Antecedent-contained Deletion Research Articles

Quantifier Scope in Russian

This paper documents certain quantifier scope ambiguities in Russian and argues that these are derived by means of a covert syntactic movement operation, Quantifier Raising (QR). By doing so, it argues against the popular “frozen scope” view of Russian (Ionin 2003) by showing that optional, non-local QR past vP level must be available in the language in exactly the contexts where it is available in English. Syntactic evidence for the parallelism comes from Inverse Linking, Antecedent Contained Deletion and other contexts that have been argued to point to the existence of QR in English. Additionally, evidence for the availability of subject Reconstruction for scope in Russian (contra Ionin 2003) strongly suggests that the mechanism deriving ambiguity in basic SVO sentences in Russian must involve subject Reconstruction plus object QR to vP, exactly as has been argued to be necessary for English in Johnson (2000; 2001). The conclusion that Russian possesses QR with the same properties known from English suggests that the availability of QR in certain languages or certain constructions should not be tied to the unavailability of Scrambling since Russian also exhibits Scrambling. This conclusion thus carries important implications for other languages still currently taken to be similarly scope frozen due to the availability of Scrambling.

The Revised Split Relative Clause for Antecedent-Contained Deletion

This paper aims to provide an analysis of antecedent-contained deletion (ACD) which is compatible both with the copy theory of movement and with the cyclic Spell-Out hypothesis. For this, this paper revises Ding’s (2013) split relative clause account for ACD. The core of the revision is to assume that the relative clause containing the ellipsis site (RCE) is adjoined not to vP but to VP. This assumption is supported by three pieces of empirical evidence presented in Sato (2013). This paper’s proposal can account for the correlation between ACD and NPI, NP-contained ACD. Under this paper’s analysis, the RCE and the normal relative clause are differentiated with respect to the adjunction site. This phenomenon may be interpreted as the consequence of last resort. This paper has no explanation for Tiedeman’s puzzle. But we may find the possibility of its resolution in Chomsky’s (2004) afterthought account. Although this paper’s proposal is not perfect, ACD can be given a better explanation by it.

How to Neutralize a Finite Clause Boundary: Phase Theory and the Grammar of Bound Pronouns

A bound pronoun in the subject position of a finite embedded clause renders the clause boundary relatively transparent to relations ordinarily confined to monoclausal, control, and raising configurations. For example, too/ enough-movement structures involving a finite clause boundary are degraded in sentences like * This book is too long [for John to claim [that Bill read ___ in a day]] but improved when the finite clause has a bound pronominal subject as in ? This book is too long [for John1to claim [that he1read ___ in a day]]. This bound pronoun effect holds across a wide range of phenomena including too/ enough-movement, tough-movement, gapping, comparative deletion, antecedent-contained deletion, quantifier scope interaction, multiple questions, pseudogapping, reciprocal binding, and multiple sluicing; we confirm the effect via a sentence acceptability experiment targeting some of these phenomena. Our account has two crucial ingredients: (a) bound pronouns optionally enter the derivation with unvalued ϕ-features and (b) phases are defined in part by convergence, so that under certain conditions, unvalued features void the phasal status of CP and extend the locality domain for syntactic operations.

Antecedent‐Contained Deletion in Japanese: Support for the VP‐Ellipsis Analysis

AbstractIn this article, I demonstrate that antecedent‐contained deletion (ACD) exists in Japanese, thereby supporting the V‐stranding VP‐ellipsis analysis, and that the infinite regress resolution with such ACD cases must be made by overt movement. I suggest that this way of infinite regress resolution is a universal property, demonstrating that English ACD cases can be analyzed along these lines. I also argue that what seems like a case of Japanese ACD involves argument ellipsis, even though it violates the anti‐c‐command requirement, and that the apparent violation is resolved by adopting the late insertion of adjuncts and the derivational definition of c‐command.

Even doesn’t move but associates into traces: A reply to Nakanishi 2012

Nakanishi (Nat Lang Semant 20(2):115–136, 2012) presents a novel argument for the so-called scope theory of English sentential even (VP-even), based on examples with antecedent-contained deletion (ACD). Nakanishi’s argument is based on the assumption that even cannot associate with a focus which has moved out of its LF scope. I show that this assumption is incorrect, defusing Nakanishi’s argument. I propose that when even associates with a focus which has moved out of its surface scope, it actually associates with focused material in the lower copies of movement (trace positions). I show that a closer look at ACD examples of Nakanishi’s type in fact yields a new argument against the scope theory. I conclude that English sentential even must always be interpreted in its pronounced position. The patterns of focus association with even presented here constitute a new argument for the copy theory of movement.

