This volume is a collection of eleven of Richard Kayne's recent articles (presented in order of composition between 2006 and 2010). They deal with micro-parametric variation stemming from the distribution of silent elements (largely within the Determiner Phrase) and potential implications of his antisymmetry hypothesis (Kayne Reference Kayne1994). The empirical focus is generally Germanic (mainly English) and Romance (mainly French and Italian), but the analyses have wider theoretical implications (especially those given in Chapters 6, 8, 9 and 10). All of the articles have been published elsewhere, as detailed in the acknowledgements (xi). However, as only two of the chapters have appeared as journal articles, those interested in comparative syntax will certainly welcome this volume as an accessible collection of Kayne's recent work.
As always, Kayne focuses on finding connections where none appear to exist and on finding differences where none are apparent. The volume is packed with interesting empirical observations and fresh approaches to fundamental syntactic questions. As space prevents an in-depth discussion of individual articles, I first briefly summarise the eleven chapters and then make some general comments about the main ideas emerging from the volume.
In Chapter 1, ‘Some preliminary comparative remarks on French and Italian definite articles’, Kayne focuses on what might be called polidefiniteness effects in French, illustrated in (1).
(1) lelivrelepluscourtthebookthemostshort‘the shortest book’(9, ex. (47))
Building on a proposal by Alexiadou & Wilder (Reference Alexiadou, Wilder, Alexiadou and Wilder1998) for Greek, he suggests that these phrases are relatives of sorts, derived via predicate fronting from a small clause to the specifier of the Complementiser Phrase (SpecCP), followed by the merger of the definite article and remnant small clause preposing, as outlined in (2).
(2) [DP [SC [le livre] ti ]j le [CP [plus court]i C tj]](10, ex. (48))
The superficially distinct Italian counterpart, Kayne argues, contains a null definite article, but is otherwise syntactically identical. This difference is then linked to the more general fact that Italian has bare plurals, whereas French does not, or rather to a P(honological)F(orm)-parameter determining whether a definite D(eterminer) with a filled specifier is pronounced (no in Italian, yes in French), which also accounts for additional differences between French and Italian.
Chapter 2, ‘Several, few and many’, investigates further the role of covert elements in DP. The claim is that few and many are adjectives associated with a null/overt NUMBER/number head (capitalisation indicates that a head is not pronounced) and, similarly, that the mass-modifiers much and little are associated with AMOUNT. The null structure is required by a Universal Grammar constraint which imposes a maximum of one interpretable syntactic feature per lexical item (32). The chapter discusses a plethora of empirical differences between indefinite modifiers in DP, such as the co-occurrence restrictions illustrated in (3)–(4).
(3) too/as/so/how/very few/many books(41, ex. (99))
(4) *too/*as/*so/*how/*very several books(41, ex. (100))
On the basis of differences of this kind, it is proposed that -several lexicalises the structure ‘MORE THAN A several NUMBER’.
Chapter 3, ‘A note on the syntax of numerical bases’, is a discussion of ‘multiplicative’ numerals, such as three hundred. First, Kayne proposes that English approximative multiplicative numerals contain a null nominal suffix equivalent to French -aine ‘-ish’, which forces the presence of the case-marker of:
(5) hundred-AINE-s *(of) books(61, ex. (42))
(6) two hundred (*of) books(61, ex. (43))
But matters are more complex than this. Differences between English several and its French counterpart plusieurs lead Kayne to posit a different nominal suffix -NSFX even for cases like (6) (64). Further evidence for this position comes from Romanian, which requires a preposition with -NSFX.
It is in Chapter 4, ‘On parameters and on principles of pronunciation’ that Kayne is most explicit about his view of cross-linguistic variation. He claims that it is possible to maintain a strong version of what has been called the Borer–Chomsky conjecture (that parametric variation concerns differences in the featural specifications of functional heads) if we allow functional classes, such as feminine gender, to contain the list of nouns which they select. The rest of the chapter focuses on a phase-based account of non-pronunciation, whereby material in the phase edge is not spelled out for principled reasons. Taking this as an axiom, Kayne sketches how this account could be extended to all instances of syntax-sensitive non-pronunciation. The implication is that differences between languages in this respect reduce to differences in movement operations. It should be noted, though, that this approach is not explored further in the book.
