Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS (linguistics.oxfordre.com). (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 22 October 2017

Copular Constructions in Syntax

Summary and Keywords

Copular sentences (sentences of the form A is B) have been prominent on the research agenda for linguists and philosophers of language since classical antiquity, and continue to be shrouded in considerable controversy. Central questions in the linguistic literature on copulas and copular sentences are (a) whether predicational, specificational, identificational, and equative copular sentences have a common underlying source; and, if so, (b) how the various surface types of copular sentences are derived from that underlier; (c) whether there is a typology of copulas; and (d) whether copulas are meaningful or meaningless.

The debate surrounding the postulation of multiple copular sentence types relies on criteria related to both meaning and form. Analyses based on meaning tend to focus on the question of whether or not one of the terms is a predicate of the other, whether or not the copula contributes meaning, and the information-structural properties of the construction. Analyses based on form focus on the flexibility of the linear ordering of the two terms of the construction, the surface distribution of the copular element, the restrictions imposed on the extraction of the two terms, the case and agreement properties of the construction, the omissibility of the copula or one of the two terms, and the connectivity effects exhibited by the construction.

Morphosyntactic variation in the domain of copular elements is an area of research with fruitful intersections between typological and generative approaches. A variety of criteria are presented in the literature to justify the postulation of multiple copulas or underlying representations for copular sentences. Another prolific body of research concerns the semantics of copular sentences. In the assessment of scholarship on copulas and copular sentences, the article critiques the ‘multiple copulas’ approach and examines ways in which the surface variety of copular sentence types can be accounted for in a ‘single copula’ analysis. The analysis of copular constructions continues to have far-reaching consequences in the context of linguistic theory construction, particularly the question of how a predicate combines with its subject in syntactic structure.

Keywords: copulas, predication, specification, equation, connectivity effects

1 Copular Sentences in Linguistic Theory

Copular sentences (i.e., sentences of the form A is B) have had a prominent place on the research agenda for linguists and philosophers of language since the earliest days of Western linguistic thought. Perhaps surprisingly, given their apparent simplicity, copular sentences continue to stir up considerable controversy. There is ongoing debate regarding all of the fundamental questions posed by copular constructions, in (morpho)syntax as well as in semantics. Among the prominent points of disagreement in the linguistic literature on copulas and copular sentences are the questions of (a) whether there is a typology of copular sentences (predicational, specificational, identificational, equative) with different underlying representations or instead a common source for all surface varieties of copular constructions, and, if there is to be a single underlier, (b) how the various types of copular sentences are derived from it; (c) whether copulas are meaningful, thematic-role assigning verbs or meaningless ‘dummies’; and (d) whether there is a typology of different copulas.

In this article, the term “copular sentence” will not include sentences featuring be in which one of the terms is a verbal constituent (a gerund, as in The president is sleeping, or a past-participial or to-infinitival verbal expression, as in The president is (expected) to make an important speech later today), nor will it include possessive constructions (The book is mine) or locative existentials (The book is on the table, There is a book on the table). The reason not to discuss these explicitly is a practical and conventional one rather than a principled stance. An approach that takes the copula uniformly to be a meaningless mediator of a predication relation (which is one of the prevailing perspectives on the copula in the literature) can accommodate these kinds of be-constructions under the same rubric as the be-sentences generally discussed under the header of “copular sentences.”

For a subject that is at once so fundamental and so laden with controversy, it is impossible to do full justice to the vast literature that has made contributions to the syntax and semantics of copulas and copular sentences. This article will highlight what the authors consider to be the main issues, provide brief discussion of the main points of view regarding each of them, and give references to the primary literature for further pursuit of these matters.

The article is organized as follows. Section 2 lays out the various different types of copular sentences that have been identified in the literature and discusses the meaning-based and formal criteria that have been used to establish distinctions in the realm of copular constructions. Section 3 looks more microscopically at morphosyntactic variation in the domain of copular elements. Although this article focuses primarily on the morphosyntax of copulas and the sentences they build, it is necessary to address the main questions that they raise from a semantic perspective. This is the topic of section 4. The discussion in sections 2–4 mostly take the shape of an eyewitness report, but section 5 presents a brief critical analysis of the generative linguistic literature on copulas and copular sentences and places this literature in the broader context of linguistic theory construction, with particular emphasis on the question of how a predicate and its subject combine in syntactic structure.

2 Copular Sentence Types

For sentences of the general format A is B, a variety of different types can be distinguished on the surface. The literature on copular sentences has generally excelled at making taxonomical distinctions within the class of copular sentences.

2.1 Types of Taxonomical Approaches

In the A is B format, B is very often a predicate of A, its meaning attributing a property to the referent of A. If pseudoclefts with a copular wh-clause ([what A is] is B—e.g., what Dr. Jekyll is is handsome) and sentences like honest is honest are set aside, one can generalize that whenever B is a nonnominal element (for instance, an adjectival phrase), it is invariably a predicate of A: there is no other role that handsome in (1a) could play. But when B is nominal and A is a referential expression (see (1b,c)), the question of the function of the postcopular noun phrase immediately becomes a matter of debate. Since a handsome man and my best friend can indubitably be used as referential arguments of a verbal predicate (a handsome man/my best friend just arrived), it may not be obvious what the predicate of (1b,c) is: while Dr. Jekyll is certainly the subject of predication, the predicate could be either the postcopular constituent or the projection of be. In the latter approach, the postcopular noun phrase can be treated as an argument of the copular clause, with the copula itself as the predicate head—an analysis that is often deemed necessary for copular sentences of the type in (1d), where both dependents of the copula are referential expressions.

(1)

Copular Constructions in Syntax

It is important to note that predications of the type in (1b,c) survive without a copula, for instance below epistemic verbs such as consider, just like (1a) but in contrast to (1d):

(2)

Copular Constructions in Syntax

The copula thus does not appear to be an indispensable ingredient of predications such as (1b,c), at least on the surface. One could respond to this in either of two ways. One would be to set up a structure for the copula-less versions of (2b,c) in which the complement of consider contains a silent token of to be. The alternative would be to countenance the possibility that noun phrases such as a handsome man and my best friend can be used not just as arguments but also as predicates.

2.1.1 A Taxonomy of Copulas

The first line of approach to the questions posed by (1) and (2) conjointly gives rise to a typology of copular elements, minimally containing the following members:

(3)

Copular Constructions in Syntax

For (1a), where handsome is itself inherently predicative, there is no role to play for the copula other than that of providing support for the tense of the clause. This may be called the support copula (cf. do-support). In (1b)–(1c), with the postcopular noun phrases treated as referential expressions, there would be a predicational copula. The equative construction in (1d) would feature the equative copula. And for the versions of (2b,c) that lack to be, one would postulate a silent copula establishing the predication below consider.

More members of the typology of copular elements could be postulated to accommodate more fine-grained distinctions (e.g., by adding a specificational copula to the taxonomy; see section 2.1.2 for this notion).1 The trend in the generative literature, pursuing a maximally simple theory, has been to collapse all surface copulas into a single type, namely, type (3a), the meaningless support copula. The question of precisely what the copula supports, however (e.g., inflectional material, empty syntactic heads, event-related semantic material), is a matter of vigorous debate (see, e.g., Doron, 1986; Déchaine, 1995; Carnie, 1997; Rothstein, 1999; Schütze, 2004; den Dikken, 2006; Benmamoun, 2008. Under the support approach, sentences of the type in (1a) contain one predicate (handsome), rather than two (is and handsome).

While a treatment of the variation in the family of copular sentences in terms of a dense taxonomy of copular types is generally not fruitful, the morphosyntactic distribution of different copular elements can serve as a diagnostic for the structure of these different copular sentence types. Section 4 revisits different perspectives in the literature on whether there is evidence for the existence of a copula of identity (3c) alongside the support copula (3a).

2.1.2 A Taxonomy of Copular Sentences

Higgins (1979) endowed the linguistics literature with syntactic and semantic grounds for distinguishing a four-way typology of copular sentences:

(4)

Copular Constructions in Syntax

Declerck (1988) considers the possibility that there might be even more types of copular sentences, highlighting in particular the ‘definitional copular sentence.’

In the nomenclature of this taxonomy, (1a) and (1b) both unambiguously represent the predicational type (4a), and (1d) is unmistakably of the equative type (4d). For (1c), no simple categorization is possible a priori: much depends on the information-structural context in which (1c) is used. If (1c) serves as an answer to the question What is Dr. Jekyll to you?, it is a predicational copular sentence. If it answers the question Who is your best friend?, it is specificational—it supplies a value for the variable that your best friend represents.

The list of examples of copular sentences in (1) does not include a specimen of what Higgins calls the ‘identificational copular sentence.’ Sentences like That (man) is my best friend are mentioned by Higgins under this rubric. But Higgins himself notes that his identity statements and identificational copular sentences are reducible to one single type (see section 2.2). Identificational copular sentences will not be discussed at any length in this article.

The typology of copular sentences is a function of (a) semantic and information-structural properties (referentiality and discourse function), (b) syntactic properties (reversibility, agreement, case), and (c) phonological properties (prosody). This article will concentrate on (a) and (b), mentioning prosodic clues on the side whenever they are helpful.

2.2 Meaning-based Criteria for the Classification of Copular Sentences

2.2.1 Referentiality

The semantic ingredient most commonly used in the classification of copular sentences is referentiality (although there is disagreement in the literature on the proper definition of ‘referential’ in this context, which complicates matters, cf. Partee, 1986; Jacobson, 1994; Heycock and Kroch, 1999; Sharvit, 2003; Mikkelsen, 2005; Heller, 2005; Romero, 2005; Comorovski, 2008). Using referentiality as the criterion, the two extremities of the typology of copular sentences are relatively easy to circumscribe: copular sentences of the form NP1 is NP2 (where ‘NP’ stands for a nominal phrase of unspecified size and complexity) involve the predicational type when NP2 is nonreferential, denoting a property attributed to NP1;2 the equative type is involved when both NP1 and NP2 are referential.3 A useful test for referentiality is modification by a nonrestrictive relative clause introduced by who: only referential noun phrases can be so modified (see (5)); as (6) shows, NP2 in (1d) accepts modification by a nonrestrictive clause with who, whereas such modification fails for NP2 in (1b) (see discussion in Heycock and Kroch, 1999).

