Children's acquisition of English "high" negation: A window into the logic and composition of bias in questions

Chapter 11 Children’s acquisition of English “high” negation: A window into the logic and composition of bias in questions Rebecca Woods Newcastle University Tom Roeper University of Massachusetts Amherst In our investigation of biased “high” negation questions (NegQs) and negative tag structures, we present production data from early child acquisition of English and judgements from adult English. We use this data to demonstrate that the structures of negative tags and NegQs are distinct despite similarities in their interpretation, spelling out how distinct structures lead to differences in use and acquisition. We also highlight the remarkable ability of very young children to manipulate a discourse context shared with another person using increasingly fine-grained syntactic structures. 1 Introduction In this chapter we argue that data from non-canonical biased questions in early language acquisition can greatly enhance and refine our formal understanding of such questions and how we conceptualise the integration of propositional and contextual information in the minds of language users. Children’s production has not, to our knowledge, been considered in theoretical accounts of biased structures and their meanings; nor have biased structures received much attention in existing acquisition literature, except from the point of view of non-target structures in child syntax (e.g. Guasti et al. 1995). We therefore make an empirical Rebecca Woods & Tom Roeper. 2025. Children’s acquisition of English “high” negation: A window into the logic and composition of bias in questions. In Tue Trinh, Anton Benz, Daniel Goodhue, Kazuko Yatsushiro & Manfred Krifka (eds.), Biased questions: Experimental results and theoretical modelling, 331–369. Berlin: Language Science Press. DOI: 10.5281/zenodo.17158196 Rebecca Woods & Tom Roeper contribution by focusing on English-acquiring children’s production of “high” negation structures in naturalistic speech settings. We also make a theoretical contribution by revealing structural differences between NegQs and negative non-matching1tag questions on the basis of this data, as well as claiming a structural distinction between rising and falling negative tag questions. Let us first present the core data for this chapter. (1) illustrates negative polar questions (NegQs).2All examples in (1) are taken from the CHILDES database (MacWhinney 2000); corpus names are given in parentheses. (1) a. Isn’t that funny? Child 2 (Valian), 1;9 b. There, don’t you see it? Ross (MacWhinney), 2;4 c. Doesn’t it feel good, ma? Victor (Gleason), 2;4 d. Isn’t this mine? Isn’t this mine? Barbara (Belfast), 2;7 In NegQs, the clitic negation n’t is, we claim, not propositional but metalinguistic (see Goodhue 2022a,b for a similar claim, building on intuitions by Ladd 1981). This metalinguistic negation is structurally and semantically distinct from the negation we see in superficially similar negative tag structures like (2). In (1), metalinguistic negation scopes above the propositional content of the utterance, negating the typical interpretation of an interrogative-typed clause, namely, that the speaker is ignorant as to the truth of the proposition and expects their addressee to be knowledgeable.3In (2), n’t is interpreted as propositional negation that scopes under an interrogative clause-type operator, negating some proposition pwithin that clause. (2) a. Close to Rachel’s feet, wasn’t it? Anne (Manchester), 1;11 b. Now it needs ironing, doesn’t it? Gail (Manchester), 2;3 c. We saw some at the zoo, didn’t we? Joel (Manchester), 2;6 The structures in (2) consist of positive anchors – affirmative declarative clauses that first introduce the proposition at issue – and negative tags, the structure 1This means that the polarity of the tag and its associated declarative clause (its anchor) do not match. In the context of our chapter, all anchors have positive polarity. 2As some NegQs, like (1a), are string-equivalent to negative polar exclamatives, we handchecked all potentially ambiguous strings in context to determine whether the string was used as a NegQ. See Section 2.1 for more details. 3Details to follow in Section 3.2 and Section 3.3.3. 332 11 Children’s acquisition of English “high” negation of which is a key proposal in this chapter.4We do not have prosodic information for these structures, as the original audio is no longer available, but following Dehé & Braun’s (2013) work on the British component of the International Corpus of English, we assume that they, like most reverse-polarity tags in adult English, are likely composed of two intonational contours; one for the anchor and one for the tag. Such tags are often referred to in the literature as nuclear negative tag structures (following Ladd 1981). The negation that they contain is, we argue, typical propositional negation. We will propose structures for NegQs and negative tag questions assuming a speech act syntactic framework as in (3), which we assume is present in the left periphery of all root utterances. (3) [SpeechActP [PerspectiveP [CP [TP….]]]] cf. Woods (2021) This structure contains two discourse-related syntactic projections, SpeechActP and PerspectiveP. SpeechActP, the highest syntactic node in (3) (and therefore in (1–2)), contains a discourse commitment operator. All root utterances contain one of these operators, which mark to what and to whom the speaker5is committed. This is very like the concept of the Common Ground as articulated by Stalnaker (1978, 2002), Gunlogson (2003, 2008) and others, but the commitmentbased approach we use here (for further elaboration see e.g. Krifka 2015, Geurts 2019) differs from intentionalist, belief-oriented approaches to speech acts, so we will briefly explain what discourse commitment operators express. In this chapter we focus on operators that express one of two types of commitment between the speaker, the addressee, the propositional content and the discourse context. One type is found in utterances, in the making of which a speaker commits in the discourse context to acting as though some proposition is true (essentially, assertions).6This will be referred to in shorthand as the speaker commits to the proposition. The second type is found in utterances, in the making of which the speaker commits to the realisation of a goal, namely the addressee committing 4We do not consider matching tags – i.e. negative anchors with negative tags – in this chapter, though we do find a very small number of them in the dataset. 5We recognise here that terminology such as “speaker” and “speech act” can be exclusionary and ableist. We use the terms “speaker” and “speech act” throughout because our primary empirical data is spoken child English, but we believe that our theoretical assumptions and claims about interlocutor-information relationships can apply whether or not the interlocutors in question use spoken or signed languages. 6In Gu & Roeper (2011) and Roeper (2016), a slightly different perspective is advanced about implicit arguments, general point of view and what the affirmation of a speech act involves. Of course, a dialogues carries implications and invokes inferences beyond calculation of Common Ground or commitments and the pursuit of assent or confirmation. 