This paper is a syntactic investigation into the Double Modals might could, might would and might should in two varieties of American English, an Arkansas dialect and a Tennessee idiolect. Based on inversion and negation facts, the paper concludes that Double Modals in the Arkansas dialect are adverb-modal constructions as, like adverbs, might cannot undergo inversion in interrogative contexts, and cannot occur with clitic negation. In the Tennessee idiolect, on the other hand, Double Modals receive a modal-modal analysis with might as an untensed modal based on the fact that it can invert in interrogatives but cannot occur with clitic negation.

In this paper we examine the distribution and interpretation of quantifiers formed by indeterminate pronouns and disjunction denoting suffixes. The bulk of the paper is devoted to the analysis of Korean where these quantifiers receive an unexpected universal interpretation. We argue that this interpretation can be attributed to the presence of a covert distributive operator. We develop a Hamblin semantics for these cases and show that the interpretations in question fall naturally out of the system. Finally we show that the same patterns can be found in languages such as Chinese and Malayalam, thereby reinforcing the general proposal.

In the adult grammar of English, the phonetically null subject of a temporal adjunct clause is interpreted as referring to the main clause subject NP. This rule is frequently not acquired by children until they are into the school years. A between-subject, test-retest experiment with four-to-five-year old children found no evidence that simple exposure to the temporal adjunct construction promoted use of the adult rule. However, some support was found for an effect of exposure coupled with a morphological clue to the rule.

It has been suggested that a high degree of phonetic reduction serves as supporting evidence for analysing a multi-word phrase as a single lexical unit. I suggest that this does not mean that when we encounter a phrase that appears to function as a pragmatic unit — such as I don't know — and it is highly reduced, we consider the phonetic reduction accounted for by a single-unit analysis. I first show that Dutch verbal phrases which share the pragmatic function of initiating prepositioned self-initiated self-repair also share the phonetic characteristic of a high degree of phonetic reduction, even though a single-unit analysis is only appropriate for some. I then show that similar phrases are not phonetically reduced in the context of dispreferred turns, even though a single-unit analysis is available.

In this paper I consider the notion of contact as a motivation for language change. I present evidence from a community, which, given its changing social conditions over the last half-century, we would predict to be a prime candidate for exhibiting contact-induced language change mechanisms. In a dialect contact situation, we would anticipate reduction of localised features due to levelling and increase in features external to the traditional local variety as a result of diffusion. However, instead of the predicted pattern of change, in this community we observe retention of localised features, suggesting that contact can, in some cases, cause resistance to language change.

This paper combines interactional and phonetic approaches to examine the phonetics-interaction interface of call closings in naturally-occurring talk-in-interaction. Specifically, it investigates the phonetic and interactional properties of one device, namely 'Multi-Unit First Closing Turns' (such as 'yes + okay then'), which British English speakers regularly employ to manoeuvre the call from some on-topic talk into the closing sequence. Such shifts are managed by these turns having a two-unit design in which the first unit attends to the preceding sequence and serves to close it down and the second unit offers and makes relevant the subsequent action of call closure. In addition to commonalities in interactional function, these turns have systematic phonetic designs with various phonetic parameters such as pitch, loudness and glottalisation patterning together in different ways. The most striking phonetic regularity observed, however, is the common occurrence of a click in the boundary between the two units. This paper contributes to our understanding of the organisation of naturally-occurring talk-in-interaction and demonstrates the fruitfulness of conducting context-bound, interactional and phonetic investigations hand-in-hand.