Post Hoc Analysis Decisions Drive the Reported Reading Time Effects in Hackl, Koster-Hale & Varvoutis (2012)

Abstract Hackl, Koster-Hale & Varvoutis (2012; hereafter HKV) provide data that suggest that in a null context, antecedent-contained deletion (ACD) relative clause structures modifying a quantified object noun phrase (NP) are easier to process than those modifying a definite object NP. HKV argue that this pattern of results supports a quantifier-raising (QR) analysis of both ACD structures and quantified NPs in object position: under the account they advocate, both ACD resolution and quantified NPs in object position require movement of the object NP to a higher syntactic position. The processing advantage for quantified object NPs in ACD is hypothesized to derive from the fact that—at the point where ACD resolution must take place—the quantified NP has already undergone QR, whereas this is not the case for definite NPs. Here, we question these conclusions. In particular, our analyses of HKV’s reading time data reveal several unreported choice points, errors and concerns regarding multiple comparisons in the original HKV data analysis. Importantly, most other plausible ways to analyze these data that we describe here result in the crucial interaction being non-significant. Putting this observation together with the failure to observe the crucial interaction in Gibson & Levy (2016), we conclude that the experiments reported by HKV should not be viewed as providing evidence for the ACD quantifier-raising processing effect.

Focus association into copies and the scope of even

Nakanishi 2012 presents a novel argument for the so-called “scope theory” of English VP-even, based on examples with antecedent-contained deletion (ACD). Nakanishi’s argument is based on the assumption that even cannot associate with a focus which has moved out of its scope. I show that this assumption is incorrect, defusing Nakanishi’s argument. I propose that when even associates with a focus which has moved out of its scope, it actually associates with focused material in the lower copies of movement (trace positions). I show that a closer look at ACD examples of Nakanishi’s type in fact forms a new argument against the scope theory. I conclude that English VP-even must always be interpreted in its pronounced position. The patterns of focus association with even presented here constitute a new argument for the copy theory of movement.

Phases and argument ellipsis in Japanese

This article investigates the nature of Japanese null arguments. Although it has been more or less standardly assumed that Japanese null arguments are empty pronouns, recent literature has shown that they can also be derived via argument ellipsis based on the fact that they can yield readings that pronouns generally cannot support (Oku in A theory of selection and reconstruction in the minimalist program, 1998; Saito in Lang Res 43: 203–227, 2007; Takahashi in: Miyagawa and Saito (eds) The Oxford handbook of Japanese linguistics, Oxford University Press, New York, 2008a, in Linguist Inq 39:307–326, 2008b, among others). Most of the literature on argument ellipsis has discussed its availability, while less attention has been paid to the issue of the conditions under which ellipsis of arguments is possible. This article focuses on this issue. Specifically, building on Abe’s (in: Hoshi (ed.) The dynamics of the language faculty: Perspectives from linguistics and cognitive neuroscience, Kuroshio Publishers, Tokyo, 2009) observation that argument ellipsis is not freely available, I argue that argument ellipsis is locally phase-constrained (cf. Chomsky in Martin et al. (eds) Step by step: Essays on minimalist syntax in honor of Howard Lasnik, MIT, Cambridge, 2000; in Kenstowicz (ed) Ken Hale: A life in language, MIT, Cambridge, 2001, and seq.), developing a particular phase-based approach to argument ellipsis which is shown to have consequences for Abe’s (in: Hoshi (ed.) The dynamics of the language faculty: Perspectives from linguistics and cognitive neuroscience, Kuroshio Publishers, Tokyo, 2009) anti-c-command generalization regarding the availability of argument ellipsis and Takahashi’s (in: Laurencot (eds) University of Connecticut working papers in linguistics 5, University of Connecticut, Storrs, 1996a; in: Uchibori and Yatsushiro (eds) University of Connecticut working papers in linguistics 7: Papers in honor of Mamoru Saito, University of Connecticut, Storrs, 1996b) “Antecedent-Contained Deletion” in Japanese. The approach developed in this article thus provides a tool which contributes to our understanding of the context in which the ellipsis-indicating readings are/are not available with Japanese null arguments.

Antecedent contained deletions in native and non-native sentence processing

Abstract We report the results from an eye-movement monitoring study investigating native (L1) and non-native (L2) speakers’ real-time processing of antecedent-contained deletion (ACD), a type of verb phrase ellipsis in which the ellipsis gap forms part of its own antecedent. The resulting interpretation problem is traditionally thought to be solved by quantifier raising, a covert scope-shifting operation that serves to remove the gap from within its antecedent. Our L2 group comprised advanced, native German-speaking L2 learners of English. The analysis of the eye-movement data showed that both L1 and L2 English speakers tried to recover the missing verb phrase after encountering the gap. Only the native speakers showed evidence of ellipsis resolution being affected by quantification, however. No effects of quantification following gap detection were found in the L2 group, by contrast, indicating that recovery of the elided material was accomplished independently from the object’s quantificational status in this group.