333 Rebecca Woods & Tom Roeper to acting as though some proposition is true (essentially, questions).7This will be referred to in shorthand as the speaker commits to resolving the issue of some proposition. These operators, assert and question, are often realised in English as different intonation contours on the right edge of the utterance, so for our purposes, on the tag part of negative tag structures.8 These operators are not unique to NegQs or tag structures, but these noncanonical question structures are excellent proving grounds for their presence and effects. We will claim that NegQs are fundamentally question structures containing interrogative clauses that generate the commitments typical of questions – i.e. that the speaker is committed to the goal of resolving an issue, via an expectation that the addressee will provide an answer. Note that this implies that NegQs will typically be used when the speaker believes this expectation will be met; an assumption made by speech act theorists from Searle (1969) to Farkas (2022) and many in between. Negative tag structures, on the other hand, are split into two types. assert negative tag structures are fundamentally assertions and behave as such in discourse, while question negative tag structures are fundamentally questions in their discourse commitments. We move down the structure in (3), now, to the second discourse-related projection, PerspectiveP, which encodes interlocutor perspective. That is, elements in PerspectiveP mark and modify from whose perspective we should understand the propositional content of the CP. The range of operators that can merge in PerspectiveP is much larger, ranging from representations of the speaker and addressee, to modal and logical operators. In the case of NegQs, we will argue that metalinguistic negation is merged in PerspectiveP; for negative tag structures, we argue for representations of the speaker and addressee. With these two projections, SpeechActP and PerspectiveP, and the elements that merge in them, we can make specific claims about the source of bias in NegQs and tag structures; that is to say, how the speaker expresses their prior knowledge or beliefs while still looking to elicit a response from their addressee. We will claim that the bias in a NegQ arises from the metalinguistic negation and 7Note that there are issues here around pragmatic recursion or recursion of interlocutor goals, which could have interesting consequences for theories of pragmatic development and evolution of communication. We do not have space to deal with these issues here and leave them for future work. 8See Heim (2019) for an approach to mapping prosody and speech acts that does not assume speech act operators, but demonstrates a relationship between pitch excursion and commitment; namely that rising pitch contours correlate with reduced speaker commitment to the proposition. 334 11 Children’s acquisition of English “high” negation its interaction with interrogative clause typing. In assert negative tag structures, bias arises from the fact that the user actually does assert that a particular proposition is true from their perspective. In question negative tag structures, bias arises through the interaction of the anchor proposition and an addressee operator in PerspectiveP (details to follow in Section 3.1.3.1). We will demonstrate that these proposals predict, correctly, that the bias in question tag structures is more similar to that in NegQs that to assert tag structures, without being identical or derived from NegQ bias. In addition to our syntactic claims, we use Farkas & Bruce’s (2010) Table model, as updated by Farkas (2022), to fully model the pragmatic characteristics of these utterances that fall in part out of our syntactic analyses. These frameworks allow us to demonstrate our claims that NegQs and negative tag structures are syntactically different (contra Sailor 2012, Jamieson 2018, a.o.) and pragmatically distinct in how they generate and communicate bias. Note that we do not give in this chapter formal semantic denotations for metalinguistic negation or the operators we propose. In this chapter we will demonstrate the meanings of these operators using paraphrase and leave formalisations for future work. Our account is motivated by empirical evidence from the developmental path of English-acquiring children that we present in this chapter: children acquire and use negative tag structures before they begin to use NegQs. This evidence strongly supports an account of these constructions that is parsimonious in its lexical array in order to explain very early, target-like acquisition of negative tag structures. However, it is fine-grained with respect to the high left-periphery of the clausal structure to allow children to express the complex relationships between interlocutors and propositional material that they develop the ability to conceptualise. We therefore exhort theorists to take account of acquisition data wherever possible rather than relying on adult production and intuitions, not least so that we do not try to reduce to specialised operators that which we can achieve with a compositional, acquirable concept of syntactic structure and its interfaces with other modules. The chapter is structured as follows. In Section 2 we give a quantitative overview of how children produce negative tag structures and NegQs before looking deeper into the qualities of these utterances. We present our proposal for the syntax of negative tag structures (Section 3.1) and NegQs (Section 3.2), then in Section 3.3 demonstrate how our proposal captures the child’s developmental path. In Section 3.3 we also introduce new diagnostics pertaining to response and assent patterns that further support our syntactic claims. We then summarise in Section 4. 335 Rebecca Woods & Tom Roeper 2 “High” negation in child English 2.1 Quantitative data The use of “high” negation in child English follows a child’s first use of tense, auxiliaries (sentence-medially and sentence-initially) and fronted wh-elements. However, it is still used quite early in acquisition despite the complexity of the meanings attributed to it and the fact that it requires fronting of the tensed auxiliary and fronting (or base-generation) of clitic negation. We used the Wang CHILDES browser9to search 44 UK and US English corpora for “high” negation. We searched the CHI tier for instances of BE, DO and HAVE auxiliaries with clitic negation10 in children up to age 4, returning over 20,000 hits. We then removed all instances of negative declaratives, utterances with nonovert or indecipherable subjects, song lyrics, and one example of a misattribution of an adult utterance to CHI. We also separated off imitations, wh-questions, negative anchors with positive tags, lone tags without a clear anchor within 5 lines, negative polar questions without inversion, and any unclear structures. This left us with 633 instances of true “high” negation structures from 67 children across 24 corpora including negative tag structures with positive anchors, negative imperatives with overt subjects, and auxiliary-initial structures containing “high” negation. These latter we then tagged for the act that was being performed by the structure: biased polar question (NegQ), negative polar exclamative, or ‘persuasion’ question.11 We did this by hand using features of individual utterances (e.g. the absence of a gradable predicate predicts that the structure is not an exclamative) and up to 5 lines of discourse preceding and following the utterance to judge the utterance in context. We also included an “other” category for imperatives with overt subjects, as these are not used as question acts, in addition to instances where other aspects of syntax left us unsure as to the act 9https://naclo.cs.umass.edu/childes-search/, currently maintained by Christa Bowker. The corpora are pulled from CHILDES (https://childes.talkbank.org/, MacWhinney 2000). To avoid overlong citations in our in-text examples, please see the appendix for a guide to references for the CHILDES corpora cited in this chapter. 10isn’t, wasn’t, aren’t, weren’t, don’t, didn’t, doesn’t, haven’t, hasn’t, hadn’t, ain’t. 11Persuasion questions are polar questions used to exhort the addressee to do something – they are similar in effect to an imperative. Imagine a parent trying to get out of the house who says to their child “Can’t you just put your shoes on already!” This utterance is neither a question requiring a response, nor an exclamative in the typical, surprise at some exceeded degree, sense. They are termed ‘suggestion’ questions by Romero & Han (2004), who also mention them briefly. 