(5)

Copular Constructions in Syntax

(6)

Copular Constructions in Syntax

But fitting the specificational copular sentence into a reference-based typology is not a straightforward matter. For Higgins (1979), neither NP1 nor NP2 of a specificational copular sentence can be referential, since specificational copular sentences are not ‘about’ either of the two NPs. He therefore introduces additional terminology to label the constituents of specificational sentences, as ‘superscriptional’ (defining the heading of a list) and ‘specificational’ (specifying the content of the list). The intuitions behind these terms are formalized in more contemporary semantic approaches treating the superscriptional term as an intensional object, whose value in the evaluation world is provided by the specificational term (e.g., Sharvit, 2003; Schlenker, 2003; Romero, 2004; Comorovski, 2008; see section 4).

2.2.2 Subject and Predicate

A second criterion figuring in the semantic classification of copular sentence types is the distinction between subject and predicate. For copular sentences of the type in (1a) and (1b), it is not difficult to agree that the postcopular constituent serves as part of the predicate of the precopular one. In (1d), on the other hand, it is not obvious that either of the two noun phrases is predicated of the other. Once again, the specificational copular sentence lands somewhere in the middle: the status of my best friend in (1c) as a predicate has been hotly debated.

Blom and Daalder (1977) (who express their indebtedness to den Hertog, 1903) take the extreme view that all copular sentences, including equative ones, are underlyingly represented as predications, with one of the two NPs functioning as the hyperonym (predicate) of the other (the hyponym). For Blom and Daalder (1977), a predicate nominal, while nonreferential, can be interpreted referentially on the basis of the language user’s non- or extragrammatical knowledge. Declerck (1988, pp. 92–93) dismisses an approach to the classification of copular sentences based on hyperonymy and hyponymy, stressing that “[s]pecificational sentences do not express a hyponymy relation.” For Higgins, the essence of the specificational relationship is that the contents of the concept expressed by the superscriptional NP are specified, which “naturally means that the NP denoting the concept and the NP specifying its contents cannot differ in generality or specificity” (Declerck, 1988, p. 93).

2.2.3 Information Structure

Blom and Daalder (1977) also draw on the information-structural properties of the terms of copular sentences (as well as their associated prosodic contours) to make the distinction between specificational and predicational copular sentences. They point out that in a specificational copular sentence, but not in a predicational one, the focused element is contained in the hyponym, which for them is the underlying subject of predication.

There is indeed a strong tendency for the underlying subject, or ‘value,’ of a specificational sentence to be the focus of the construction, conveying new information and being prosodically prominent. The superscriptional term in precopular position, on the other hand, is a topic. Note the asymmetry between two replies in the brief discourse in (7):

(7)

Copular Constructions in Syntax

For Mikkelsen (2005), den Dikken (2006), and Partee (2010), the fixed Topic-Comment information structure of specificational copular sentences is a correlate of their predicate-initial syntax (see also Guéron, 1993). But as Collins (1991, p. 119) points out, while “in the majority of basic pseudo-clefts the relative clause/theme [i.e., variable] contains (at least some) new information… there are several factors which work together to attenuate the newness of the item or items so marked intonationally.” The example in (8) in shows that newness does not need to be located in the value of a specificational copular sentence.

(8)

Copular Constructions in Syntax

The focused element of a specificational copular sentence is typically an exhaustive focus. Collins notes, however, that exhaustivity (aka exclusivity or uniqueness) is often at best a cancelable implicature (p. 32). (10) provides an example of a specificational copular sentence in which exhaustivity is explicitly overridden by among other things.

(9)

Copular Constructions in Syntax

While the exhaustive nature of the focus is a crucial component of semantic analyses of specificational sentences assimilating them to question-answer structures (e.g., Schlenker, 2003), most authors concede, with Collins, that the exhaustivity entailment or implicature can be canceled, perhaps on a par with “mention-one” questions, (10).

(10)

Copular Constructions in Syntax

2.3 Syntax-based Criteria for the Classification of Copular Sentences

On syntactic grounds, it is often easier to make clear-cut distinctions between predicational, specificational, and equative copular sentences. Some of the syntactic diagnostics that will be reviewed in this section deliver interesting results for all three types of copular sentence, setting predicationals apart from both specificationals and equatives, while others concentrate primarily on contrasting predicational and specificational copular sentences, leaving equatives untouched.

2.3.1 Linear Order

Probably the most familiar syntactic diagnostic is the fact that, poetic license aside, predicational copular sentences (in languages that are not generally predicate-initial) show a rigid subject–predicate order, whereas the two terms of specificational and equative copular sentences can typically change places freely. Thus, compare the examples in (11) to those in (1), above:

(11)

Copular Constructions in Syntax

2.3.2 Copula Distribution

A second important diagnostic differentiating between predicational copular sentences, on the one hand, and specificational and equative copular sentences on the other centers around the copula.4

In predicational constructions, the copula is not an absolute necessity; in embedded environments in which verbal and verbless constructions are in free variation in principle, predicational constructions do not require be:

(12)

Copular Constructions in Syntax

(2d) already showed that equative constructions do strictly require the copula. The set of examples in (13) and (14) amplify this conclusion on the basis of a larger set of copula-omission environments, and demonstrates as well that the copula is obligatory regardless of the relative order of the two terms. Section 3 provides additional examples from languages which, unlike English, do not categorically require an overt copula in finite predicational environments, but do require an overt copular element in equative environments.

(13)

Copular Constructions in Syntax

(14)

Copular Constructions in Syntax

Specificational constructions find themselves halfway in between predicational and equative copular sentences when it comes to copula distribution. When the value precedes the superscriptional term, the copula is freely omissible in the contexts under discussion (recall (2c)). However, when the superscriptional term comes first, a form of the copula is required.

(15)

Copular Constructions in Syntax

(16)

Copular Constructions in Syntax

This is the first of a family of diagnostics that yield different results for specificational copular sentences with value-initial orders and those with value-final orders. One general response to this split between two types of specificational copular sentence is to treat the two varieties differently, assimilating value-initial specificationals to predicationals and putting value-final specificationals in the same category as equatives. The alternative approach is to treat the class of specificational copular sentences uniformly at an underlying level of analysis and to derive one of its surface forms (the value-final order) in the syntax via an operation that reverses the initial relative order of value and superscriptional term—an operation that will then be held responsible for the obligatory occurrence of the copula (Moro, 1997; den Dikken, 2006).

2.3.3 Extraction Restrictions

Constraints on the establishment of filler-gap dependencies (here referred to as ‘extraction restrictions’) form a third domain in which predicational copular sentences behave systematically differently from the other two main types of copular sentences. We see these restrictions at work in two different ways.

First, predicational copular sentences freely allow the predicate to contain a gap linked to a filler in the same clause:

(17)

Copular Constructions in Syntax

In equative copular sentences, by contrast, the second noun phrase resists extraction, as shown by the example in (18b) (modeled on a sentence constructed by Heycock and Kroch, 1999):

(18)

Copular Constructions in Syntax

As in the previous section, specificational copular sentences behave differently with respect to this diagnostic depending on the relative order of their two terms: the second noun phrase can contain a gap only when it is the superscriptional term. The examples in (19) and (20) are modeled on sentences originally constructed by Moro (1997).

(19)

Copular Constructions in Syntax

(20)

Copular Constructions in Syntax

The examples reviewed in this subsection so far concentrate on extraction from the second term of the copular sentence. Extraction of that term gives rise to a very similar dichotomy:

(21)

Copular Constructions in Syntax

(22)

Copular Constructions in Syntax

(23)

Copular Constructions in Syntax

Moro (1997) presents an account of the extraction restrictions shown in (20b) and (23b) couched in then-current Principles-and-Parameters technology, in particular the Subjacency Condition and the Empty Category Principle: a photo of the wall in (20a) is an opaque domain (whence (20b)) and not in a ‘properly governed’ position (whence (23b)). Den Dikken (2006) presents an analysis of these data in terms of a ‘minimalist’ theory of locality (based on Chomsky, 1995), and covers the equative cases in (18b) and (22) as well.

Extraction from and of the precopular expression yields a different pattern of results. Subextraction from the precopular noun phrase is impossible, regardless of whether it is the superscriptional term or the value, as illustrated in (24).

(24)

Copular Constructions in Syntax

It is sometimes claimed (see, e.g., Heggie, 1988; Moro, 1997) that A′-extraction of the superscriptional term from precopular position is categorically impossible as well, evincing a restriction more severe than what is imposed on A′-extraction of structural subjects. However, while it is true that all versions of (25b) (based on (25a)) are ungrammatical, this is unlikely to be the reflex of a syntactic ban on A′-movement of the precopular superscriptional term: as (25c) and (26b) demonstrate jointly, such extraction is grammatical in other contexts (questions with a so-called ‘D-linked’ or presuppositional wh-expression, as in (25c), or appositive relatives, as in (26b)), provided that the complementiser that is omitted, avoiding a Comp-trace violation.

(25)

Copular Constructions in Syntax

(26)

Copular Constructions in Syntax

Extraction from and of the precopular term of a binominal copular sentence thus seems to pattern by and large like extraction from and of noun phrases in the structural subject position in general. This confirms that when the superscriptional term occurs in precopular position, it occupies an A-position (contra Heggie, 1988)—more specifically, the structural subject position. The fact that all versions of (25b) are sharply ungrammatical most likely has an extrasyntactic cause: in (25b), how good a player and me compete for the pragmatic role of identificational focus, which is unique per sentence. Relative clauses and questions with D-linked wh-items avoid this information-structural conflict, and come out grammatical. Purely syntactic analyses of (25b) (such as the one advanced in Moro, 1997) would not be able to account for this pattern. While the syntax can take charge of the pattern displayed by (17)–(24), the restrictions at work in (25)–(26b) are best left for information structure to elucidate.

2.3.4 Case

We now move on to case, which is better examined beyond the boundaries of English. In predicational copular sentences, the postcopular predicate may show a case form that is concordial with that of its subject (so-called ‘case concord,’ as in Russian (27a)) or an oblique form (the instrumental case in Russian (27b)).

(27)

Copular Constructions in Syntax

Pereltsvaig (2007) points out that the two sentences in (27) are not in free variation: (27b) says only that Ivan is no longer a soldier; but (27a) induces a lifetime effect: Ivan has to be dead for (27a) to be true. Likewise, the specificational counterpart of (27), where the predicate is in precopular position, may feature instrumental case on the predicate in the past and future tenses (Geist, 2007; Partee, 2010), (28). In Russian equatives, on the other hand, both expressions must take nominative case, which could be taken as evidence that neither expression is predicative, (29).

(28)

Copular Constructions in Syntax

(29)

Copular Constructions in Syntax

The asymmetry between (28) and (29) with respect to case and the presence of the demonstrative element èto points to a split in Russian between predicational and specificational sentences on the one hand, and equatives on the other.

