This paper examines tensions between the syntax of control and semantic resource sensitivity. Structure sharing of controller and control target leads to apparent RESOURCE DEFICIT under certain circumstances. An analysis is presented using Glue Semantics for Lexical Functional Grammar. It demonstrates that structure sharing and resource sensitivity can be reconciled without giving up or relaxing either notion. It is shown that the analysis can handle either property or propositional denotations for controlled complements. The analysis is extended to finite controlled complements, which raise the opposing problem of RESOURCE SURPLUS. A solution is proposed and its typological implications discussed. The syntax and semantics of control as structure sharing is compared to a recent anaphoric control analysis by Dalrymple (2001). Based on facts of exhaustive and partial control, the present analysis is argued to be superior.

With this issue of the Journal Maggie Tallerman steps down from the editorial team. Maggie has been the review editor since 1994 and we are indebted to her for the enormous contribution that she made to the Journal. We are very pleased to welcome Kerstin Hoge as the new review editor.

A multiple-proposition theory is proposed, according to which a single utterance can express several distinct propositions. An utterance of “Jon, the idiot, missed the train” expresses the official proposition THAT JON MISSED THE TRAIN and the background proposition THAT JON IS AN IDIOT. It is argued that only the official proposition affects the truth-value of the utterance, while the background proposition(s) may be treated similarly to pragmatic presuppositions. The multiple-proposition theory helps us to defend the thesis that, in an utterance such as “Jon1 promised to come but the idiot1 missed the train”, the epithet the idiot must be understood as an attributive anaphor, for it inherits its value from Jon and attributes to Jon the property of being an idiot. It is argued that epithets in such constructions should be viewed in a way similar to quasi-indicators, insofar as the latter are also attributive anaphors. The quasi-indicator she herself in “Jane1 believes that she (herself)1 is rich” inherits its value from Jane and attributes an I-thought to Jane. In focusing on the similarities and differences between epithets and quasi-indicators, we can explain some alleged difficulties, difficulties which suggest that epithets cannot be anaphoric (e.g. “*Jon1 claimed/said/thinks/ … that the bastard1 was honest”). It is shown that the ungrammaticality of sentences like these rests on the fact that the epithet is embedded in what should be a de se attribution (e.g. “Jon1 claimed/said/thinks that he (himself)1 was honest”) while its nature is to contribute to the expression of a de re attribution. This helps us to understand the ungrammaticality of “*Jane1 said/thinks/promised/ … that the idiot1 will come” vs. the grammaticality of “Jane1 said/thinks/promised/ … that she (herself)1 will come” on the one hand, and, on the other hand, the ungrammaticality of “Jon1 ran over a man who was trying to give him (*himself)1 directions” vs. the grammaticality of “Jon1 ran over a man who was trying to give the idiot1 directions”.

Singapore English is a contact language with a constant linguistic substratum and superstratum. It lends itself to an interesting case study on how linguistic neologisms emerge out of a pool of competing features from the typologically distinct languages active in the contact ecology. This paper investigates the aspectual system of Singapore English and that of Chinese, the main substrate language, and of English, the lexical-source language. Despite the presence of competing aspectual categories from the two languages, the aspectual system of Singapore English is essentially the Chinese system filtered through the morphosyntax of English. Substrate influence is systemic, and the competing grammatical subsystems do not mix.

Following Hale & Reiss' paper on the Subset Principle (SP) in phonology, we draw attention here to some unsolved problems in the application of SP to syntax acquisition. While noting connections to formal results in computational linguistics, our focus is on how SP could be implemented in a way that is both linguistically well-grounded and psychologically feasible. We concentrate on incremental learning (with no memory for past inputs), which is now widely assumed in psycholinguistics. However, in investigating its interactions with SP, we uncover the rather startling fact that incremental learning and SP are incompatible, given other standard assumptions. We set out some ideas for ways in which they might be reconciled. Some seem more promising than others, but all appear to carry severe costs in terms of computational load, learning speed or memory resources. The penalty for disobeying SP has long been understood. In future language acquisition research it will be important to address the costs of obeying SP.

