speakers, NI clearly belongs to a range of informal styles of the predominately working class and is what Labov would refer to as a “social marker,” which is a linguistic item that “may lie below the level of conscious awareness, [but] will produce regular responses on subjective reaction tests” (Labov 1972b: 314). The idea here is that speakers are aware that the linguistic item in question sets the manner of speaking apart somehow from other regional varieties or social styles, and that the items may carry particular social evaluative information as well – i.e., be seen as prestigious, stigmatized, or convey some other dimension of social meaning. This description is quite apt for NI, which drew many survey responses labeling it as informal, to be used with family and friends, improper, the way rednecks talk, the way uneducated people talk, rural and so on. Survey participants were quite aware of the Negative Inversion and tended to have very strong feelings about it.175 This situation provides a straightforward means of accounting for emphasis with NI in terms of style-shifting. It has long been known that code-switching and style-shifting can be used to convey emphasis, among a

175 This situation with NI is very similar to that which Labov (2012: 16) describes for the -in(g) variable found in unstressed, word-final syllables of words such as fishin’, mornin’, etc. The quote from Labov below is long, but it is worth quoting in full to illuminate the comparison. In the classroom or on the pulpit, people will attribute the use of the -in’ form to laziness, ignorance, or just plain rascality. Yet the high value we put on the -in’ form in other contexts is not hidden from public view. When we see the large illuminated sign, DUNKIN’ DONUTS, we recognize the claim that dunkin’ donuts taste better than dunking donuts. The -in’ form, as we have seen, is associated with home language, and DUNKIN’ DONUTS calls upon the general belief that home cooking is better than commercial cooking [. . .]. People throughout the country use the -in’ form more when they are speaking informally, less when they are speaking formally. People with more power, education, and money use -in’ less often than people who rank lower on these dimensions of social life. We sum up these findings by saying that that variable (ING) is stratified by style and social class, evaluated by a social consensus, and available for public discussion. See also Campbell-Kibler (2007) and the numerous references therein on social meanings associated with -in(g) variation.

154

Chapter 5 Negative inversion and emphatic meanings

wide range of other discourse meanings. For example, Gumperz (1970: 8–10) and Gumperz and Hernández-Chavez (1972: 97) discuss a speaker’s shifting from standard English to AAVE to convey emphasis. Labov (1972a: 804) reports a similar situation, in which a speaker switches from standard English to AAVE for purposes of emphasis. Similarly, Brown and Levinson (1978) discuss several cases of switching involving various languages in which emphasis or involvement is conveyed as a result. Gumperz (1982) discusses a range of cases and languages in which extra-propositional information is conveyed in a similar manner. In all of these cases, it is not that the linguistic features themselves are necessarily more emphatic in one code than another; rather, it is the very act of switching that generates the emphasis meaning. As Gumperz (1982: 94–45) writes, “Perhaps the closest analogue to this view of what code-switching does can be found in Paul Grice’s discussion of conversational implicature.” Essentially, the shift from one code to another invites addressee inferences as to speaker purpose in so doing. In fact, this understanding of style is built into Irvine’s (2001) conception of style in general. According to Irvine, “[Styles] are part of a system of distinction, in which a style contrasts with other possible styles, and the social meaning signified by the style contrasts with other social meanings” (2001: 22).176 Thus, in making a stylistic move, a speaker contrasts with other potential stylistic moves she might have made.177 The implicit question in context, then, is why did she make the choice she did? Emphasis is one possible answer to that question, and it is straightforward to conceive of speakers of Texas English generating emphasis in the same way with use of an NI, which clearly functions as a socially marked linguistic form. Now, it is worth mentioning here that the concept of style-shifting is not well understood in its entirety. As has been noted frequently in the literature, it is not always clear the extent to which a complete “shift” from one dialect or register to another actually occurs in a style-shift situation. Wolfram and Schilling (2016: 283) write, for example:

176 See also Coupland (2007: 2): “Reading the meaning of a style is inherently a contrastive exercise. You have to find those [observed things] ‘different’ in order to see them as having some stylistic significance. This is the old principle of meaning depending on some sort of choice being available.” Also, Hickey (1993: 574): “The basic principle in any stylistics or study of style is that there must be more than one way of doing or saying something, what is done or said being different from how it is done or said.” And Akhmanova (1976: 3): “The concept of style presupposes the existence of objects which are essentially identical but which differ in some secondary, subservient feature or features.” 177 Ervin-Tripp (2001: 49) also discusses dialect feature shift functions such as “change of domain or stance, emphasis, or emotion.”

5 History of emphatic claims

155

Because there is no clear dividing line between register/genre shifting and dialect shifting, or between dialect shifting and code-switching, and because people rarely simply “turn off” one dialect or register and “turn on” another, it is perhaps more fruitful to think about stylistic variation – as with social and ethnic-group based variation – in terms of stylistic repertoires rather than register or dialect per se. Again repertoire can be taken to refer to the collection of linguistic features that each individual has at his or her disposal at any given moment, to be employed as needed for different social, interactional, and personal reasons, rather than conceiving of people shifting into and out of abstract entities like “Latino English” or a “legalese register [. . .].”

This observation has a clear place in the present work. I have been referring to style-shifting thus far, allowing the assumption of a complete shift from style 1 to style 2. In reality, though, it is not necessary for a speaker to perform a complete shift in this way to create the emphasis in question. Using an NI within otherwise standard speech can have the effect of indexing a second code and so can achieve the same result as a complete switch. Consider Brown and Fraser (1979: 48). If in bilingual communities choice of code marks formality, in monolingual, monodialectal speech communities the same kinds of information about formality of scene are conveyed by switching between levels of style. Here the switch is not between discrete, independently isolable codes, but between varieties of a single language; a great proportion of the codes overlap and the different styles may be signaled by a relatively few items which are loaded with social significance as markers of that particular style [. . .]. [emphasis mine].

Thus, it isn’t necessary to fully switch from one style or code to another; rather, it is possible to create the desired result performatively by using a clear dialectally and stylistically marked form such as NI.178 A second means of conveying emphasis with NI does not depend on styleshifting at all; rather, it can be derived directly from the Israel-emphatic content posited of the construction. Let’s return briefly to part of the Israel quote above. In using a negative polarity item, according to Israel, a speaker: expresses an attitude toward her audience, saying ‘you are the sort of person with whom I would express myself in this way.’ [. . .] Emphasis, on the other hand, is a sign of intensity and speaker involvement. As such, it serves as a marker of camaraderie and solidarity with the hearer.

178 Cf. Poplack (1980) on “emblematic code-switching” and Gumperz and Hernández-Chavez (1972: 90) on individual forms functioning as ethnic identity markers. See also Urciuoli (1996), cited in Fought (2006), in which a single feature, in this case use of habitual be, causes a speaker to be identified as African American. Cf. also Irvine and Gal’s (2000: 37) discussion of “iconization,” and Eckert’s (2012: 94) discussion of distinguishing linguistic features of a given population and the uses to which such features can be put in the sociolinguistic marketplace.

156

Chapter 5 Negative inversion and emphatic meanings

With respect to Negative Inversion, if it conveys social information of speaker camaraderie and solidarity with the hearer, it is only natural that speakers would feel more comfortable making strong claims in a context like this – i.e. in a context in which solidarity has already been established. LePage (1968: 192) tells us that speakers “construct their system of verbal behavior to resemble that common to the group or groups with which [they] wish from time to time to be identified.”179 I propose that NI is such an identifier for the groups in question. Thus, if a speaker uses NI, she has signaled that she is speaking from within the group, from a place of comfort, which also invites the hearer to infer additional strengthening of the NI proposition. One might pause and ask here how the findings mentioned above, in which NI is described as improper, the way rednecks talk, the way uneducated people talk, informal and so on, squares with the idea of NI serving also as a marker of camaraderie and solidarity. This rather puzzling pairing of sentiments turns out to be common with vernacular features and varieties, however, and has been discussed since at least Labov’s (1972b) work on covert and overt prestige. Similarly, creole languages are very frequently described in both ways by their speakers. For example, Rickford (1983: 156), partially quoting Reisman (1970: 40), writes “Creole [in Antigua] violates English standards of ‘order, decorum, quietness, and authority,’ but in which people in fact ‘take great joy.’” See also Salmon and Gómez Menjívar (2016) and Gómez Menjívar and Salmon (2018) for similar findings with respect to Belizean Kriol. Thus, it is not surprising to find nonstandard forms such as NI considered as incorrect but at the same time as expressing familiarity and camaraderie. In fact, the latter qualities are very likely brought about as a result of the former. Thus, if a speaker uses a stigmatized form such as NI in a comfortable situation, it is possible for the stigmatized form to take on a positive meaning of camaraderie in the process, perhaps not unlike what happens when slurs are reclaimed by marginalized communities as in-group speech.180 In other words, a speaker’s use of NI requires a significant amount of trust in her speech partner. If a speaker can trust another speaker to judge her in the desired way in her use of NI, then NI can come to indicate a kind of shared trust or intimacy. The third means of generating emphasis I will discuss here abstracts away from the social meanings so far considered and relies instead upon the basic pragmatic mechanism of choice and invited inference. Thus, speakers of the

179 In later work, Le Page and Tabouret-Keller (1985: 14) define a concept they call an act of identity, which is a linguistic choice through which “people reveal their personal identity and their search for social roles.” See also Bell (1984) on speech accommodation theory. 180 See Curzan (2014) on reappropriation of pejorative terms.

5 History of emphatic claims

157

linguistic varieties under consideration here have other options for expressing the negative propositions in question: i.e. speakers can choose NI or its noninverted counterpart, both of which are present in their linguistic varieties. For example, consider (217), which is part of a transcript from jazz trumpeter Buddy Anderson, describing composition styles of Charlie Parker and Dizzy Gillespie, quoted in Gillespie and Fraser (1975: 119). The first bolded sentence is in the non-inverted word order, Nobody could X. In the very next line, however, Anderson switches to NI order, Couldn’t nobody X. Both word orders are thus clearly available to him. (217) γ Bird [i.e. Parker] paid strict attention to what people did, and if he found anything they did that struck him, he brought that into his thing. That’s the reason he had so many ‘outs’ because he borrowed from everybody he could, and he’d embellish it and what not. But Diz is damn near all Diz, and it’s a little bit studied, but nobody could do it but Diz. They couldn’t copy Diz. They copied Charlie, but couldn’t nobody copy Diz. I mean it took a long time. The same is seen to be true in (218), which is a gospel lyric from Pastor John P. Kee’s 1990 song ‘Can’t nobody do me.’ Pastor Kee, who is from North Carolina, alternately sings Can’t nobody do me and Nobody can do me several times at the end of the song, with both the inverted and non-inverted word orders clearly available to him.181 (218) γ I’ve got You, Jesus I won’t let go Can’t nobody do me No, no Nobody can do me Oh yes, can’t nobody do me The choice between these two grammatical forms in (217) and (218) is, as Bakhtin says above in the chapter epigraph, a stylistic act. It is an interactive, rational

181 Examples like (218) are problematic for Green’s (2014) account of NI strength, in which the NI makes a semantically wider, stronger claim than the non-inverted counterpart. On Green’s account, then, Pastor Kee’s back and forth here would have a strong claim with the NI, then a weaker claim with the non-inverted counterpart, and then a return to a strong claim with the next NI, resulting in something like indecision. This is clearly not the intent of the song, however, as readers can judge for themselves here: .

158

Chapter 5 Negative inversion and emphatic meanings

move. Speakers make a choice for reasons of style and information. It is also more than this, however, as it can invite inferences above and beyond the sentential propositional meaning. One way of conceptualizing this is with Horn’s (1984, 2004) Division of Pragmatic Labor.182 Here is Horn (1984: 22) in a description of the process: The use of a marked (relatively complex and/or prolix) expression when a corresponding unmarked (simpler, less ‘effortful’) alternate expression is available tends to be interpreted as conveying a marked message (one which the unmarked alternative would not or could not have conveyed).

The idea is that if a speaker uses a marked expression to describe a situation, it is likely that the speaker believes something is marked about the situation itself. This in turn invites addressees to calculate conversational implicatures above and beyond the truth conditional content. Horn (2004:16) illustrates this process, as in (219): (219) a. He stopped the machine. b. He got the machine to stop. c. There is something unusual about the way he got the machine to stop. While (219a–b) are truth-conditionally consistent, use of (219b), which is more prolix, likely would convey a conversational implicature that there is something unusual in the way he got the machine to stop, as in (219c). Horn is concerned here primarily with markedness as a product of prolixity; however, markedness can be achieved many ways.183 NI is clearly grammatically marked in the relevant dialects. Further, it appears in fewer contexts than its non-inverted counterpart: i.e., the non-inverted counterpart is acceptable in every context and speech genre, even informal ones, while NI is acceptable in only a subset of those. Accordingly, it is straightforward to assume that a speaker’s use of a marked NI form can invite the

182 See McCawley (1978) for description of a similar reasoning process. 183 For example, Comrie (1976:111) defines markedness much more generally: The intuition behind the notion of markedness in linguistics is that, where we have an opposition between two or more members [. . .], it is often the case that one member is felt to be more usual, more normal, less specific than the other (in markedness terminology it is unmarked, the others marked). See especially Haspelmath (2006) for discussion of the many kinds of markedness which have been discussed in the linguistics literature.

5 History of emphatic claims

159

addressee to draw inferences with respect to a marked situation, including the speaker’s intention or state of mind in using it. It is only a small step from there to inferring emphasis or a stronger reading of a statement. Along similar lines, NI is marked compared to its non-inverted counterpart by dint of its conventional semantic content. As seen in Chapter 4, NI is socially marked in the Labovian sense, which is not the case with the non-inverted form. Based upon this extra conventional, but non-truth conditional, content, we find another straightforward means of inviting a conversational implicature of emphasis. In the introduction to the present book, I discuss Horn and Abbott’s (2012) argument that the conventional implicature content of uniqueness associated with the definite article the can be used to invite conversational implicatures of familiarity. For present purposes, it is straightforward to argue that a conversational implicature of emphasis can be invited through the use of an NI, as a result of its conventionally associated social content. To sum up, it is clear that the emphasis frequently associated with NI must be pragmatic in nature. The survey questions reported above show that emphasis doesn’t always accompany NI, and the Horn diagnostics and my own native speaker intuitions confirm this. What remains, then, is to show how emphasis can be derived pragmatically from the use of NI. I have shown above three straightforward and very plausible ways of doing so.

5.3 NI as linguistic variable? A question that has not been raised in this chapter or elsewhere in this book directly is that of the Labovian linguistic variable and whether or how this concept is relevant to the treatment of NI given here. First described in Labov (1966, 1972b), the linguistic variable has played a central role in variationist sociolinguistics since that time. As originally envisioned by Labov, linguistic variables were “alternate ways of saying ‘the same thing’” (1972b: 188), or, two forms that had equivalent meanings. It would thus be tempting to argue that NI is in such a relation with its non-inverted counterpart; I do not do so in this present work, however, for a few reasons. On the one hand, it is not clear in the literature if or in what way the linguistic variable concept extends to grammatical features beyond the phonological level.184 As Terkourafi (2011: 344) writes in the introduction of her article on the linguistic variable and pragmatics:

184 The most commonly studied variables have been phonological or morphological, with perhaps the most widely discussed example the pronunciation of (ING) in words like fishing or

160

Chapter 5 Negative inversion and emphatic meanings

This article revisits the thorny question of the linguistic variable’s applicability beyond phonology, where it has proved most fruitful, to pragmatic variation in particular. Raised in the 1970s, this question was hotly debated by researchers until the early 1990s, when it was set aside with inconclusive results [. . .].185

A second reason for not bringing the linguistic variable discussion into the present account of NI is that the status of the concept itself has grown somewhat unclear over the last several decades.186 In Labov’s original discussion of the linguistic variable, it was necessary for the variants in question to have the same truth-conditional meaning, and this was an important defining criterion. This argument has been questioned over the years, though, with some researchers, such as Lavandera (1978), Cheshire (1987), Buchstaller (2006), and others, arguing that truth-conditional equivalence is inadequate, and that it is more accurate to discuss linguistic variables in terms of discourse functional equivalence. Operationalizing this latter insight, Terkourafi (2011) argues that variables must encode the same “procedural” meaning, which is a term of art from relevance theory (Sperber and Wilson, 1986/1995), taken to direct addressees’ interpretations of an utterance. For present concerns, it is not necessary to argue one way or the other if NI and its non-inverted counterpart fit into one of these ways of characterizing linguistic variables. It is enough to say, as I do here, that NI is truth-conditionally equivalent to its non-inverted counterpart, and that it encodes a range of social meaning, which is conventional and non-truth conditional, which is not similarly encoded by its counterpart.187 This sets NI apart from its non-inverted counterpart, and so provides a semantic basis for deriving the pragmatic implicatures of emphasis with which this chapter is concerned.

singing. Here, the surface form varies between the standard velar nasal [-ɪŋ] or the alveolar nasal [-ɪn], with no apparent effect on truth-conditional meaning, but with the former assumed to evoke standard or formal situations and the latter informal or uneducated and so on. 185 See also Walker (2010) for extensive discussion of this problem. 186 See Wolfram (1991), which raises foundational questions as to exactly what researchers are looking at when they look at the linguistic variable. 187 On the other hand, as was seen above, in the AAVE under consideration in Green (2014) and Green and Sistrunk (2015), NI and its non-inverted counterpart are assumed to have different truth conditional meanings, so an argument toward the two sentence types being sociolinguistic variables there would need to proceed under the functional equivalency approach of Lavendera (1978), Buchstaller (2006), etc., if at all. See Guy et al. (1986: 45–47) for discussion of the kinds of theoretical problems that can arise though with “meaningful variation.”

5 History of emphatic claims

161

5.4 Conclusion This chapter has argued that among the Texas survey participants, Negative Inversion sentences are not emphatic in ways that have been previously assumed in the literature. Instead, NI is argued to be a sociolinguistic marker of speaker involvement, language variety, and style. The chapter argues that the emphatic affect often felt with these constructions is not a product of syntax or semantics as much as an interaction of Gricean (1975) pragmatics and this sociolinguistic content. NI is a Labovian social marker in the sense that use of it clearly marks the social class and informal style of the speaker as well as her beliefs about her audience. It is argued here that the emphatic qualities of the construction can be derived from the resulting social dynamics of its use. This strategy yields important dividends, as it respects the intuitions of those scholars who have argued that the construction is not emphatic; at the same time, it also accounts for the emphatic affect that very frequently does accompany the construction. It also provides a means of pragmatically deriving the reinterpretation of Green (2014) and Green and Sistrunk (2015), in which the construction is argued to be semantically wider. Further, as an investigation of Gricean pragmatics, style-switching, and social meaning, this chapter is a significant contribution to a largely unexplored area of sociolinguistic-pragmatics in general. Finally, deriving emphasis from the social content of NI and conversational dynamics is preferable in its economy. Grice (1989: 47) reminds us of a principle he might call “Modified Occam’s Razor,” which tells us that “senses are not to be multiplied beyond necessity.” This is exactly what is attempted in the present chapter. It is clear that the social content of NI is there already. It is conventional. The dynamics of conversation in general are well-established, if not well-understood. If we can assume these two things, then the emphasis of NI comes for free and need not be built into the constructional semantics, which is taken up in the final chapter of this work.

