This paper builds on the system of building predication dependencies set out in Adger and Ramchand (2003), providing a semantic characterisation of EPP features (Chomsky 2000) wherein they are treated as syntactic [Λ] features on argument-related heads that map directly to the semantics as a predicate abstraction operator λ. [Λ] features enter into binding dependencies via Agree with [Id] features on V, which are interpreted as variables; they may also be seen here as a featural reinterpretation of θ-roles. The interaction of [Λ] and [Id] is responsible both for (re-)introducing arguments, and for interpreting the structures so created. An extension of the system to cover other kinds of depenencies, such as quantification and control, is sketched.

The problems that repetition poses for utterance interpretation have been the subject of some analysis in the pragmatics literature. Sperber and Wilson's claim that 'the effects of repetition on utterance interpretation are by no means constant', is shown to be particularly apposite when we turn our attention to repetitions produced in naturally-occurring talk. Repetitions are complex phonetic objects whose design has received little analytic attention. As part of an ongoing study of how phonetics relates to the dynamic evolution of meaning within the sequential organisation of talk-in-interaction, we present an analysis of a particular kind of self-repetition.

The practice of repetition we are concerned with is clausal rather than lexical, and exhibits a range of syntactic forms (have another go tomorrow ... have another go tomorrow, it might do ... it might do, it's a shame ... it's a shame). The approach we adopt emphasises the necessity of exploring participants', rather than analysts', understandings of pragmatic inferences in talk and attempts to prejudge as little as possible the relevance of phonetic (prosodic) parameters. The analysis reveals that from a phonetic point of view speakers draw on a range of phonetic features and relationships between features which include tempo and loudness as well as pitch in designing these repetitions. From a pragmatic point of view, it reveals that these repetitions function to close sequences of talk.

Our findings raise a number of theoretical and methodological issues surrounding the prosody-pragmatics interface and participants' understanding of naturally occurring discourse.

In spite of the substantial literature dedicated to it, the status of French subject clitics is still an unresolved issue within morpho-syntactic theory. Two main analyses have been proposed and defended over the past three decades: one advocating that French subject clitics are syntactic arguments bearing a theta-role and the other viewing such clitics as inflectional morphemes on the verb. This papaer demonstrates that the empirical basis motivating the morphological analysis of French subject clitics is much narrower than has been assumed in the literature and shows that the implementation of such an analysis faces numerous theoretical and empirical difficulties. It concludes that the limited similarities between the behaviour of French subject clitics and that of morphemes should be treated as accidental rather than as decisive factors in favour of a morphological analysis.

This paper discusses the formation of suppletive versus non-suppletive verbal stems in Modern Greek within Distributed Morphology (Halle and Marantz 1993). I argue that the formation of non-suppletive stems occurs at the morphological component, whereas the formation of suppletives ones in the vocabulary via morpholexical rules. From a theoretical point of view, I propose that the phonological forms of the vocabulary items — corresponding to suppletive items — stored in the vocabulary are altered by rules which involve reference to morphological and syntactic features. The application of such rules in the vocabulary also proves that its nature is not restricted to a repository of the mapping between morphophonological and morphosyntactic features but it is now extended to an active subcomponent participating in word formation. Consequently word formation is seen as a process requiring the obligatory interaction of syntax-morphology-phonology and the rules applied in the vocabulary.

This paper offers a descriptive and comparative view of quantifiers formed by the combination of wh-words and morphemes denoting disjunction and conjunction in Korean, Japanese and Malayalam. This paper brings together a variety of apparently loosely connected facts about wh-composite quantifiers in the three languages. We examine and compare their distributions and precise interpretations. The observations on wh-composite quantifiers provide a good contrast to English universal quantifiers, existential quantifiers and polarity item any. This crosslinguistic view shows how different languages, using the available morphological resources, realise quantifier paradigms and also help us understand better the ontology and meaning of polarity items, such as English any, which is otherwise very difficult to define.