Chapter 5, ‘A short note on where vs. place’, is the book's most narrowly focused chapter, which essentially makes the case that where is a (pronominal) determiner, whereas place is simply a noun. Once again a plethora of interesting observations are made, though the text is very sparse in this chapter, making the proposal difficult to follow at times, a point to which I return below.
Chapter 6, ‘Expletives, datives and the tension between morphology and syntax’, proposes that there is a (deictic) D-head, which can combine with null PLACE, THING or PERSON to give a locative/deictic reading. This proposal is extended to cover existential constructions in a range of European languages, the idea being that there is base-generated in a ‘big’ DP with the overt nominal associate. The latter then raises out of DP before the DP-remnant moves past it to the subject position. This proposal has a number of implications. Notably, the remnant DP in subject position will contain a copy of the associate, making Spec–head agreement possible. Moreover, the fact that the same lexical item serves a locative, deictic and expletive function in a number of languages is explained. The remainder of the chapter addresses potential problems, such as certain differences between Italian, French and English concerning the distribution of the respective equivalents of there.
Chapter 7, ‘Some silent first-person plurals’, addresses the behaviour of French on (literally) ‘one’. Kayne argues that the different interpretations of on arise due to covert syntactic differences. Thus first-person plural interpretations of on are triggered by the presence of a null NOUS. Empirical evidence for this proposal comes from binding facts, the distribution of the floated quantifier tous and the distribution of overt co-referential pronouns. Kayne then turns to more general differences between first-, second- and third-person pronouns, and argues that the Italian first-person plural object clitic ci is the same element as the locative clitic, extending his discussion of this clitic in Chapter 6.
Chapter 8, ‘A note on auxiliary alternations and silent causation’, explores intricate differences between auxiliary selection in French and Italian, in particular in relation to anticausatives such as ‘grow’. The analysis provided links this difference to facts about passive causative constructions in the two languages, arguing, surprisingly, that anticausatives are covertly causative (154).
Chapter 9, ‘Antisymmetry and the lexicon’, claims to derive the noun/verb categorial distinction, often taken as a primitive, from antisymmetry (Kayne Reference Kayne1994). The proposal is that singleton set formation can be used to solve the ‘bottom pair problem’ associated with Chomsky's (Reference Chomsky and Webelhuth1995) Bare Phrase Structure. Crucially, only categories lacking uninterpretable features can undergo singleton set formation and so a basic asymmetry is predicted between lexical heads with such features and those without: verbs vs. nouns, respectively. Although verbs appear to form an open class, Kayne argues that this is an illusion created by productive incorporation of N(oun) into v (see Hale & Keyser Reference Hale, Keyser, Hale and Keyser1993), so that all verbs are actually light verbs. The remainder of the chapter explores some of the implications of this proposal, focusing primarily on the implied absence of nominal complements, and proposing that all apparent nominal complements must be rethought as relative clauses.
Chapter 10, ‘Why isn't this a complementiser?’, explores why it is the distal demonstrative that which grammaticalises as a C-element in English (and other Germanic languages), rather than the proximate demonstrative this. Building on the claims made in Chapter 9, Kayne proposes that embedded clauses are actually relative clauses introduced by a relative pronoun that. The remainder of the chapter focuses on justifying (i) that that is a relative pronoun and (ii) why this fails to grammaticalise as a relative pronoun. The reason why that surfaces in all these contexts, rather than this is because this only surfaces in a low post-D position, where a null first-person element fills Spec DP. This null person element blocks extraction of N from DP, ruling out the formation of a relative.
Finally, Chapter 11, ‘Towards an analysis of French hyper-complex inversion’ (written in collaboration with Jean-Yves Pollock), addresses a curious variant of complex inversion where the post-verbal clitic agrees in gender with an object clitic rather than the subject.
(7) Celalagêne-t-elle?(229, ex. (8))this.mher.fbother-t-her.f
Kayne proposes a complex DP analysis of hyper-complex inversion (reminiscent of that in Chapter 6), whereby the clitic is base-generated in a constituent with the object clitic. In his terms, the need for an overt non-clitic preverbal subject in hyper-complex inversion stems from thematic requirements, and the restriction to third-person object clitics is attributed to the claim that only third-person clitics are D-heads in French (240). Finally, the possibility is raised, though not explored, that all instances of phi-feature agreement might actually have a similar syntax, where a large DP is split up via movement.