Chapter 6 Negative inversion is a formal idiom A reasonable approach would be to focus attention on the core system, putting aside phenomena that result from historical accident, dialect mixture, personal idiosyncrasies, and the like. – Noam Chomsky and Howard Lasnik, ‘The Theory of Principles and Parameters’

6 Introduction Based on the accumulation of claims made in the preceding chapters, this final chapter argues that a reasonable approach to understanding NI is to consider it a formal idiom (FI), as originally described in Fillmore, Kay, and O’Connor (1988) and developed in numerous later works.188 The NI has an idiomatic syntactic frame and encodes specific semantic meanings: i.e. that of speaker involvement and the additional social marker content as described above in Chapters 4 and 5. At the same time, the rigid syntactic frame contains variable slots that are constrained to allow only particular forms of subjects and auxiliary verbs. Thus, NI as seen here is partly generative and partly idiomatic, but unlikely to be part of the core referred to above in the epigraph. Taken as a whole, the FI account allows us to account for the structure, meaning, and usage requirements that make up the entirety of the Negative Inversion construction. Along the way in this chapter, critiques are given of previous syntactic accounts of the construction. The NI has received significant attention since Labov et al. (1968) and continues to be discussed in numerous articles, conference proceedings, and very recent dissertations, such as Blanchette (2015, CUNY) and Matyiku (2017, Yale). Almost all previous accounts have relied upon generative grammar frameworks, assuming generally the same empirical description of the construction, and running up against essentially the same puzzles.189 This chapter shows that these approaches to this particular construction are ultimately unmotivated, or else are arranged to account for different empirical objects, and so this

188 For a range of works on FIs, see, for example, Michaelis and Lambrecht (1996), Kay and Fillmore (1999), Goldberg and Jackendoff (2004). See also Culicover’s (1999) work entitled Syntactic Nuts, though the term “formal idiom” is not actually used in that work. I have chosen to discuss NI in the general terms of Fillmore et al.’s FI because it is well known and can be translated into numerous other frameworks straightforwardly. 189 The only exception of which I am aware is Weldon (1994), which is couched in the headdriven phrase structure framework of Pollard and Sag (1994). https://doi.org/10.1515/9781501512346-007

6 Introduction

163

part of the chapter clears a path forward for a more inclusive and efficient way of thinking about the construction. There have been two prior accounts focusing specifically on Texas Englishes: viz, Foreman (1999) and Matyiku (2017). Most of the other theoretical work on the constructions focuses on AAVE or Appalachian English. In the discussion below, I will discuss the Foreman and Matyiku accounts specifically; however, it is often necessary to look closely at the proposals given for the construction in other varieties as well, as they have played an important role in the basic theoretical understanding of it over the years. The chapter begins, then, with a discussion of Labov et al. (1968), which was the first contemporary discussion of NI and which provides the foundation for most accounts that have been given since that publication. Finally, the conclusions presented in this chapter should not be taken as an argument against a generative grammar in general. In all grammars, whether generative or otherwise, there are some objects that due to rigidity of form and/or meaning will have to be assumed to be idiomatic. I am assuming that the particular NI construction of this book is one of those objects. In principle, a generative grammar account of the NI here is possible, though I don’t know how it would be arranged or what would motivate it. Some way would need to be devised to account for the subject restrictions and the social meaning as well, all of which falls out perfectly naturally on the post-existential idiom account provided here.

6.1 The nature of previous negative inversion analyses As is pointed out succinctly in White-Sustaita (2010: 429), researchers tend to take one of two approaches in accounting for the grammatical structure of Negative Inversion. In White-Sustaita’s words: An auxiliary inversion analysis argues that the word order is derived via movement of the auxiliary to the left periphery, whereas an existential analysis argues that the word order is an artifact of deletion of the expletive subject, paralleling there-insertion existential constructions.

Both of these approaches have their origins in Labov et al. (1968), which, as I show below, argued that two different kinds of approach were necessary to account for the range of data. Most treatments of NI since Labov et al. (1968) have taken the former tack, i.e. positing some type of movement and subject-auxiliary inversion. In this present book, and in Salmon (2018a), I argue that the latter approach is actually the more fitting one. Let’s look now at Labov et al.’s two original proposals.

164

Chapter 6 Negative inversion is a formal idiom

6.2 The Labovian two-solution account Labov et al. (1968) is a book-length work in two volumes, which compares the non-standard speech of African Americans in Central Harlem with standard English. The two volumes span approximately 700 pages and cover a wide range of topics from phonology and grammar to narrative and language use in the community in general. Approximately five pages of Volume 1 are dedicated to Negative Inversion. As mentioned above, however, the two solutions proposed there by Labov and his colleagues have strongly influenced every account of the construction that has been given subsequently. This first mention of the construction in contemporary linguistics literature is given below, with the first description by Labov et al. It is worth reproducing here in its entirety: (220) In an early narrative of the danger of death, a Negro man from the SEENYC sample, raised in New York City, said of a gunshot: Didn’t nobody see it; didn’t nobody hear it. These sentences are puzzling because their semantic interpretation does not fit the known rules of [standard English] or [white non-standard English]. [. . .] The construction of [. . .] is that of a question, yet the intonation contour and the context plainly signal a declarative statement. It would seem at first glance that these sentences represent deep-seated differences in the organization of [African American non-standard] grammar, which affect meaning and the fundamental syntactic apparatus. However, the discussion in the following pages will show that this is not the case. In the next two sections, I provide an overview of the two solutions Labov et al. provide for the two kinds of data they discuss. 6.2.1 The Labovian expletive deletion account The first set of examples with which Labov et al. are concerned are given below in (221–224) [Labov et al. ex. 350–353]. The authors write that as far as the “formal structure and derivation are concerned, it is possible to relate them to sentences of the form Nothin’ ain’t happening’ or of the form It ain’t nothin’ happening’.”190 The first possibility would involve a movement operation to

190 In the AAVE variety of Labov’s study, expletive it is preferred in existential sentences rather than the there used in other English varieties. The italicized sentence here glosses as ‘There ain’t nothing happening.’

6 Introduction

165

result in the inversion effect; the second possibility, which Labov et al. describe as “the simplest,” is to simply delete the expletive it, rather than positing syntactic movement. (221) Ain’t nothin’ happenin’ ‘n shit. ‘There isn’t anything happening and shit.’ (222) Ain’t nobody complainin’ but you, man. ‘There isn’t anybody complaining but you, man.’ (223) Ain’t nobody gon’ let you walk all around town to find somebody to whup them. ‘There isn’t anybody who is going to let you walk all around town to find somebody to whup them.’ (224) Ain’t no white cop gonna put his hands on me. ‘There isn’t any white cop who is gonna put his hands on me.’ The authors go on to note, however, that there are some sentences that completely resist a syntactic movement account, as there is no non-inverted form from which to move. Thus, consider (225) [Labov et al. ex. 358]. While, (225) is a perfectly acceptable structure, its non-inverted counterpart in (226) is not acceptable. (225) Ain’t nothin’ you can do for ‘em. ‘There isn’t anything you can do for them.’ (226) *Nothin’ ain’t (that) you can do for ‘em. ‘There isn’t anything you can do for them.’ For sentences such as (226), it seems as if an expletive deletion account is necessary for Labov et al.191 Let’s look now at the second group of data.

191 Most researchers do not actually consider these sentences to be proper Negative Inversion at all, considering them instead as “the equivalent of there-insertion existential constructions” with the expletive lost phonologically due to normal conversational deletion. See Chapter 2, above, and White-Sustaita (2010: 432).

166

Chapter 6 Negative inversion is a formal idiom

6.2.2 The Labovian inversion movement account The authors begin the next section with example (227) [Labov et al. ex. 271], writing, “this is plainly negative inversion, somehow related to Nobody didn’t see it” as in (228). (227) Didn’t nobody see it, didn’t nobody hear it. ‘Nobody saw it, nobody heard it.’ (228) Nobody didn’t see it. ‘Nobody saw it.’ The authors then go on to provide comparable examples with don’t, didn’t, and can’t which they suggest cannot be accounted for by expletive deletion of it. Sentences such as (229–231) with expletives in residence, were ungrammatical in the variety Labov et al. were working with. As the authors write, “It ain’t cannot account for did, was, modals, so we need negative inversion anyway” (Labov et al. 1968: 286: a’).192 (229) *It don’t nobody break up a fight. ‘Nobody breaks up a fight.’ (230) *It didn’t nobody see it. ‘Nobody saw it.’

192 It is worth noting here that sentences (229–231), when reconfigured with expletives there or they, as in (i–iii), are in fact acceptable to many older speakers of the Texas Englishes under consideration in this book: including some of the Texas AAVE speakers. See Chapters 2 and 3 above and the discussions of Modal Existential sentences in those chapters. (i)

There don’t nobody break up a fight. ‘Nobody breaks up a fight.’

(ii) There didn’t nobody see it. ‘Nobody saw it.’> (iii) There can’t nobody tag you then. ‘Nobody can tag you then.’ As such, Labov et al.’s argument against an expletive deletion account for these sentences does not work for the Texas groups under consideration here.

6 Introduction

167

(231) *It can’t nobody tag you then. ‘Nobody can tag you then.’ Putting Labov et al.’s inversion hypothesis into more modern generative syntactic terms, Sells, Rickford, and Wasow (1996: 599) write that the account involves “head movement of a negative auxiliary from INFL, the head of IP, to COMP, the head of CP,” essentially a derivation in which the negative auxiliary raises from its canonical position in the verb phrase to a position higher than the subject. Labov et al. suggest that the inversion movement is triggered by the presence of an affect feature – essentially the affect-attraction of Klima (1964). For Klima, affect is a “grammatico-semantic feature” (Klima 1964: 313) common to negative terms and various syntactic configurations, which among other things, was thought to trigger subject-auxiliary inversion and license indefinite quantifiers like ever and any. Labov et al. look to this affect feature as motivation for the inversion movement, and they posit inversion as an optional rule. Sells et al. (1996) note, however, that affect does not seem to be the proper trigger for the inversion question, as positive sentences that seem similarly affective do not undergo inversion.193 Similar claims are also found in Foreman (1999) and White-Sustaita (2010), among others. Thus, while Labov et al.’s motivations for the inversion movement are not accepted by most researchers, the basic idea that there is some kind of movement in which the negative auxiliary raises over the subject is still accepted by most, as we see in the many movement-based accounts that have been given since.194 This basic illustration sets the stage for the next fifty years of research on the construction. To my knowledge, there have only been three accounts of Negative Inversion – in any varieties – that do not rely on syntactic movement. It is worth discussing these in the following section, as the account given in the present work also falls on this non-movement side of Labov et al.’s solutions.

6.3 Non-movement accounts after Labov et al. (1968) The earliest account of which I am aware that does not posit a movement analysis for Negative Inversion is Martin (1992). This dissertation covers a wide range

193 See discussion in Chapter 5. 194 See Labov (1972a), Foreman (1999), White-Sustaita (2010), Green (2011, 2014), Zanuttini and Bernstein (2014), Blanchette (2015), and Blanchette and Collins (2018). See especially Matyiku (2017) and the many sources therein.

168

Chapter 6 Negative inversion is a formal idiom

of topics but dedicates only three pages to Negative Inversion. As mentioned in Chapter 2, Martin’s concern in his discussion of NI is its behavior with respect to licensing negative polarity items. According to Martin, if inversion movement were to occur in NI, then negative polarity items should not be licensed in embedded clauses due to intervening syntactic constituents that would result from the inversion. Yet, as we saw in Chapter 2, this is not the case: NPIs are fine in embedded clauses of Negative Inversion. Here is (65), repeated below. (65)

Don’t nobody say that dealers sellin’ drugs in the school yard {no more / anymore}. ‘Nobody says that dealers are selling drugs in the school yard anymore.’

Martin reasons thus that no inversion movement must occur. Rather, as described above in Chapter 3, Martin posits a phonological deletion rule, which deletes the phonological form of an unstressed expletive subject. This is essentially all we are told on the subject by Martin, who then moves on to other concerns. Green (2014), however, notes that Martin’s suggestion will not work for the variety of AAVE with which she is concerned, as that variety (like that of Labov et al.) does not allow expletive subjects in Negative Inversion at all.195 However, Green continues “To Martin’s credit, the expletive destressing rule is plausible for other varieties of American English, such as Alabama English, because they allow expletives in a range of negative inversion sentences” (2014: 122). As was mentioned above in Chapter 3, though, the picture as to whether expletive subjects are allowed in NI in the Texas Englishes under consideration here is a bit more complicated. Recall, the youngest participants described in Chapter 3 mostly did not allow NI with expletive subjects (i.e. Modal Existentials), while the oldest participants did allow them – especially the Anglo and Chicano, with a smaller number of African Americans. Thus, it would seem that by Green’s reasoning, Martin’s rule could work for many of the older Texas participants, but not for the younger. Although this sounds problematic at first blush, I suggest in this book that this is very close to what does happen. As I argue in Chapter 3, for the older participants, it appears that the NI construction just is an expletive NI – i.e. a Modal Existential – that can undergo synchronic conversational deletion in allegro speech, much as negative there-existentials do. In the diachronic picture, though, with respect to the younger participants, the expletive subject is absolutely gone, much as with Green’s particular AAVE population. This diachronic

195 As noted in Chapter 2, however, this claim with respect to AAVE and NI and expletive subjects is not universally true.

6 Introduction

169

view of the expletive subject clears up a lot of the confusion that has surrounded it in the past. Parrott (2000), also writing on AAVE, similarly posits an account of NI that does not involve inversion movement; rather, he proposes that AAVE has a phonologically null weak expletive, which is a lexical item that is “in most respects identical to the English weak expletive there” (Parrott 2000: 419). In other words, Parrott suggests that AAVE has a silent there. NI is then derived from expletive sentences via various syntactic Merge operations that need not concern us here. Like Martin’s (1992) account, Parrott’s approach does offer the benefit of speaking to the definite subject effects (Milsark, 1974; Chapter 3 above) that have been claimed for NI: i.e. if the NI is a type of existential sentence then it should not be surprising for it to exhibit definiteness effects that are similar to those of prototypical existential sentences. On the other hand, Parrott’s account suffers from the need to posit null elements that don’t seem to be motivated on any other grounds.196 The final non-movement account to mention is that of Sells et al. (1996). In this approach, the authors assume that NI and negative existential sentences are essentially generated in the same manner as the underlying forms of TE sentences: i.e. the subject is generated internally to the underlying verb phrase. This is essentially a generative syntax approach, which relies upon optimality theory (Prince and Smolensky, 1993/2004) constraint rankings to dictate what forms are allowed to surface. Thus, Sells et al. assume a sentence like (232) [Sells et al. ex. 29] is generated with “nothing at all in the surface subject position (SpecIP)” (1996: 605). (232) Ain’t nothin’ happenin’. ‘There is nothing happening.’ Similarly, (233) [Sells et al. ex. 30] is also generated with the subject nobody generated internal to the underlying verb phrase, which is indicated by the brackets. (233) Can’t [nobody tag you then]. ‘Nobody can tag you then.’ Ultimately, the authors write that “AAVE will differ from [standard English] in allowing SpecIP to go unfilled.” What this latter quote means essentially is that

196 See Pullum (1991) on justifications for positing null elements in general.

170

Chapter 6 Negative inversion is a formal idiom

in AAVE, the authors suggest that the Extended Projection Principle of Chomsky (1982), which requires clauses to have an NP in subject position, is violable, where it was claimed to be inviolable in standard English.197 A series of optimality theoretic constraints are then ordered to predict the surface form. In this way, the authors attempt to bring together both of Labov’s solutions into one, to account for the negative existential sentence in (232) and the NI in (233) with one system. In the present work, as noted above and in Chapter 3, negative existential sentences such as (232), which occur in many dialects of English, are considered to be simply the product of synchronic conversational deletion. As such, there is no reason that an account of NI should seek to describe negative existentials as well; all that is needed is a general theory of conversational deletion.

6.4 Movement accounts of negative inversion motivated by disambiguation Two important accounts of NI have argued that the primary syntactic relation between the NI and its non-inverted counterpart is that the former is a disambiguated variant of the latter, and that disambiguation is actually a motivating force for the syntactic movement(s) described in those accounts. Foreman (1999, 2015) claimed that Negative Inversion sentences in West Texas English exhibit a particular scope pattern when they appear with universally quantified subjects. Essentially, Foreman (1999, 2015) wrote that non-inverted sentences such as (234) are “ambiguous between the negation wide scope (‘It is not the case that every person went to the party’) and the negation narrow scope (‘For all people, none went to the party.’)” (11). (234) Everybody didn’t go to the party. On the other hand, with NIs such as (235), Foreman (1999, 2015) writes “the negation must scope over the subject” (11), which would gloss as something like “It is not the case that all people went to the party,” and which would allow for some people to have gone to the party.

197 The requirement that clauses have an NP in subject position – i.e. Chomsky’s EPP – has been a central tenet of generative grammar, beginning in the 1980s. It has also been the subject of significant criticism and is sometimes denied altogether. See Bošković (2002) for broad discussion of the EPP, its significance, and criticism.