336 11 Children’s acquisition of English “high” negation involved, context suggested a different reading, or context didn’t help to differentiate possible readings. Examples of each of the categories are shown below.12 (4) TAG: Got got a small boy haven’t we Mummy Anne (Manchester), 1;11 (5) IMPERATIVE: Don’t you pee pee in the big girl pants. Eve (Brown), 1;11 (6) BIASED Q: There, don’t you see it? Ross (MacWhinney), 2;4 (7) EXCLAMATIVE: Isn’t it sweet. Anne (Manchester), 2;5 (8) PERSUASION: Mommy, don’t you think we could play? Abe (Kuczaj), 3;4 (9) OTHER: Mommy, isn’t this a house or apartment? Abe (Kuczaj), 3;6 We found that tag structures were by far the most common structures containing high negation in the corpus (457 instances). NegQs were next most common (74), followed by imperatives with overt subjects (38) and negative polar exclamatives (36). There were just 3 examples of ‘persuasion’ questions. Tag structures were also used earliest (38/51 utterances before age 2;6), followed by NegQs (7/51) and exclamatives (1/51). The full breakdown of act by age is shown below, in this table adapted from Woods & Roeper (2021: 765): Table 1: High negation questions by age and act TagQ NegQs NegExcl Persuasion Other Total <2;0 7 2 0 0 1 10 2;0–2;5 31 5 1 0 4 41 2;6–2;11 268 25 5 1 31 330 3;0–3;5 94 15 14 1 12 136 3;6–3;11 57 27 16 2 14 116 Total 457 74 36 4 62 633 Emerging in Table 1 is an acquisition path that we aim in the rest of this chapter to capture: tag structures emerge before (and in greater numbers) NegQs, which emerge before negative polar questions and persuasion questions. This holds across children, as illustrated in Table 1, but also within individuals. Table 1 also suggests that children acquiring English have a sophisticated and nuanced understanding of different types of negation that interacts with different relationships 12For access to the resulting database of English “high” negation, please contact the first author. 337 Rebecca Woods & Tom Roeper between interlocutors and the propositional material they are trying to share. This understanding develops and changes over a short space of time. The rest of this chapter aims to make sense of and account for the first steps in the acquisition path in Table 1. We will focus on unpacking the syntax and pragmatics of negative tag structures and NegQs in child production and in adult English. We first take a qualitative look at the earliest NegQs and negative tag structures in the child data (Section 2.2). As mentioned above, we will argue that our acquisition data supports an analysis of negative tag structures whereby the tag is not simply an elided NegQ, principally because negative tag structures precede NegQs in acquisition. We provide an analysis for NegQs that combines insights from Krifka (2015), Goodhue (2022a) and Holmberg (2016) to capture our data and form the basis for a minimal and plausibly acquirable analysis of “high” negation structures (Section 3). 2.2 Qualitative data Existing work on negative tag structures and NegQs is clear that sophisticated discourse management skills are required to use and interpret such structures. Negative tags require the user to recognise conventional uses of particular syntactic structures and model the cognitive state, however shallowly, of their interlocutor (see, e.g. Sadock 1974, Ladd 1981, Asher & Reese 2007, Reese & Asher 2006, Malamud & Stephenson 2015 a.m.o.). NegQs also require the user to recognise marked uses of syntactic conventions, in this case combining negation and polar interrogatives to express a bias that they hold. The data in Table 1, therefore, demonstrate that children around the age of 2 are already sophisticated conversationalists. Some of them are aware of the possibility that their beliefs are not shared by others and they are capable of expressing this through the choice of linguistic structures that they employ. Take as an example the two NegQs used before age 2, which are used after the child’s assumptions are put into doubt by a previous utterance. (10) 02b.cha (Valian), 1;9 MOT: did you play marbles with cousin George? CHI: yeah! [laughs] MOT: that’s funny? CHI: isn’t that funny? (11) Joel (Manchester), 1;11 MOT: tell Caroline what you’re gonna have for your dinner. 338 11 Children’s acquisition of English “high” negation INV: what are you gonna have for your dinner? CHI: don’t you know? In both cases, the child’s interlocutor asks a question that causes them to question some previously held belief; they then ask a NegQ to check whether the propositional content of the belief can still hold. These uses of NegQs chime with adult uses of the same structures. The earliest negative tag structures in our corpus behave a little differently. Negative tags can be used to request confirmation of a proposition that the speaker believes to be true, as in (12). However, they can also be used when the speaker is more certain of the proposition, but wishes to “hedge” (in Ladd’s (1981) terms) or seeks only “acknowledgement” of the proposition by their addressee (in Asher & Reese’s (2007) terms), as in (13). (12) Context: A is fairly sure that B wants to go to the park, but they’re slow to put their shoes on at the door. A: You want to go to the park, don’t you? (13) Context: A and B have discussed going out after lunch. At 12.30pm, B is by the front door, shoes on, with bucket and spade in hand. A: You want to go the park, don’t you. In the earliest part of our dataset, acknowledgement-type uses appear to be more common.13 In (14), a child of 1;11 appears to be looking simply to gain her mother’s attention using a negative tag structure, in the middle of a period of monologuing (numbers in brackets represent pauses in seconds). In this instance, during a period of toy play, it is thought that knowledge about the smallness of the boy is shared knowledge: (14) Debbie (Wells), 1;11 CHI: Gotto pick it up. Throw it out. Pick it up. Throw it out. xxx (14). Got a boy. Got a got a small boy, haven’t we Mummy? We’ve got a big girl (2). xxx get a big girl. Look Mum I’m nearly getting big. MOT: You are getting big, mm. 13We identify the uses of tag structures in our dataset on the basis of context alone; none of the corpora below have audio files attached, so prosodic information is not available to us. We do not infer prosodic information from punctuation in the transcript (i.e. the use of turn-final . vs ?to represent intonation) because this is not consistent across corpora. In any case, it is unclear what the prosody of tag structures is when they are produced by children, as this has not yet been studied. 339 Rebecca Woods & Tom Roeper Table 2: Canonical and non-canonical uses of clause types in English Clause type Canonical use Non-canonical use Declarative Assertion Question (e.g. rising declaratives); Request (with modal auxiliaries); Exclamative (with intonation); Command (with attitude verbs e.g. want) Interrogative Question Request (e.g. with please); Exclamative (e.g. negative polar exclamatives); Assertion (e.g. fuck-inversion; Sailor 2020) Imperative Command Request (e.g. with please) 2023) to the interpretation of embedded clauses. Moreover, embedded typed clauses are easily dissociable from their canonical speech acts. For example, interrogative clauses under response verbs like know are not interpreted as open questions, but rather as something like the answers to that question (e.g. Lahiri 2002, Uegaki 2015). Returning to the role of clause types in negative tag structures, the conjunction of different clause types in a negative tag structure means that it points both to a set of possible worlds and a set of sets of possible worlds, where the former is a subset of the latter. Though this seems to be redundant, it is important to note there are (at least) two perspectives at play when a negative tag structure is used. Let us start with an assert tag structure. If a proposition is simultaneously true and maybe true for the same person, the questioning of the proposition seems redundant, or even contradictory. However, we noted in Section 2.2 and Section 3.1.2 that assert tag structures tend to be used when the speaker recognises that the addressee might already know the proposition to be true. This runs against the core felicity condition of an assertion: that the speaker believes that the addressee does not already believe that some proposition is true (Searle 1969, Farkas 2022). However, that proposition might not have been accepted publicly as true by the addressee, whether because they have previously refused to accept it or it has not been addressed directly by the addressee.16 In either case, assert 16An interesting case of such uses of tag questions with falling intonation is in the case of predicates of