The volume contains a number of interesting contributions both in relation to recalcitrant syntactic problems (expletive-associate constructions, polidefiniteness, auxiliary selection, pronominal doubling, the structure of numerals and prenominal modifiers, and ‘complementation’), and in relation to more general theoretical debates (parametric variation, the syntax–PF interface, and the nature of lexical categories). Despite the fact that this is a collection of stand-alone articles, there is a degree of coherence to the volume, and many of the articles are quite closely related. In the remainder of this review, I discuss some of the main proposals in the volume and raise some potential challenges to them.
The following axioms appear to be central to the approach taken by Kayne in this volume:
(a) a lexical item can realise only one syntactic feature (32);
(b) homophonous functional items are syntactically identical (Chapters 6, 7, 10);
(c) the non-pronunciation of a lexical item X must be licensed: either by some universal PF-mapping mechanism (i.e. phasal transfer, Chapter 4), or by some language-specific rule (Chapters 1, 3).
While (a) is explicitly stated, Kayne himself notes that it remains problematic in the absence of any independent definition of ‘syntactic feature’. This is no trivial problem. If these features alone are used to motivate null categories such as NUMBER, AMOUNT, etc., then a large part of the syntactician's task is to identify the syntactic features in a given language (or universally). It is by no means clear what Kayne has in mind here. For example, based on the condition in (a), the following structure is proposed for colour adjectives:
(8) a red COLOR car(32, ex. (28))
But, as Jackendoff (Reference Jackendoff1987: 380) has noted, no known grammatical operation is sensitive to colour, and so it is not clear in what sense the latter can be considered ‘syntactic’. This important issue is therefore in need of clarification, as is the connection between syntactic features and semantic interpretation.
Axiom (b) appears to be implicitly assumed in Kayne's discussion of English there and the equivalent Romance clitics as well as in the discussion of this/that in Chapter 10. Though the results in this volume show that the strategy has certain merits, its conceptual basis is not explicitly discussed and so it might best be considered a methodological decision. It remains unclear, as a result, whether functional lexical items can ever change category as is often assumed in diachronic work.
Finally, (c) is exploited in various ways in the volume, most elegantly in Chapter 1, where various asymmetries between French and Italian are reduced to a PF-difference of this kind. Kayne appears to waver on the question of whether the licensing in question is a universal or language-specific mechanism, though. In Chapter 4 he proposes that it is material in the phase edge which is not pronounced, implying that variation between languages in this respect should be modelled by differences in movement operations. If, on the one hand, syntax-sensitive (non-)pronunciation is a language-specific matter, as proposed in Chapters 1 and 3, the implication is that much of narrow syntax might also be constant across languages, the differences between languages stemming from externalisation (cf. Berwick & Chomsky Reference Berwick, Chomsky, Di Sciullo and Boeckx2011). On the other hand, if syntax-sensitive (non-)pronunciation is determined by universal (phase-based) mechanisms, then narrow syntax must differ substantially between languages, notably in relation to movement operations (as also implied by Kayne Reference Kayne1994). It remains unclear to what extent it is possible or desirable to pursue both these possibilities.
To say that the volume is coherent is not to say that it is without internal contradictions of this kind, as is to be expected in a collection of articles spanning several years (here, 2006 to 2010). One problem with the volume, given this fact, is the lack of any introduction and/or conclusion. Such an addition could have resolved issues such as those discussed above and made explicit Kayne's assumptions and core research questions.
A further criticism one might raise in relation to some articles in this volume concerns the connection between data and theory. While strong empirical support is presented in some chapters for the precise analyses posited (as in Chapters 1, 5 and 8), at other times, the empirical predictions of the analyses are not all that clear and the empirical case is not made fully explicit (as in Chapters 2 and 3). Perversely, this is at times due to an overload rather than a lack of data, as Kayne keeps the discussion of the data presented to a minimum, requiring some effort on part of the reader to reconstruct his argument.
Despite these challenges, though, the volume is a mine of useful observations and interesting proposals which will no doubt stimulate further research. Kayne raises far more questions than he answers: could Agree reduce to Move (243, recalling Hornstein Reference Hornstein2009); could the verb/noun distinction be derived from antisymmetry (Chapter 9); should lexical gaps, such as that of shallow in French, be interpreted in parametric terms (Chapter 4); could non-pronunciation be derived from phasal spell-out (Chapter 4)? It is hoped that these questions will be taken up by Kayne and others in future research. If this is the case, then the volume is sure to make an important contribution to research in comparative theoretical syntax.