6 Introduction

171

(235) Didn’t everybody go to the party ‘It is not the case that everybody went to the party.’ For Foreman (1999, 2015), scope ambiguities of the type discussed for (234) are not possible for NIs such as (235): the negation unambiguously scopes over the subject. Foreman (1999, 2015) takes this observation as evidence that the negativized auxiliary verbs in NI are raising “for scope reasons” (11). Thus, Foreman (1999, 2015) reasons that the NI construction is motivated by the need to resolve scopal ambiguities found in the non-inverted counterpart. As Foreman (1999, 2015) writes in 2015 “[NI] is motivated by semantic interpretation. The noninverted counterparts of [NI] either have different meanings, or potentially ambiguous ones, that the Negative Inversion resolves” (142). In Foreman’s (1999, 2015) movement-based account, a second negative phrase is posited in the syntactic structure: above the subject position of these sentences, that is above AgrS-P. I label this NegP2 to distinguish it from the NegP originally proposed in Pollock (1989). It is this position [. . .] to which the Negative Aux moves in NI structures. Thus a sentence Cain’t nobody do that would have the following structure: [NEGP [NEG Cain’ti] [AGRS-P nobody ti. [NEGP [NEG tI] [VP do that]]]]

As Foreman (1999, 2015) goes on to write: “the subject nobody has raised out of its VP internal position into the syntactic subject position, Spec, AgrS-P. The Negative Aux has then raised over it into the head of NegP” (12). Despite the elegance of Foreman’s (1999, 2015) syntactic solution, it is one that is ultimately unmotivated for the present data. As we saw above in Chapter 2, the ambiguity mismatch Foreman (1999, 2015) hopes to resolve with the NI derivation appears to not actually exist, at least conventionally. Recall example (236), in which, pace Foreman (1999, 2015), the negation is coerced into a narrow scope reading by the context and intonation.198 (236) Speaker A: Hey man. How was the party last night? Speaker B: Not so good. Didn’t some people even bother comin’ out. Hell, what am I sayin’. Didn’t everybody come out. There wudn’t a soul there last night. It was a total bust. ‘Nobody came out.’ It has been known at least since Jackendoff (1972) that speakers use intonation and prosodic cues to disambiguate sentences involving negation and universal

198 This example appears as (50) in Chapter 2.

172

Chapter 6 Negative inversion is a formal idiom

quantifiers such as (234).199 It should be no surprise to find similar ambiguity possibilities also dependent upon intonation and prosody in the closely related NI as in (236). In addition, the fact that intonation and context steer interpretation in this way is itself a further argument against the movement-based disambiguation hypothesis. Language is unruly, unpredictable, and often inefficient; still, however, if the ambiguity of a universal sentence like (234) can be resolved via intonation, it would be surprising for speakers to resort to the complex derivation of a syntactically marked, highly stigmatized construction like the NI in order to disambiguate – especially when there are more straightforward, unmarked, prosodic means of doing so. In the present work, disambiguation plays no role in the conventional aspects of the account. The NI is simply a dialectal and stylistic variant: though, it is marked for a range of social meanings and is used to accomplish specific pragmatic purposes. A further problem for disambiguation accounts can be seen in NI examples with weak quantifiers, such as some, in which there is no difference in ambiguity between the NI form and the non-inverted form. Here, both the NI in (237a) and the non-inverted (237b) show the same ambiguous possibilities, which is problematic for a scope-based motivation such as Foreman’s (1999, 2015) . (237) a. Didn’t some of the students show up. ‘Some of the students didn’t show up.’ b. Some of the students didn’t show up. As such, there is no need or way to motivate a disambiguation account with such examples. Finally, it is worth mentioning that throughout Foreman’s (1999, 2015) account, he argues that NI shows strong structural parallels to English sentences with pre-subject not, as in (238) below [Foreman’s ex. 32a–f and judgements], and that this contributes to, for example, both sentence types ruling out definite subjects. (238) a. *Ain’t Jack seen the baby yet. ‘Jack hasn’t seen the baby yet.’

199 See also Syrett et al. (2014) for an experimental take on the question of intonation and ambiguity in such sentences, as well as an extensive discussion of previous literature on the topic.

6 Introduction

173

b. *Not Jack has seen the baby yet. ‘Jack hasn’t seen the baby yet.’ c. *Won’t the student answer the question. ‘The student won’t answer the question.’ d. *Not the student will answer the question. ‘The student won’t answer the question.’ e. *Cain’t their dogs do that. ‘Their dogs can’t do that.’ f. *Not their dogs can do that. ‘Their dogs can’t do that.’ As we have seen above in Chapter 2, however, NI does in fact permit definite subjects in the proper pragmatic contexts. Thus, in (239), we see an NI with a demonstrative subject, which is acceptable; the same subject is not acceptable, however, in a pre-subject not sentence, as in (240). (239) Cain’t ☞ + their dogs do that. ‘Their ☞ + dogs can’t do that.’ (240) *Not ☞ + their dogs can do that. ‘Their ☞ + dogs can’t do that.’ Thus, Foreman’s (1999, 2015) strong association of NI and pre-subject not sentences is also problematic for his larger account. Blanchette (2015) raises a further problem for disambiguating accounts of NI. Blanchette is ultimately concerned with Appalachian English, not Texas English, but the criticism is also applicable to the Texas varieties.200 The concern Blanchette raises has to do with subjects of NI that are morphologically negative, as in (241) [Blanchette’s (36)]. (241) Didn’t nobody live in there then. ‘Nobody lived in there then.’ According to Blanchette, sentences such as (241) receive a negative concord interpretation, with both -n’t and the negative subject marking the same negation. Assuming this, “if the structure of [241] is such that a negative operator

200 Example (241) comes from a corpus of Appalachian English, but the sentence is perfectly acceptable in the Texas varieties under consideration in the present work as well.

174

Chapter 6 Negative inversion is a formal idiom

moves over the quantificational subject nobody, then it is unclear how this movement could be construed as disambiguating, given that both elements mark the same negation” (Blanchette 2015: 114). Such examples are then further complications for disambiguation types of accounts. Horn (2015) raises a different challenge for disambiguation approaches, reminding us that many NIs in fact have no non-inverted counterpart, as in (242a–b) [Horn’s ex. 51a–b, with Horn’s judgments], and thus there is nothing to disambiguate. (242) a. Cain’t but two people sit in back. (#But two people cain’t sit in the back.) ‘Only two people can sit in back.’ b. Didn’t anybody go. (*Anybody didn’t go.) ‘Nobody went.’ Presumably, then, for a movement-based account, there must be something more driving the movement than the need to disambiguate.201 Foreman (1999) and Matyiku (2017) are the major competing accounts of NIs, which as we have seen, cannot be extended to the present data.

6.5 Movement accounts of negative inversion with other motivations Another movement account to consider is White-Sustaita’s (2010) proposal for NI in AAVE.202 An immediate mismatch, however, between White-Sustaita’s account and the needs of the present data is that according to Matyiku (2017), White-Sustaita predicts that NI should appear only with negative indefinite subjects, or n-words such as nothing, nobody, and so on. Thus, NIs such as (243) and (244), which have positive subjects, are not allowed to occur in White-Sustaita’s system.

201 Matyiku’s (2017) account of NI begins with assumptions similar to those of Foreman (1999), but it provides a much more in-depth treatment of the construction. Like Foreman, though, Matyiku’s movement account is ultimately motivated by the need to resolve scope ambiguities of the kind discussed above in (234) and (235). As we have seen, these ambiguities do not occur for the participants in this work, and so Matyiku’s account is not a good fit for the present data. Zanuttini and Bernstein’s (2014: 162) account of NI follows Matyiku’s account. 202 White-Sustaita does not provide information on the regional variety of AAVE with which she is concerned. According to Lars Hinrichs (p.c.), however, the NI data in White-Sustaita’s article was gathered in New Orleans.

6 Introduction

175

(243) Didn’t many people show up last night. ‘Many people didn’t show up last night.’ (244) Didn’t every member of the team get to play. ‘Every member of the team didn’t get to play.’ These sentences are perfectly acceptable in the Texas Englishes of the present work, though. See Matyiku (2017: 260) for further discussion of White-Sustaita’s NI proposal. An additional account that does not meet the needs of the present data is that of Green (2014) and Green and Sistrunk (2015). Green is also concerned with AAVE, though like White-Sustaita, Green does not identify the specific population under consideration. Green’s account is set up to describe the emphatic nature of her data – i.e. the absolute negation. She posits movement of a negated auxiliary to a position higher than the subject, landing in an articulated CP structure, reminiscent of Rizzi (1997). This position, she writes, is “associated with the left periphery [and] is linked to the absolute negation/strong domain reading” (Green 2014: 126). As we saw above in Chapters 2 and 5, however, the data in the present study is not required to receive an “absolute negation/strong domain reading.” Recall that in Chapter 5, above, sentences like (206) [Green’s (20)], in which there are exceptions appended to the NI, are found to be “strange or contradictory” in Green’s account. This is due to Green’s requirement that NI express an absolute negation with no exceptions. (206) #Don’t nobody ride bus number 201 – just the three people who live in the country. Most of the students in this class ride bus number 99. This sentence and others like it, however, were acceptable to my African American, Anglo, and Chicano participants. As such, it seems that Green’s system describes a different phenomenon perhaps in a different language variety than the ones with which we are concerned here.

6.6 Summary of previous accounts We see, then, that there is no satisfying account of NI on the market for our present purposes. As we have seen, all previous accounts – whether addressing AAVE or Anglo Englishes – make empirical predictions that do not fit the data in question. In the next sections, I propose a very different kind of structural

176

Chapter 6 Negative inversion is a formal idiom

account for the NI data. Almost all of the previous accounts in the literature have been framed within generative syntax. I argue here that a more complete picture of NI, and one that is much simpler, can be presented in terms of Fillmore, Kay, and O’Connor’s (1988) formal idioms, which allows us to talk about the structure of NI as well as the associated semantic and social meanings and pragmatic usage conditions.

6.7 Negative inversion as a formal idiom As we have seen throughout this work, Negative Inversion has a very rigid and recognizable structure in some ways but allows for a range of variability in others. With very little exception, NI requires a negative inflected auxiliary to appear clause-initially. There is also the requirement that the subject NP refer to an entity that is hearer-new in the sense of Ward and Birner (1995). This is a conventional requirement that the subject NP interact with the context in a specific pragmatic way. Negative Inversion also carries the specific intonational pattern of the declarative. Based on the surface syntax alone, however, we would expect the rising intonation of a polar question; yet, as is well known in the literature, NI carries a falling tone of declaration. Finally, NI is restricted to particular styles and speech genres and carries a range of associated social meanings. As such, NI presents a good case for what Fillmore, Kay, and O’Connor (1988) refer to as a formal or lexically open idiom. In the next section, I introduce the formal idiom as presented in that landmark work. Formal idioms: Background For Fillmore, Kay, and O’Connor, formal idioms contrast with lexically filled idioms. Lexically filled idioms are frozen phrases with specific meanings that allow little room for structural variation. Prime examples are phrases such as spill the beans and let the cat out of the bag. These idioms are non-compositional, must be learned one at a time, and allow little lexical variation: *spill the seeds and *let the kitten out of the sack. Formal or lexically open idioms, on the other hand, are “frozen” in some ways and generative in others. Fillmore Kay, and O’Connor (1988: 510) provide many well-known examples of formal idioms, perhaps the best known of which is the incredulity response construction (ICR), given in (245). Such constructions have some frozen idiomatic components and some components that are variable. As can be seen in (245), the frozen aspects of the ICR include: (i) a subject in objective case, (ii) a bare stem verb form, (iii) rising intonation, and (iv) a specific

6 Introduction

177

pragmatic context in which the speaker is responding to a dubious prior claim. All of this is built into the construction, in Fillmore, Kay, and O’Connor’s account: i.e., they are not predictable from general rules of grammar. Beyond these requirements, the constructions allow for considerable flexibility in the other constituent components. Importantly, however, assuming a typical Fregean composition process, the meaning and use requirements of the whole cannot be derived from the sum of the parts.203 (245) Incredulity Response Construction204 Him be a doctor?! Your brother help me?! Him write a novel about the Spanish Inquisition?! *He be a doctor? *Him was a doctor? *Him be a doctor.

[Subject in nominative case] [Verb in agreement form rather than bare form] [With declarative falling intonation]

Fillmore et al. provide numerous other examples of these flexible idioms, such as the correlative comparative in (246) [Fillmore et al.’s (1)]. According to Fillmore et al., The X-er, the Y-er in (246) is an idiom of sorts, that is “used for expressing a correlation between an independent variable and a dependent variable” and which is highly productive. These authors describe the productivity of the construction in the following manner: In spite of the fact that it is host to a large number of fixed expressions, the form has to be recognized as fully productive. Its member expressions are in principle not listable; unlimitedly many new expressions can be constructed within its pattern, their meanings constructed by a means of semantic principles specifically tied to this construction. (Fillmore et al. 1988: 507)

(246) The X-er, the Y-er The more carefully you do your work, the easier it will get. As Fillmore et al. suggest, constructions in this form can be multiplied without limit. Here are several more in (247) for good measure that were constructed by the author.

203 But see Kay and Michaelis (2012) for arguments that constructional composition can actually be consistent with the Fregean principle. 204 See early work on this construction in Akmajian (1984). See also Lambrecht (1990, 1994) and Michaelis (2009).

178

Chapter 6 Negative inversion is a formal idiom

(247) The taller you are, the thinner you look. The better you jump rope, the better reader you’ll be. The more books you have, the more boxes you need. The faster runner you are, the more shoes you have to buy. Like the ICR given above in (245), The X-er, the Y-er has basic constructional requirements, violation of which render the construction ungrammatical. For example, the articles must be definite. Indefinite articles result in an ill-formed sentence: (248) *A taller you are, a thinner you’ll look. Similarly, the X and Y variables must be in comparative form or the result is also ungrammatical, as in the non-comparative (249a) and the superlative (249b), which are equally ill-formed: (249) a. *The tall you are, the thin you are. b. *The tallest you are, the thinnest you are. The reader is invited to consult Fillmore et al. for a full account of the construction. Meanwhile, it is obvious that parts of this construction are idiomatic, while other parts are endlessly variable. In (250) [Fillmore et al.’s (14)], we see several other constructions that are simultaneously frozen and variable. Each construction in (250) is a token of an underlying formal idiom construction; each example has aspects of its structure and meaning that are “frozen” and aspects of it that are variable, much the same as we see with The X-er, the Y-er and much the same as we see with Negative Inversion. (250) a. b. c. d. e. f. g. h. i. j. k. l.

There goes Charlie again, ranting and raving about his cooking. Look who’s here! what with the kids going off to school and all. Why not fix it yourself? He’s not half the doctor you are. Much as I like Ronnie, I don’t approve of everything he does. He may be a professor, but he’s an idiot. Him be a doctor? What do you say we stop here? It’s time you brushed your teeth. One more and I’ll leave. No writing on the walls!

6 Introduction

179

m. That’s not big enough of a box. n. It satisfied my every wish. Fillmore et al. (1988: 511) go on to write that formal idioms typically receive a negative answer to both of the following questions: (i) Does the expression exhibit properties that are fully predictable from independently known properties of its lexical makeup and its grammatical structure? (ii) Does the expression deserve to be listed in a general phrasal lexicon of the language, and treated as a fixed expression? The NI of the present book also receives negative answers to these questions and so fits neatly into the formal idiom category. In response to (i), for example, it is not fully predictable that the initial auxiliary of an NI must be negativized. At the same time, in response to (ii), it is clear that NIs are not fixed expressions. They cannot be listed or learned individually; despite those component requirements of the construction that are rigid, the number of possible instances of it is infinite. In the following section, I lay out the complete case for the claim that NI is in fact a formal idiom, with both frozen and infinitely generative components.

6.8 Conventional requirements of negative inversion The conventional requirements I discuss in this section are of two basic kinds. Broadly, I discuss aspects of social meaning associated with the construction itself as well as its restriction to various regional styles and registers. I then discuss the internal grammatical requirements of the construction, such as the negativized auxiliary, declarative intonation, and information structural requirements on the post-verbal subject. Much of this information has already been provided in preceding chapters of this book. In this present section, discussion of the individual aspects of the construction will be truncated when possible, with pointers included to the various sections in which the particular component in question is described. We begin with discussion of social and evaluative meaning as well as the place of the NI in the regional and social varieties in question, and then move forward to discuss the other non-trivial components of the construction. 6.8.1 Style and social meaning Negative Inversion is a well-known regional marker of US Southern varieties and is also a marker of AAVE around the US. It is also a Labovian social marker within these regional and ethnic groupings, as it belongs primarily to working-class

180

Chapter 6 Negative inversion is a formal idiom

speech in these groups.205 Finally, it is also a stylistic marker, associated with informal speech genres. The NI is thus not simply a dialectal variant of a noninverted counterpart; nor is it simply a less ambiguous variant of some noninverted form. It is a highly distinctive marker of social and regional identity. As such, use of an NI is socially and regionally bound, meaning it is appropriate in some speech situations and inappropriate in others. This situational boundedness is a strong indication that the form is a constructional object – that is, it has meaning and restrictions that are not predictable from the form alone. Lakoff (1987: 470) writes, “one of the major ways we have of identifying grammatical constructions is to find cases where pragmatic conditions are associated with syntactic conditions.”206 We can extend this reasoning straightforwardly to the present situation. The NI in (251a) carries social meaning and associations of regional and stylistic identity. The truth-conditionally equivalent, non-inverted form in (251b) does not. (251) a. Wouldn’t many people buy that car. ‘Many people wouldn’t buy that car.’ b. Many people wouldn’t buy that car. Though (251a–b) are truth conditionally equivalent, (251b) is acceptable in far more situations than its counterpart in (251b), even taken within a single regional variety. On the formal idiom account given in the present work, I assume this social meaning and discourse restriction is part of the lexical content of the construction itself. This is shown clearly in Chapter 5, above, in which Gricean (1975) meaning diagnostics such as detachability, non-cancelability, and noncalculability, show the NI social meanings to be conventional, though nontruth conditional. The primary result here for the present work is that the social meaning in question is semantic, thus associated with the construction. Further, this social meaning is not derivable from the meanings of the constituent parts of the NI. That is, there is no individual constituent in (251a) to which we can assign the social meaning in question. None of the individual words or phrases in (251a) contributes social meaning. Further, the SAI order itself does not convey social meaning. Thus, example (251a), if intended as a question, would not be socially

205 As I have maintained throughout this book, NI can certainly be used by middle and upper classes; however, the NI is generally assumed to be a working-class feature. 206 See also Östman and Trousdale’s (2013: 484) discussion of constructions, in which they write: “in addition to talking about the internal relations between attributes like ‘word class’ and values like ‘noun’ in a construction, we also need to be able to specify when, where, how, and why something can be said.”

6 Introduction

181

marked. The social content of the NI therefore must be attached to the construction as a whole. 6.8.2 Intonational contour As many researchers have noted in the past, despite its SAI syntax, NI has the falling intonation of a declarative sentence. Here are a few representative quotes: Labov et al. (1968: NN) [AAVE] The construction of [the NI] is that of a question, yet the intonation and the context plainly signal a declarative statement. Sells et al. (1996: 591) [AAVE] [NIs] have the inverted form of questions but the falling intonation and sentence meaning of (emphatic) declaratives. Foreman (1999: 2) [West Texas English] NI sentences have the subject-Aux inverted word order of yes/no questions. However, they are interpreted as declarative sentences [. . .] and receive the falling intonation of a declarative. White-Sustaita (2010: 429) [AAVE] NI receives a declarative reading, and prosodically, NI constructions co-occur with a falling intonation. Matyiku (2017: 7) [West Texas English] Although the auxiliary precedes the subject, a clause containing negative auxiliary inversion is a declarative and receives the falling intonation of a declarative.