In specificational and equative copular sentences, the postcopular noun phrase often bears a morphological case that is hard to associate with a particular case assigner. In English, this is seen in value-final specificationals that have a pronominal value: though (31a) occurs in highly formal language, (31b) is far more natural. In equative hypotheticals with two pronominal terms, as in (32), the use of the nominative in postcopular position is even more peculiar in present-day English than it is in (31).

(30)

Copular Constructions in Syntax

(31)

Copular Constructions in Syntax

(32)

Copular Constructions in Syntax

In Dutch, specificational copular sentences with a postcopular pronominal value do not give rise to an unexpected case form; the pronoun must be nominative. But Dutch equative hypotheticals show a case picture that is highly reminiscent of the one presented by English (31) and (32).

(33)

Copular Constructions in Syntax

(34)

Copular Constructions in Syntax

For the English accusative occurring in (32b), a common assumption in the literature is that it is the default case: by hypothesis, the postcopular position of an equative or of a specificational copular sentence whose superscriptional term occupies the precopular position is not a structural Case position, so the only case that the postcopular term can be given is default accusative. The Dutch equative hypotheticals in (34) show that this approach is insufficiently general: default case in Dutch is nominative, so the occurrence of a nonnominative pronoun as the second term of (34b) calls for further thought. See Sigurðsson (2006) for detailed discussion of the empirical facts of case in equative hypotheticals, as well as for an analytical perspective.

Specificational pseudoclefts give rise to an interesting case pattern in languages such as German. In specificational pseudoclefts that have the value in postcopular position, the value can show what is called ‘case connectivity’ with the verb in the wh-clause. We see this in (35b). It is interesting to note that while (35a) does, (35b) does not allow reversal of the two terms.

(35)

Copular Constructions in Syntax

2.3.5 Agreement

The contrast between English and Dutch in the previous subsection does not confine itself to the case form of the second term but manifests itself also in the agreement form of the copula: third singular in English (31) and first singular in Dutch (33) (agreeing with the postcopular value). The two differences between English and Dutch in the realm of value-final specificational copular sentences are in all likelihood correlated.

The agreement properties of specificational copular sentences have attracted an enormous amount of attention in the literature (Moro, 1997; Costa, 2004; Sigurðsson, 2006; Heycock, 2012). In Italian, Catalan, Spanish, and Portuguese, for example, agreement is with the postcopular focus (value), and not with the superscriptional term. Kinande (Bantu) similarly exhibits agreement with the postcopular focus (Hedberg and Schneider-Zioga, 2015). French patterns with English; however, it involves an additional complication in the form of the pronominal element ce in subject position.

(36)

Copular Constructions in Syntax

The agreement facts in specificational sentences are not always clear-cut. We refer the reader to Heycock (2012) for extensive discussion of variation in the agreement exhibited by the copula, especially across Germanic.

2.3.6 Ellipsis

Omitting the copula in constructions in which it has a syntactic antecedent elsewhere in the sentence is generally quite free in English and similar languages. Thus, in the second conjuncts of the sentences in (36) and (37), the copula can readily be left out; the relative order of the two nominal phrases exerts no influence on ellipsis of the copula.

(36)

Copular Constructions in Syntax

(37)

Copular Constructions in Syntax

But in the particular subclass of copular sentences formed by specificational pseudoclefts, ellipsis of the matrix copula shows a peculiar sensitivity to (a) the nature of the verb in the wh-clause, and (b) linear order. As Higgins (1979) notes, in specificational pseudoclefts in which the wh-clause is itself copular, the copula cannot be elided when the wh-clause precedes the constituent specifying the value, although copula omission is perfectly acceptable when the order of the two terms is reversed. We see this in the contrast between (38a) and (38b). But copula ellipsis is not categorically impossible in wh-clause-initial specificational pseudoclefts: there is no significant contrast in (39).

(38)

Copular Constructions in Syntax

(39)

Copular Constructions in Syntax

The ungrammaticality of (38b) without the second token of is is remarkable not just because of the grammaticality of (37b) with ellipsis of the copula but also in light of the fact in (40), where pronouns are substituted for the reflexives in (39b) (so that predicational rather than specificational constructions are involved), the output of copula omission is fine.

(40)

Copular Constructions in Syntax

The distribution of copula omission thus gives us a diagnostic for the predicational or specificational nature of copular sentences with a wh-clause as the initial term. But this diagnostic does not generalize to copular sentences with two unambiguously nominal terms: even the unambiguously specificational (36b) is grammatical without copula in the second conjunct.

There is no match, therefore, between the nonfinite copula omission facts catalogued in section 2.3.2, above, and the pattern in (36)–(40). In particular, there is no descriptive generalization to the effect that a word order that resists copula omission in nonfinite contexts will also forbid copula ellipsis in finite contexts. A comparison of the paradigm in (36)–(39) with that in (41)–(44) makes this clear. (On (42), see Williams, 1983. On (43), see Ross, 2000.)

(41)

Copular Constructions in Syntax

(42)

Copular Constructions in Syntax

(43)

Copular Constructions in Syntax

(44)

Copular Constructions in Syntax

Den Dikken (2005, p. 52) provides further discussion of patterns of ellipsis and omission of the subject and predicate across copular sentence types.

2.3.7 Connectivity

As some of the facts reviewed in the previous subsection show, specificational pseudoclefts can include an element in the value for the operator in the wh-clause that is in a dependency with a constituent (usually the subject) of the wh-clause:

(45)

Copular Constructions in Syntax

For Higgins (1979), the presence of such dependencies formed an important diagnostic to tease apart predicational and specificational readings of copular sentences of which one of the terms is a wh-clause. While the examples in (45) are unambiguously specificational, with the wh-clause serving as the superscriptional term and the AP headed by important as its value, the sentence in (46a) is only interpretable with coreference between him and John if important to him is construed as the predicate of the wh-clause. Concomitantly, while (45) shows a free alternation between two linear orders, (46b) is ungrammatical.

(46)

Copular Constructions in Syntax

The so-called connectivity (or connectedness) effect seen in (45) does not stand alone: specificational pseudoclefts give rise to a variety of such effects, including bound variable anaphora (47), ‘emotive should’ connectivity (48), opacity connectivity (49), and polarity-item licensing (50). (For a full inventory of connectivity effects in specificational pseudoclefts, see den Dikken, 2005.) With the exception of polarity-item licensing, connectivity effects manifest themselves entirely regardless of the relative order of the two terms of the pseudocleft.

(47)

Copular Constructions in Syntax

(48)

Copular Constructions in Syntax

(49)

Copular Constructions in Syntax

(50)

Copular Constructions in Syntax

These connectivity effects have given rise to a tremendous amount of analytical activity in the generative framework, motivating many to either start out with the dependent expression (e.g., himself) and its licenser (e.g., John) in the same clause and drive them apart in the course of the derivation, as was common in early generative approaches (see Moreau, 1971; Grosu, 1972, 1973; Chomsky, 1972; Verkuyl, 1972, p. 146; Akmajian, 1979), or to bring them closer together in the interpretive component, collapsing the biclausal surface structure into a monoclausal representation at some level of logical form (Bošković, 1997; Heycock and Kroch, 1999).

Binominal copular sentences (i.e., copular sentences in which neither of the two terms itself harbors a clause into or out of which the surface form can be (re)constructed) exhibit some of the same kinds of connectivity effects (Sharvit, 1999):

(51)

Copular Constructions in Syntax

(52)

Copular Constructions in Syntax

For the connectivity seen in binominal copular sentences such as those in (51) and (52), semantic accounts have been proposed (Jacobson, 1994; Sharvit, 1999) that facilitate the observed interpretive relations without syntactic c-command. Sharvit’s proposal also accounts for some cases of apparent connectivity in predicational copular sentences. Those accounts carry over directly to the pseudocleft connectivity cases in (47)–(49), rendering syntactic derivations in which the italicized elements start out or end up in a local c-command relation unnecessary.

Polarity-item connectivity (as seen in (50)), on the other hand, presents a different picture. Here precisely the same effect cannot be reproduced in binominal copular sentences, (53).

(53)

Copular Constructions in Syntax

The sensitivity of the NPI-connectivity effect to linear order also sets it apart from the anaphor, bound-variable, ‘emotive should,’ and opacity connectivity effects reviewed previously, in both pseudocleft and binominal environments.

(54)

Copular Constructions in Syntax

(55)

Copular Constructions in Syntax

Connectivity effects, particularly NPI-connectivity, have motivated an analysis of specificational pseudoclefts as underlyingly bisentential constructions akin to question-answer pairs, with nonpronunciation or ellipsis of repeated material in the focus (see Bach and Peters, 1968; Clifton, 1969; Emonds, 1970; Ross, 1972, 2000; Seuren, 1985; Kayne, 1998, p. 26; and more recently, den Dikken et al., 2000 and Schlenker, 2003). For a sentence such as (56a), the bisentential approach posits (56b) as its underlier.

(56)

Copular Constructions in Syntax

Under this approach, both the superscriptional term and the value are treated as underlyingly propositional. In English, for example, the superscriptional wh-clause patterns syntactically with indirect questions rather than with free relatives (on the wh-initial order, see den Dikken et al., 2000; on the reverse order, see O’Neill, 2015b). Caponigro and Heller (2007) observe, however, that not all of these facts are replicated crosslinguistically. The bisentential model of pseudoclefts finds additional support from the fact that the structure in (56b) (and its counterparts in languages like French, German, and ASL; Caponigro and Davidson, 2011) can be pronounced in full. Although the seminal work of Higgins (1979) declares such fully bisentential pseudoclefts ‘irremediably anacoluthic,’ corpus and experimental research finds that they occur robustly in English (Lambrecht and Ross-Hagebaum, 2006; Massam, 2013; Guz, 2015; O’Neill, 2015b).

3 Copular Elements

While the preceding sections have focused on the classification of copular constructions, the discussion now turns to a crosslinguistic survey of variation in the distribution and form of copular elements. A good deal of research in typology, morphosyntax, and semantics examines the distribution of different copular elements within and across languages. Variation in copular form cannot be considered independently from variation in the structure of copular sentences, so these issues are considered together in the present section.

3.1 Absent Copulas

While Standard American English requires the copula to be expressed overtly in finite clauses with nonverbal predicates, many languages do not. For example, Chol (Mayan) relates subject and predicate directly, without using a copular element, (57).