personal taste. For an example, imagine that a co-worker mentions that she has a new neighbour, and blushes while saying so. You might say: He’s attractive, isn’t he? to determine the cause of the blushing (see Malamud & Stephenson 2015). In this case the speaker 346 11 Children’s acquisition of English “high” negation tag structures are used when the issue of the proposition is not publicly settled for the addressee. Therefore, if the anchor reflects the speaker’s commitments, but the tag reflects the speaker’s perception that the addressee lacks public commitment to the proposition in this discourse context, there is no longer any redundancy in the structure or its use. We can consider the negative tag structure to carry a conventional implicature (in the sense of Potts 2005), in that the speaker chooses to assert their proposition using a negative tag structure because they also want to communicate how they view the addressee’s commitments. We therefore update (18) to reflect this by adding a speaker operator into the PerspectiveP of the declarative clause and an addressee operator into the PerspectiveP of the interrogative clause to represent their differing worlds of evaluation.17 (26) assert tag structure: second pass a. Lucy is coming, isn’t she.↘ b. [SAP [[PerspP speaker [CP decl [IP Lucy is coming ]]][⋀][[PerspP addressee [CP q isn’t [IP she tisn’t coming ]]]]↘] What about question tag structures? Here we claim again that the two parts of the tag structure relativise to different interlocutors. The felicity conditions of canonical questions include speaker ignorance and addressee competence – i.e. that the speaker doesn’t know the answer and the addressee does – as well as addressee compliance – i.e. that the addressee will provide the true answer (Farkas 2022). However, there are a number of non-canonical questions in which speaker ignorance is weakened because the speaker may already have evidence for the true answer, e.g. rising declaratives (Gunlogson 2003). Given that this type of non-canonical question operationalises a steep rise in intonation, which we also see in question tag structures (Dehé & Braun 2013: 140), it is possible that the same process is active here: the speaker reflects their perception of addressee competence in the anchor18 and their need for addressee compliance (due to their is making a guess as to the addressee’s point of view, but it is important that it is presented as an assertion by the speaker for the addressee to publicly weigh in on. Such cases require more independent work – thanks to Dan Goodhue for bringing them to our attention. 17We are agnostic as to the exact mechanisms for how the world of evaluation interacts with material in the CP. There are many approaches that are compatible with our proposal, including Tsoulas & Kural (1999), Speas & Tenny (2003) and Schwarz (2012), i.a. Approaches such as Sigurðsson (2014) are not compatible with our proposal as in that case, the speaker/addressee features are lower in the clausal hierarchy than e.g. the clause-typing head. 18This analysis is actually quite similar to Malamud & Stephenson’s (2015) approach to rising tags whose polarity matches that of the anchor, as they claim that such tags place the commit347 Rebecca Woods & Tom Roeper own uncertainty) in the tag. Therefore, (19) can be updated for question tag structures as follows: (27) question tag structure: second pass a. Lucy is coming, isn’t she?↗ b. [SAP [[PerspP addressee [CP decl [IP Lucy is coming ]]][⋀][[PerspP speaker [CP q isn’t [IP she tisn’t coming ]]]]↗] A difference to note between the proposal here and rising declaratives is that, in the case of rising declaratives, the addressee is publicly known to know the proposition (Gunlogson 2003: 84–85). This needn’t necessarily hold in the case of a tag structure (see also Hepburn & Potter 2011), hence the interrogative tag formally requests that the addressee make public their commitment to the proposition. These amendments to the proposal may lead a reader to question whether high level speech act operators are necessary when there are also perspectival operators which appear to be doing the same job – in other words, why can’t the difference between assert and question tag structures boil down to the speaker/addressee operators and their relationship to the different types of clause? One reason that speaker/addressee operators and their interaction with clause types cannot alone explain the different structures is intonation. There is evidence from languages like Korean that intonation expresses that a particular response pattern is expected – i.e. that a specific speech act is being performed (Ceong 2019). Korean also represents interlocutor perspective and clause typing separately through specific verbal morphology, as shown in (28–29), in which the arrows represent final falling or rising intonation (examples from Ceong 2017: 12–13). (28) Meysi-lul Messi-acc mannameetsspasttadeclkocomp- ↘ speaker-commitment ‘I said I met Messi.’ Reinforcing assertion ment for the proposition in the addressee’s future commitments and not those of the speaker. However, they also mitigate the speaker’s commitment to the proposition in their account of mismatching tags, by suggesting that the speaker only commits provisionally to the proposition, on the proviso that the addressee confirms its truth. Moreover, they do not address how the structure of the tag and anchor result in these discourse effects. As we do not address matching tags in this chapter, and Malamud & Stephenson (2015) restrict themselves to rising tags, more work is to be done on how intonation contour and polarity interact. 348 11 Children’s acquisition of English “high” negation (29) Meysi-lul Messi-acc mannameetsspasttadeclkocomp- ↗ addressee-commitment ‘Are you saying you met Messi?’ Echo question This justifies the inclusion of specific speech act operators that are separate from representations of perspective. Moreover, intonation does not appear to accompany perspectival shifts in the same way. To our knowledge, shifting phenomena such as monstrous indexicals (e.g. in Uyghur, Shklovsky & Sudo 2014) do not trigger specific intonational contours. Additionally, we might expect a completely different intonational contour for assert tag structures than we actually see if speaker/addressee operators were the locus of intonation, given that (we argue) they contain an addressee operator scoping over an interrogative clause, just like a typical information-seeking question might. Another reason for including speaker/addressee operators as well as speech act operators is that we believe we have evidence that these are just two examples from a large range of operators, including overt items, can be hosted in PerspectiveP in different constructions, and that these other operators are compatible with the intonational contours introduced by assert/question speech act operators. We present this evidence now, as we turn to discuss NegQs in more detail. 3.2 Metalinguistic negation Recall that we presented our NegQ structure in (25), repeated here: (25) a. Isn’t Lucy coming? b. [SAP question [PerspectiveP Is+n’t [CP q tis [IP Lucy tis coming ]]]] In (25), the clitic negation n’t is above the level of the proposition. Evidence for this stems back to Ladd’s (1981) observation that NPIs are not licensed in NegQs. (30) Isn’t Jane coming {too/#either}? Many linguists before us have proposed that some negation is metalinguistic as it scopes over some object that is bigger than the proposition alone (Horn 1985, 1989; Wood 2014; Holmberg 2016, a.o.). We join this tradition in claiming that negation in PerspectiveP negates, from the speaker’s perspective, that the typical interpretation of a typed clause holds. This is similar to Krifka’s (2015; see also Cohen & Krifka 2014) concept of denegation of speech acts, whereby negation over a speech act “prunes [the]legal developments” (Krifka 2015: 330) of some speech act in a discourse; in other words, it prevents certain discourse 349 Rebecca Woods & Tom Roeper continuations that would usually stem from canonical use of some clause type from being licit. How does this fall out? We propose that negation deployed in PerspectiveP indicates that the speaker rejects the interrogativity