As these researchers and others make clear, the falling intonation is a necessary component of Negative Inversion and distinguishes the NI from other superficially similar syntactic strings, such as yes/no questions as in (252a) or negative inverted imperatives as in (252b).207 (252) a. Don’t anybody answer that phone ↑? Yes/No question ‘Does anybody answer that phone?’ b. Don’t anybody answer that phone ↓↓. Imperative ‘I order that nobody answer that phone.’

207 I do not provide an analysis of the intonations of these various sentence types. The up and down arrows included here are impressionistic and indicate rising and falling intonation. The double arrows on the imperative indicates the strong down motion of that sentence type as described in Bolinger (1989: 150). As Bolinger writes, “The speech act of commanding is more beholden to the intonation than to the syntax.”

182

Chapter 6 Negative inversion is a formal idiom

c. Don’t anybody answer that phone ↓. Negative Inversion ‘Nobody usually answers that phone.’ It has long been known that specific intonation patterns can be conventionally associated with constructions. For example, the incredulity response construction described above in (245) is known to require a rising intonation. Similarly, sarcasm constructions such as Is a sauna hot?, as described in Fried and Östman (2004), depend on a particular rising intonation for their exclamative reading.208 Hoffman’s (2013) treatment of the correlative comparative construction (i.e. The X-er, the Y-er) also requires a specific rise-fall intonation. Along the same lines, a paratactic construction described in Culicover and Jackendoff (1999) also requires a rise-fall, as in (253) [Culicover and Jackendoff’s (24)].209 (253) Mary listens to the Grateful Dead ↑, she gets depressed ↓. ‘If/When Mary listens to the Grateful Dead, she gets depressed.’ Like these constructions and many others, NI comes with an obligatory intonational pattern, which in this case is the falling intonation of a declarative, with a tonic accent on a subject syllable. As I am arguing that the NI is a formal idiom, it is necessary to assume the intonation is a conventional, idiomatic component of the construction. In the next sections, I consider conventional aspects of the NI that are internal to the sentence, beginning with the requirement of the initial auxiliary. 6.8.3 The initial verb must be an auxiliary As we have seen in numerous examples throughout this book, the initial verb of an NI must be an auxiliary. Lexical verbs are completely disallowed, as in (255). (254) Didn’t nobody lift that rock. ‘Nobody lifted that rock.’

208 Readers might be more familiar with other well-known instances of this construction, such as, Is the pope Catholic?, Is a frog’s ass watertight?, Is a preacher sober on Sunday?, or, Does a bear shit in the woods?, all of which require specific intonation, which differs from the interrogative intonation one might expect if these were instances of prototypical yes/no questions. 209 Another construction with a strong intonational component, which to my knowledge has not been mentioned at all in the literature, is the “What the X?!” construction, where X can be hell, hay, deuce, fuck, what, flick, and so on. The intonation in question is a sharp rise on what, a fall on the definite article, and a second, lesser rise on the X constituent.

6 Introduction

183

(255) *Lifted nobody that rock. ‘Nobody lifted that rock.’ The auxiliary in question appears clause-initially for the most part. Older speakers can allow an expletive to precede the auxiliary as in (256), and it is also possible for various adverbs, as in (257a–c) to do so. The adverbs can be separated from the NI with comma intonation, but they can also reside on the same intonational tier as the rest of the construction. We also find discourse markers such as but, in (257d).210 (256) There can’t nobody lift that rock. ‘There is nobody who can lift that rock.’ (257) a. Clearly(,)didn’t nobody lift that rock. ‘Clearly, nobody lifted that rock,’ b. Maybe(,)didn’t anybody want to go. ‘Maybe, nobody wanted to go.’ c. Possibly(,)couldn’t nobody get in there. ‘Possibly, nobody could get in there.’ d. I guess I knew right then how little chance I had to win, But can’t nobody say I didn’t try. ‘But nobody can say I didn’t try.’ 6.8.4 The auxiliary verb must be negativized The initial auxiliary verb (whether modal, semi-modal, etc.) must be negativized. Examples such as (258a) are unacceptable. (258) a. *Cannot many people lift that rock. ‘Many people cannot lift that rock.’ b. Cain’t many people lift that rock. ‘Many people can’t lift that rock.’ In the non-inverted counterpart, however, as in (259), the auxiliary can appear with or without negative inflection, with no discernible difference in meaning.

210 Example (257d) comes from the chorus of Texan Delbert McClinton’s (2009) song ‘Can’t nobody say I didn’t try.’

184

Chapter 6 Negative inversion is a formal idiom

(259) a. Many people can’t lift that rock. b. Many people cannot lift that rock. It should be mentioned that while this negative pattern is obligatory for NI, it is not required in other SAI sentences. For example, consider the various SAI constructions below in (260), which are adapted from (Goldberg 2006: 166). All of the SAI constructions can occur without negation at all. Further, they exhibit different patterns as to whether the auxiliary must be negativized or not, if negation does in fact occur. (260) a. Can she go? Yes/No questions %Cannot she go? Can’t she go? b. Had she gone, they would be here by now. Counterfactual conditionals Had she not gone, they would be here by now. Hadn’t she gone, they would be here by now. c. Seldom had she gone there . . . Initial negative adverbs Seldom had not she gone there. Seldom hadn’t she gone there. d. May a million fleas infest his armpits! Wishes/curses *May not a million fleas infest his armpits! *Mayn’t a million fleas infest his armpits! e. Boy did she go! Exclamatives %Boy did not she go! Boy didn’t she go! f. He was faster at it than was she. Comparatives *He was faster at it than was not she. *He was faster at it than wasn’t she. g. Neither do they vote. Negative conjunct %Neither don’t they vote. *Neither do not they vote. h. So does he. Positive rejoinder %So doesn’t he. *So does not he. It appears, then, that the negativization requirement on the NI is not a feature of SAI in general. It appears to be specific to the NI. The only exception to this negative-inflection requirement of which I am aware is with two specific semi-modal

6 Introduction

185

auxiliaries where the negation of the auxiliary would violate pre-existing morphotactic constraints on the auxiliaries themselves. Thus, consider (261) and (262), repeated from Chapter 2 above, in which NIs with auxiliaries better and best are not allowed to be negativized. (261) a. Better not anybody bother him before he gets his coffee. ‘Nobody should bother him prior to him drinking his coffee.’ b. *Bettern’t anybody bother him before he gets his coffee. (262) a. Best not anybody go in there right now. ‘Nobody should go in there right now. b. *Bestn’t anybody go in there right now. The auxiliary forms of better and best simply do not accept negative inflection in any environment, as we can see in (263) and (264). (263) a. You should not go. b. You better not go. c. You best not go. (264) a. You shouldn’t go. b. *You bettern’t go. c. *You bestn’t go. The morphotactic requirements of these particular auxiliaries strongly restrict their appearances in the NI construction. It is thus not surprising that the kind of NI we see in (261a) and (262a) have not been mentioned before in any previous NI literature: they are possible forms, but marginal ones. The formal idiom requirement in general can then be stated that the inverted auxiliary must be negativized, unless there is a conflict with the morphotactic requirements of a constituent component.211 6.8.5 Post-verbal negative inversion subjects As was shown in Chapters 2 and 3, NI subjects are constrained in various ways so that not all NP configurations are allowed. Researchers have long believed the constraints to be formal – i.e. definites are prohibited, etc. – as I have shown, though, subject NPs are allowed to bear either definite or indefinite marking if

211 See McCawley (1988) for discussion of negotiating grammatical conflicts in this way.

186

Chapter 6 Negative inversion is a formal idiom

the context is right. As such, the sensitivity must be to something other than the grammatical form of the NP. I argue instead in this work that the constraints are found in the interaction of the discourse function of the NI sentence itself and the information status of the subject NP. In doing so, I am following a line of thought that has been developed in several works to account for the similar puzzle with TE sentences. Given the close identity that has been developed in this work thus far with NI and existentials, this move should not be surprising. Essentially, I will argue here that the NI is conventionally associated with a felicity condition that it introduce a novel referent or reintroduce a referent that might have been forgotten. A twist here, though, is that NI is negative, so the introduction in question is to the absence of an entity or kind of entity. This felicity condition on the construction itself then serves to restrict the kinds of NPs that can appear in the focus-subject position. This is a line of reasoning that is very familiar from the existentials literature, and it originates, I believe, in Lakoff (1987: 545): The function of the [existential] is to focus the hearer’s awareness on the referent of the construction. If the construction is to serve this function, then either the hearer must not have been aware of the referent, or must have forgotten about it. [. . .] If the [subject] noun phrase refers to a specific entity, then the definite article indicates that the hearer is already aware of it. This violates the functional condition on the construction. Hence, the noun phrase cannot be both specific and definite, unless it is serving a reminding function.

Similar sentiments are found across the literature. Below we hear from additional representative authors who tie discourse function to subject type. As we can see, there are variations on exactly what the basic discourse function of the construction is taken to be, but in general, TEs are argued to introduce some entity into the discourse or bring it to the awareness of the addressee, and this function or condition on felicitous use is linked to the character of the subject NP, which effectively blocks most NPs that are definite.212

212 In discussion of the nature of there- existential discourse function, Szekely (2015: 2) writes that the function of there-existentials in discourse is: primarily to introduce new entities. There are of course, many ways to present an object, such that it is possible to go on to say something about it. “Introducing new entities” therefore encompasses a number of different speech acts, including asserting the existence of an entity (for example, Yes, Virginia, there is a Santa Claus); asserting the presence, or absence, of an entity of a particular kind (for example, upon looking into a bare cupboard: There’s no coffee!); issuing a reminder about an entity’s presence (for example, Who can we get to fix the sink? Well, there’s John.

6 Introduction

187

The existential construction is characteristically used to introduce addressee-new entities into the discourse, and for this reason the displaced subject NP is usually indefinite. Ward, Birner, and Huddleston (2002: 1396) [E]xistential sentences serve primarily to introduce a novel referent into the discourse – one fitting the description provided by the pivot nominal. McNally (2011: 1832) [The existential] introduces a novel referent into the discourse, blocking an expression describing the pre-existing referent. Kim (2013: 21)

So, there is a strong precedent for linking sentence function or felicity with the status of the subject NP in the existentials literature. I propose to do the same for the NI of the present book, and there are good reasons for doing so. First, as I have shown in Chapter 3, above, the set of subject NPs allowed in NI overlaps almost perfectly with the set of subject NPs allowed in there-existentials. In addition, I also show in Chapter 3 that there is strong reason to believe that the NI is a diachronic descendent of a close cousin of TE sentences: i.e. the Modal Existential, which has a very similar structure. Starting from these two pieces of information, I argue that the puzzle of how to account for NI subjects can be easily understood based on what others have already claimed about TEs. Let’s consider a somewhat longer quote now on TEs from Abbott (1993: 41), which goes directly to the heart of the issue in this section as well as raising the question of discourse function of negative TEs, which is closer to the Modal Existential and thus the NI [emphasis mine]: [T]he function of existential sentences is to draw the addressee’s attention to the existence and/or location of the entity or entities denoted by the focus NP. Existential sentences are thus claimed, following Hetzron (1975) and Bolinger (1977), to have an essentially presentation function. (See also the discussion in Lakoff 1987). Negative existentials call the addressee’s attention to the absence of some entity or entities. This function follows naturally from the syntactic and semantic properties described above. [. . .] It seems to follow automatically from the foregoing that NPs which do not presuppose the existence of their referents should be perfectly natural in the focus position of an existential, and of course this is the case. [. . .] On the other hand, NPs which do presuppose existence require special contextualization and their existentials cannot initiate a discourse. Many definite NPs fall into the latter category [. . .].

Negative TEs, then, “call the addressee’s attention to the absence of some entity or entities.” This is closer to what is needed in discussion of MEs, such as (265). (265) There can’t nobody lift that rock. ‘There is nobody who can lift that rock.’ The ME discourse function is essentially that of a negative TE – i.e. calling attention to a negative state of affairs; but, the ME is also enriched with the modal

188

Chapter 6 Negative inversion is a formal idiom

perspective of the speaker. Huddleston and Pullum (2002: 173) define modality as “concerned with the speaker’s attitude toward the factuality or actualization of the situation expressed by the rest of the clause.” This basic definition can be straightforwardly adapted to the ME, so that it presents a negative state of affairs but also expresses the speaker’s attitude toward the factuality or actualization of that state of affairs. The TE in (266a) draws attention to an entity or its existence. In (266b), attention is drawn to the state of affairs in which a cat is in the kitchen, and in (266c) a state of affairs in which there is an absence of a cat in the kitchen. (266) a. There is a cat. b. There is a cat in the kitchen. c. There is no cat in the kitchen. An ME, on the other hand, draws an addressee’s attention to a state of affairs while at the same time expressing the speaker’s modal belief toward that state of affairs. In (267a–b), for example, we see modals expressing differing degrees of strength or commitment to the claim that no cat is in the kitchen. (267) a. There shouldn’t no cat be in the kitchen. ‘There should not be a cat in the kitchen.’ b. There can’t no cat be in the kitchen. ‘There cannot be a cat in the kitchen.’ The NI in (268) – which is just an ME that has lost its expletive subject – has this same discourse function: it draws attention to a negative state of affairs while expressing a modal perspective toward it. (268) Can’t no cat be in the kitchen. ‘No cat can be in the kitchen.’ This discourse function, in this present account, is considered to be part of the conventional constructional content of the Negative Inversion. This assumption of the conventional association of function or felicity condition also finds a strong precedent in the TE literature. Lakoff’s (1987) treatment of existentials, for example, is situated in the framework of cognitive grammar, which sees grammatical constructions as “direct pairings of parameters of form with parameters of meaning” (Lakoff 1987: 464). Lakoff builds the discourse function of the existential directly into the conventions of the construction itself. Working in the framework of generative syntax and formal semantics, McNally

6 Introduction

189

(1992) discusses not the discourse function of the existential per se, but rather a felicity condition on its use, which entails a discourse function. McNally’s felicity condition holds that “expletive there has come to be associated with the condition that a new discourse referent must be introduced” (McNally 1992: 200). In other words, if the TE does not introduce a new discourse referent, use of the sentence will be infelicitous. McNally assumes this felicity condition is conventionally associated with the there expletive itself, rather than with the entire construction as is the case in Lakoff’s account. Regardless, it is still a conventional association. Returning to the NI, I am arguing in this work that the NI is simply an ME that has lost its expletive over time. Recall the generational data in Chapter 3 which supports this claim, in which older speakers have sentences such as (269a) and (269b), while younger speakers have only sentences such as (269b). (269) a. There can’t no cat be in the kitchen. b. Can’t no cat be in the kitchen. There is no expletive there with NI such as (269b) for younger speakers; however, the expletive commonly was there in previous generations. A McNally style account, then, in which the felicity condition is conventionally bound to the expletive itself, would have little to say about NIs – at least in a synchronic sense. However, in a diachronic sense, in which the expletive has deleted across the generations, it is plausible that the felicity condition remained behind on the sentence type even as the expletive itself deleted, as a remnant – a general process which is common in diachronic linguistics.213 In this case, the discourse function or felicity condition would just be associated with the syntactic frame of the NI itself, which is common for formal idioms in general. Recall Fillmore, Kay, and O’Connor’s discussion above of the incredulity response construction. There, a felicity condition that is conventionally associated with the formal idiom requires the ICR to appear as a response to a proposition that the speaker finds to be dubious.

213 Recall, for example, the mention of French vowel nasality discussed above, in which certain French vowels retain their nasal character long after the nasal consonant to which they originally assimilated has dropped off.

190

Chapter 6 Negative inversion is a formal idiom

6.8.6 Summary The aforementioned components – i.e. the social content, the intonation, the negativized auxiliary, and the discourse function and information status of the subject NP – are idiosyncratic, conventionalized aspects of the construction. Beyond these, the Negative Inversion frame is lexically open, allowing for a verb phrase and subsequent modifiers of practically any form. Like Fillmore, Kay, and O’Connor’s (1988) formal idioms, NI is partially conventional and partially open. Two other factors point to the idiomization of the NI over time. As was shown in Chapter 2, it was possible till at least the early-mid 20th century for Modal Existentials to appear with full negation; i.e. it was not required for the auxiliary to be negativized, as in (270). (270) As a matter of fact, there cannot anybody go in there and let horses for what they are letting them now and have to haul in their feed. ‘There is nobody who can go in there and let horses [. . .].’ This older form is not a possibility with contemporary speakers I surveyed for this book – not even the older participants – which suggests a construction in flux in the 20th century. Secondly, there is the loss of the expletive subject itself across the last few generations, as described in detail in Chapter 3. If, as I am arguing, the NI has become a formal idiom, it is not surprising that it would lose non-crucial grammatical elements such as the expletive subject in a reanalysis of the surface structure. Both of these changes – i.e. the loss of the expletive subject and the requirement that the auxiliary bear negative inflection – can be seen as supporting arguments that the NI has become a type of idiomatic construction, involving the tightening of form and the shedding of non-crucial components.

6.9 Conclusion This chapter has outlined what I believe to be the obligatory, conventional components of the NI construction. These are the minimal components of the construction that must be lexically specified – other constituents appearing in tokens of NI are general and need not be elaborated upon here. The account has followed the literature on there-existentials closely in its discussion of NI. It might be surprising at first blush that the account has gone in this direction rather than in the direction of other SAI constructions, such as yes/no questions

6 Introduction

191

and the rest of those mentioned in Section 6.8.4 above. I assume, however, that the resemblance of NI to those other SAI constructions is an accident, rather than a family likeness. As I have argued in this book, the “inversion” apparent in NI is really not an inversion at all but is a deletion of an expletive subject, which leaves behind what only looks like an inversion. It is of course possible that the significant presence of SAI in English grammar in general has softened resistance to NI in some way or made its existence less marked via analogy: i.e. NI surface form is similar to other common SAI. However, I argue that the NI bears those other SAIs little relation, which is an opinion that is shared by many researchers, including Sells et al. (1996), Foreman (1999), and Green (2014), and more. The family likeness of NI can be found in the existential branch, not the SAI branch. Thus, a fuller constructional account of NI, which I do not provide here, would ultimately link NI through inheritance hierarchies with other sub-types of existential construction.214 More detail than this, however, is far beyond the needs of the present work.