It is well-known that English in its history changed from predominantly object-verb (OV) in Old English to categorically verb-object (VO) in Modern English. Van der Wurff (1999) made the important observation that the change from OV to VO did not affect all objects at the same time: negative and quantified objects continued to appear in pre-verbal position in Early Modern English, after non-negative non-quantified (henceforth 'positive') objects were all post-verbal. Van der Wurff proposes that up until the beginning of the 15th century, OV word order for all three types of objects was derived in the same way, and he suggests that 15th century English and Modern Icelandic pattern alike in permitting pre-verbal negative and quantified objects in particular syntactic contexts. Using quantitative evidence, in this paper we show, contra van der Wurff (1999), that negative, quantified and positive objects behaved differently and were therefore derived differently during the Old English period. Moreover, it is not the case that the parallel with Icelandic was in place at an earlier stage: we show that negative and quantified objects in Old English do not conform to the pattern of Modern Icelandic.

This paper examines the patterns of use of questions containing the string est-ce que in three varieties of French, as spoken by Canadians, Belgians and French people. The patterns found strongly suggest that the syntax of these 'locutionary' questions is of two distinct types. Canadian speakers have added to the inventory of simple wh-words complex words such as qu'est-ce que and où est-ce que and such forms appear freely in both matrix and embedded contexts. At the other end of the spectrum Belgian speakers provide clear evidence that inversion is productive in their variety of spoken French and that locutionary strings involve such inversion, explaining why these forms are essentially absent from embedded contexts. The speakers from France show a pattern which resembles each of the other varieties in some respects but is taken to represent a restricted use of the inversion strategy, explaining why it is but rarely used in embedded contexts.

This paper looks in some detail at the status of the Extended Projection Principle (EPP) satisfaction. Evidence towards the direction that EPP can no longer express the requirement that a certain position in the clausal architecture ([Spec, IP]) is 'reserved' for DP-lexical subjects is brought in from subjectless clauses, imperatives, object shift and Null-Subject Languages.

The fact that various XPs (i.e. PPs, VPs, etc.) may appear in [Spec, IP] forces us to reconsider the featural requirement EPP expresses. This more general status of the EPP is theoretically supported by recent Minimalist advances (Chomsky 2002) through which EPP is viewed as the only formal trigger for movement in the narrow syntactic component. Additional theoretical arguments that support the more general requirement EPP expresses are derived from recent assumptions which associate EPP-driven movement with discourse-related/semantic effects (cf. Chomsky 2002). I argue, in line with Chomsky (2002), that even though only syntactic features are allowed to operate in NS, yet the expressions present in the derivation may receive an appropriate semantic interpretation (edge phenomena/topic, focus, specific, etc.) by the Semantic Component simply by being dislocated for EPP satisfaction reasons.

Based on Holmberg (2000), I test the hypothesis that EP comprises a D- and a P(honological)-feature. For EPP to be deleted, a phonologically overt syntactic category needs to bear an interpretable D-feature, in order to raise to [Spec, IP] to satisfy both features in one go. This work entertains the assumption that EPP may be present in every single functional projection, since it is the only feature that induces movement in NS.

The empirical basis to test the above theoretical assumptions is offered by verb-initial orders in Null-Subject Languages, and more specifically from Greek. These verb-initial orders in Greek are viewed as the result of VP-Preposing for EPP satisfaction reasons, an analysis that accounts not only for the different word orders, but also for the variability in focus/stress assignment patterns attested in these orders. In effect, I dispense with the generation of Topic and Focus Projection in the clausal architecture, and I also argue that no discourse-related features are allowed to drive operations in syntax, thus allowing NS to retain a certain degree of autonomy.

Since Chomsky (1981), the standard assumption has been that Exceptional Case Marking (ECM) verbs take a reduced complement clause (S'-deletion or TP-complementation). In Tanaka (2002), I argued against this assumption, and showed that ECM verbs in Japanese take a full-fledged CP, and the constructions in questions involve raising to object (RTO). In this paper, I argue that this account, RTO out of a CP, extends to English. In particular, I demonstrate that RTO complement clauses behave similarly to CPs headed by that or for, and not to reduced clauses such as small clauses, with regard to the possibility of heavy NP shift and antecedent contained deletion, either of which can shift a category out of a CP complement clause. As a consequence, I show that antecedent contained deletion crucially involves heavy NP shift.