of the CP – in other words, that they do not believe that alternatives to the proposition are true. This is an indirect, weak method of expressing belief in the truth of the proposition, hence it is not at odds with the question force and corresponding felicity conditions of the NegQ (the speaker’s commitment to resolving the issue and their expectation that the addressee will provide the information required). We paraphrase (25) as below: (31) I the speaker am committed to resolving the issue [QUESTION]of my not believing that there is a plausible alternative proposition to 𝑝 [NEG-in-PerspectiveP]in the set 𝑝or not 𝑝[CP Q] This means that biased meanings in tag structures cannot fall out from the same mechanism as NegQs because there is no metalinguistic negation in negative tag structures and therefore no rejection of interrogativity in question tag structures. Bias in question tag structures is generated by the interaction of the addressee operator with the proposition: the speaker expresses that they perceive the addressee as knowing a proposition and as asking them to publicly express this commitment. In assert tag structures, bias is created by asserting the proposition in the anchor. As such, the bias in negative tag structures is not homogenous, and in a question tag structure it is indirect – it is achieved by the speaker looking to resolve the issue of whether they’re right to perceive the addressee as believing the proposition to be true. This could be the root of ‘deferential’ readings of some tag structures (as in Lakoff 1975), as the speaker appears to be adopting the addressee’s perspective in uttering the declarative clause, but this is an example of discourse inference that we expect to be outside the core, syntactically articulated speech act mechanisms we are presenting here. If we are right so far, we make predictions for both adult and child English in terms of response patterns to NegQs and negative tags. Our approach predicts that adult NegQs should have a response pattern like, but not identical to, canonical questions and unlike assertions. This is because the speaker does not directly assert belief in the positive proposition; this belief is only implied by the speaker’s rejection of the interrogativity of CP. assert tag structures will differ from NegQs as acceptable responses to assert tag structures should be almost identical to that of a canonical assertion, given that they contain a declarative clause under an assert operator. question tag structures, finally, are predicted 350 11 Children’s acquisition of English “high” negation to provoke response patterns like those of canonical questions, even though they are not neutral questions, because they contain an interrogative clause under a question operator. We return to this prediction in Section 3.3.2 and Section 3.3.3. Our approach also predicts that negative tag structures should be acquired earlier and more accurately than NegQs by English-acquiring children. This is because the negation in tag structures is propositional, therefore lower in the syntactic tree, than the metalinguistic negation of NegQs (following logic first propounded by Rizzi 1993/1994 in his Truncation Hypothesis and updated in Friedmann et al. 2021). Metalinguistic negation also scopes over a more complex object (a CP) and requires more complex computation. It is also likely to be harder to acquire because it is realised in the same way phonologically as cliticised propositional negation, so the child must create two categories for the same phonological realisation. We return to this prediction in Section 3.3.1. 3.3 Strengths of our account We turn now to showing how our proposal in Section 3.1 accounts for both the child and adult data. We will show that our proposal is compatible with aspects of Asher & Reese’s (2007) discourse-driven account, and many aspects of Holmberg’s (2016) syntactic approach, at least as far as question tag structures are concerned. Section 3.3.1 demonstrates that children are not as target-like in their production of NegQs as in their production of negative tags, with respect to how clitic negation is realised and how they encode bias outside of tag structures. Section 3.3.2 shows that negative tag structures differ from NegQs in terms of their response patterns, and that differences obtain between the two types of negative tag structure, further supporting our claim that neither is derived from NegQs. Section 3.3.3 demonstrates how these response patterns fall out from the structures of the different constructions, using Farkas’s (2022) update of Farkas & Bruce’s (2010) Table model. 3.3.1 Child “high” negation structures – are they target-like? We have already seen in Section 2 that negative tag structures are used early and with adult-like syntax. Section 2.1 showed that NegQs are produced later, but we will now show that they display more evidence of non-target-like syntax. In 306 negative tag structures up to and excluding age 3, there are only 8 errors concerning using the correct auxiliary and two where a full DP subject is used (we do not report here on tense and agreement errors, as we consider these orthogonal). However, in 32 NegQs up to and excluding age 3, 10 of them contain 351 Rebecca Woods & Tom Roeper errors concerning the auxiliary (again, excluding errors of tense and agreement), which hosts negation. In fact, they are all errors of auxiliary doubling like in (32): (32) Do they don’t eat people up? Nina (Suppes), 2;9 Some of these examples are also plausibly different from NegQs as the bias appears to be towards the negative proposition; an adult-like paraphrase of (32) might be Do they not eat people up? In this case, we have even fewer examples of target-like NegQs in child speech despite clear evidence that they are able to conceive of, and try to express, biased meanings in questions with bias towards a positive proposition. For more on auxiliary doubling errors, see Woods & Roeper (2020), who connect this type of error directly to early attempts by children to express biased meanings, where the child is biased towards the negative proposition. Given the above, our proposal already accounts for the child data more effectively than approaches in which the tag part of a negative tag structure is an elided NegQ, however that is construed. 3.3.2 Response patterns It has been established since Sadock (1971, 1974) that both negative tag structures and NegQs show evidence of questionhood (e.g. using the Tell me… test) and assertionhood (e.g. the After all… test; see also Asher & Reese 2007). However, as Asher & Reese’s (2007) work suggests, different types of negative tag structures, and NegQs, invite slightly different response patterns. Our proposal also predicts different response patterns and some work on this by Holmberg (2016) already exists, so we devote this section to response patterns. Let’s lay out the data. We take our three structures of interest and compare them with canonical assertions and canonical polar questions, as well as a modalised assertion, given that some proposals for tag questions postulate a modal operator in the anchor (e.g. Bill & Koev 2025 [this volume]). We will use the basic proposition Bilal is coming. We also take the following possible responses: polarity particles with matching elided propositions (Yes he is; no he isn’t) to model polarity-based answers, polarity particles with non-matching elided propositions to model truth-based answers (Yes he isn’t)19, “agreement indicators” right and so he is (as dubbed and investigated by Holmberg 2016) and silent acceptance, represented by [silence]. Note that we have used Holmberg’s (2016) judgments, 19See Holmberg (2016: Chapter 4.1), Jones (1999), and references therein for more on truth-based or polarity-based answering systems. 352 11 Children’s acquisition of English “high” negation which we agree with, except in the modalised assertion condition (which he does not discuss). The data are as follows: (33) assert tag A: Bilal is coming, isn’t he.↘ B: Yes (he is); #Yes (he isn’t); No (he isn’t); Right; So he is; [silence]. (34) question tag A: Bilal is coming, isn’t he?