214 Lakoff (1987), for example, provides a constructional account of existential sentences that posits a central existential construction, which is then inherited by numerous other noncentral, sub-types of existential. In this way, Lakoff accounts for the syntactic, semantic, and pragmatic commonalities shared across all of the existential sub-types. These existential subtypes then have the basic qualities of Lakoff’s central construction but also individual qualities, due to their membership in other inheritance networks, which sets them apart from each of the other sub-types. In such a system, NI would be merely a sub-construction of some more basic, central existential construction. See Goldberg (1995: Chapter 3) for much more discussion on constructional inheritance and means of relating constructions to one another.

Conclusion Can’t nobody say I didn’t try. – Delbert McClinton, 2009

Introduction In this concluding chapter I reiterate that I have been arguing for an essentially there-existential analysis of NI all the way through the book. Existential analyses have been presented for this construction since the beginning in Labov et al. (1968) but have also been found lacking for various reasons over the intervening decades. The existential account presented here differs from all of the preceding accounts in that it builds in a diachronic component as well as a significant update in the empirical picture; these two things, I argue, make all of the difference. Alongside this diachronic picture of the NI, I also understand it as an idiomatic construction. It is a sentence type that is loaded down with the social meaning of Chapters 4 and 5 as well as the post-verbal subject restriction of Chapter 3 and the rigid formal issues with the auxiliary and negation of Chapters 2 and 6. At this point, seeing the NI as an idiom is just the simplest, most straightforward way to view it. Not everything need be built into the constructional form, however: both the subject restriction and the emphatic character of the construction are ultimately got for free via the presentational sentence function and standard Gricean pragmatic interaction, respectively. In this final chapter, I provide a final overview of this book. I also discuss inquiries which were not made in this book and which are very wide open for future research.

It’s all existential Variations on the existential analysis of NI have been presented in the literature since Labov et al. (1968), which is described in detail above in Chapter 6. In addition to the Labovian account, Martin (1992) suggests an existential analysis in which NI is a result of post-syntactic phonological expletive deletion. This explains for Martin, among other things, the similarity of the Milsark (1974) definiteness effects seen in NI. Parrott (2000) argues that AAVE has a lexical item that has most of the properties of SE expletive there, but that in AAVE it is silent. Thus, for Parrott, NI has a silent expletive there, which explains the https://doi.org/10.1515/9781501512346-008

It’s all existential

193

presence of the definiteness effects. White-Sustaita (2010) argues that all NIs have an existential meaning which serves as pragmatic motivation for the specific syntactic movement of her account. In this book, I assume simply that the existential-indicating expletive there has fallen off for many speakers, especially younger ones, over the last two to three generations. Let’s look now at the advantages and disadvantages that have been raised with respect to existential accounts of NI over the years and consider how they stack up against the present account. White-Sustaita (2010: Section 3) provides a solid starting place for this task, as she discusses advantages and disadvantages of both inversion and existential analyses made up to that time. She begins with inversion analyses and notes that an advantage of such approaches is they can account “for the ban on expletive subjects with NI constructions in present-day [African American English]” (White-Sustaita 2010: 438).215 This detail, she notes, would be difficult to account for on the various existential analyses that had been given in the literature thus far. However, as I have shown in Chapter 2, the ban on expletives is not complete in AAVE or in either of the other two ethnolects discussed in this book. Specifically, expletive constructions (i.e. ME) are much more common among older speakers than younger. As I write in Chapter 3, among older speakers, the NI is more likely a product of surface level conversational deletion – i.e. older speakers actually have ME which undergoes conversational deletion in context, resulting in surface NI. Among younger speakers, the expletive is generally fully deleted, but even here there are exceptions. So, inversion analyses provide no advantage in this area that straightforward conversational deletion cannot handle. The next objection to existential analyses White-Sustaita mentions is that “[inversion analyses] explain why some dialects permit negative existential constructions without an expletive.” Thus, sentences like (271), which are not true NI, appear in many dialects in which NI does not appear. (271) Ain’t nobody home right now. ‘There isn’t anybody home right now.’ As such, White-Sustaita suggests there must be two different mechanisms at work: i.e. an inversion mechanism to account for true NI and an existential analysis to account for negative existentials as in (271). Under this assumption, it would be possible for a dialect to have only one of the grammatical constructions.

215 This concern is also the main reason Labov et al. (1968) and Martin and Wolfram (1998) give for rejecting an existential analysis.

194

Conclusion

As I described above in Chapter 3, conversational deletion of the expletive works for both sentence types – not to mention many others. The negative existential in (271) undergoes synchronic CD, while the NI is a product of diachronic deletion, which is mostly complete for younger generations but not so for older speakers. More recently, as we have seen above in Chapters 2 and 6, inversion analyses are frequently motivated by scope disambiguation (Foreman 1999, 2015; Zanuttini and Bernstein 2014; Matyiku 2017) and emphatic force (Green 2014; Green and Sistrunk 2015). As we also saw in those chapters, though, especially Chapter 6, those motivations do not hold true with the present data. So, there is very little to recommend any of the inversion analyses proffered in the literature up to this point. On the other hand, the benefits of the existential analysis provided in the present book are many. First, no researcher across the entire canon of NI literature has been able to determine a meaning or usage difference between MEs like (272a) and NIs like (272b). (272) a. There can’t nobody hold me down. ‘Nobody can hold me down.’ b. Can’t nobody hold me down. ‘Nobody can hold me down.’ They are almost certainly the same construction modulo the application of CD. This is straightforwardly accounted for on the existential account given here. Secondly, as I show in significant detail in Chapters 2 and 3, NI and TEs have exactly the same subject distribution, except for few NP, which appears in the latter but not the former. However, as I also point out in those chapters, few NP is a strong positive polarity item, which doesn’t appear in the scope of negation in any sentence type. This too is a strong benefit of the present existential account. Third, as is well known, and is discussed above in Chapters 2 and 6, many NIs don’t have a well-formed non-inverted counterpart; especially those with non-negative subjects. For movement-based inversion analyses, this presents a further complication. On the existential account, though, it is no complication at all. The expletive can delete as easily from a sentence with a well-formed counterpart like (273) as it can in a situation like (274), in which the result will be an NI that does not have a well-formed counterpart. (273) There can’t many people lift that rock. Negative Inversion: Can’t many people lift that rock Non-inverted Counterpart: Many people can’t lift that rock.

It’s all existential

195

(274) There can’t anybody lift that rock. Negative Inversion: Can’t anybody lift that rock. Non-inverted Counterpart: *Anybody can’t lift that rock. Fourth, and most importantly, and an argument which in my opinion is just irrefutable, is the fact that NI exhibits the same definiteness effects that TEs do. As far as I know, there is no other construction in the English language besides TE which exhibits definiteness effects in this way. It would be a strange indeed if the two constructions were not closely related. On the existential account provided here, this relation is expected. At the end of the book, then, the account of NI provided here is quite simple. As I will remind the reader in the final few sections of chapter summary to come, all that’s needed is a basic account of TEs, some sociolinguistic facts, old-fashioned Gricean pragmatics, and perhaps most important of all, a fresh empirical perspective on the construction. In Chapter 2, I provide an up-to-date empirical portrait of NI. This includes the by-now familiar definite subject examples, weak quantifier examples, and more. This chapter is very important in the book, because it allows the account given here to break free of the weight of all previous accounts. Frequently in the history of linguistics, received characterizations of data have become cemented into the thinking about that data, sometimes for generations. Milsark’s definiteness effects are this way. Many researchers quote the Milsarkian generalization still as if it were consensus, even though by now we know it no longer is. Von Fintel and Gillies (2010) refer to what they call the “mantra” that epistemic must is weak – a claim that has been repeated so often, that as those authors write, “Usually, the weakness intuition is simply announced without much argument [. . .].” Gibson and Fedorinko (2013) list several other cases in the syntax literature where inadequate data characterizations have led to wrong turns in theorizing.216 I believe this kind of effect has occurred with NI theorizing in the past: i.e. the rigidified assumptions that the data is a certain way have restricted the kinds of accounts of it that can be given. The new empirical account given here frees NI from the weight of that history. Chapter 3 provides the backbone of the existential account. This chapter shows that there are definiteness effects associated with NI, but they are not what has previously been believed. They are, however, very similar to the

216 The supra-regional myth of a monolithic AAVE described in Wolfram (2007) fits here as well. It is well known now that AAVE varies across regions, yet it is often treated as though it does not.

196

Conclusion

definiteness effects observable in TEs, which, as researchers like Prince (1992), Abbott (1993), Ward and Birner (1995), and others have shown, are also not what has previously been believed. This chapter then makes the diachronic argument via the apparent-time construct that NI is very likely the descendant of MEs that have lost their expletive over the last two generations. Chapter 4 takes up the question of constructional stigma. It probes the kinds of social meanings associated with the construction, finding that NI fits well with what is by now a well observed pattern in AAVE, creoles, and other vernacular languages. On one level, the form is highly stigmatized, yet on another level, it is used to assert identity in a group and so also contributes connotations of familiarity. Thus, it is similar to what Rickford and Traugott (1985: 252) describe as the “paradoxical combination of negative and positive attitudes which are found in communities where pidgin and creole varieties of English are spoken.” In what I believe is a significant contribution to both the fields of pragmatics and sociolinguistics, this chapter then tests the social, sociolinguistic meanings of NI in a wide range of diagnostics from the semantics and pragmatics literature. A kind of kinship is shown here between the social meaning often discussed in the sociolinguistics literature and the conventional implicature meaning often discussed in the semantics and pragmatics literature. This finding provides a clear path to the pragmatic account of emphasis given in Chapter 5. In Chapters 4 and 5 I show that NI and its non-inverted counterpart are truth-conditionally equivalent, differing only in the social meanings that are attached to the former. It was shown in Chapter 4 that at least some of the NI social meanings are conventionally associated with the NI form. Based on this conventional association of social meaning, I argue that an NI can be used to convey conversational implicatures of emphasis and so explain the variation in researchers’ understandings of this part of the NI construction. This manner of reasoning – from conventional to conversational implicature – is well established in the semantics and pragmatics literature. As mentioned above, Horn and Abbott (2012) show that non-truth conditional, conventional implicature content of uniqueness associated with the definite article the can be used to conversationally implicate familiarity, and so account for multiple aspects of the with minimal content specified in the lexicon. The same goal obtains with the NI: i.e. it is preferable to specify limited lexical content and then generate other aspects of the NI’s meaning via pragmatic reasoning. The final chapter of the book provides in-depth argumentation that previous movement-based analyses of NI syntax are unmotivated. The Foreman (1999)inspired accounts mentioned above are shown to be inadequate, because there are many cases in which the NI form does not disambiguate. Similarly, the emphatic force account of Green (2014) is inadequate, because the emphasis in

Summary

197

question with the data here is very clearly pragmatic. From this point, Chapter 6 sketches a formal-idiom account, loosely situated in the framework of Fillmore, Kay, and O’Connor (1988). The motivation for doing so is not to question generative frameworks in general; rather, the NI has enough rigid requirements of form and meaning that it seems more straightforward at this point to argue instead that it is a kind of idiom. It is more of a peripheral item than one that demands an account in a generative core. Chomsky and Lasnik (1993/1995: 20) write that “A reasonable approach [to the study of language] would be to focus attention on the core system, putting aside phenomena that result from historical accident, dialect mixture, personal idiosyncrasies, and the like.” The NI described in the present work would seem to be best put aside, then, from Chomsky and Lasnik’s approach. In the formal-idiom account given in Chapter 6, a description is given of those aspects of the NI’s form and meaning which are not predictable from other aspects of the grammar of the construction, including the social meaning, intonation, initial auxiliary, and negative inflection. Here also, an explanation is given for the information status requirements of the NI subject NP. Essentially, it needs to not conflict with the discourse function of the NI, which requires it to introduce an absence that is novel in some way. This is an argument that runs exactly parallel to those made with respect to TEs, as described in Chapter 6. Finally, I situated this description of NI as formal idiom in Fillmore et al.’s framework because it is easily translatable to many of the other construction-based frameworks, as surveyed, for example, in Hoffman and Trousdale (2013).

Summary In short, then, in this book I have argued that a formal-idiom account of NI in the Texas ethnolects considered herein provides tidy solutions to many of the questions that have surrounded the construction since its entry into mainstream linguistic literature in the late 1960s. I argue first and foremost that the definiteness effect that has been discussed since Martin (1992) is an artifact of the construction’s diachronic relation to the Modal Existential construction, and that it still follows the same basic organizing principles required of definite subjects in contemporary TE sentences. Similarly, the question of why inversion occurs with NI has a clear answer: it doesn’t. The deletion of the expletive subject that results in the sentence-initial modal contrives the appearance of inversion, but none actually occurs. Relatedly, the question of why comparable positive sentences don’t undergo inversion also has its answer. The aforementioned expletive deletes in order to accentuate the sentence-initial negation on the modal: if the sentence is positive, with no negation, the motivation for

198

Conclusion

deletion diminishes. Another question: Why is the NI considered to be a forceful assertive sentence type by some researchers but not others? This question finds its answer in Gricean pragmatics and Labovian social markers. The construction marks social relations with the hearer, but if uttered in the proper contexts that relation marker can be leaned upon to make a strong or forceful statement pragmatically. So, depending on what kinds of contexts are imagined by the researcher when investigating this feature, the NI might appear to be forceful and assertive, or not. Or, depending on the nature of the corpus and conversation the analyst investigates, an NI might be forceful, or it might simply mark in-group social relations. A lingering question is why must the auxiliary of the NI obligatorily host negative inflection? I have no solid answer here, but only a conjecture which returns us to Jespersen and the need to put negation early in a sentence. Negative inflection on the auxiliary and loss of the expletive puts the negation about as close as it can get to the front of a sentence without completely altering the sentence type. Add to this the idiomization over time and there is a plausible story for the necessary negative inflection.

Future directions of research There are many aspects of the present study I would have liked to explore further given world enough and time. First and foremost, I would like to survey many additional senior speakers of Texas AAVE. This was a demographic to which I had only limited access in the time I had at my disposal. Their input would be especially welcome in developing the claims made in Chapter 3 regarding the diachronic relation of ME and NI. In general, I also did not consider gender differences in attitudes toward NI or in usage of it. I would very much like to do so in the future. Does it continue to follow well-known patterns in sociolinguistics in which strong vernacular forms are viewed more positively by men than women? I do not know. I also did not question at all usage specifics of genre or register, or even of social class. I would like to know more as well about the urban and rural thoughts on the construction. What kinds of patterns will emerge there with more data from each environment? Also, it is well known that Texas urban areas are filling rapidly with people from other parts of the US. Are these newcomers adopting NI as they have with other aspects of Texas speech? In their “National Geographic Survey of Texas Dialects” Guy Bailey and Jan Tillery found that many newcomers to the state do pick up some stereotypically Texas features, such as y’all and fixin’ to, while rejecting others. I would love to know

Pragmatics and the “syntactico-semantic straitjacket”

199

how NI fits in here. Is it slowly disappearing with the urbanization of Texas and large influx of newcomers, or will it persist as a marker of social and regional identities? Already for many speakers it seems to have reached a Labovian stereotype status – the subject of memes and a source of humor and a possible indication that it is on its way out of the language. Turning to the semantics and pragmatics, I have not attempted to provide a formal account of the conventional but non-truth conditional social meaning found in NI. In what ways does this content interact with propositional content, and how should it be put together into a formal semantics? The multidimensional logic of Potts (2005) and further developments of this logic such as that in McCready (2010) are likely good places to begin this task, as these logics were designed to account for multi-dimensional conventional implicature content. There are also clear places to begin this task in the various literatures of construction grammar. Further, more and sharper work is needed to clarify what exactly the dimensions are in terms of the NI social content, and what the relations are between its ineffability and the indexical fields of Eckert (2008b). This seems an obvious and important point of contact at which pragmaticists and sociolinguists might “have a beer,” as Cameron and Schwenter (2013) would say.

Pragmatics and the “syntactico-semantic straitjacket” The intersection of Gricean pragmatics and sociolinguistic work of any stripe is one that is not frequently inhabited. Nor am I aware of any other work showing a kinship between Gricean conventional implicature and Labovian social meaning. In this present work, this intersection of thought and practice has provided straightforward solutions to puzzles that have long confounded syntactic accounts of Negative Inversion – accounts which Bar-Hillel (1971: 401) might say try to force pragmatic matters into a “syntactico-semantic straitjacket.” This long and at times overly specific disquisition has shown us quite a lot about one negative construction; one hopes that in its specificity, however, it continues to contribute to larger, positive generalizations.