(57)

Copular Constructions in Syntax

Chol is not alone in lacking a copula. For example, sign languages typically lack copulas (although see Guðbjörnsdóttir et al., 2015 for arguments that a copula has developed in an Icelandic Sign Language). In other languages, the copular element is optional, as in Cantonese (Yip and Matthews, 1994, cited in Pustet, 2003) and Turkish (Kornfilt, 1997).

More common, however, is that the copula is absent in some subset of grammatical environments. The grammatically conditioned nonoccurrence of the copula in languages which do express a copula in some contexts will henceforth be referred to as ‘zero-copula’ (cf. the copula-less clauses in (12)). Whether languages that never exhibit copular elements have a zero-copula in their grammatical inventory is a controversial question. Alternations between overt and zero-copulas provide valuable insights into the structure of copular constructions and predication more generally.

3.1.1 Tense

One set of accounts of the distribution of zero-copula in finite clauses hinges on the close structural relationship between Tense and the minimal syntactic domain instantiating predication. These accounts fall under the rubric of the support copula (recall (3a)). The assumption generally made by proponents of this approach is that the copula supports the content of the T head of the clause (Baker, 2003; Schütze, 2004, among others). For Baker (2003), the host of tense inflection must be a lexical rather than a functional element. When a lexical predicate is not available (i.e., in sentence types where predicates are not rendered lexical by a process of conflation with the functional head Pred, cf. Bowers, 1993; Hale and Keyser, 1993), the verbal copula is co-opted to perform the duty of hosting tense. Within the tense-support model, a different view is espoused by Doron (1986), Doherty (1996), Becker (2004), den Dikken (2006), and Benmamoun (2008), among others, who deny that the support copula is a lexical element, arguing instead that it is a functional element spelling out the inflectional features of T.

Tense-support analyses predict that a language lacking tense should not induce copula insertion, as Coon (2013) points out. Indeed, Coon (2013) proposes that languages lacking inflectional tense should be zero-copula languages, a prediction borne out, for instance, in Chol and Tagalog (although see Richards, 2009 for an argument that Tagalog has a copula).

As one might expect given the findings for tenseless languages, the presence of a specific tense value often conditions the distribution of zero-copulas. Russian (Partee, 2000; Geist, 2007), Hebrew (Rapoport, 1987; Doron, 1986), Arabic (Benmamoun, 2008), Michoacán Nahuatl (Sischo, 1979, cited in Pustet, 2003), African American Vernacular English (e.g., Labov, 1995), and many other languages have a zero copula only in the present tense, presumed to be the unmarked tense, and thus, the tense least in need of ‘support.’

(58)

Copular Constructions in Syntax

(59)

Copular Constructions in Syntax

(60)

Copular Constructions in Syntax

Stassen (2003, p. 68) objects to the tense-support analysis of the copula on empirical grounds, observing that there are many languages where the overt copula and overt tense inflection are not correlated. Sinhala (Indo-European), for example, uses no copula to mediate nonverbal predication, but it nevertheless has overt tense marking with verbal predicates. Croft (2002) likewise offers the example of Cambodian (Austroasiatic), which lacks tense marking but nevertheless has an overt copula in nonverbal predications.

3.1.2 Person

Another inflectional category implicated in the availability of the zero-copula is person, where the copula is less likely to be retained in the third-person environment. Pustet (2003) point out that third person often “teams up” with present tense as an unmarked category that favors a zero-copula, either variably or categorically. This interaction between tense and person is found in Hungarian (Finno-Ugric; Pustet, 2003), Tarma Quechua (Quechuan; Adelaar, 1977), and AAVE (Green, 2002).

3.1.3 Predicate Type

A crucial domain of crosslinguistic variation in the occurrence of an overt copula is the category and/or lexical semantics of the predicate itself. Informally, typological studies of copular sentences, such as Pustet (2003), observe an implicational hierarchy of predicate types that require a copula. For Pustet (2003), who builds on criteria developed in Givón (1979), the most salient lexical semantic feature responsible for the distribution of copulas is time-stability, cast as an iconic cognitive category. Event concepts are not time-stable, because they denote rapidly changing entities in the real world, while entity concepts are highly time-stable, because they denote entities that do not change from the point of view of human perception. The central observation in this line of work (cf. Croft, 2002) is that the more time-stable the predicate type is, the more marked it is. The formal manifestation of this markedness is the added structural complexity contributed by the copula.

More time-stable predicate types tend to be encoded with a copula, while less time-stable predicate types tend not to combine with a copula. This implication can be recast in traditional categorial terms, assuming that lexical categories are associated with semantic prototypes: NOMINALS > ADJECTIVALS > VERBALS. A verbal predicate usually does not require a copula in order to combine with its subject, whereas if a language has a copula at all, it is likely to occur with nominal predicates.

The following examples bear out the implicational hierarchy. English is a language that copularizes with adjectival and nominal predicates, but not verbal predicates. In Bambara, by contrast, the copula occurs with all three predicate categories. In Basque, there is a split within the verbal category: some verbs take a copula, while others do not, patterning with adjectival and nominal predicates. In Japanese, the split occurs in the adjective category, with some adjectives patterning with nominal predicates and some with verbal predicates. Burmese is a language which requires the copula only with nominal predicates.

The lexical semantic nature of the predicate is also central in Stassen’s (2003) crosslinguistic analysis of intransitive predication. Stassen examines how four basic predicate types (events, properties, classes, and locations) are encoded, and how these predicate types and encoding strategies interact with the distribution of copulas and tense marking. Although the studies in the typological tradition briefly surveyed above distance themselves from the support analysis of the copula, which finds more favor among generative linguists, research in both frameworks finds that the distribution of tense marking interacts in crucial ways with the presence or absence of copular elements.

3.1.4 Copular Construction Type

Since the category and lexical semantics of the predicate conditions the availability of the zero-copula, it is unsurprising that the type of copular sentence also plays a role. Predicational copular sentences allow predicates from a variety of categories, so, naturally, they exhibit the most liberal distribution of zero-copulas. In specificational and equative copular sentences, however, the predicate is generally nominal; moreover, it is generally definite. These nominal predicates fall near the extreme end of the time-stability hierarchy; thus, they are expected to favor overt copulas. We also expect to find similar interactions with morphological categories like tense and agreement.

The latter prediction is straightforwardly borne out for Standard English. Recall that in predicational clauses embedded in ECM or raising contexts, the nonfinite copula to be is not required, while in specificational and equative clauses, the nonfinite copula must occur.

(61)

Copular Constructions in Syntax

(62)

Copular Constructions in Syntax

Many languages with zero-copula in finite clauses exhibit the same pattern, especially when the predicational and equative constructions are contrasted. Consider Hebrew for example (Doron, 1986; Heggie, 1988; Rothstein, 2001), where the pronominal copula is optional in predicational clauses but obligatory in equatives.

(63)

Copular Constructions in Syntax

Languages differ when it comes to the specificational sentence type. For example, Russian specificational sentences pattern with predicational clauses but not equatives in disallowing the (pronominal) copula in the present tense (64) (Geist, 2007), while Hebrew, Garifuna (65),5 and Haitian Creole (66) require the copular element in specificational clauses (DeGraff, 1992; Déprez, 2003).

(64)

Copular Constructions in Syntax

(65)

Copular Constructions in Syntax

(66)

Copular Constructions in Syntax

3.2 Multiple Copula Systems

Another source of variation in the domain of the copular element itself is the existence of multiple copular elements in many languages. In such languages, different copular elements are specialized for different morphological, syntactic, and semantic environments. Copular elements in multiple copula systems are the result of grammaticalization, often of a lexical verb, such as stand, do, make, go, come (Stassen, 2003), or see (Creissels and Taine-Cheikh, 2015).

3.2.1 Tense, Aspect, Mood, and Evidentiality

Even in English, which is not traditionally understood to have multiple copulas (semi-copulas aside), the members of the copula’s suppletive paradigm, originating from separate verbal paradigms (beon vs. wesan), are specialized for finiteness and tense distinctions. Such suppletion is not uncommon in Indo-European languages. Using data from child language and a comparison with AAVE, Becker (2004) argues that the finite copular forms (e.g., is) in fact belong synchronically to a separate paradigm from the nonfinite copular verb be. For related discussion of multiple copulas in Celtic languages, see Carnie (1995), Doherty (1996), Zaring (1996), Adger and Ramchand (2003), Borsley and Roberts (2005), and McCloskey (2005).

In multiple copula systems, a distinction in mood or evidentiality can also condition the distribution of different copular elements. For example, Turkish includes affixal copular markers -(y)dI, -(y)mIș, and -(y)sA, which indicate past tense, indirect evidentiality, and conditional mood, respectively (Göksel and Kerslake, 2004). Lhasa Tibetan offers another example, contrasting copulas of “personal, factual, [and] testimonial” evidence (Hill, 2013).

3.2.2 Predicate Type

Just as the categorial and lexical semantic properties of the predicate correlate with the availability of zero-copula, they also tend to condition the distribution of different copular elements in multiple-copula languages. Aspectual distinctions among predicates can be responsible for the choice of copula. A commonly cited example of such a multiple-copula system is that of Spanish, where, roughly, individual-level predicates use the copula ser and stage-level nonverbal predicates use the copula estar (from Latin stāre ‘to stand’; compare also Italian). Many predicates can occur with either copula, giving rise to different interpretations (see Welch, 2012 for discussion of a similar alternation in Northern Dene languages).

(67)

Copular Constructions in Syntax

Predicate class also conditions the choice of copula in many Bantu languages. Locative predicates often require a distinct copula from other predicate types (McWhorter, 1995). In Kinyarwanda, for example, the copula ri occurs with locatives, while ni occurs with other predicate types (Jerro, 2015). This distinction is also sensitive to person, as it is neutralized with first- and second-person subjects. Still finer-grade distinctions are found in Bambara, which exhibits four specialized copular elements, (68) (Pustet, 2003, 2.75–2.78).

(68)

Copular Constructions in Syntax

3.2.3 Copular Construction Type

Once again, the type of copular construction plays a role in the distribution of different copular elements in multiple-copula systems. If any distinction is made, it is typically between the predicational and equative environment. Russian, for example, has a verbal copula that may occur in predicational clauses in the nonpresent tenses and a pronominal copula that occurs obligatorily in equative clauses (Geist, 2007).