↗ B: Yes (he is); #Yes (he isn’t); No (he isn’t); Right; So he is; #[silence]. (35) NegQ A: Isn’t Bilal coming? B: Yes (he is); %20 Yes (he isn’t); No (he isn’t), ??Right; #So he is; #[silence]. (36) Assertion A: Bilal is coming. B: Yes (he is); #Yes (he isn’t); No (he isn’t); Right; So he is; [silence]. (37) Modal assertion A: Bilal might be coming. B: Yes #(he is); #Yes (he isn’t)21; No (he isn’t); Right; #So he is; [silence]. (38) Polar question A: Is Bilal coming? B: Yes (he is); #Yes (he isn’t); No (he isn’t); #Right; #So he is; #[silence]. This paradigm demonstrates that an assert tag structure patterns just like an assertion in terms of licit responses to it. question tag structures differ in that they require a response – silence is not appropriate – but they differ from canonical polar questions by allowing the agreement indicators (see also Holmberg 2016: 183). NegQs pattern most closely with canonical polar questions, with some gradability or dialectal variation with respect to truth-based responses and confirmational right. Why should response patterns differ between negative tag structures and NegQs when existing assertionhood and questionhood tests suggest that they are 20Based on author intuitions, this is fine in UK English but not in US English. 21While He isn’t is fine as a response to a modalised assertion, it is the use of yes here that leads to infelicity. 353 Rebecca Woods & Tom Roeper the same? We think that existing tests are not granular enough when it comes to understanding the nuances around bias and interlocutor perspective in these two structures: in short, speaker belief that the proposition may be true is not the end of the story. Recall that we spelled out the difference between speaker belief and speaker commitment in our motivation of an extended left periphery in Section 3.1. We have devised a test to demonstrate this distinction with respect to tag structures and NegQs which we call the Don’t you agree…? test. Using the English verb agree, it is possible either to agree with an asserted proposition or with the person who expresses that proposition. These are expressed using different pronouns to point to these different referents. We use agreement with the asserted proposition (that) to reflect matching commitments to the proposition in the shared discourse context and agreement with the individual (me) to reflect matching perspectives. What we find is that it is possible to ask whether an addressee agrees with the proposition, but not the perspective (i.e. the speaker), following a NegQ (39) or a question tag structure (40), but with either following an assert tag structure (41): (39) Isn’t Jane a good choice? [silence from addressee]Don’t you agree with that/#me? (40) Jane’s a good choice, isn’t she?↗[silence from addressee]Don’t you agree with that/#me? (41) Jane’s a good choice, isn’t she.↘[silence from addressee]Don’t you agree with that/me? Moreover, Don’t you agree… is completely incompatible with canonical polar questions and entirely natural with canonical assertions. (42) Is Jane a good choice? #Don’t you agree with that/me? (43) Jane is a good choice. Don’t you agree with that/me? In these examples the pie is cut slightly differently again, but along lines predicted by our analysis: our structures of interest with question operators pattern together, and the structures with assert operators pattern differently. In short, NegQs and question tag structures do not offer up a proposition asserted by the speaker to agree with, despite having some assertion-like properties (see Sadock 1971, 1974, Asher & Reese 2007, i.a.). This suggests that assert tag structures differ from NegQs in that they express speaker commitment in addition to foregrounding some proposition. This chimes 354 11 Children’s acquisition of English “high” negation with recent work by Ceong (2019), Krifka (2014, 2023), Wiltschko & Heim (2016, 2021), Woods (2021) and Woods & Vicente (2021) that commitment is a part of grammar over and above (doxastic) belief. It also highlights that longstanding diagnostics for assertion can obscure nuanced differences between different types of non-canonical speech act. One last new diagnostic. We borrow and extend a test that Asher & Reese (2007: 10) apply to NegQs to our three structures of interest – the prior beliefs test. Note that boldface is used to indicate stress on the auxiliary verbs in T and we include arrows here to indicate intonation contours. (44) I have no prior beliefs on the matter. I just want to know that Lucy is coming, isn’t she?↗ (45) I have no prior beliefs on the matter. I just want to know #isn’t Lucy coming? (46) I have no prior beliefs on the matter. I just want to know that #Lucy is coming, isn’t she.↘ Another distinction emerges – where a question tag structure is compatible with the speaker claiming no prior beliefs, NegQs and assert tag structures are not. (44) in particular supports our claims about question tag structures that the “assertion flavour” of them is derived and not directly attributable to the speaker. We represent the different profiles of our three structures of interest in Table 3, where we mark whether the different types of “high” negation structure pattern with canonical assertions, canonical questions, or neither, in our three diagnostics. While assert tag structures are predominantly assertion-like, and question tag structures/NegQs are predominantly question-like, the latter two diverge in different ways and all structures also diverge from both questions and assertions too. All of these nuances are captured by our proposal. Table 3: Negative tag and NegQ profiles assert tag question tag NegQ Response patterns Assertions Neither Questions Don’t you agree… Assertions Questions Questions Prior beliefs Neither Questions Neither 355 Rebecca Woods & Tom Roeper The entry in the speaker’s discourse commitments in Table 10 means that the speaker commits to there being no plausible alternative to pgiven their knowledge and beliefs. This is not logically equivalent to asserting pbecause the truth of pcould be left undefined, therefore it is only implied that the speaker must, therefore, believe pto be true. In Horn-style terms, the speaker is registering their objections to accepting ¬pin the face of some possible evidence for it. This follows traditional accounts of metalinguistic negation that claim that “rectification or correction is a necessary part of the interpretation of […]metalinguistic negation” (Kay 2004: 689, discussing Horn 1985). However, because NegQs fold metalinguistic negation into a question structure, the burden of rectifying or correcting falls on the addressee rather than on the user of metalinguistic negation themselves. Because the speaker does not formally commit to p, the inclusion of DC𝐴𝑑∪ {¬𝑝}as a canonical response to the NegQ is formally licit and does not logically contradict the speaker’s public commitments. If we are correct, the analyses above demonstrate more precisely why negative tag structures are acquired before NegQs. Negative tag structures are transparent in that their at-issue content directly follows from their surface structure. NegQs, on the other hand, are not transparent as their at-issue content is “less than” the phonologically expressed material. To put this another way, metalinguistic negation is phonologically expressed in the middle of (indeed, it is cliticised to) propositional, at-issue material, but is not itself propositional or at-issue. It is therefore a complex task for the child to separate out at-issue and not-at-issue expressions that are phonologically tightly bound together. 4 Summary In this chapter we focused on the production of English nuclear negative tag structures and NegQs by very young children and adults, as well as their contribution to discourse. Our empirical contributions are as follows. We created a dataset containing over 600 utterances of “high” negation structures by 67 English-acquiring children, demonstrating that negative tag structures precede and outnumber NegQs. Qualitatively, the dataset shows that children use negative tag structures accurately both in terms of adult-like syntax and discourse contribution. We also discussed a particular non-target-like NegQ that uses auxiliary doubling, noting that target-like NegQs with positive bias are even more rare than the dataset suggests. 