References Abbott, Barbara. 1993. A pragmatic account of the definiteness effect in existential sentences. Journal of Pragmatics 19. 39–55. Abbott, Barbara. 2003. A reply to Szabo’s ‘Descriptions and uniqueness.’ Philosophical Studies 113. 223–231. Abrusán, Márta. 2016. Presupposition cancellation: Explaining the ‘soft-hard’ trigger distinction. Natural Language Semantics 24. 165–202. Akhmanova, Olga. 1976. Linguostylistics: Theory and Method. The Hague: Mouton. Akmajian, Adrian. 1984. Sentence types and the form-function fit. Natural Language and Linguistic Theory 2. 1–23. Appendix to the Journals of the Senate and Assembly of the Thirty-Ninth Session of the Legislature of the State of California. Vol. 2. 1912. Sacramento: Superintendent of State Printing. Bach, Kent. 1999. The myth of conventional implicature. Linguistics and Philosophy 22. 367–421. Bailey, Guy. 2002. Real and apparent time. In J.K. Chambers, Peter Trudgill, & Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 312–332. Malden, Mass: Blackwell. Bailey, Guy, Natalie Maynor & Patricia Cukor-Avila. 1991. The Emergence of Black English: Text and Commentary. Amsterdam: John Benjamins. Bakhtin, Mikhail. 1986. The problem of speech genres. In Caryl Emerson & Michael Holquist (eds.), Speech Genres and Other Late Essays, by Mikhail Bakhtin, 60–102. Austin: University of Texas Press. Ball, Catherine. 1991. The Historical Development of the ‘It’-Cleft. Philadelphia, PA: University of Pennsylvania dissertation. Bar-Hillel, Yehoshua. 1971. Out of the pragmatic wastebasket. Linguistic Inquiry 2. 401–407. Barron, Anne & Klaus Schneider. 2009. Variational Pragmatics. Amsterdam: John Benjamins. Baugh, John. 1984. ‘Steady’: progressive aspect in black English vernacular. American Speech 59. 1–12. Bayley, Robert & Otto Santa Ana. 2004. Chicano English grammar. In Edgar Schneider & Bernd Kortmann (eds.), A Handbook of Varieties of English: Morphology and Syntax. Vol. 2, 167–183. Berlin: Mouton de Gruyter. Beaver, David, Itamar Francez & Dmitry Levinson. 2005. Bad subject: (Non-)canonicality and NP distribution in existentials. Semantics and Linguistic Theory 15. 19–43. Beaver, David & Bart Geurts. 2014. Presupposition. In Ed. Zalta (ed.), The Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/archives/win2014/entries/presup position/ (accessed 12 Dec 2018). Bell, Allen. 1984. Language style as audience design. Language in Society 13. 145–204. Bender, Emily. 2007. Socially meaningful syntactic variation in sign-based grammar. English Language and Linguistics 11. 347–381. Berlin, Isaiah. 1953. The Hedgehog and the Fox. London: Weidenfeld & Nicolson. Bickerton, Derek. 1975. Dynamics of a Creole System. Cambridge: Cambridge University Press. Blakemore, Diane. 2001. Relevance and Linguistic Meaning: The Semantics and Pragmatics of Discourse Markers. Cambridge: Cambridge University Press. Blakemore, Diane. 2011. On the descriptive ineffability of expressive meaning. Journal of Pragmatics 43. 3537–3350. https://doi.org/10.1515/9781501512346-009

202

References

Blanchette, Francis. 2015. English Negative Concord, Negative Polarity, and Double Negation. New York, NY: City University of New York dissertation. Blanchette, Francis & Chris Collins. 2018. On the subject of negative auxiliary inversion. Journal of Canadian Linguistics 64. 1–30. Boberg, Charles. 2010. The English Language in Canada: Status, History and Comparative Analysis. Cambridge: Cambridge University Press. Bolinger, Dwight. 1975. Aspects of Language. 2nd edn. New York: Harcourt. Bolinger, Dwight. 1977. Meaning and Form. London: Longman. Bolinger, Dwight. 1989. Intonation and Its Uses. Stanford, CA: Stanford University Press. Bošković, Željko. 2002. A-movement and the EPP. Syntax 5. 167–218. Bowie, David. 2005. Language change over the lifespan: A test of the apparent time construct. University of Pennsylvania Working Papers in Linguistics 11 (2).Article 3. 45–58. Bresnan, Joan. 1970. An argument against pronominalization. Linguistic Inquiry 1. 122–123. Brown, Penelope & Colin Fraser. 1979. Speech as a marker of situation. In Klaus Scherer & Howard Giles (eds.), Social Markers in Speech, 33–62. Cambridge: Cambridge University Press. Brown, Penelope & Stephen Levinson. 1978. Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Buchstaller, Isabelle. 2006. Diagnostics of age-graded linguistic behavior: The case of the quotative system. Journal of Sociolinguistics 10. 3–30. Cameron, Richard & Scott Schwenter. 2013. Pragmatics and variationist sociolinguistics. In Robert Bayley, Richard Cameron, & Ceil Lucas (eds.), The Oxford Handbook of Sociolinguistics, 464–483. Oxford: Oxford University Press. Campbell-Kibler, Katherine. 2007. Accent, (ING) and the social logic of listener perceptions. American Speech 82. 32–64. Cheshire, Jenny. 1987. Syntactic variation, the linguistic variable, and sociolinguistic theory. Linguistics 25. 257–282. Cheshire, Jenny. 2005. Age and generation-specific use of language. In Ulrich Ammon, Norbert Dittmar, Klaus Mattheier & Peter Trudgill (eds.), Sociolinguistics: An Introductory Handbook of the Science of Language and Society. 2nd. edn., 1552–1563. Berlin: Mouton de Gruyter. Chierchia, Gennaro & Sally McConnell-Ginet. 1990. Meaning and Grammar. Cambridge, MA: MIT Press. 2nd edn. 2000. Chierchia, Gennaro. 2006. Broaden your views. Implicatures of domain widening. Linguistic Inquiry 37. 535–590. Chomsky, Noam 1982. Some Concepts and Consequences of the Theory of Government and Binding. Cambridge, MA: MIT Press. Chomsky, Noam & Howard Lasnik. 1993. The theory of principles and parameters. In Joachim Jacobs, Arnim Von Stechow & Wolfgang Sternefeld (eds.), Syntax: An International Handbook of Contemporary Research. Vol. 1, 506–569. Berlin: De Gruyter. Reprinted in N. Chomsky, The Minimalist Program. Cambridge, MA: MIT Press, 1995. Collins, Chris. 2016. Not even. Natural Language Semantics 24. 291–303. Collins, Peter. 2001. Some discourse functions of existentials in english. In Proceedings of the 2001 Conference of the Australian Linguistic Society. http://www.als.asn.au/proceedings/ als2001/collins.pdf (accessed 26 Sept 2018). Comrie, Bernard. 1976. Aspect: An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: Cambridge University Press.

References

203

Coupland, Nikolas. 2007. Style: Language Variation and Identity. Cambridge: Cambridge University Press. Curme, George. 1931. A Grammar of the English Language. Vol. 3: Syntax. Boston: Heath. Croft, William. 1991a. Syntactic Categories and Grammatical Relations: The Cognitive Organization of Information. Chicago: University of Chicago Press. Croft, William. 1991b. The evolution of negation. Journal of Linguistics 27. 1–27. Cruse, David Alan. 1986. Lexical Semantics. Cambridge: Cambridge University Press. Cukor-Avila, Patricia & Guy Bailey. 1996. The spread of urban AAVE: A case study. In Jennifer Arnold, Renée Blake, Brad Davidson, Scott Schwenter & Julie Solomon (eds.), Sociolinguistic Variation: Data, Theory, and Analysis. Selected papers from NWAV 23, 469–85. Stanford, CA: CSLI. Cukor-Avila, Patricia & Guy Bailey. 2001. The effects of the race of the interviewer on sociolinguistic fieldwork. Journal of Sociolinguistics 5. 254–70. Cukor-Avila, Patricia & Guy Bailey. 2015. Rural Texas African American Vernacular English. In Sonja Lanehart (ed.), The Oxford Handbook of African American Language, 181–200. Oxford: Oxford University Press. Cukor-Avila, Patricia, Lisa Jeon, Patricia Rector, Chetan Tiwari, & Zak Shelton. 2012. Texas–It’s like a whole nuther country: Mapping Texans’ perceptions of dialect variation in the Lone Star State. Texas Linguistics Forum 55. 10–19. Culicover, Peter. 1999. Syntactic Nuts. Oxford: Oxford University Press. Culicover, Peter & Ray Jackendoff. 1999. The view from the periphery: The English comparative correlative.” Linguistic Inquiry 30. 543–571. Curzan, Anne. 2014. Fixing English: Prescriptivism and Language History. Cambridge: Cambridge University Press. Dahl, Östen. 1979. Typology of sentence negation. Linguistics 17. 79–106. Dancygier, Barbara. 1998. Conditionals and Prediction: Time, Knowledge and Causation in Conditional Constructions. Cambridge: Cambridge University Press. Dase, Amy & Stephanie Katauskas. 2011. For love of country: The Killeen base oral history project. U.S. Army Garrison, Fort Hood Archeological Resource Management Series Research Report No. 58, 1–598. Austin, TX: Prewitt and Associates. http://www.nicap.org/docs/541001kill een_oralhistory.pdf (accessed8 Feb 2019). Denison, David & Alison Cort. 2010. Better as a verb. In Hubert Cuyckens, Kristin Davidse & Lieven Vandelanotte (eds.), Subjectification, Intersubjectification and Grammaticalization, 349–383. Berlin: de Gruyter Mouton. Duffley, Patrick & Pierre Larrivée. 2010. Anyone for non-scalarity? English Language and Linguistics 14. 1–17. Eckert, Penelope. 2004. The meaning of style. Proceedings of the Eleventh Annual Symposium about Language and Society 47. 41–53. Eckert, Penelope. 2008a. Where do ethnolects stop? International Journal of Bilingualism 12. 25–42. Eckert, Penelope. 2008b. Variation and the indexical field. Journal of Sociolinguistics 12. 453–476. Eckert, Penelope. 2012. Three waves of variation study: The emergence of meaning in the study of sociolinguistic variation. Annual Review of Anthropology 41. 87–100. Ervin-Tripp, Susan. 2001. Variety, style-switching, and ideology. In Penelope Eckert & John Rickford (eds.), Style and Variation, 44–56. Cambridge: Cambridge University Press.

204

References

Fasold, Ralph & Walt Wolfram. 1970. Some linguistic features of Negro dialect. In Ralph Fasold & Roger Shuy (eds.), Teaching Standard English in the Inner City, 41–86. Washington, D. C.: Center for Applied Linguistics. Feagin, Crawford. 1979. Variation and Change in Alabama English: A Sociolinguistic Study of the White Community. Washington: Georgetown University Press. Field, Fredric. 2011. Bilingualism in the USA: The Case of the Chicano-Latino Community. Philadelphia: John Benjamins. Fillmore, Charles, Paul Kay & Mary Catherine O’Connor. 1988. Regularity and idiomaticity in grammatical constructions: The case of ‘let alone.’ Language 64. 501–538. von Fintel, Kai. 1992. The modal-existential construction. Formal Linguistics Society of Midamerica 3. http://mit.edu/fintel/fintel-1992–modal-existentials.pdf (accessed 5 Mar 2019). von Fintel, Kai & Anthony Gillies, 2010. Must. . .stay. . .strong! Natural Language Semantics 18. 351–383. The five least literate cities in America. (2014, February 24). Time. https://time.com/9549/the5-least-literate-cities-in-america/ (accessed 25, Nov 2019). Foley, Neil. 1999. The White Scourge: Mexicans, Blacks, and Poor Whites in Texas Cotton Culture. Berkeley, CA: University of California Press. Foreman, John. 1999. Syntax of negative inversion in non-standard English. West Coast Conference on Formal Linguistics 17. 205–219. Stanford: CSLI. Foreman, John. 2015. The syntax of ain’t. In Patricia Donaher & Seth Katz (eds.) Ain’thology: The History and Life of a Taboo Word, 113–155. Newcastle upon Tyne: Cambridge Scholars Press. Fought, Carmen. 2003. Chicano English in Context. New York: Palgrave. Fought, Carmen. 2006. Language and Ethnicity. Cambridge: Cambridge University Press. Francez, Itamar. 2007. Existential Propositions. Palo Alto, CA: Stanford University dissertation. Frege, Gottlob. 1892. Über Sinn und Bedeutung. In Zeitschrift für Philosophie und philosophische Kritik. Translated as “On sense and reference” by Max Black in Translations from the Philosophical Writings of Gottlob Frege, Peter Geach and Max Black (eds. & trans. ) Oxford: Blackwell, 3rd edn., 1980. 56–78. Frege, Gottlob. 1918/1956. The thought: A logical inquiry. Mind 55. 289–311. Fried, Mirjam & Jan-Ola Östman. 2004. Construction grammar: A thumbnail sketch. In Mirjam Fried & Jan-Ola Östman (eds.), Construction Grammar in a Cross-language Perspective, 11–86. Amsterdam: John Benjamins. Geurts, Bart. 2007. Really fucking brilliant. Theoretical Linguistics 33. 209–214. Gibson, Edward & Evelina Fedorenko. 2013. The need for quantitative methods in syntax and semantics research. Language and Cognitive Processes 28. 88–124. Gillespie, Dizzy & Al Fraser. 1975. To Be, or Not. . .to Bop. Minneapolis: University of Minnesota Press. Goffman Erving. 1963. Stigma: Notes on the Management of Spoiled Identity. London: Penguin. Goldberg, Adele. 1995. Constructions: A Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. Goldberg, Adele. 2006. Constructions at Work. Oxford: Oxford University Press. Goldberg, Adele & Ray Jackendoff. 2004. The English resultative as a family of constructions. Language 80. 532–567. Gómez Menjívar, Jennifer & William Salmon. 2018. Tropical Tongues: Language Ideologies, Endangerment, and Minority Languages in Belize. Chapel Hill: University of North Carolina Press. Green, Georgia. 1976. Main clause phenomena in subordinate clauses. Language 52. 382–397.

References

205

Green, Georgia. 1985. The description of inversions in generalized phrase structure grammar. Berkeley Linguistics Society 11. 117–45. Green, Lisa. 1998. Aspect and predicate phrases in African-American vernacular English. In Guy Bailey, John Baugh, Salikoko S. Mufwene, & John Rickford (eds.), African-American English: Structure, History, and Use, 37–68. New York: Routledge. Green, Lisa. 2001. Negative concord and negative inversion in African American English. Ms. University of Texas. Green, Lisa. 2002. African American English: A Linguistic Introduction. Cambridge: Cambridge University Press. Green, Lisa. 2006. Negative inversion. Presentation, University of Texas at Austin, 17 April. Green, Lisa. 2011a. Force, focus and negation in African American English. Presentation at LSA Annual Meeting. Green, Lisa. 2011b. Language and the African American Child. Cambridge: Cambridge University Press. Green, Lisa. 2014. Force, focus, and negation in African American English. In Raffaella Zanuttini & Laurence Horn (eds.), Micro-syntactic Variation in North American English, 115–142. Oxford: Oxford University Press. Green, Lisa & Walter Sistrunk. 2015. Syntax and semantics in African American English. In Sonja Lanehart (ed.), The Handbook of African American Language, pp. 355–370. Oxford: Oxford University Press. Greenblatt, Stephen. 1980. Renaissance Self-Fashioning: From More to Shakespeare. Chicago: University of Chicago Press. Grice, H. Paul. 1975. Logic and conversation. In Peter Cole & Jerry Morgan (eds.), Syntax and Semantics. Vol. 3: Speech Acts, 41–58. New York: Academic Press. Grice, H. Paul. 1978. Further notes on ‘Logic and conversation.’ In Peter Cole (ed.), Syntax and Semantics. Vol. 9: Pragmatics, 113–128. New York: Academic Press. Grice, H. Paul. 1989. Studies in the Way of Words. Cambridge, MA: Harvard University Press. Gumperz, John. 1970. Verbal strategies in multilingual communication. Working Paper 36. Berkeley: Language Behavior Research Laboratory, University of California. Gumperz, John. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Gumperz, John & Eduardo Hernández-Chavez. 1972. Bilingualism, bidialectalism, and classroom interaction. In Courtney Cazden, Vera John & Dell Hymes (eds.), Functions of Language in the Classroom, 84–110. Prospect Heights, IL: Waveland Press. Gundel, Jeanette. 1999. Topic, focus, and the grammar-pragmatics interface. University of Pennsylvania Working Papers in Linguistics 6 (1). Article 14. 1–18. Guy, Gregory, Barbara Horvath, Julia Vonwiller, Elaine Daisley & Inge Rogers. 1986. An intonational change in progress in Australian English. Language in Society 15. 23–52. Haegeman, Liliane. 2000. Negative preposing, negative inversion, and the split CP. In Laurence Horn & Yasuhiko Kato (eds.), Negation and Polarity: Syntactic and Semantic Perspectives, 21–61. New York: Oxford University Press. Haspelmath, Martin. 2006. Against markedness (and what to replace it with). Journal of Linguistics 42. 25–70. Haugh, Michael. 2014. Im/Politeness Implicatures. Berlin: De Gruyter Mouton. Heinämäki, Orvokki. 1972. Before. Chicago Linguistic Society 8. 139–151. Hetzron, Robert. 1975. The presentative movement, or why the ideal word order is V.S.O.P. In Charles Li (ed.), Word Order and Word Order Change, 347–388. Austin: University of Texas Press.

206

References

Hickey, Leo. 1993. Stylistics, pragmatics, and pragmastylistics. Revue Belge de Philologie et d’Histoire 71. 573–586. Hoeksema, Jack. 1989. Review of E. Reuland & A. ter Meulen (eds.), The Representation of (In) definiteness (Cambridge, MA, 1987). Language 65. 115–125. Hoffman, Thomas. 2013. Abstract phrasal and clausal constructions. In Thomas Hoffman & Graeme Trousdale (eds.), The Oxford Handbook of Construction Grammar, 307–328. Oxford: Oxford University Press. Hoffman, Thomas & Graeme Trousdale. 2013. The Oxford Handbook of Construction Grammar. Oxford: Oxford University Press. Horn, Laurence. 1972. On the Semantic Properties of Logical Operators in English. Los Angeles, CA: University of California Los Angeles dissertation. Horn, Laurence. 1984. Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In Deborah Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications, 11–42. Washington, DC: Georgetown University Press. Horn, Laurence. 1989. A Natural History of Negation. Chicago: University of Chicago Press. Horn, Laurence. 1991. Given as new: When redundant information isn’t. Journal of Pragmatics 15. 305–328. Horn, Laurence. 2000. Pick a theory (not just any theory): Indiscriminatives and the free choice indefinite. In Laurence Horn & Yasuhiko Kato (eds.), Negation and Polarity: Syntactic and Semantic Perspectives, 147–192. New York: Oxford University Press. Horn, Laurence. 2004. Implicature. In Laurence Horn & Gregory Ward (eds.), The Handbook of Pragmatics, 3–28. Blackwell: Malden, MA. Horn, Laurence. 2013. I love me some datives: Expressive meaning, free datives, and Fimplicature. In Daniel Gutzmann & Hans-Martin Gärtner (eds.), Beyond Expressives: Explorations in Use-conditional Meaning, 151–199. Leiden: Brill. Horn, Laurence. 2015. Negative inversion(s). Paper presented at The Pragmatics of Grammar: Negation and Polarity. Université de Caen. Horn, Laurence. 2018a. Words in edgewise. Annual Review of Linguistics 4. 1–19. Horn, Laurence. 2018b. Negation and word order in the footsteps of neg-first. Paper presented at Nayfest. University of Maryland. Horn, Laurence & Barbara Abbott. 2012. (in)definiteness and implicature. In William P. Kabasenche, Michael O'Rourke & Matthew H. Slater (eds.), Reference and Referring, 325–356. Cambridge, MA: MIT Press. Horn, Laurence & Heinrich Wansing. 2017. Negation. The Stanford Encyclopedia of Philosophy. In Ed Zalta, (ed.). https://plato.stanford.edu/archives/spr2017/entries/negation/ (accessed 7 Feb. 2017). Huang, Yan. 2007. Pragmatics. Cambridge: Cambridge University Press. Huddleston, Rodney & Geoffrey Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Irvine, Judith. 2001. Style as distinctiveness: The culture and ideology of linguistic differentiation. In John Rickford & Penelope Eckert (eds.), Style and Sociolinguistic Variation, 21–43. Cambridge: Cambridge University Press. Irvine, Judith & Susan Gal. 2000. Language ideology and linguistic differentiation. In Paul Kroskrity (ed.), Regimes of Language, 35–83. Santa Fe, NM: School of American Research Press. Israel, Michael. 1996. Polarity sensitivity as lexical semantics. Linguistics and Philosophy 19. 619–666.