QR Out of a Tensed Clause: Evidence from Antecedent‐Contained Deletion

AbstractThis paper presents an argument based on evidence from experiments featuring Antecedent‐Contained Deletion (ACD) sentences situated in carefully‐manipulated discourse contexts, that covert movement is not grammatically constrained by tense. ACD is a form of Verb Phrase Ellipsis in which ellipsis is embedded in its antecedent. Under an account appealing to Quantifier Raising, the quantificational phrase containing the ellipsis site raises to a VP‐external position, allowing the VP to become the antecedent. When ACD is embedded in a non‐finite clause, such sentences are ambiguous, since multiple VPs can serve as an antecedent. However, when ACD is embedded in a finite clause, the range of interpretations has been claimed to be restricted, because of an independent ‘clause‐bounded’ movement constraint on Quantifier Raising. However, there are exceptions to this generalization. I present evidence from an experimental investigation of finite‐clause‐embedded ACD sentences, relying on Cecchetto (2004), to demonstrate that under the right discourse conditions, the supposedly unavailable Matrix reading surfaces robustly, at a percentage that is surprising if the constraint were rooted in the grammar. I argue that these results call into question the source of this locality restriction, and propose that it has nothing to do with an arbitrary grammatical constraint on movement.

A Pragmatic Account of Complexity in Definite Antecedent-Contained-Deletion Relative Clauses

Hackl, Koster-Hale & Varvoutis (2012; HKV) provide data that suggest that in a null context, antecedent-contained-deletion (ACD) relative clause structures modifying a quantified object noun phrase (NP; such as every doctor) are easier to process than those modifying a definite object NP (such as the doctor). HKV argue that this pattern of results supports a ‘quantifier-raising’ (QR) analysis of both ACD structures and quantified NPs in object position: under the account they advocate, both ACD resolution and quantified NPs in object position require movement of the object NP to a higher syntactic position. The processing advantage for quantified object NPs in ACD is hypothesized to derive from the fact that—at the point where ACD resolution must take place—the quantified NP has already undergone QR whereas this is not the case for definite NPs. Although in other work it is shown that HKV’s reading time analyses are flawed, such that the critical effects are not significant (Gibson et al. submitted), the effect in HKV’s acceptability rating is robust. But HKV’s interpretation is problematic. We present five experiments that provide evidence for an alternative, pragmatic, explanation for HKV’s observation. In particular, we argue that the low acceptability of the the / ACD condition is largely due to a strong pressure in the null context to use a competing form, by adding also or same. This pressure does not exist with quantified

Antecedent contained deletion in the domain of a raised object

AbstractWhen a verb takes two complements, the first complement, typically an accusative object, c-commands the second complement, and the second complement can host a relative clause containing antecedent contained deletion (ACD). This simple fact, which, upon careful considerations, reveals that English accusative objects raise to a higher position than previously thought. The theoretical claim is that objects raise to check Case to the outer vP-spec. We argue that the v consists of two separate heads, the higher v

The syntax–semantics interface

The study of the syntax–semantics interface is concerned with linguistic phenomena that are the product of interactions between principles of syntactic organization and principles of semantic interpretation. Such interactions abound in natural language and can be found in all subsystems of the grammar. This paper examines a particular subclass of such phenomena revolving around quantificational expressions. The central concern of the discussion are the grammatical mechanisms that mediate between the syntactic position a quantifier appears in and the semantic import it has on the sentence meaning. Of particular interest are cases where a quantifier is interpreted in a position that is different from the position it seems to occupy in the syntax. A leading hypothesis to explain cases of this sort, which exemplify a general property of natural language called displacement, is that they are the product of overt or covert movement operations. Empirical support for this approach is presented in the form of correlations between three grammatical phenomena – Quantifier Scope, Antecedent Contained Deletion, and Extraposition – which receive a uniform account under the above hypothesis.