The copula of specificational copular clauses, ever the in-between case, may pattern with either the predicational or equative types. For example, as in (64), neither predicational nor specificational clauses in Russian can take the pronominal copula. In Lakota, on the other hand, Pustet (2003) reports that the copula hécha occurs in predicational contexts, while the copula é is used for specificational and equative sentences. Likewise, two copulas of Kinande, ni and yo, can be distinguished along the lines of predicational versus specificational function, the latter being analyzed as occupying a focus head in Hedberg and Schneider-Zioga (2015). Michaelis et al. (2013) describe Saramaccan as unusual among creoles for having two copulas, ɗε and ɗa, which occur in predicational and specificational environments, respectively. Veenstra (2015) argues that ɗε can also emerge in some specificational clauses, analyzing it not strictly as a predicational copula but as a last-resort mechanism for resolving a selectional clash between υ and a fully categorized predicate.

The existence of such multiple copula systems seems to present a challenge to the simple support analysis of the copula, and an advantage for a rich taxonomy of copulas (recall section 2.1.1). However, this is not the only way to interpret the facts. Since the sentence types under consideration are structurally different, so that the copula associates with different amounts and types of functional structure, it might be expected, under a late morphological insertion perspective (e.g., Distributed Morphology, Halle and Marantz, 1993), that different lexical items with different inflectional features would be suited to different contexts.

3.3 Categories of Copular Elements

A final source of variation in the form of the copular element is the word class of the copula itself. Inflected verbal copular elements are the most familiar cases, discussed at length in section 2, primarily on the basis of Germanic and Romance languages. The verbal copula is implicated in the support analyses cited above, which hinge on a relationship between V/v and T.

Many languages also use nominal copulas, either in addition to or instead of verbal ones. Particles related to discourse markers form the third major category of copular elements. A final minor category, which will not be discussed further here, is the affixal copula, found in Turkish (Kornfilt, 1997) and Blackfoot (Frantz, 1991), among others.

3.3.1 Pronominal copulas

The discussion has already presented several examples of pronominal elements serving as copular elements. In Zoque (van Gelderen, 2009), Hebrew, Russian, and Irish (Carnie, 1995, 1997), for example, a third person pronoun or demonstrative can have a copular function. In Irish, the pronominal copular augment obligatorily co-occurs with the defective verbal copula is in equative and specificational clauses, (69).6

(69)

Copular Constructions in Syntax

The Mandarin Chinese copula shì likewise developed from a demonstrative element (Li and Thompson, 1977). For a detailed discussion of the diachronic development of nominal copulas, see van Gelderen (2009).

3.3.2 Particle Copulas

Another common diachronic trajectory for copulas relates them to focus particles (Stassen, 2003; Harris and Campbell, 2001). For example, the copulas of Garifuna and Haitian Creole (DeGraff, 1992; Déprez, 2003) also occur synchronically in focus-marking contexts. Kinande, as discussed in sections 2 and 3, uses a focus marker as a copula in specificational contexts. This pattern is common in Bantu languages (Heine and Reh, 1984), where a specificational cleft construction may be biclausal, with a verbal copula, or monoclausal, with a focus-particle (see comparison of Kîîtharaka, Abels and Muriungi, 2008, and Shona in Zentz, 2016).

Topic-marking particles are also candidates for copular elements in many languages. In Rotuman, for instance, the particle ne occurs in a range of Topic-Comment structures, including simple nonverbal predicational clauses (den Dikken, 2006). O’Neill, 2015a argues that English is has a topic-particle use in the double-is construction and related sentence-types, (70), while Massam (2013) associates such uses of the English copula with focus-marking.

(70)

Copular Constructions in Syntax

Similarly, the copula exhibits hybrid behavior in the fully bisentential pseudoclefts in (71): its inflection is like that of a verbal copula, while its syntactic and pragmatic distribution patterns with discourse particles (O’Neill, 2015b).

(71)

Copular Constructions in Syntax

4 Semantics of Copular Constructions

Bertrand Russell’s often-quoted words about the semantics of copular constructions, echoing Western debates about the function of the copula, serve as a useful starting point for a brief discussion of the semantics of copular constructions: “It is a disgrace to the human race that it has chosen the same word is for those two such entirely different ideas as predication and identity—a disgrace which a symbolic logical language of course remedies” (Russell, 1919, p. 172).

The lack of consensus about the structure of copular sentence types, surveyed in section 2, cannot be disentangled from the debate about the meaning of these sentence types. Studies of the semantic underpinnings of the copular sentence taxonomy are generally concerned with two questions: (a) what are the semantic types of the two relata in the different construction types, and (b) what is the semantic contribution, if any, of the copula in the different construction types? The central debate in this literature concerns whether the evidence points to the existence of multiple copulas or whether a simple meaningless copula is sufficient. This section will briefly summarize the main findings of this literature. For a more comprehensive review of the semantics of copular constructions, see Mikkelsen (2011).

4.1 Semantic Types of the Two Relata

Copular constructions are distinguished largely by the types of expressions they relate. The status of the relata in the predicational copular sentence type is uncontroversial. As section 2.2.1 showed, the predicational construction relates a referential or quantified expression and a predicate of type <e,t>. In a type-theoretic model based on the types of expressions discussed in Partee (1986), the subject and predicate can compose directly, (72).

(72)

Copular Constructions in Syntax

Several tests show that the subject of a predicational copular clause is a referential element (except when it is a quantified expression). Referential elements are unique in supporting modification by a nonrestrictive relative clause (recall (5)). Mikkelsen (2005) shows that in English and Danish, referential subjects agree with fully inflected pronouns in tag questions (Mikkelsen, 2005), (73), whereas the putative inverted predicates of specificational constructions can only antecede the neuter pronoun, (74).

(73)

Copular Constructions in Syntax

(74)

Copular Constructions in Syntax

The status of the relata in the equative copular construction is more contentious. Intuitively, an equative construction, or identity statement, must relate equivalent expressions; after all, that is what identity means.

(75)

Copular Constructions in Syntax

As in section 2.2.1 with (6), both terms in such equatives are eligible for nonrestrictive modification, suggesting that they are both referential (Rothstein, 1995; Heycock and Kroch, 1999). An fresh example is provided in (76).

(76)

Copular Constructions in Syntax

Although honest generally occurs as a predicative expression, even (75c) may involve two referential expressions if honest serves as a nominalized property as a result of a type-shifting operation (Partee, 1986; Chierchia and Turner, 1988). At a superficial level of analysis, there is no semantic asymmetry between the two expressions in an equative copular sentence.

Despite the identity relation expressed by the equative construction, the two relata in an equative copular sentences are asymmetrical both in syntax and in information function. If indeed both expressions in an equative construction are referential, they may pick out their referents in different ways. Some analyses of equatives deny that both dependents of the copula are referential expressions. Concerning themselves primarily with the syntactic derivation of equatives, Moro (1997) and den Dikken (2006) treat the precopular expression as an inverted predicate. For den Dikken (2006), the expression Cicero, which is itself referential in an equative like (75a), is contained within a null-headed predicate. The three major copular construction types thereby involve the same basic predicational core. On this account, a sentence like (75a) could be paraphrased as a specificational sentence, as in (77).

(77)

Copular Constructions in Syntax

The semantic counterpart to this proposal posits that one of the apparently referential expressions undergoes a type-shift, deriving a predicative expression (Partee, 1986). The type-shifter ident applies to the entity-type expression, for example, Cicero, formalized in (78), and yields a predicate denotation, namely, the singleton set of individuals that are equal to Cicero, which can then be applied to the subject Tully, as in (79).

(78)

Copular Constructions in Syntax

(79)

Copular Constructions in Syntax

Semantics-based accounts of equatives, like the type-shifting approach above, have generally treated at least one of the relata as higher-type expressions. The information function of equative statements corroborates this view, suggesting that their semantics is not purely extensional, but instead involves a layer of intensionality. For instance, identity statements like those exemplified above are useful in situations where an interlocutor is acquainted with an entity or individual under two different descriptions and unaware that the descriptions pick out the same individual, or acquainted with an entity but unaware that it is associated with two different descriptions or names. An approach to this asymmetry is to treat at least one of the expressions as a nonreferential or intensional object. For Jacobson (1994), both relata are functions from individuals to individuals (<e,e>). Others build intensionality into the semantics of only one of the expressions, treating it as an intensional individual or individual concept (Romero, 2005; Comorovski, 2008), an intensional property (Sharvit, 1999), or an intensional proposition or question (Schlenker, 2003).

The status of the two expressions in the overtly asymmetrical specificational construction is, as always, more controversial. Whether the superscriptional term is a predicate, a referential expression, or something else remains a matter of debate. If the superscriptional term is simply a predicate in an unusual structural position (Williams, 1983; Heggie, 1988; Moro, 1997; Mikkelsen, 2005; den Dikken, 2006), then the specificational sentence type collapses with the predicational type. As in (74), Mikkelsen’s (2005) pronominalization test points to a predicative analysis of superscriptional terms like the tallest girl. Likewise, as in (77)–(78), even an expression that is superficially referential may have a predicate-denotation as a result of covert syntactic structure or a type-shifting function (Partee, 1986; Jacobson, 1994; den Dikken, 2006).

If instead there is evidence that the putative predicate has the same logical type as the subject (Jacobson, 1994; Heycock and Kroch, 1999; Sharvit, 1999; Schlenker, 2003), then the specificational category collapses with the equative category. Heycock and Kroch (1999) defend this perspective, in light of the fact that the superscriptional term in sentences like (80a) patterns with the referential expressions exemplified in the equatives (recall (76)) rather than with their uncontroversially predicative counterparts, as in (80b).

(80)

Copular Constructions in Syntax

Table 1 shows the types of “subject” and “predicate” in the three sentence types, highlighting the controversial status of the less referential constituent of the specificational sentence. (The table assumes the prototypical Topic-Comment order for the specificational sentence.)

Table 1: Logical types of relata in specificational copular sentences

Analyses of Specificational Sentences

Precopular

Postcopular

Higgins (1979), Williams (1983), Partee (1986), Heggie (1988), Moro (1997), Mikkelsen (2005), den Dikken (2006)

<e,t>

e

Heycock and Kroch (1999)

e

e

Jacobson (1994)

<e,e>

<e,e>

Romero (2005), Comorovski, 2008

<s,e>

e

Sharvit (1999)

<s<e,t>>

<s<e,t>>

Schlenker (2003)

<s<s,t>>

<s,t>

4.2 Multiple-be vs. One-be Approaches

Russell’s (1919) intuition that the copula contributes different meanings in predicational and equative sentences sets the stage for a multiple-be approach to the copula in modern semantic theories. Moreover, languages with multiple copulas conditioned by the meaning or function of different predicates would certainly seem to have an arsenal of specialized copulas (section 3). Whether or not these intuitions are rooted in the actual semantics of the copula remains an open question. The principal approaches to this question are closely tied to the different options for analyzing the semantics of the two relata in specificational and equative constructions, as discussed in the preceding subsection.