References

207

Israel, Michael. 2011. The Grammar of Polarity. Cambridge: Cambridge University Press. Jackendoff, Ray. 1972. Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press. Jespersen, Otto. 1909–49. A Modern English Grammar on Historical Principles. London: Allen and Unwin. Jespersen, Otto. 1917. Negation in English and Other Languages. Copenhagen: Høst. Kadmon, Nirit. 2001. Formal Pragmatics. Malden, MA: Blackwell. Kadmon, Nirit & Fred Landman. 1993. Any. Linguistics and Philosophy 16. 353–422. Karttunen, Lauri. 1973. Presuppositions and compound sentences. Linguistic Inquiry 4. 169–193. Karttunen, Lauri & Stanley Peters. 1979. Conventional implicature. In ChoonKyu Oh & David Dinneen (eds.), Syntax and Semantics. Vol. 1. Presupposition, 1–56. New York: Academic Press. Kay, Paul & Charles Fillmore. 1999. Grammatical constructions and linguistic generalizations: The ‘what’s X doing Y’ construction. Language 75. 1–33. Kay, Paul & Laura Michaelis. 2012. Constructional meaning and compositionality. In Claudia Maienborn, Klaus von Heusinger & Paul Portner (eds.), Semantics: An International Handbook of Natural Language Meaning. Vol. 3. 2271–2296. Berlin: De Gruyter Mouton. Kecskes, Istvan. 2012. Sociopragmatics and cross-cultural and intercultural studies. In Keith Allan & Kasia Jaszczolt (eds.), The Cambridge Handbook of Pragmatics, 599–616. Cambridge: Cambridge University Press. King, Ruth. 2013. Morphosyntactic variation. In Robert Bayley et al. (eds.), The Oxford Handbook of Sociolinguistics, 445–463. Oxford: Oxford University Press. Kim, Jong-Bok. 2013. On the structure of the existential construction: A hybrid analysis. Studies in Modern Grammar 72. 17–39 Klima, Ed. 1964. Negation in English. In Jerry Fodor & Jerrold Katz (eds.), The Structure of Language: Readings in the Philosophy of Language, 246–323. Englewood Cliffs, NJ: Prentice Hall. Krifka, Manfred. 1995. The semantics and pragmatics of polarity items. Linguistic Analysis 25. 209–257. Kruisinga, Etsko. 1931. A Handbook of Present-day English. 5th edn. Vol. 1. Part 2. Groningen: Noordhoff. Labov, William. 1963. The social motivation of a sound change. Word 19. 273–309. Labov, William. 1966. The Social Stratification of English in New York City. Washington, D.C.: Center for Applied Linguistics. 2nd edn., 2006. Cambridge University Press. Labov, William. 1972a. Negative attraction and negative concord in English grammar. Language 48. 773–818. Labov, William. 1972b. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Labov, William. 1972c. Some principles of linguistic methodology. Language in Society 1. 97–120. Labov, William. 2001. Principles of Linguistic Change. Vol. 2. Malden, MA: Wiley-Blackwell. Labov, William. 2012. Dialect Diversity in America: The Politics of Language Change. Charlottesville: University of Virginia Press. Labov, William et al. 1968. A Study of the Non-standard English of Negro and Puerto Rican Speakers in New York City. Vol. 1: Phonological and Grammatical Analysis. Final Report, Cooperative Project No. 3288, United States Office of Education. Labov, William, Sharon Ash & Charles Boberg. 2006. The Atlas of North American English: Phonetics, Phonology and Sound Change. Berlin: De Gruyter Mouton.

208

References

Ladusaw, William. 1979. Polarity Sensitivity as Inherent Scope Relations. Austin, TX: University of Texas at Austin dissertation. Lakoff, George. 1987. Women, Fire, and Dangerous Things. Chicago: University of Chicago Press. Lambrecht, Knud. 1990. ‘What, me worry?’ Mad Magazine sentences revisited. Berkeley Linguistics Society 16. 215–228. Lambrecht, Knud. 1994. Information Structure and Sentence Form: Topic, Focus and the Mental Representations of Discourse Referents. Cambridge: Cambridge University Press. Lanehart, Sonja. 2015. The Handbook of African American Language. Oxford: Oxford University Press. Lavendera, Beatriz. 1978. Where does the sociolinguistic variable stop? Language in Society 7. 171–182. Leech, Geoffrey. 1983. The Principles of Pragmatics. New York: Longman. LePage, Robert. 1968. Problems of description in multilingual communities. Transactions of the Philological Society 67. 189–212. LePage, Robert & Andrée Tabouret-Keller. 1985. Acts of Identity: Creole-based Approaches to Language and Ethnicity. Cambridge: Cambridge University Press. Levinson, Dmitry. 2008. Licensing of Negative Polarity Particles ‘Yet,’ ‘Anymore,’ ‘Either,’ and ‘Neither’: Combining Downward Monotonicity and Assertivity. Palo Alto, CA: Stanford University dissertation. Levinson, Stephen. 1983. Pragmatics. Cambridge: Cambridge University Press. Levinson, Stephen. 2000. Presumptive Meanings: The Theory of Generalized Conversational Implicatures. Cambridge, MA: MIT Press. Lippi‐Green, Rosina. 2012. English with an Accent: Language, Ideology, and Discrimination in the United States. 2nd edn. London: Routledge. Lumsden, Michael 1988. Existential Sentences: Their Structure and Meaning. London: Routledge. Macaulay, Ronald. 1975. Negative prestige, linguistic insecurity, and linguistic self-hatred. Lingua 36. 147–161. McCawley, James. 1978. Conversational implicature and the lexicon. In Peter Cole (ed.), Syntax and Semantics 9: Pragmatics, 245–258. Academic Press: New York. McCawley, James. 1988. The comparative conditional construction in English, German, and Chinese. Berkeley Linguistics Society 14. 176–187. McCawley, James. 1991. Contrastive negation and metalinguistic negation. Chicago Linguistic Society 27. 189–206. Martin, Stefan. 1992. Topics in the Syntax of Nonstandard English. College Park, MD: University of Maryland dissertation. Martin, Stefan & Walt Wolfram. 1998. The sentence in African-American Vernacular English. In Guy Bailey, John Baugh, Salikoko S. Mufwene, & John Rickford (eds.), African-American English: Structure, History, and Use, 11–36. New York: Routledge. Matthewson, Lisa. 2004. On the methodology of semantic fieldwork. International Journal of Linguistics 70. 369–415. Matyiku, Sabina. 2017. Semantic Effects of Head Movement: Evidence from Negative Auxiliary Inversion. New Haven, CT: Yale University dissertation. Matyiku, Sabina. 2011. Negative inversion. Yale Grammatical Diversity Project: English in North America. Updated by Tom McCoy (2015) and Katie Martin (2018). http://ygdp.yale. edu/phenomena/negative-inversion (accessed 02 April 2019).

References

209

McCready, Eric. 2010. Varieties of conventional implicature: Evidence from Japanese. Semantics and Pragmatics 3.8. 1–57. McNally, Louise. 1992. An Interpretation for the English Existential Construction. Santa Cruz, CA: University of California Santa Cruz dissertation. Reprinted: New York: Garland, 1997. McNally, Louise. 2011. Existential sentences. In Klaus Von Heusinger, Claudia Maienborn & Paul Portner (eds.), Semantics: An International Handbook of Natural Language Meaning, 1829–1848. The Hague: De Gruyter. Metcalf, Allan. 1974. The study of California Chicano English. International Journal of the Sociology of Language 2. 53–58. Michaelis, Laura. 2009. Sign-based construction grammar. In Bernd Heine & Heiko Narrog (eds.), The Oxford Handbook of Linguistic Analysis, 155–176. Oxford: Oxford University Press. Michaelis, Laura & Knud Lambrecht. 1996. Toward a construction-based model of language function: The case of nominal extraposition. Language 72. 215–247. Milsark, Gary. 1974. Existential Sentences in English. Cambridge, MA: Massachusetts Institute of Technology dissertation. Minutes and Testimony of the Joint Legislative Committee Appointed to Investigate the Public Service Commissions. 1916. In Documents of the Senate of the State of New York Legislature, 139th Session. Vol. 32, No. 42, Part 5. Albany: Lyons Printers. Montejano, David. 1987. Anglos and Mexicans in the Making of Texas, 1836–1986. Austin, TX: University of Texas Press. Montgomery, Michael. 2006. Notes on the development of existential they. American Speech 81. 132–145. Montgomery, Michael & Joseph Hall. 2004. A Dictionary of Smoky Mountain English. Knoxville: University of Tennessee Press. Napoli, Donna Jo. 1982. Initial material deletion in English. Glossa 16. 85–111. Newmeyer, Frederick. 1998 Language Form and Language Function. Cambridge, MA: MIT Press. Östman, Jan-Ola & Graeme Trousdale. 2013. Dialects, discourse, and construction grammar. In Thomas Hoffmann & Graeme Trousdale (eds.), The Oxford Handbook of Construction Grammar, 476–490. Oxford: Oxford University Press. Parrott, Jeffrey. 2000. Negative inversion in African American Vernacular English: A case of optional movement? Western Conference on Linguistics 11. 414–427. Pollard, Carl & Ivan A. Sag. 1994. Head-driven Phrase Structure Grammar. Chicago: University of Chicago Press. Pollock, Jean-Yves. 1989. Verb movement, UG, and the structure of IP. Linguistic Inquiry 20. 365–424. Poplack, Shana. 1980. Sometimes I’ll start a sentence in Spanish y termino en Español: Toward a typology of code-switching. Linguistics 18. 581–618. Potts, Christopher. 2005. The Logic of Conventional Implicature. Oxford: Oxford University Press. Potts, Christopher. 2007. Into the conventional implicature dimension. Philosophy Compass 2. 665–679. Potts, Christopher. 2015. Presupposition and implicature. In Shalom Lappin & Chris Fox (eds.), The Handbook of Contemporary Semantic Theory, 2nd edn., 168–202. Oxford: WileyBlackwell. Poutsma, Hendrik. 1916. A Grammar of Late Modern English. Part 2. The Parts of Speech, Section 1B. Groningen: Noordhoff. Preston, Dennis. 1989. Perceptual Dialectology: Nonlinguists’ Views of Areal Linguistics. Dordrecht: Foris.

210

References

Preston, Dennis. 1996. Where the worst English is spoken. In Edgard Schneider (ed.), Focus on the USA, 297–360. Amsterdam: Benjamins. Preston, Dennis. 2006. Language with an attitude. In Jack Chambers, Peter Trudgill & Natalie Schilling-Estes (eds.) The Handbook of Language Variation and Change, 40–66. Malden, MA: Blackwell. Preston, Dennis. 2015. The South: Still different. In Michael Picone & Catherine Davies (eds.), New Perspectives on Language Variety in the South: Historical and Contemporary Approaches, 311–326. Tuscaloosa: University of Alabama Press. Preston, Dennis. 2018. Changing research on the changing perceptions of Southern U.S. English. American Speech 93. 471–496. Prince, Ellen. 1988. The discourse functions of Yiddish expletive es + subject-postposing. IPrA Papers in Pragmatics 2. 176–179. Prince, Ellen. 1992. The ZPG letter: Subjects, definiteness, and information-status. In Sandra Thompson & William Mann (eds.), Discourse Description: Diverse Analyses of a Fundraising Text, 295–325. Amsterdam: John Benjamins. Prince, Alan & Paul Smolensky. 1993/2004. Optimality theory: Constraint interaction in generative grammar. Technical Report, Rutgers University and University of Colorado at Boulder, 1993. Revised version published by Blackwell, 2004. Malden, MA. Pullum, Geoffrey. 1991. English nominal gerund phrases as noun phrases with verb-phrase heads. Linguistics 29. 763–799. Putnam, George & Edna O’Hern. 1955. The status significance of an isolated urban dialect. Language 31. V-32. Reichardt, Mary. 1996. Introduction. In Mary Reichardt (ed.), A Mary Wilkins Freeman Reader, vii–xxi. Omaha: University of Nebraska Press. Reisman, Karl. 1970. Cultural and linguistic ambiguity in a West Indian Village. In Norman Whitten & John Szwed (eds.), Afro-American Anthropology, pp. 129–144. NY: The Free Press. Reynolds, Bill. 1994. The /z/ → [d] variable rule in negative auxiliary contractions in Southern states English. American Speech 69. 361–372. Rickford, John. 1983. Standard and non-standard language attitudes in a creole continuum. Society for Caribbean Linguistics, Occasional Papers, 16. St. Augustine, Trinidad. Reprinted in Nessa Wolfson & Joan Manes (eds.), Language of Inequality, 145–160. The Hague: Mouton, 1985. Rickford, John. 1987. The haves and have nots: Sociolinguistic surveys and the assessment of speaker competence. Language in Society 16. 149–177. Rickford, John & Elizabeth Closs Traugott. 1985. Symbol of powerlessness and degeneracy, or symbol of solidarity and truth? Paradoxical attitudes towards pidgins and creoles. In Sidney Greenbaum (ed.) The English Language Today, 252–261. Oxford: Pergamon. Rizzi, Luigi. 1997. The fine structure of the left periphery. In Liliane Haegeman (ed.), Elements of Grammar, 281–337. Dordrecht: Kluwer. Rochemont, Michael. 1978. A Theory of Stylistic Rules in English. Amherst, MA: University of Massachusetts dissertation. Rochet, Bernard. 1976. The Formation and the Evolution of the French Nasal Vowels. Berlin: Walter de Gruyter. Rohrbaugh, Gene. 1994. Focus and the licensing of negative polarity items. Proceedings of the Mid-America Linguistic Conference. 224–235. Lawrence, KS. Rullmann, Hotze. 1996. Two types of negative polarity items. North East Linguistic Society 2. 335–350.

References

211

Russell, Bertrand. 1905. On denoting. Mind 14. 479–493. Sadock, Jerrold. 1978. On testing for conversational implicature. In Peter Cole (ed.), Syntax and Semantics. Vol. 9: Pragmatics, 281–297. New York: Academic Press. Sailor, Craig. 2009. Tagged for Deletion: A Typological Approach to VP Ellipsis in Tag Questions. M.A. thesis, UCLA. Salmon, William. 2011. Conventional implicature, presupposition, and the meaning of must. Journal of Pragmatics 43. 3416–3430. Salmon, William. 2017. Meaning and use of negative inversions in Texan and African-American English. Berkeley Linguistics Society 43. 285–300. Salmon, William. 2018a. Negative auxiliaries and absent expletives. Journal of Pragmatics 130. 51–66. Salmon, William. 2018b. Camouflage and style across two varieties of US English. Style 52. 404–420. Salmon, William & Jennifer Gómez Menjívar. 2016. Language variation and dimensions of prestige in Belizean Kriol. Journal of Pidgin and Creole Languages 31. 316–360. Sapir, Edward. 1931. Dialect. Encyclopedia of the Social Sciences 5. 123–126. New York: Macmillan. Sanders, Willease. 1978. Selected Grammatical Features of the Speech of Blacks in Columbia, S.C. Columbia, SC: University of South Carolina dissertation. Santa Ana, Otto. 1993. Chicano English and the nature of the Chicano language setting. Hispanic Journal of Behavioral Sciences 15. 3–35. Schneider, Edgar. 1989. American Earlier Black English. Tuscaloosa: The University of Alabama Press. Schütze, Carson. 1996. The Empirical Base of Linguistics: Grammaticality Judgments and Linguistic Methodology. Chicago: University of Chicago Press. Sells, Peter, John Rickford & Thomas Wasow. 1996. An optimality theoretic approach to variation in negative inversion in AAVE. Natural Language and Linguistic Theory 14. 591–627. Siemund, Peter. 2013. Varieties of English: A Typological Approach. Cambridge: Cambridge University Press. Smitherman, Geneva. 2006. Word from the Mother: Language and African Americans. New York: Routledge. Sobin, Nicholas. 2003. Negative inversion as nonmovement. Syntax 6. 183–212. Spears, Arthur. 1982. The Black English semi-auxiliary ‘come.’ Language 58. 850–872. Spears, Arthur K. 2015. African American Standard English. In Sonja Lanehart (ed.), The Oxford Handbook of African American Language, 786–799. Oxford: Oxford University Press. Sperber, Dan & Deirdre Wilson. 1986. Relevance: communication and cognition. Oxford: Blackwell. 2nd edn. 1995. Syrett, Kristin, Georgia Simon & Kirsten Nisula. 2014. Prosodic disambiguation of scopally ambiguous sentences in a discourse context. Journal of Linguistics 50. 453–493. Szekely, Rachel. 2015. Truth Without Predication. The Role of Placing in the Existential ThereSentence. London: Palgrave. Terkourafi, Marina. 2011. The pragmatic variable: Toward a procedural interpretation. Language in Society 40. 343–372. Thomas, Erik. 1997. A rural/metropolitan split in the speech of Texas Anglos. Language Variation and Change 9. 309–332. Thomas, Erik. 2007. Phonological and phonetic characteristics of African American Vernacular English. Language and Linguistics Compass 1. 450–475.