362 11 Children’s acquisition of English “high” negation We also demonstrated using adult judgements that negative tag structures in English divide into two types, both of which are also distinct from NegQs, in terms of interpretation and in terms of response pattern. We propose two new diagnostics, the don’t you agree? test and the prior beliefs test that further refine our understanding of speaker commitment and belief in the deployment of negative tag structures and NegQs. Our empirical findings feed our theoretical claims. Contra much of the existing literature (see Holmberg 2016 and Krifka 2015 for exceptions), we argue that English nuclear negative tag structures are simple speech acts that are complex at the clausal level. They do not consist of an assertion combined with a NegQ – indeed, to speak of complex speech acts creates problems further down the theoretical line in terms of predicting and understanding responses to such acts. Negative tag structures are a declarative clause conjoined with an interrogative clause containing propositional negation, and this whole is interpreted in one of two ways depending on (a) the perspective attributed to each clause (speaker or addressee) and (b) the speech act operator that scopes over the whole (assert or question). In contrast, English NegQs consist of an interrogative clause scoped over by metalinguistic negation and a question speech act operator. We demonstrated how clauses, perspectives and speech act operators interact using Farkas’s (2022) version of Farkas & Bruce’s (2010) Table model. Bias is created in nuclear negative tag questions via the relationship of speech act operator, perspective and proposition in the anchor, whereas in NegQs it arises from the metalinguistic rejection of the interrogativity of the CP – in other words, the speaker expresses that they believe there to be no plausible alternative to the main proposition. A number of areas for future study remain, most pertinently the prosody of tag structures in child speech and in their input. We hope, however, that the breadth of the predictions made by the proposals here, and the empirical evidence that we have been able to offer in this chapter, provide support for the enterprise of trying to understand how children acquire different types of speech acts and will invite energetic debate. Abbreviations acc accusative Ad addressee comp complementiser DC discourse commitments decl declarative NegQ interrogative Q “high” negation question PerspP Perspective Phrase Pol(P) Polarity (Phrase) ps projected set SA(P) Speech Act (Phrase) Sp speaker 363 Rebecca Woods & Tom Roeper Acknowledgements We would like to thank: the participants at Leibniz-ZAS’s Biased Question workshop in February 2021 (not least Robin Connelly for his lack of participation at 5 days old); the participants of the North East Syntax Seminar (May 2022, Newcastle), especially Anders Holmberg and George Tsoulas; colleagues in the online meetings of the UMass-based Language Acquisition Research Center; two anonymous reviewers for their careful and helpful comments; Cory Bill, Dan Goodhue, Johannes Heim and Martina Wiltschko for informal discussion; and Tue Trinh and Kazuko Yatsushiro for their (very patient!) editorship. Guide to CHILDES corpora references and DOIs, in order of first mention Manchester corpus: Theakston et al. (2001). DOI: https://doi.org/10.21415/T54G6D Valian corpus: Valian (1991). DOI: https://doi.org/10.21415/T5ZS3T MacWhinney corpus: MacWhinney (1991). DOI: https://doi.org/10.21415/T5JP4F Gleason corpus: Masur & Gleason (1980). DOI: https://doi.org/10.21415/T5101R Belfast corpus: Henry (1995). DOI: https://doi.org/10.21415/T5VG79 Brown corpus: Brown (1973). DOI: https://doi.org/10.21415/T5HK5G Kuczaj corpus: Kuczaj (1977). DOI: https://doi.org/10.21415/T5H30R Higginson corpus: Higginson (1985). DOI: https://doi.org/10.21415/T5S31M Suppes corpus: Suppes (1974). DOI: https://doi.org/10.21415/T5WS4K Tardif corpus: DOI: https://doi.org/10.21415/T5TK58 Wells corpus: Wells (1981). DOI: https://doi.org/10.21415/T5T60K References Asher, Nicholas & Brian Reese. 2007. Intonation and discourse: Biased questions. In Shinichiro Ishihara, Stefanie Jannedy & Anne Schwarz (eds.), Interdisciplinary Studies on Information Structure (ISIS) 8 (Working papers of the SFB 632), 1–38. Potsdam: Universitätsverlag. Bill, Cory & Todor Koev. 2025. Bias in tag questions. In Tue Trinh, Anton Benz, Daniel Goodhue, Kazuko Yatsushiro & Manfred Krifka (eds.), Biased questions, 145–178. Berlin: Language Science Press. Brown, Roger. 1973. A first language: The early stages. Cambridge: Harvard University Press. 364 11 Children’s acquisition of English “high” negation Ceong, Hailey Hyekyeong. 2019. The morphosyntax of clause typing: Single, double, periphrastic, and multifunctional complementizers in Korean. Victoria: University of Victoria. (Doctoral dissertation). Ceong, Hyekyeong Hailey. 2017. Firstand second-hand speech contrasts as a syntactic domain: The grounding spine. Slides from a talk presented at the Workshop at the 12th Biannual Meeting of the Association of Typology at the Australian National University in Canberra, December 15 2017. Cohen, Ariel & Manfred Krifka. 2014. Superlative quantifiers and meta-speech acts. Linguistics and Philosophy 37. 41–90. Coniglio, Marco & Iulia Zegrean. 2012. Splitting up force: Evidence from discourse particles. In Liliane Haegeman Lobke Aelbrecht & Rachel Nye (eds.), Main clause phenomena: New horizons, 229–256. Amsterdam: John Benjamins. Dehé, Nicole & Bettina Braun. 2013. The prosody of question tags in English. English Language and Linguistics 17(1). 129–156. DOI: 10.1017/S1360674312000342. Farkas, Donka F. 2022. Non-intrusive questions as a special type of non-canonical questions. Journal of Semantics 39(2). 295–337. DOI: 10.1093/jos/ffac001. Farkas, Donka F. & Kim Bruce. 2010. On reacting to assertions and polar questions. Journal of Semantics 27. 81–118. DOI: 10.1093/jos/ffp010. Friedmann, Naama, Adriana Belletti & Luigi Rizzi. 2021. Growing trees: The acquisition of the left periphery. Glossa: A Journal of General Linguistics 6(1). 1– 38. DOI: 10.16995/glossa.5877. Geurts, Bart. 2019. Communication as commitment sharing: Speech acts, implicatures, common ground. Theoretical Linguistics 45(1-2). 1–30. DOI: 10.1515/tl2019-0001. Goodhue, Daniel. 2022a. Bias in high negation questions as a quantity implicature in commitment space semantics. Slides presented at the 23rd Amsterdam Colloquium, 19th December 2022. Retrieved 12th April 2023 from https://danielgoodhue.com/s/ac.pdf. Goodhue, Daniel. 2022b. Isn’t there more than one way to bias a polar question? Natural Language Semantics 30(4). 379–413. DOI: 10.1007/s11050-022-09198-2. Gu, Chloe Chenjie & Tom Roeper. 2011. Acquisition arguments for a General Point of View as an alternative to relativism and subjectivity. Ms., University of Massachusetts Amherst. Guasti, Maria-Teresa, Rosalind Thornton & Kenneth Wexler. 1995. Negation in children’s questions: The case of English. In 19th Annual Boston University Conference on Language Development, 228–239. Somerville: Cascadilla Press. Gunlogson, Christine. 2003. True to form: Rising and falling declaratives as questions in English. New York: Routledge. 365 Rebecca Woods & Tom Roeper Gunlogson, Christine. 2008. A question of commitment. Belgian Journal of Linguistics 22(1). 101–136. DOI: 10.1075/bjl.22.06gun. Heim, Johannes. 2019. Commitment and engagement: The role of intonation in deriving speech acts. University of British Columbia. (Doctoral dissertation). DOI: 10.14288/1.0380772. Henry, Alison. 1995. Belfast English and Standard English: Dialect variation and parameter setting. New York: Oxford University Press. Hepburn, Alexa & Jonathan Potter. 2011. Recipients designed: Tag structures and gender. In Susan A. Speer & Elizabeth Stokoe (eds.), Conversation and gender, 135–152. Cambridge: Cambridge University Press. Higginson, Roy P. 1985. Fixing-assimilation in language acquisition. Washington State University. (Doctoral dissertation). Hill, Virginia. 2013. Vocatives: How syntax meets with pragmatics. Leiden: Brill. Holmberg, Anders. 2016. The syntax of yes and no. Oxford: Oxford University Press. Horn, Laurence R. 1985. Metalinguistic negation and pragmatic ambiguity. Language 61. 