Quantification and ACD: Evidence from Real-Time Sentence Processing

Quantifiers, unlike proper names or definite descriptions, cannot be given the semantics of referring expressions. This fact has triggered a long-standing debate in formal semantics and syntax as to the combinatorial means by which quantifiers are integrated into a sentence. The present paper contributes to this debate through an investigation of quantifier comprehension during real-time sentence processing. We present evidence showing that two potentially independent processes—the integration of a quantifier in object position and the resolution of antecedent-contained deletion (ACD)—are linked. Our data show, more specifically, that the resolution of a downstream ACD site is facilitated during real-time sentence processing if the upstream DP hosting the ACD site is quantificational but not if it is definite. We discuss these findings in the context of a QUANTIFIER RAISING-based approach and a type-shifting-based approach to quantifier integration. We argue that facilitation of ACD resolution by an upstream quantifier is only expected by theories, such as the QUANTIFIER RAISING approach, which employ the same mechanism for both processes. We then compare the QUANTIFIER RAISING-based account with a non-grammatical experience-based approach to our data, which attempts to explain the findings in terms of corpus frequencies. Although we cannot rule out such an alternative at this stage, we offer reasons to believe that an account that exploits QUANTIFIER RAISING has an explanatory advantage.

The scope of even and quantifier raising

This paper addresses the question of whether the preverbal even (VP-even) embedded in a nonfinite clause can take wide scope (e.g., Bill refused to even drink WATER). The paper presents novel evidence for wide scope VP-even that is independent of the presuppositions of even. The evidence is based on examples of antecedent-contained deletion (ACD), where embedded VP-even associates with a nominal constituent (or part of it) that raises out of the embedded clause via quantifier raising. Assuming that even must c-command the focus that it associates with, the case at issue forces VP-even to have wide scope, and further shows that VP-even in NPI-licensing contexts is not necessarily an NPI.

Competence, Performance, and the Locality of Quantifier Raising: Evidence from 4-Year-Old Children

We revisit the purported locality constraint on Quantifier Raising (QR) by investigating children's and adults' interpretation of antecedent-contained-deletion (ACD) sentences, where the interpretation depends on the landing site targeted by QR out of an embedded clause. When ACD is embedded in a nonfinite clause, 4-year-old children and adults access both the embedded and the matrix interpretations. When ACD is embedded in a finite clause, and the matrix interpretation is generally believed to be ungrammatical, children and even some adults access both readings. These findings allow for the possibility that the source of QR's reputed locality constraint may instead be extragrammatical, and they provide insight into the development of the human sentence parser.

An annotated corpus for the analysis of VP ellipsis

Verb Phrase Ellipsis (VPE) has been studied in great depth in theoretical linguistics, but empirical studies of VPE are rare. We extend the few previous corpus studies with an annotated corpus of VPE in all 25 sections of the Wall Street Journal corpus (WSJ) distributed with the Penn Treebank. We annotated the raw files using a stand-off annotation scheme that codes the auxiliary verb triggering the elided verb phrase, the start and end of the antecedent, the syntactic type of antecedent (VP, TV, NP, PP or AP), and the type of syntactic pattern between the source and target clauses of the VPE and its antecedent. We found 487 instances of VPE (including predicative ellipsis, antecedent-contained deletion, comparative constructions, and pseudo-gapping) plus 67 cases of related phenomena such as do so anaphora. Inter-annotator agreement was high, with a 0.97 average F-score for three annotators for one section of the WSJ. Our annotation is theory neutral, and has better coverage than earlier efforts that relied on automatic methods, e.g. simply searching the parsed version of the Penn Treebank for empty VP’s achieves a high precision (0.95) but low recall (0.58) when compared with our manual annotation. The distribution of VPE source–target patterns deviates highly from the standard examples found in the theoretical linguistics literature on VPE, once more underlining the value of corpus studies. The resulting corpus will be useful for studying VPE phenomena as well as for evaluating natural language processing systems equipped with ellipsis resolution algorithms, and we propose evaluation measures for VPE detection and VPE antecedent selection. The stand-off annotation is freely available for research purposes.

