Fragments

This chapter discusses two kinds of fragments: phrases such as DPs and PPs that occur apparently unembedded in a sentence, without any overt antecedent, and fragment answers to questions. It presents non-sententialist (Stainton 2006, Progovac 2006) and sententialist approaches (focusing especially on Merchant’s 2004 movement-then-deletion account), then considers the main evidence for each approach, and how the opposing approach has tried to address this evidence. Arguments discussed in favour of non-sententialism are anticonnectivity effects, differences in the behaviour of fragment answers and their overtly sentential counterparts with regard to presupposition inheritance, and questions about the psychological plausibility of silent syntax; arguments in favour of sententialism are various kinds of connectivity effects—case, island effects, binding, and requirements for complementizers and prepositions in fragment answers.

Null Complement Anaphora

In this chapter, the different positions regarding NCA in the literature are discussed and the analysis proposed in Hankamer and Sag (1976) is favored. NCA is taken to be a type of deep anaphor as opposed to a surface anaphor, in Hankamer and Sag’s proposed typology. It behaves like a deep anaphor in that it can take pragmatic antecedents, missing antecedents, and does not seem to require strict syntactic parallelism with its antecedent. In addition, NCA does not allow overt or covert extraction out of it, which provides evidence in favor of the view that NCA is to be represented as an element with no internal syntactic structure. It is also shown that the types of predicates that select NCA seem to be lexically determined, since no reliable pattern has yet been found to be able to predict which predicates select it.

Ellipsis and Prosody

This chapter investigates the relation between ellipsis and prosody. Prosody is most frequently described as “the organizational structure of speech” (Beckman 1996). It has been defined in terms of three independent factors of the phonological representation: “intonation, phrasal rhythmic patterning and prosodic phrasing” (Selkirk 1995: 550). Prosody plays an important role in the interpretation of elliptical utterances and bridges the gap between what is overtly expressed and what is understood. There are three specific research areas where prosody has been claimed to be relevant: first, prosody-related licensing of ellipsis; second, prosody-related conditions of recoverability of deletion; third, prosodic effects with respect to the question of whether there is structure in the ellipsis site. In this area, research has focused on prosodic conditions on extraction and locality. This chapter is structured accordingly. It first summarizes the prosodic system of English and then reviews the research on how prosody bears on the central issues of the theory of ellipsis (licensing, recoverability, structure in the ellipsis site). The discussion shows that the interaction of prosody and the absence of sound contributes to the understanding of the theories of sound–meaning correspondence.

Dutch

This chapter discusses ellipsis in Dutch and the dialects of Dutch. It provides detailed information on the major types of ellipsis as they have been presented in Part III of this handbook: gapping and stripping, predicate ellipsis (VP-ellipsis and pseudogapping), Conjunction Reduction and Right-Node Raising, sluicing, fragments, nominal ellipsis, Comparative Deletion, and Null Complement Anaphora. It discusses the main insights from the literature as well as new observations with respect to these constructions. The final section shows that the Dutch dialects display an enormous amount of variation concerning ellipsis constructions. In particular, it examines the variation in NP-ellipsis with possessive, demonstrative, and adjectival remnants and variation with respect to sluicing.

ellipsis in inquisitive semantics

Work in inquisitive semantics has developed an alternative-rich notion of semantic content which is uniform across questions and assertions. This chapter explores the ramifications of this view for the theory of ellipsis. It examines these issues primarily by focusing on the analysis of a particular ellipsis process in English: sluicing. Empirically, it reviews a number of arguments in favor of an account of sluicing incorporating inquisitive semantics, most notably cases where truth-conditionally equivalent sentences have differential behavior for sluicing. Theoretically, the chapter demonstrates how a theory of ellipsis building on Merchant (2001) but based on an inquisitive semantics helps address this data. Beyond this, it briefly discusses motivations for extending this sort of approach to other ellipsis processes and compares the proposed account with other potential ways of incorporating inquisitive semantics into a theory of ellipsis.

Ellipsis and Discourse

Despite the considerable attention paid to elliptical phenomena in the literature, the conditions under which a representation of an utterance may serve as a suitable antecedent for interpreting a subsequent ellipsis remain highly contentious and poorly understood. Focusing primarily on VP-ellipsis, this chapter surveys evidence that informs this debate, focusing on issues bearing on the discourse context in its role of making meanings available for ellipsis recovery. It argues that the evidence more strongly favors discourse-based explanations over syntactic ones, although problems for discourse-based theories do remain.

Ellipsis and Computational Linguistics

This chapter considers approaches to ellipsis within computational linguistics. It begins with the structure of the ellipsis site—a topic that has received little attention in computational linguistics. There are two prominent accounts of the recovery of ellipsis: that of Lappin and McCord (1990) and Dalrymple et al. (1991). The topic of licensing follows. This topic has not been directly addressed in the computational literature, but the chapter covers two related issues of direct computational interest: the identification and generation of ellipsis occurrences. This is followed by three additional topics: identifying the antecedent, dialogue, and text mining.

Japanese

This chapter provides an overview of, while examining various proposals for, ellipsis in Japanese. Fragments are examined, and it is claimed to be reasonable to assume that stripping, sluicing, and ellipsis in comparatives are a uniform phenomenon while short answers are distinct. It is also argued that the properties of Right-Node Raising can be best captured by a non-constituent string deletion analysis. Three approaches to null arguments are examined, and shortcomings in each are discussed. N’-deletion is then explored and claimed to be ambiguous between two structural possibilities: ellipsis and non-ellipsis. VP-ellipsis, gapping, and pseudo-gapping are also touched upon. One significant aspect of the ellipsis phenomena in Japanese illustrated in this chapter is that the presence and absence of a case-marker plays a crucial role, with case-marked and non-case-marked fragments being analyzed as instances of surface anaphora and deep anaphora, respectively. This indicates the importance of focusing on case-marked versions in the syntactic investigation of these phenomena.

Ellipsis and Acquisition

Ellipsis provides a quintessential acquisition challenge—deleted material is an obvious instance of poverty of the stimulus. This chapter reviews the naturalistic and experimental evidence to evaluate how far UG allows immediate correct projection of VP and NP ellipsis, based on Merchant’s (2014) E-features. It develops an acquisition theory whereby a child moves, step by step, from Open Interfaces to isomorphic Strict Interfaces between LF and syntax which entails full reconstruction. The approach both utilizes pro as an initial default and explains why it remains in the final grammar for particularly complex constructions. It follows an economy principle: minimize recovery by contextual inference. The analysis is extended to sluicing, ACD, fragments, and the projection of QUD, and engages English, Portuguese, German, Dutch, Italian, and African-American English.

Ellipsis

This chapter surveys a variety of approaches to ellipsis, classifying them by how they answer the questions of structure (‘Is there structure inside ellipsis sites?’) and identity (‘What kind of identity, if any, holds between an ellipsis and its antecedent?’). It reviews the evidence for and against structure in ellipsis, from lower infinitivals and predicate answers. While the results are mixed, on the whole, the preponderance of the evidence suggests that the analysis of the full range of facts of ellipsis requires reference to syntactic structure.