212

References

Thomas, Erik. 2015. Prosodic features of African American English. In Sonja Lanehart (ed.), The Oxford Handbook of African American Language, 420–435. Oxford: Oxford University Press. Thomas, Erik & Guy Bailey. 2015. Segmental phonology of African American English. In Sonja Lanehart (ed.), The Oxford Handbook of African American Language, 403–419. Oxford: Oxford University Press. Thrasher, Randolph. 1974. Shouldn’t Ignore These Strings: A Study of Conversational Deletion. Ann Arbor, MI: University of Michigan dissertation. Tonhauser, Judith, David Beaver, Craige Roberts & Mandy Simons. 2013. Toward a taxonomy of projective content. Language 89. 66–109. Urciuoli, Bonnie. 1996. Exposing Prejudice: Puerto Rican Experiences of Language, Race, and Class. Boulder, CO: Westview Press. Wald, Benji. 1996. Substratal effects on the evolution of modals in East LA English. In Jennifer Arnold et al., (eds.), Sociolinguistic Variation: Data, Theory, and Analysis: Selected Papers from NWAV 23 at Stanford, 513–530. Stanford, CA: CSLI. Walker, James. 2010. Variation in Linguistic Systems. New York: Routledge. Ward, Gregory & Betty Birner. 1995. Definiteness and the English existential. Language 71. 722–742. Weir, Andrew. 2012. Left-edge deletion in English and subject omission in diaries. English Language and Linguistics 16. 105–29. Weiner, E. Judith & William Labov. 1983. Constraints on the agentless passive. Journal of Linguistics 19. 29–58. Weldon, Tracey. 1994. An HPSG account of negative inversion in African American Vernacular English. Ms., Ohio State University. White-Sustaíta, Jessica. 2010. Reconsidering the syntax of non-canonical negative inversion. English Language and Linguistics 14. 429–455. Wilkinson, Kyle. 2008. Yeomen, Sharecroppers, and Socialists: Plain Folk Protest in Texas, 1870–1914. College Station, TX: Texas A & M Press. Wilson, August. 1986. Fences: A Play. New York: Plume. Wolfram, Walt. 1991. The linguistic variable: Fact and fantasy. American Speech 66. 22–32. Wolfram, Walt. 2004. The grammar of urban African-American Vernacular English. In Bernd Kortmann & Edgar Schneider. Handbook of Varieties of English. Vol. 2: Morphology and Syntax, 111–132. Berlin: Mouton de Gruyter. Wolfram, Walt. 2007. Sociolinguistic folklore in the study of African American English. Language and Linguistics Compass 1. 450–475. Wolfram, Walt & Natalie Schilling-Estes. 1998. American English. Malden, MA: Blackwell. Wolfram, Walt & Natalie Schilling. 2016. American English: Dialects and Variation. 3rd edn. Malden, MA: Blackwell. Wolfram, Walt & Donna Christian. 1976 Appalachian Speech. Arlington, VA: Center for Applied Linguistics. Zanuttini, Raffaella & Judy Bernstein. 2014. Transitive expletives in Appalachian English. In Raffaella Zanuttini & Laurence Horn (eds.), Micro-syntactic Variation in North American English, 143–177. Oxford: Oxford University Press. Ziv, Yael. 1982. Another look at definites in existentials. Linguistics 18. 73–88. Zwicky, Arnold & Geoffrey Pullum. 1983a. Cliticization vs. inflection: English n’t. Language 59. 502–513. Zwicky, Arnold & Geoffrey Pullum. 1983b. Deleting named morphemes. Lingua 59. 155–75.

Index /z/ → [d] in Southern US English 35, 100 [t] release of – prestige 110 Abbott, Barbara 5, 26, 159 accommodation theory 156 act of identity 156 affect 6, 140, 148 African American Standard English 16, 115 African American Vernacular English, AAVE – and camouflage 60–61 – grammatical features of 16 – in contact with Anglo and Chicano Englishes 25–27 – in east-central Texas 15 – in New York City 14, 15, 27, 139 – in Palo Alto, California 15, 54 – regional variation of 15, 50 – zero copula in 51 age – apparent time and 95 – of survey participants 91, 94 – variation of negative inversion based on 91 Anglo (definition of) 16 Anglo English – grammatical features of 17 – in California 20, 26 – in contact with African American and Chicano Englishes 25 – in Texas 16–18, 91–93 – social class and 17 any – free choice 58 – negative polarity item 58, 60, 141, 146, 148 apparent-time construct 3, 5, 10, 95, 108, 196 assimilation 28, 35, 100, 103 auxiliary verb 1, 2, 11, 22, 32, 35 Bailey, Guy 7, 15, 16, 20, 62, 89, 95, 108 barely meaning just recently 23 Bernstein, Judy 13, 45, 52, 61, 66, 89, 90, 98

https://doi.org/10.1515/9781501512346-010

bidialectalism 27 Birner, Betty 3, 4, 5, 15, 44, 55, 68, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79 Blanchette, Frances 1, 2, 12, 42, 43, 45, 47, 48, 52, 55, 56, 66, 69 Bolinger, Dwight 4, 20, 99 borrowing – affective 110 camouflage – habitual be 21, 49, 155 – indignant-come 61 – negative inversion 60–61 – progressive aspect steady 61 – remote-time been 61 Chicano (definition of) 19 Chicano English – grammatical features of 18 – in California 13, 18–19, 20, 21, 23, 31 – in contact with African American and Anglo Englishes 25 – in contact with Spanish 9 – in Texas 94 – relation to Spanish 18 cliticization 36 code-switching 6, 18, 152, 153, 154 coke/pop/soda 116 Collins, Chris 2,42, 43, 45, 46, 47, 48, 52, 55, 56, 66, 69 compensatory lengthening 28 competence could 23 compositionality 16, 157 consonant cluster reduction 19 construable 3, 44, 55, 71 construction grammar 2, 6, 19, 27, 37, 145 constructions 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 14, 16, 17, 18, 19, 20, 21, 22, 25, 28, 29, 30, 32, 33, 34, 36, 41, 42, 43, 52, 65, 67, 70, 79, 84, 86, 88, 96, 97, 99, 105, 106, 107, 109, 113, 115, 117, 119, 120, 122, 135, 136, 138, 139, 142, 149, 152, 155, 161 – components of 18, 29

214

Index

contrastive focus 85, 86 conventional implicature – Bach, Kent and 134 – f-implicature and 121 – Grice, H. Paul and 133 – Horn, Laurence and 134 – Karttunen, Laurie & Stanley Peters and 133 – Potts, Christopher and 134 conversational deletion – extraposition and 101 – it clefts and 101 – telegraphic speech and 102 – weather it and 101 conversational implicature – (non)conventionality of 132 – (non)detachability of 131 – calculability of 131 – cancellability of 129 – Reinforceability of 133 cooperative principle 7, 129 creole languages 5, 110, 114, 156 Cukor-Avila, Patricia 15, 16, 17, 30, 62, 89, 114 danger of death narrative 3 declarative negative auxiliary inversion 2, See also negative inversion definite article – conventional implicature of 5, 159 – conversational implicature and 5, 159 – familiarity and 5, 159 – truth conditional content of 5 – uniqueness and 5, 159 definiteness effects – gradient effects 68 – Milsark, Gary and 68 – negative inversion and 68–80 – Prince, Ellen and 71 – Ward, Gregory & Birner, Betty and 68, 71 Deliverance (film) – “Dueling Banjos” song 62 – as negative icon for rural communities 62 demonstratives 44, 76 dialectal crossing 14 discourse markers 22 division of pragmatic labor 158

echoic usage 36 Eckert, Penelope 8, 110, 121, 122, 155 emphatic – conventional 150–152 – Green, Lisa and 141–142, 146 – Israel, Michael and 146–149 – Labov, William and 139–140, 142–146 – motivation for negative inversion movement 6 – pragmatic 152–159 ethnolect 6, 7, 18, 31, 65, 96 existential constructions 30, see also there-existential expletive subjects – de-stressing and deletion 106 – it 3, 4 – phonologically null there and 103 – there 5 – they 5 expressive meaning 121 ex-slave narratives 25, 62, 89, 108 extended projection principle 9 eye dialect spelling – cain’t 33 – wudn’t 34 face threatening act 148 family of sentences 123, 124, 144 – presupposition 123, 124, 144 fieldwork 115 Fillmore, Charles 1, 6, 11, 15, 16, 17, 18, 28, 29 focuser like 23 Foreman, John 2, 5, 6, 9, 10, 11, 12, 13, 30, 34, 35, 43, 44, 45, 46, 47, 48, 49, 50, 52, 53, 55, 56, 57, 58, 61, 63, 66, 69, 75, 80, 103, 104, 105, 106, 138, 139, 143 formal idiom 11, 162–191 Fought, Carmen 2, 12, 13, 14, 18, 19, 20, 21, 22, 23, 24, 25, 155 Great Black Migration 26 Green, Lisa 2, 5, 7, 11, 14, 30, 36 ,43, 45, 49, 50, 52, 57, 60, 61, 62, 64, 66, 70, 75, 89, 90, 97, 119, 138, 139, 141, 142, 145, 146, 148, 151, 152, 157, 160, 161 Grice, H. Paul 3, 129, 131, 134, 135, 144, 149, 154, 161

Index

group membership 113 Gullah 105 gumband 116 Gumperz, John 6, 154, 155 Gundel, Jeanette 74 head-driven phrase structure grammar 142 Hillbilly Highway 26 honorifics 131 Horn, Laurence 2, 5, 13, 30, 58, 59, 63, 86, 98, 120, 121, 122, 127, 129, 134, 151, 152, 158, 159 idioms 1, 11, 12, 38, 162–191 imperatives 20 incredulity response construction 15, 17, 21, 28, 36, 176, 177, 182, 189 indefinite article 5, 17 indexical fields 8, 121, 122 Indian English 105 ineffability 8, 120–122 informal speech 4, 19, 31 informative value 147 inheritance hierarchy 191 invited inference 6, 140, 146, 154, 156, 158, 159 Israel, Michael 53, 58, 85, 139, 146, 147, 148, 149, 155 Jackson, Vince 70, 75, 76 Jespersen, Otto 4, 7, 68, 80, 98, 99, 100, 102, 106, 145 – Jespersen’s Cycle 145 Kay, Paul 1, 6, 11, 15, 16, 28, 29 Kecskes, Istvan 7 Labov, William 1, 2, 3, 4, 5, 6, 8, 9, 10, 11, 15, 22, 27, 32, 34, 35, 42, 50, 51, 52, 54, 79, 93, 95, 107, 112, 113, 116, 136, 138, 139, 140, 141, 145, 146, 150, 151, 153, 154, 156, 159, 160 Lakoff, George 19, 25, 27, 28, 30, 37 language attitudes – attitudes toward speakers 10 – formal 115

215

– informal 115 – paradoxical attitudes toward vernacular language 114 language change – age and variation 95 – apparent time 3, 95, see also apparenttime construct language contact 9, 19, 25, 27 language ideologies. See language attitudes language shift 106 Levinson, Stephen 6, 68, 82, 125, 126, 127, 129, 130, 131, 132, 133, 134, 154 lexical content 5, 6, 19 linguistic inferiority principle 114 linguistic insecurity 116 linguistic push-pull 114 linguistic self-hatred 112 linguistic variable – (ING) 153 – functional equivalence 160 – negative inversion and 159–160 – truth-conditional equivalence 160 markedness 158 markers. See social marker Martha’s Vineyard research (Labov, William) 95 Martin, Stefan 1, 2, 4, 6, 7, 8, 15, 43, 52, 53, 62, 64, 68, 69, 90, 97, 103, 106, 138 Matyiku, Sabina 1, 2, 12, 13, 14, 31, 33, 42, 43, 44, 45, 50, 52, 54, 55, 57, 59, 63, 66, 67, 69, 80, 81, 82, 85, 86, 106, 139, 150 maxim of manner 131 McNally, Louise 27, 28, 43, 68, 83, 85 metalinguistic negation 85 minimizers 147 modal existential construction 2, 4, 10, 29, 61–63, 68, 79, 89, See also transitive expletive construction – older and younger speakers 87–91, 106, 108, 166, 168, 190, 197 – reconstructed 97–103 Modified Occam’s Razor 161 n’t inflection 36 nasal assimilation in French 103, 189

216

Index

needs washed construction 116 negative auxiliary first 2, See also negative inversion negative auxiliary inversion 2, See also negative inversion negative existential sentences 8, 9, 10, See also nexistentials – negative inversion and 8, 107 negative inversion in standard English 115 negative inversion – (overt) complementizers 53 – absolute negation interpretation of 14, 141, 142 – affective character of 139–140 – and scope freezing 45 – anti-indiscriminative reading of 58, 59 – apparent time and 95 – but and NI subject 122 – cataphoric reference and 77 – compromisers and 143 – definite subjects in 2, 10, 12, 28, 44, 67, 69, 71, 72, 76, 79, 108 – deictics and 76 – deontic reading of 41, 42 – double modals in 41 – downtoners and 143 – dual reference and 73 – emphatic character of 1, 52, 145 – epistemic reading of 41 – false definites and 78 – hearer-new referents and 67–108 – hearer-old referents and 67–103 – hedges and 143 – in African American English 15–16 – in Alabama English 50, 61 – in Appalachian English 31, 61, 69 – in country music 65 – in embedded contexts 126 – in K-pop 110 – indefinite subjects of 13, 33, 56, 67 – list reading and 73 – modal existential and 29, 61–63, 67–108 – movement analyses of 6, 170–179 – negative concord and 22, 33, 47, 50, 51, 57, 70, 113, 115, 143, 173 – non-inverted counterpart of 56–60 – non-movement analyses of 6–9, 166–170

– propositional content of 5, 6, 8, 52, 60, 118–120, 131–135, 137, 140–142, 145, 147, 152, 159, 180, 196 – prosodic properties of 10, 11, 98 – scope of negation in 3 – semi-modal verbs and 37–39, 183–184 – subject noun phrases in 42–49 – subject-auxiliary inversion and 2, 43, 44, 53, 180, 181 – superlatives and 17, 76, 78 – Syntactico-semantic straitjacket and 8, 199 – uniquely identifiable referents and 67–108 – with weak quantifiers 4, 11, 45, 46, 66 negative polarity item 7, 38, 52–53, 58, 60, 82, 139, 141, 146, 147, 148, 155 negative strengthening 156 neg-first principle 98 non-canonical negative inversion 2, See also negative inversion non-negative inversion 64 nonstandard pronoun forms 21 nonstandard use of in and on 24 northern cities vowel shift 116 not-initial sentences 55–56 n-words 13, 50 O’Connor, Mary Catherine 1, 6, 11, 15, 16, 28, 29 optimality theory 8 Parrott, Jeffrey 1, 2, 8, 15, 37, 50, 52, 60, 103, 106, 145, 146 perceptual dialectology 17, 29 – map task and 30 perfective had 21 personal dative construction 120, 121 pin-pen merger 136 politeness 6, 7 popular negative inversion 2, See also negative inversion positive anymore 116 positive polarity item 3, 85 Potts, Christopher 8, 75, 120, 121, 122, 123, 128, 129, 133, 134, 144 pragmatics 1, 2, 3, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 19, 27, 28, 29, 30, 32, 37, 45, 47,

Index

48, 49, 52, 53, 60, 67, 68, 69, 70, 71, 74, 75, 80, 105, 108, 115, 119, 120, 122, 135, 136, 137, 139, 144, 146, 151, 152, 159, 160 – sociolinguistics and 3–7, 8, 10, 192–199 – covert 113–115 – overt 113–115 Preston, Dennis 29, 114 presupposition – as pragmatic 3, 144 – as semantic 3, 144 – backgrounded 123, 128, 144 – cancellation of 126 – conditionals and 123 – defeasibility 126, 129 – family of sentences 123, 124, 144 – holes and 124, 125 – interrogatives and 123 – negation and 124, 126 – plugs and 125 – suspension of 127 – taken for granted 123 principle of subordinate shift 93 prolixity 158 propositional content 199, See also ‘the thought,’ and ‘what is said’ quantitative value 147 quotative like 23 redundancy 133 register 7, 16, 17, 18, 84, 154 regularization of irregular verbs 21 relevance theory 120, 160 Rickford, John 2, 5, 6, 22, 114, 142, 156 Salmon, William 2, 14, 42, 43, 44, 60, 67, 69, 72, 80, 156 Sapir, Edward 113 Schilling, Natalie 31, 110, 113, 114, 152, 154 Schwenter, Scott 7, 8, 199 slur reclamation 156 social indicator 127 social marker 1, 3, 4, 7, 10, 11, 18, 136, 153, 161

217

social meaning 1, 2, 3, 4, 5, 6, 8, 10, 11, 12, 15, 18–119, 120, 122, 124, 125, 127, 131, 135–136, 139, 144, 149, 152, 153, 154, 156, 160, 161 social stereotype 135–136 socioeconomic class 30–31 sociolinguistics – interactional 6 – variationist 4, 159 sociopragmatics 7 solidarity 113, 147, 149, 156 Spanish 9, 13, 17, 18, 24 – relation to Chicano English 18 Spears, Arthur 10, 16, 60, 115 split subjects 61 stigmatization 4, 5, 9, 10, 11, 16, 17, 19, 24, 62, 93, 110, 111, 112, 113, 114, 115, 119, 136, 149, 153, 156 strength 27, 46, 148, 151, 152, 157 style 6, 10, 11, 15, 16, 17, 18, 20, 28, 30, 32, 113, 114, 118, 119, 149, 153, 154, 155, 157, 158, 161 – difference 87 style-shifting 3, 152, 153, 154, 155 – iconicity and 155 subject – of negative inversion 42–49 – of tag questions 104, 105 subject-auxiliary inversion – comparatives and 23, 184 – counterfactual conditionals and 23, 184 – exclamatives and 21, 23, 184 – in embedded questions 22, 184 – positive rejoinder and 23, 184 – wishes/curses and 23, 184 – yes/no questions and 20, 23, 29, 53, 184 tag questions – as subjecthood diagnostic 103, 106 – syntactic identity with subject 105 – yes/no questions and ellipsis 105 Texas – Abilene (city) 29, 91 – agriculture in 8 – Corpus Christi (city) 20, 25, 29, 72, 91, 111 – ethnic groups in 26

218

Index

– language varieties in 17 – Odessa (city) 29, 91 – rural and urban dialect divergence in 17 – settler history of 17 Texas English Project 16 ‘the thought’ 119 the X-er, the Y-er construction 16, 17, 21 there-existential – anchor subject 68 – cataphora and 77 – conversational deletion and 4 – deictics and 76 – dual reference and 73 – every and TE subjects 81, 86 – false definites and 78 – focus NP 68 – half and TE subjects 81, 83 – hearer-new referents and 55, 71, 77 – hearer-old referents and 71 – many of the TE subjects 83, 84 – list reading and 74 – negative inversion and 80–87 – pivot of 68 – none of the TE subjects 83 – uniquely identifiable referents and 55, 71, 77 Thomas, Erik 17, 20 transitive expletive construction 89, See also modal existential construction trust 149, 156 tu/vous pronouns 122

universal quantifier interpretation 10, 46, 47, 66 US Standard English 20 variable absence of past-tense marking 24 variationist pragmatics 4 verb-initial negative inversion 2, See also negative inversion vernacular language 5, 7–8, 110, 115 Ward, Gregory 3, 4, 5, 15, 44, 55, 68, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 82 Weldon, Tracey 1, 2, 34, 35, 37, 62, 138, 139, 140, 142 Welsh English 105 ‘what is said’ 131, 132, 133, 134, 144, 147 what the X construction 21 White-Sustaíta, Jessica 2, 4, 6, 13, 14, 15, 49, 56, 57, 61, 62, 63, 64, 87, 89, 90, 99, 100, 107, 145 Wolfram, Walt 2, 4, 15, 31, 43, 60, 62, 90, 97, 110, 113, 114, 152, 154, 160 working-class speech – informal speech and 19 – stigmatization of 93 Yale Grammatical Diversity Project 33, 43, 61, 67, 69, 70, 97, 138 Zanuttini, Raffaella 3, 13, 45, 52, 61, 66, 89, 90, 98