<i>One</i>-replacement and the label-less theory of adjuncts

One-replacement and the label-less theory of adjuncts Yosuke Sato 1. Introduction The proper treatment of adjuncts has been a central issue in syntactic theory since the 1980s. In the X′-theoretic analysis in the Government and Binding (gb) Theory (Chomsky 1981, 1986), they were commonly treated as optional elements attached to an intermediate or maximal projection of the head they modify. However, this treatment has been shown to be no longer tenable in the more recent Bare Phrase Structure (bps) Theory for several conceptual reasons (Chomsky 1995: ch. 4), which renders the status of adjuncts all the more puzzling. Recently, however, Hornstein and Nunes (2008) (henceforth H&N) have proposed a new theory of adjuncts that conforms to the precepts of the bps. It proposes that adjuncts need not be labeled for purposes of syntactic computation in contrast to complements and specifiers, which require the result of concatenation to be labeled. This squib presents new evidence for H&N’s theory from one-replacement. It has been commonly held that replacement is subject to the constituency requirement. However, Radford (1988) and Culicover and Jackendoff (2005) observe that the anaphoric one may pick up a discontinuous constituent as its antecedent, contrary to the conventional view. I show that this interpretation, though puzzling under the X′-theoretic analysis of onereplacement, naturally falls into place under H&N’s theory of adjuncts. 2. Hornstein and Nunes’s (2008) Label-Less Theory of Adjuncts H&N propose that both complements and specifiers require the result of concatenation to be labeled for the purposes of semantic interpretation at the Conceptual– Intentional System while adjuncts may only require concatenation. This theory may sound at odds with the view in the gb era that special modes of licensing (e.g., Chomsky-adjunction, special labeling algorithms) are in order for the proper integration of adjuncts into a phrase structure, unlike complements and specifiers. H&N argue, however, that their alternative view is preferred over the gb view under the Neo-Davidsonian semantic framework (Parsons 1990, Schein 1993).Within this framework, it is complement/specifier elements that need grammatical pivots (such as subject-of and object-of ) to properly serve as participants of an event denoted by the predicate; adjuncts, on the other hand, can directly modify the event without such aid. This contrast can be seen in the Neo-Davidsonian semantic representation for the sentence in (1a), shown in (1b). [End Page 416] 1. a. John ate the cake in the yard. b. ∃e [eating (e) & subject (John, e) & object (the cake, e) & in-the-yard (e)] (H&N:70) H&N claim that this difference between complements/specifiers and adjuncts at the semantic interface motivates their different modes of integration into a phrase structure. Specifically, adopting the idea from Hornstein (2009) that the so-called Merge operation (Chomsky 1995) is decomposed into concatenation and labeling, they propose that the composition of a complement/specifier requires both concatenation and labeling whereas that of an adjunct may only require concatenation. This difference is shown in (2). 2. a. [X X∧Y] b. [X X∧Y] ∧ Z c. [X [X X∧Y] ∧ Z] ] (exx. 2a, 2b from H&N: 65–66) In (2a), X is concatenated with Y. The result of this concatenation is then labeled as X to yield a complex atomic unit accessible for further concatenations. In (2b), Z is concatenated with the atomic unit in (2a), but the concatenate is not followed by labeling and hence cannot be accessed for further concatenations. This derivation, thus, gives substance to one recent view, expressed in one way or another, that adjuncts “dangle off” the syntactic workspace (Uriagereka 1998, Chomsky 2004). Example (2c), on the other hand, illustrates the case where the concatenation of Z with [X X∧Y] is followed by labeling.1 H&N show that this label-less theory of adjuncts brings favorable empirical payoffs in several areas, including the invisibility of adjuncts to focus projection (Gussenhoven 1984), the non-inhibitory effect of adjuncts to affix-hopping (Chomsky 1957, Bobaljik 1995), the ellipsis resolution in the antecedent-contained deletion (May 1985), and the “discontinuous” interpretation of the anaphoric do so-proform. I briefly illustrate...

QR in Child Grammar: Evidence from Antecedent-Contained Deletion

We show that 4-year-olds assign the correct interpretation to antecedent-contained deletion (ACD) sentences because they have the correct representation of these structures. This representation involves Quantifier Raising (QR) of a Quantificational Noun Phrase (QNP) that must move out of the site of the verb phrase in which it is contained to resolve a case of verb phrase ellipsis. Furthermore, not only do children provide clear justifications for such sentences with ACD, but they treat ACD sentences differently from sentences with coordinated conjunction, a plausible alternative if they lacked QR. The findings have implications for the interpretation of experimental results in which children appear to lack this grammatical operation, and instead point to extragrammatical factors as the source of this pattern.

WCO, ACD and what they reveal about complex demonstratives

This squib presents a rebuttal to two of King’s (Complex demonstratives: A quantificational account. Cambridge, Mass.: MIT Press, 2001) arguments that complex demonstratives are quantifier phrases like every man. The first is in response to King’s argument that because complex demonstratives induce weak crossover effects, they are quantifier phrases. I argue that unlike quantifier phrases and like other definite determiner phrases, complex demonstratives in object position can corefer with singular pronouns contained in the subject DP. Although complex demonstratives could undergo LF-movement, the ruling out by weak crossover is empirically undetectable. The second rebuttal is in response to King’s argument that because complex demonstratives allow antecedent-contained deletion, they are quantifier phrases. I present data showing that along with quantifier phrases, complex demonstratives pattern with proper names in allowing ACD with restrictive modification, but usually not with non-restrictive modification.