There is a broad consensus that in the predicational context, the copula is vacuous: it simply passes up the denotation of the predicate, which combines directly with the subject.

(81)

Copular Constructions in Syntax

The predication relation can obtain in the absence of an overt copula (e.g., in a bare small clause) or with an overt copula (e.g., in a tensed clause), with no discernible difference in the meaning of the relation (although see Rothstein, 1999 for another view). To accommodate the inverse order of specificational sentences, Williams (1983) proposes that the vacuous copula may take its arguments in either order.

For equatives and specificationals, there is a debate as to whether the copula itself contributes any meaning. Multiple-be analyses are generally those which hold that the two relata in the equative clause must be of the same type; in other words, that neither is a predicate of the other (Russell, 1919; Higgins, 1979; Akmajian, 1979; Higginbotham, 1987; Jacobson, 1994; Sharvit, 1999; Schlenker, 2003; Mikkelsen, 2005; Romero, 2005, etc.). Indirect evidence for a meaningful copula is the fact that specificational and equative clauses are more likely than predicational clauses to require an overt copula (see the discussion in section 2.3.2). If the copula is a meaningful element in specificational and equative copular sentences, an important question is what its meaning(s) would be, such that it is suitable for use in both copular sentence types. One answer is found in Romero (2005), who argues on the basis of a subtle pattern of interpretation in specificational sentences that the copula in such cases is an identity predicate with an extra world argument that selects an intensional object as its structural subject.

The one-be approach, motivated by parismony, takes the vacuous predicational copula to be basic (e.g., Williams, 1983; Partee, 1986, 2010; Adger and Ramchand, 2003; den Dikken, 2006). In these approaches, the vacuous predicational copula is basic (cf. (3a)), and the specificational and equative meanings are derived through type-shifting, either of one of the copula’s dependents, as in (78), or of the copula itself (Geist, 2007). Partee proposes that the copula has the basic vacuous type in (82).

(82)

Copular Constructions in Syntax

In equatives (and possibly specificationals, depending on the analysis of the superscriptional term), where two referential expressions must combine in the absence of an identity predicate, the type-shifter ident is marshaled to resolve the type mismatch. Recall that ident maps a referential expression onto the singleton set that contains it, (83).

(83)

Copular Constructions in Syntax

This type-shift enables the two expressions in an equative to combine by simple function application.

In Geist’s analysis of Russian equatives, ident does not apply to the eventual predicate; rather, it composes with the vacuous predicational copula itself, to yield a two-place predicate of identity, (84) (Geist, 2007, p. 89).

(84)

Copular Constructions in Syntax

Geist (2007) interprets the obligatory copular material that tends to occur in equatives, such as the obligatory presence in Russian of the pronominal copula èto, as overt evidence for the application of ident.

5 Critical Analysis of Scholarship

Regardless of their granularity and irrespective of whether they focus on syntax or semantics, all approaches to the taxonomy of copular sentences that assume a variety of copula types face the same kind of basic criticism for languages such as English: all the audible members of the typology of copulas laid out in (3) sound exactly the same (that is, as be and its inflectional forms). If one’s typology of copulas also includes (3d), the zero-copula, an additional concern is that this element may very well be a figment of the theory, not independently supportable.

If, in light of this criticism, all surface copulas are to be collapsed into a single type, the best option will be to reduce them all to type (3a), the meaningless support copula. For (3a), there is no plausible alternative unless one treats sentences of the type in (1a) as containing two predicates (is and handsome), for which no evidence is forthcoming. A support analysis of the copula is defensible even for languages exhibiting multiple copulas. In a late-insertion morphological framework (e.g., Distributed Morphology; Halle and Marantz, 1993), formally distinct copulas can be specialized for insertion in different morphosyntactic environments, without implying that copulas have lexical content. This leads one to the conclusion that a treatment of the variation in the family of copular sentences in terms of a taxonomy of copular types is not fruitful.

Conceding that the copula is uniformly a meaningless support morpheme is not necessarily tantamount, however, to assuming (a) that all copular sentence types involve the same categories and semantic types of relata; (b) that they have the same underlying structure; or (c) that they uniformly involve a predication relation between the two terms of the construction. First of all, there are clear differences with respect to the category and meaning of the different predicate types involved in copular constructions, from adjectival and locative predicates in simple clauses to indirect questions in certain pseudoclefts, as in Ross (1972), den Dikken et al. (2000), Schlenker (2003), and O’Neill (2015b). If the vacuous copula is polymorphic with respect to the types of the two expressions it relates, as suggested by Partee (2010), it plays the same role in all of these diverse contexts. With respect to (b), the question typically concerns whether the superscriptional-initial order of the specificational copular sentence is base-generated (Williams, 1983; Mikkelsen, 2005) or derived via inversion (Moro, 1997; den Dikken, 2006; Partee, 2010).

Finally, considering (c), the issues examined in section 4 will be revisited as they bear on the inversion question. For equative copular sentences, one could treat the physical copula itself as a support morpheme but resort to an equative predicate for the semantics of identification—perhaps the equivalent of the logical operator ‘=’ (Heycock and Kroch, 1999, section 3.3)—or to a type-shifter applied either to a referential element to yield a predicate (Partee, 1986), or to the copula itself to yield ‘=’ (Geist, 2007). That equative predicate or type-shifter may require licensing, given that it is abstract in English and many other languages, which could account for the obligatory overtness of the support copula in equative copular sentences (recall (13)–(14) in section 2.3.2).

To account for the obligatory overtness of the copula in specificational copular sentences in which the superscriptional term is in initial position, one could either follow an approach along lines similar to what is suggested in the previous paragraph for equatives or assume a predicative underlying representation with inversion of the predicate around its subject and derive the obligatory copula as an integral ingredient of the inversion process (Moro, 1997; den Dikken, 2006). This approach can in principle be extended to equative copular sentences. Postulating a null equative predicate head (Heycock and Kroch, 1999) or a type-shifter (Partee, 1986) and requiring inversion for its licensing (as in den Dikken, 2006) would achieve both, and could account in a uniform way for the syntax of equatives and specificational copular sentences with an initial superscriptional term. Determining the optimal analysis of both awaits further investigation.

On a more general plane, the formal variation among copulas and copular constructions, both within and across languages, provides a useful testing ground for hypotheses about clause structure. The broadest and most fundamental question that copulas and copular constructions bear on is the question of how a subject and its predicate combine. Moro (2000) assumes that subject–predicate combinations are underlyingly symmetrical structures, and goes on to argue that the syntax needs to dissolve this underlying symmetry by raising either the subject or the predicate up in the course of the derivation. This ‘dynamic antisymmetry’ hypothesis has given rise to a significant amount of research in its wake, including Chomsky’s (2008 et seq.) influential discussion of the labeling problem posed by structures in which two phrasal categories are merged together (so-called [XP YP] configurations).

Relevant in connection with the structural signature of subject–predicate combinations is the status of the functional elements that occasionally occur between the subject and its predicate, such as the underlined words in I regard you as my best friend, they took me for a fool, and he looks/seems like the best candidate. If the structural relation between a subject and its predicate is assumed to be symmetrical in the general case, but if words like as, for, and like are treated as spell-outs of a functional element that breaks the symmetry, there ought to be no need for symmetry-breaking movement of either the subject or the predicate whenever one of these words shows up. Predicate raising does indeed seem impossible in their presence: in contrast to the best candidate seems to be John, *the best candidate seems <like> John <like> is ungrammatical, regardless of where like is positioned. But it remains a matter of debate in the literature what the analysis of elements such as as, for, and like should be (see especially Marelj and Matushansky, 2015 for recent discussion of opposing views).

The syntax of predication remains a fertile area for linguistic analysis, with important repercussions for theory construction. Beyond the broad consensus that the subject–predicate dichotomy is real and fundamental to syntax and semantics, very little is beyond debate. Even such profound matters as the (a)symmetry of predication structures and the status and the syntactic and semantic function(s) of copular elements continue to be controversial. This article has presented the general trends in the data and the analytical literature. The reader is strongly encouraged to pursue the many references included throughout the discussion for important details that this overview has not addressed at any length.

Further Reading

Bowers, J. (1993). The syntax of predication. Linguistic Inquiry, 24(4), 591–656.Find this resource:

Carnie, A. (1997). Two types of non-verbal predication in Modern Irish. Canadian Journal of Linguistics, 42, 57–74.Find this resource:

den Dikken, M. (2005). Specificational copular sentences and pseudoclefts. In M. Everaert and H. C. v. Riemsdijk (Eds.), The Blackwell companion to syntax (pp. 292–409). Malden, MA: Blackwell.Find this resource:

den Dikken, M. (2006). Relators and linkers: The syntax of predication, predicate inversion, and copulas. Cambridge, MA: MIT.Find this resource:

Heggie, L. (1988). The syntax of copular constructions. (Unpublished doctoral dissertation). University of Southern California, Los Angeles.Find this resource:

Heller, D. (2005). Identity and information: Semantic and pragmatic aspects of specificational sentences. (Unpublished doctoral dissertation). The State University of New Jersey, Rutgers.Find this resource:

Heycock, C., and Kroch, A. (1999). Pseudocleft connectedness: Implications for the LF interface level. Linguistic Inquiry, 30(3), 365–397.Find this resource:

Higgins, F. (1979). The pseudo-cleft construction in English. New York: Garland.Find this resource:

Mikkelsen, L. (2005). Copular clauses: Specification, predication and equation. Amsterdam: John Benjamins.Find this resource:

Mikkelsen, L. (2011). Copular clauses. In C. Maienborn, K. von Heusinger, and P. Portner (Eds.), Semantics: An international handbook of natural language meaning, vol. 2 (pp. 1805–1829). Berlin: Mouton de Gruyter.Find this resource:

Moro, A. (1997). The raising of predicates: Predicative noun phrases and the theory of clause structure. Cambridge: Cambridge University Press.Find this resource:

Moro, A. (2000). Dynamic antisymmetry. Cambridge, MA: MIT.Find this resource:

Partee, B. (1986). Ambiguous pseudoclefts with unambiguous be. In Proceedings of the North East Linguistics Society, vol. 16 (pp. 354–366). Amherst, MA: GLSA.Find this resource:

Partee, B. (2010). Specificational copular sentences in Russian and English. Russian in Contrast. Oslo Studies in Language, 2(1), 25–49.Find this resource:

Pustet, R. (2003). Copulas: Universals in the categorization of the lexicon. Oxford: Oxford University Press.Find this resource:

Ross, J. R. (1972). Act. In D. Davidson and G. Harman (Eds.), Semantics of natural language (pp. 70–126). Reidel: Dordrecht.Find this resource:

Rothstein, S. (2001). Predicates and their subjects. Dordrecht: Kluwer.Find this resource:

Schlenker, P. (2003). Clausal equations (a note on the connectivity problem). Natural Language & Linguistic Theory, 21(1), 157–214.Find this resource:

Sharvit, Y. (1999). Connectivity in specificational sentences. Natural Language Semantics, 7(3), 299–339.Find this resource:

Stassen, L. (2003). Intransitive predication. Oxford: Oxford University Press.Find this resource:

Williams, E. (1983). Semantic vs. syntactic categories. Linguistics and Philosophy, 6(3), 423–446.Find this resource:

References

Abels, K., and Muriungi, P. (2008). The focus marker in Kîîtharaka: Syntax and semantics. Lingua, 118(5), 687–731.Find this resource:

Adelaar, W. F. (1977). Tarma Quechua: Grammar, texts, dictionary. Lisse: Peter de Ridder Press.Find this resource:

Adger, D., and Ramchand, G. (2003). Predication and equation. Linguistic Inquiry, 34(3), 325–359.Find this resource:

Akmajian, A. (1979). Aspects of the grammar of focus in English. New York: Garland.Find this resource:

Bach, E., and Peters, S. (1968). Pseudo-cleft sentences. Unpublished manuscript, University of Texas.Find this resource:

Baker, M. C. (2003). Lexical categories: Verbs, nouns and adjectives. Cambridge: Cambridge University Press.Find this resource:

Becker, M. (2004). Is isn’t be. Lingua, 114(4), 399–418.Find this resource:

Benmamoun, E. (2008). Clause structure and the syntax of verbless sentences. Cambridge, MA: MIT.Find this resource:

Blom, A., and Daalder, S. (1977). Syntaktische theorie en taalbeschrijving. Muiderberg: Coutinho.Find this resource:

Borsley, R. D., and Roberts, I. (2005). The syntax of the Celtic languages: A comparative perspective. Cambridge: Cambridge University Press.Find this resource:

Bošković, Ž. (1997). Pseudoclefts. Studia Linguistica, 51(3), 235–277.Find this resource:

Bowers, J. (1993). The syntax of predication. Linguistic Inquiry, 24(4), 591–656.Find this resource:

Caponigro, I., and Davidson, K. (2011). Ask, and tell as well: Clausal question-answer pairs in ASL. Natural Language Semantics, 19(4), 323–371.Find this resource:

Caponigro, I., and Heller, D. (2007). The non-concealed nature of free relatives: Implications for connectivity in specificational sentences. In C. Barker and P. Jacobson (Eds.), Direct compositionality (pp. 237–263). Oxford: Oxford University Press.Find this resource:

Carnie, A. (1995). Non-verbal predication and head-movement. (Unpublished doctoral dissertation). Massachusetts Institute of Technology, Cambridge, MA.Find this resource:

Carnie, A. (1997). Two types of non-verbal predication in Modern Irish. Canadian Journal of Linguistics, 42, 57–74.Find this resource:

Chierchia, G., and Turner, R. (1988). Semantics and property theory. Linguistics and Philosophy, 11(3), 261–302.Find this resource:

Chomsky, N. (1972). Studies on semantics in generative grammar. The Hague: Mouton.Find this resource:

Chomsky, N. (1995). The minimalist program. Cambridge: Cambridge University Press.Find this resource:

Chomsky, N. (2008). On phases. In J.-R. Vergnaud, R. Freidin, C. P. Otero, and M. L. Zubizarreta (Eds.), Foundational issues in linguistic theory: Essays in honor of Jean-Roger Vergnaud (pp. 133–166). Cambridge, MA: MIT.Find this resource:

Clifton, E. (1969). The English pseudo-cleft. Unpublished manuscript, Massachusetts Institute of Technology.Find this resource:

Collins, P. C. (1991). Cleft and pseudo-cleft constructions in English. London: Routledge.Find this resource:

Comorovski, I. (2008). Intensional subjects and indirect contextual anchoring. In J. Guéron and J. Lecarme (Eds.), Time and modality (pp. 39–57). Dordrecht: Springer.Find this resource:

Coon, J. (2013). Predication, tenselessness, and what it takes to be a verb. In H.-L. Huang, E. Poole, and A. Rysling (Eds.), Proceedings of the 43rd annual meeting of the North East Linguistics Society (pp. 77–90). Amherst, MA: GLSA.Find this resource:

Costa, J. (2004). Subjects in Spec,vP: Locality and agree. In A. Castro, M. Ferreira, V. Hacquard, and A. P. Salanova (Eds.), Collected papers on Romance syntax, MIT Working Papers in Linguistics (MITWPL). Cambridge, MA: MIT.Find this resource:

Creissels, D., and Taine-Cheikh, C. (2015). From ‘see’ verbs to copulas: A little-known grammaticalization path. Paper presented at Copulas Across Languages. June 18–19, 2015. University of Greenwich.Find this resource:

Croft, W. (2002). Typology and universals. Cambridge: Cambridge University Press.Find this resource:

Déchaine, R.-M. (1995). One be. In M. den Dikken (Ed.), Linguistics in the Netherlands, vol. 12 (pp. 73–88). Amsterdam: John Benjamins.Find this resource:

Declerck, R. (1988). Studies on copular sentences, clefts and pseudo-clefts. Berlin: Walter de Gruyter.Find this resource:

DeGraff, M. (1992). The syntax of predication in Haitian. In Proceedings of the 22nd annual meeting of the North East Linguistics Society (pp. 103–117). Amherst, MA: GLSA.Find this resource:

den Dikken, M. (2005). Specificational copular sentences and pseudoclefts. In M. Everaert and H. C. v. Riemsdijk (Eds.), The Blackwell companion to syntax (pp. 292–409). Malden, MA: Blackwell.Find this resource:

den Dikken, M. (2006). Relators and linkers: The syntax of predication, predicate inversion, and copulas. Cambridge, MA: MIT.Find this resource:

den Dikken, M., Meinunger, A., and Wilder, C. (2000). Pseudoclefts and ellipsis. Studia Linguistica, 54(1), 41–89.Find this resource:

den Hertog, C. (1903). Nederlandsche spraakkunst. Amsterdam: Versluys.Find this resource:

Déprez, V. (2003). Haitian Creole se: A copula, a pronoun, both, or neither? On the double life of a functional head. In D. Adone (Ed.), Recent development in Creole studies (pp. 135–173). Berlin: Walter de Gruyter.Find this resource:

Doherty, C. (1996). Clausal structure and the Modern Irish copula. Natural Language & Linguistic Theory, 14(1), 1–46.Find this resource:

Doron, E. (1986). The pronominal copula as agreement clitic. Syntax and Semantics, 19, 313–332.Find this resource:

Emonds, J. E. (1970). Root and structure-preserving transformations. (Unpublished doctoral dissertation). Massachusetts Institute of Technology, Cambridge, MA.Find this resource:

Frantz, D. G. (1991). Blackfoot grammar. Toronto: University of Toronto Press.Find this resource:

Geist, L. (2007). Predication and equation in copular sentences: Russian vs. English. In I. Comorovski and K. Von Heusinger (Eds.), Existence: Semantics and syntax (pp. 79–105). Dordrecht: Springer.Find this resource:

Givón, T. (1979). On understanding grammar. New York: Academic Press.Find this resource:

Göksel, A., and Kerslake, C. (2004). Turkish: A comprehensive grammar. London: Routledge.Find this resource:

Green, L. (2002). African American English: A linguistic introduction. Cambridge: Cambridge University Press.Find this resource:

Grosu, A. (1972). The strategic content of island constraints. In Working Papers in Linguistics13. Columbus: Ohio State University, Department of Linguistics.Find this resource:

Grosu, A. (1973). On the status of the so-called Right Roof Constraint. Language, 49(2), 294–311.Find this resource:

Guéron, J. (1993). Beyond predication: The inverse copula construction in English. Unpublished manuscript, Université Paris X—Nanterre.Find this resource:

Guðbjörnsdóttir, S. H., Sverrisdóttir, R., Thorvaldsdóttir, K. L., and Jónsson, J. G. (2015). A sign language copula. Paper presented at Copulas Across Languages. June 18–19, 2015. University of Greenwich.Find this resource:

Guz, W. (2015). The structural non-integration of wh-clefts. English Language and Linguistics, 19(3), 477–503.Find this resource:

Hale, K., and Keyser, S. J. (1993). On argument structure and the lexical expression of syntactic relations. In K. Hale and S. J. Keyser (Eds.), The view from Building 20: Linguistics essays in honor of Sylvain Bromberger (pp. 53–109). Cambridge, MA: MIT.Find this resource:

Halle, M., and Marantz, A. (1993). Distributed morphology and the pieces of inflection. In K. Hale and S. J. Keyser (Eds.), The view from Building 20: Linguistics essays in honor of Sylvain Bromberger (pp. 111–176). Cambridge, MA: MIT.Find this resource:

Harris, A. C., and Campbell, L. (2001). Focus and universal principles governing simplification of cleft structures. In J. T. Faarlund (Ed.), Grammatical relations in change (p. 159). Amsterdam: John Benjamins.Find this resource:

Hedberg, N., and Schneider-Zioga, P. (2015). Predication, specification and information structure in Kinande. Poster presented at Generative Linguistics in the Old World (GLOW 38), April 15–18, 2015. Paris, France.Find this resource:

Heggie, L. (1988). The syntax of copular constructions. (Unpublished doctoral dissertation). University of Southern California, Los Angeles.Find this resource:

Heine, B., and Reh, M. (1984). Grammaticalization and reanalysis in African languages. Hamburg: H. Buske.Find this resource:

Heller, D. (2005). Identity and information: Semantic and pragmatic aspects of specificational sentences. (Unpublished doctoral dissertation). Rutgers, New Brunswick.Find this resource:

Heycock, C. (2012). Specification, equation, and agreement in copular sentences. The Canadian Journal of Linguistics/La revue canadienne de linguistique, 57(2), 209–240.Find this resource:

Heycock, C., and Kroch, A. (1999). Pseudocleft connectedness: Implications for the LF interface level. Linguistic Inquiry, 30(3), 365–397.Find this resource:

Higginbotham, J. (1987). Indefiniteness and predication. In E. Reuland and A. ter Meulen (Eds.), The representation of (in)definiteness (pp. 43–70). Cambridge, MA: MIT.Find this resource:

Higgins, F. (1979). The pseudo-cleft construction in English. New York: Garland.Find this resource:

Hill, N. W. (2013). h̩dug as a testimonial marker in Classical and Old Tibetan. Himalayan Linguistics, 12(1).Find this resource:

Jacobson, P. (1994). Binding connectivity in copular sentences. In M. Harvey and L. Santelmann (Eds.), Proceedings from Semantics and Linguistic Theory IV (pp. 161–178). Ithaca: Cornell University Dept. of Modern Languages and Linguistics.Find this resource:

Jerro, K. (2015). Copulas and the semantics of location. In C. Brown, Q. Gu, C. Loos, J. Mielens, and G. Neveu (Eds.), Proceedings of the 15th Texas Linguistic Society Conference (pp. 91–109). Austin: University of Texas Linguistics Department.Find this resource:

Kaisse, E. M. (1979). Auxiliary reduction and the derivation of pseudoclefts. Linguistic Inquiry, 10(2), 707–710.Find this resource:

Kayne, R. S. (1998). Overt vs. covert movements. Syntax, 1(2), 128–191.Find this resource:

Kornfilt, J. (1997). Turkish grammar. New York: Routledge.Find this resource:

Labov, W. (1995). The case of the missing copula: The interpretation of zeroes in African-American English. In L. Gleitman and M. Liberman (Eds.), Language: An invitation to cognitive science, vol. 1 (2d ed., pp. 25–54). Cambridge, MA: MIT.Find this resource:

Lambrecht, K., and Ross-Hagebaum, S. (2006). Apo koinou and intrusion: Toward a syntactic-pragmatic typology of amalgam constructions in spoken English. In Proceedings of the 4th workshop on discourse structure. University of Texas, Austin.Find this resource:

Li, C., and Thompson, S. A. (1977). A mechanism for the development of copula morphemes. In C. Li (Ed.), Mechanisms of syntactic change (pp. 419–444). Austin: University of Texas Press.Find this resource:

Marelj, M., and Matushansky, O. (2015). Mistaking for: Testing the theory of mediated predication. Linguistic Inquiry, 46(1), 43–76.Find this resource:

Massam, D. (2013). Intrusive be constructions in (spoken) English: Apposition and beyond. In Proceedings of the 2012 annual conference of the Canadian Linguistic Association. Toronto: Canadian Linguistic Association.Find this resource:

McCloskey, J. (2005). A note on predicates and heads in Irish clausal syntax. In A. Carnie, S. A. Dooley, and H. Harley (Eds.), Verb first: On the syntax of verb initial languages (pp. 155–174). Amsterdam: John Benjamins.Find this resource:

McWhorter, J. (1995). Looking into the void: Zero copula in the creole mesolect. American Speech, 70(4), 339–360.Find this resource:

Michaelis, S. M., Maurer, P., Haspelmath, M., and Huber, M. (2013). The survey of pidgin and creole languages, vol. 1. Oxford: Oxford University Press.Find this resource:

Mikkelsen, L. (2005). Copular clauses: Specification, predication and equation. Amsterdam: John Benjamins.Find this resource:

Mikkelsen, L. (2011). Copular clauses. In C. Maienborn, K. von Heusinger, and P. Portner (Eds.), Semantics: An international handbook of natural language meaning, vol. 2 (pp. 1805–1829). Berlin: Mouton de Gruyter.Find this resource:

Moreau, M.-L. (1971). Trois aspects de la syntaxe de C’EST. (Unpublished doctoral dissertation). Université de Liège.Find this resource:

Moro, A. (1997). The raising of predicates: Predicative noun phrases and the theory of clause structure. Cambridge: Cambridge University Press.Find this resource:

Moro, A. (2000). Dynamic antisymmetry. Cambridge, MA: MIT.Find this resource:

O’Neill, T. (2015a). Demystifying double-is. In University of Pennsylvania working papers in linguistics: Proceedings of the 38th Penn Linguistics Colloquium, vol. 21.1. Philadelphia: University of Pennsylvania.Find this resource:

O’Neill, T. (2015b). The domain of finiteness: Anchoring without Tense in copular amalgam sentences. (Unpublished doctoral dissertation). CUNY, New York.Find this resource:

Partee, B. (1986). Ambiguous pseudoclefts with unambiguous be. In Proceedings of the North East Linguistics Society, vol. 16 (pp. 354–366). Amherst, MA: GLSA.Find this resource:

Partee, B. (2000). Copula inversion puzzles in English and Russian. Issues in Semantics, 23, 183–208.Find this resource:

Partee, B. (2010). Specificational copular sentences in Russian and English. Russian in Contrast. Oslo Studies in Language, 2(1), 25–49.Find this resource:

Pereltsvaig, A. (2007). Copular sentences in Russian: A theory of intra-clausal relations. Dordrecht: Springer Science & Business Media.Find this resource:

Pustet, R. (2003). Copulas: Universals in the categorization of the lexicon. Oxford: Oxford University Press.Find this resource:

Rapoport, T. (1987). Copular, nominal and small clauses: A study of Israeli Hebrew. (Unpublished doctoral dissertation). MIT, Cambridge, MA.Find this resource:

Richards, N. (2009). Nouns, verbs, and hidden structure in Tagalog. Theoretical Linguistics, 35(1), 139–152.Find this resource:

Romero, M. (2004). Tense and intensionality in specificational copular sentences. In R. B. Young (Ed.), Proceedings from Semantics and Linguistic Theory, vol. 14 (pp. 271–288). Ithaca: CLC Publications, Cornell University.Find this resource:

Romero, M. (2005). Concealed questions and specificational subjects. Linguistics and Philosophy, 28(6), 687–737.Find this resource:

Ross, J. R. (1972). Act. In D. Davidson and G. Harman (Eds.), Semantics of natural language (pp. 70–126). Dordrecht: Reidel.Find this resource:

Ross, J. R. (2000). The frozenness of pseudoclefts—towards an inequality-based syntax. Unpublished manuscript, University of North Texas, Denton.Find this resource:

Rothstein, S. (1995). Small clauses and copular constructions. In A. Cardinaletti and M. T. Guasti (Eds.), Small clauses (pp. 27–48). San Diego, CA: Academic Press.Find this resource:

Rothstein, S. (1999). Fine-grained structure in the eventuality domain: The semantics of predicative adjective phrases and be. Natural Language Semantics, 7(4), 347–420.Find this resource:

Rothstein, S. (2001). Predicates and their subjects. Dordrecht: Kluwer.Find this resource:

Russell, B. (1919). Introduction to mathematical philosophy. London: Allen and Unwin.Find this resource:

Schlenker, P. (2003). Clausal equations (a note on the connectivity problem). Natural Language & Linguistic Theory, 21(1), 157–214.Find this resource:

Schütze, C. T. (2004). Why nonfinite be is not omitted while finite be is. In A. Brugos, L. Micciulla, and C. E. Smith (Eds.), Proceedings of the 28th Boston University Conference on Language Development (pp. 506–521). Somerville, MA: Cascadilla.Find this resource:

Seuren, P. (1985). Discourse semantics. Oxford: Blackwell.Find this resource:

Sharvit, Y. (1999). Connectivity in specificational sentences. Natural Language Semantics, 7(3), 299–339.Find this resource:

Sharvit, Y. (2003). Tense and identity in copular constructions. Natural Language Semantics, 11(4), 363–393.Find this resource:

Sigurðsson, H. A. (2006). The nom/acc alternation in Germanic. In J. M. Hartmann and L. Molnárfi (Eds.), Comparative studies in Germanic syntax: From Afrikaans to Zurich German (pp. 13–50). Amsterdam: John Benjamins.Find this resource:

Sischo, W. (1979). Michoacán Nahuatl. In R. Langacker (Ed.), Studies in Uto-Aztecan grammar, vol. 2, Modern Aztec grammatical sketches (pp. 307–380). Arlington, TX: Summer Institute of Linguistics.Find this resource:

Stassen, L. (2003). Intransitive predication. Oxford: Oxford University Press.Find this resource:

van Gelderen, E. (2009). Cyclical change. Amsterdam: John Benjamins.Find this resource:

Veenstra, T. (2015). Copulas as last-resort roots. Paper presented at Copulas Across Languages. June 18–19, 2015. University of Greenwich.Find this resource:

Verkuyl, H. (1972). The compositional nature of the aspects. Dordrecht: Reidel.Find this resource:

Welch, N. (2012). The bearable lightness of being: The encoding of coincidence in two-copula languages. (Unpublished doctoral dissertation). University of Calgary, Calgary.Find this resource:

Williams, E. (1983). Semantic vs. syntactic categories. Linguistics and Philosophy, 6(3), 423–446.Find this resource:

Yip, V., and Matthews, S. (1994). Cantonese: A comprehensive grammar. London: Routledge.Find this resource:

Zaring, L. (1996). “Two BE or not two BE”: Identity, predication and the Welsh copula. Linguistics and Philosophy, 19(2), 103–142.Find this resource:

Zentz, J. (2016). The biclausal status of Shona clefts. Handout of a talk given at the 2016 annual meeting of the Linguistic Society of America. January 9, 2016. Washington, DC.Find this resource:

Notes:

(1.) There is a different sense in which there may be a typology of copulas—one that recognizes, alongside the common copula be, also a variety of ‘semi-copulas’, including become, remain, seem, appear, which obviously all sound very different. The discussion will be confined to the copula proper. For semi-copulas, questions of semantic typology and distribution of overtness or silence generally do not come into the picture.

(2.) The (non-)referentiality of NP1 is best left out of a reference-based definition of predicational copular sentences; after all, one would want to be able to classify the sentences in (i) as predicational copular sentences despite the fact that NP1 is not referential.

((i))

Copular Constructions in Syntax

(3.) Higgins (1979, 263) points out that if one assumes “that referentiality is a property of expressions, not of the use of expressions, then one might also want to insist that [an identificational copular sentence] is an identity statement.”

(4.) The main text will concentrate on copula distribution. Specificational constructions with a precopular superscriptional term have also been claimed to behave differently from both reverse-order specificationals and predicationals with respect to copula contraction, with particular reference to pseudoclefts (Kaisse, 1979, 708–709).

((i))

Copular Constructions in Syntax

The robustness of the contraction diagnostic is limited, however: neither (ib) nor (iic) is sharply unacceptable, and equatives do not seem to resist copula contraction (see (iid)).

((ii))

Copular Constructions in Syntax

(5.) Garifuna examples are drawn from the fieldwork of the second author.

(6.) See Carnie (1997) for arguments regarding whether (69b) is an inverse predicational sentence—that is, a specificational structure—or an inverse equative.