121–174. Horn, Laurence R. 1989. A natural history of negation. Chicago: University of Chicago Press. Jamieson, E. 2018. Experimental evidence for the contextual acceptability of biased questions and tag questions. In ConSOLE 26: : Proceedings of the 26th conference of the student organization of linguistics in Europe, 331–349. Leiden: Leiden University Centre for Linguistics. Jones, Bob Morris. 1999. The Welsh answering system. Berlin: Mouton de Gruyter. Kay, Paul. 2004. Pragmatic aspects of grammatical constructions. In Laurence R. Horn & Gregory Ward (eds.), The handbook of pragmatics, 675–700. Malden: Blackwell. Krifka, Manfred. 2014. Embedding illocutionary acts. In Tom Roeper & Margaret Speas (eds.), Recursion: Complexity in cognition, 59–88. Heidelberg: Springer. Krifka, Manfred. 2015. Bias in commitment space semantics: Declarative questions, negated questions, and question tags. In Sarah D’Antonio, Mary Moroney & Carol-Rose Little (eds.), Semantics and Linguistic Theory (SALT) 25, 328–345. Ithaca: CLC Publications. Krifka, Manfred. 2023. Layers of assertive clauses: Propositions, judgements, commitments, acts. In Jutta M. Hartmann & Angelika Wollstein (eds.), Propositionale Argumente im Sprachvergleich: Theorie und Empirie, 115–181. Tübingen: Narr. DOI: 10.24053/9783823394105. Kuczaj, Stan. 1977. The acquisition of regular and irregular past tense forms. Journal of Verbal Learning and Verbal Behavior 16. 589–600. 366 11 Children’s acquisition of English “high” negation Ladd, D. Robert. 1981. A first look at the semantics and pragmatics of negative questions and tag questions. In Chicago Linguistic Society (CLS) 17, 164–171. Lahiri, Utpal. 2002. Questions and answers in embedded contexts. Oxford: Oxford University Press. Lakoff, Robin. 1975. Language and woman’s place. New York: Harper & Row. MacWhinney, Brian. 1991. The CHILDES Project: Tools for analyzing talk. Hillsdale: Erlbaum. MacWhinney, Brian. 2000. The CHILDES Project: Tools for analyzing talk. 3rd edn. Mahwah: Lawrence Erlbaum Associates. Malamud, Sophia A. & Tamina Stephenson. 2015. Three ways to avoid commitments: Declarative force modifiers in the conversational scoreboard. Journal of Semantics 32(2). 275–311. DOI: 10.1093/jos/ffu002. Masur, Elise Frank & Jean Berko Gleason. 1980. Parent-child interaction and the acquisition of lexical information during play. Developmental Psychology 16. 404–409. O’Connor, J. D. 1955. The intonation of tag structures in English. English Studies 36(1-6). 97–105. DOI: 10.1080/00138385508596942. Potts, Christopher. 2005. The logic of conventional implicatures. Oxford: Oxford University Press. Reese, Brian & Nicholas Asher. 2006. Prosody and the interpretation of tag questions. In Sinn und Bedeutung 11, 448–462. Barcelona: Universitat Pompeu Fabra. Rizzi, Luigi. 1993/1994. Some notes on linguistic theory and language development: The case of root infinitives. Language Acquisition 3(4). 371–393. Roeper, Tom. 2016. Propositions and implicit arguments carry a default general point of view. In Cécile Meier & Janneke van Wijnbergen-Huitink (eds.), Subjective Meaning: Alternatives to Relativism, 201–226. Berlin: De Gruyter. DOI: 10.1515/9783110402001-010. Romero, Maribel & Chung-Hye Han. 2004. On negative yes/no questions. Linguistics and Philosophy 27(5). 609–658. Sadock, Jerrold M. 1971. Queclaratives. Papers from the Regional Meeting of the Chicago Linguistic Society (CLS) 7. 223–231. Sadock, Jerrold M. 1974. Towards a linguistic theory of speech acts. New York: Academic Press. Sailor, Craig. 2012. Remarks on replies: On emphatic polarity in English. Talk presented at the Linguistic Society of America 2012 Annual Meeting. Sailor, Craig. 2020. Rethinking “residual” Verb Second. In Rebecca Woods & Sam Wolfe (eds.), Rethinking verb second, 126–149. Oxford: Oxford University Press. DOI: 10.1093/oso/9780198844303.003.0006. 367 Rebecca Woods & Tom Roeper Schmitz, Michael. 2021. Questions are higher-level acts. Academia Letters. Article 3239. DOI: 10.20935/AL3239. Schwarz, Florian. 2012. Situation pronouns as arguments of determiners. Natural Language Semantics 20(4). 431–475. Searle, John R. 1969. Speech acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. DOI: 10.1017/CBO9781139173438. Shklovsky, Kirill & Yasutada Sudo. 2014. The syntax of monsters. Linguistic Inquiry 45(3). DOI: 10.1162/LING_a_00160. Sigurðsson, Halldór Ármann. 2014. Context-linked grammar. Language Sciences 46. 175–188. Speas, Peggy & Carol Tenny. 2003. Configurational properties of point of view roles. In Anna Maria DiSciullo (ed.), Asymmetry of grammar, vol. 1: Syntax and semantics, 315–344. Amsterdam: John Benjamins. Stalnaker, Robert. 1978. Assertion. In Peter Cole (ed.), Syntax and semantics, vol. 9, 315–332. New York: Academic Press. Stalnaker, Robert. 2002. Common ground. Linguistics and Philosophy 25. 701–721. Suppes, Patrick. 1974. The semantics of children’s language. American Psychologist 29. 103–114. Theakston, Anna L., Elena V. M. Lieven, Julian M. Pine & Caroline F. Rowland. 2001. The role of performance limitations in the acquisition of verb-argument structure: An alternative account. Journal of Child Language 28. 127–152. Tsoulas, George & Murat Kural. 1999. Indexical pronouns as bound variables. In Jason D. Haugen Sonya Bird Andrew Carnie & Peter Norquest (eds.), West Coast Conference on Formal Linguistics (WCCFL) 18, 545–557. Uegaki, Wataru. 2015. Interpreting questions under attitudes. MIT. (Doctoral dissertation). Valian, Virginia. 1991. Syntactic subjects in the early speech of American and Italian children. Cognition 40. 21–81. Wells, C. Gordon. 1981. Learning through interaction: The study of language development. Cambridge: Cambridge University Press. Wiltschko, Martina. 2021. The grammar of interactional language. Cambridge: Cambridge University Press. Wiltschko, Martina, Derek Denis & Alexandra D’Arcy. 2018. Deconstructing variation in pragmatic function: A transdisciplinary case study. Language in Society 47. 569–599. DOI: 10.1017/S004740451800057X. Wiltschko, Martina & Johannes Heim. 2016. The syntax of confirmationals: A neo-performative analysis. In Evelien Keizer Günther Kaltenböck & Arne Lohmann (eds.), Outside the clause: Form and function of extra-clausal constituents, 303–340. Amsterdam: John Benjamins. 368 11 Children’s acquisition of English “high” negation Wiltschko, Martina & Johannes Heim. 2021. Acquiring the form and function of interactional language. Paper presented at the Linguistics Association of Great Britain Annual Meeting 2021, 6–9 September 2021, held online. Winefield, Helen R., Margaret A. Chandler & Darryl L. Bassett. 1989. Tag structures and powerfulness: Quantitative and qualitative analyses of a course of psychotherapy. Language and Society 18. 77–86. Wood, Jim. 2014. Affirmative semantics with negative morphosyntax: Negative exclamatives and the New England So AUXn’t NP/DP construction. In Raffaella Zanuttini & Laurence R. Horn (eds.), Micro-syntactic variation in North American English, 71–114. Oxford: Oxford University Press. Woods, Rebecca. 2021. Towards a model of the syntax-discourse interface: A syntactic analysis of please.English Language and Linguistics 25(1). 121–153. DOI: 10.1017/S1360674319000510. Woods, Rebecca & Liliane Haegeman. 2023. Removing question force from interrogative clauses: The West Flemish particle kwestje. In Tommasso Balsemin, Irene Caloi, Jacopo Garzonio, Nicolas Lamoure, Francesco Pinzin & Emanuela Sanfelici (eds.), Festschrift in honour of Cecilia Poletto’s 60th birthday, vol. 2, 821–860. Università degli Studi di Padova. DOI: 10.57711/gy76-g337. Woods, Rebecca & Tom Roeper. 2020. Rethinking auxiliary-doubling in adult and child language: How verb movement turns propositions into illocutionary acts. In Rebecca Woods & Sam Wolfe (eds.), Rethinking Verb Second, 835–861. Oxford: Oxford University Press. Woods, Rebecca & Tom Roeper. 2021. The acquisition path of “high” negation in English. In 45th Boston University Conference on Language Development, 761– 775. Somerville: Cascadilla Press. Woods, Rebecca & Luis Vicente. 2021. Metacommunicative-why fragments as probes into the grammar of the speech act layer. Glossa: A Journal of General Linguistics 6(1). 84: 1–32. DOI: 10.5334/gjgl.1169. 369