In this paper I argue that the lexeme time constitutes a lexical category of distinct senses instantiated in semantic memory. The array of distinct senses constitutes a motivated semantic network organised with respect to a central sense termed the SANCTIONING SENSE. The senses associated with time are derived by virtue of the interaction between the Sanctioning Sense, conceptual processing and structuring, and context. Hence, semantic representations, cognitive mechanisms, and situated language use are appealed to in accounting for the polysemy associated with time. The model adduced is termed PRINCIPLEDPOLYSEMY. The conclusion which emerges, in keeping with recent studies in lexical semantics, most notably Lakoff (1987), Pustejovsky (1995), Tyler & Evans (2003) and Evans (2004), is that the lexicon is not an arbitrary repository of unrelated lexemes; rather, the lexicon exhibits a significant degree of systematicity, and productivity. In order to adduce what constitutes a distinct sense, I introduce three criteria: (1) a meaning criterion, (2) a concept elaboration criterion and (3) a grammatical criterion. A further claim is that the lexicon exhibits significant redundancy. This position is at odds with SINGLE-MEANINGAPPROACHES to polysemy, which posit highly underspecified lexical META-ENTRIES, such as the generative approach of Pustejovsky (1995) or the monosemy position of Ruhl (1989). That is, I propose that lexical items constitute highly granular categories of senses, which are encoded in semantic memory (=the lexicon). This necessitates a set of criteria for determining what counts as a distinct sense without deriving a proliferation of unwarranted senses, a criticism which has been levelled at some studies of word-meaning in cognitive linguistics (e.g. Lakoff 1987).

The meter of Greek lyric poetry shows great variation within and between lines regarding the shape, number and combinations of basic metrical units. We offer a simplifying analysis in terms of markedness, in which meters are defined by distinctive violations of linguistic constraints controlling rhythm, layering, binarity, and alignment. The constraints that are distinctively violated in meter are low ranked in the phonology of Greek.

Cantonese exhibits a pattern of variation among verbs that has often been interpreted as distinguishing a category of adjectives or a subcategory of adjectival verbs. However, neither of these approaches takes into account the complex patterns of overlap among the purported categories or subcategories. To account for these patterns, we propose a multi-dimensional, feature-based analysis, whereby morphological, phonological, syntactic, and semantic features interact to determine the distribution of each verb. While all verbs bear the same syntactic category feature, there are other features that affect the distribution of verbs independently of syntactic category. For example, constructions that resemble adjectival constructions in other languages license the semantic classes of verbs that are permanent, gradable, and/or non-dynamic, while constructions that resemble verbal constructions in other languages license the semantic classes of verbs that are dynamic, non-gradable, and/or non-permanent. Typological implications of this analysis are also considered.

This article argues that of English adjective–noun constructions involving associative adjectives (‘associative AdjNs’), some originate in the lexicon and others in the syntax. While in many cases such constructions are unambiguously and for identifiable reasons located on one side or the other of the lexicon–syntax ‘divide’, variation being possible only across speakers, a range of associative AdjNs is identified which must be simultaneously, and for the same speakers, of both lexical and syntactic provenance. There is therefore no lexicon–syntax ‘divide’: the two modules overlap.

In the first half of the twentieth century, English grammar disappeared from the curriculum of most schools in England, but since the 1960s it has gradually been reconceptualised, under the influence of linguistics, and now once again has a central place in the official curriculum. Our aim is not only to document these changes, but also to explain them. We suggest that the decline of grammar in schools was linked to a similar gap in English universities, where there was virtually no serious research or teaching on English grammar. Conversely, the upsurge of academic research since the 1960s has provided a healthy foundation for school-level work and has prevented a simple return to old-fashioned grammar-teaching now that grammar is once again fashionable. We argue that linguists should be more aware of the links between their research and the school curriculum.

Kam, a Kadai language spoken in Guizhou province (People's Republic of China), has a family of intransitive possessive constructions with the word order ‘Possessor–Verb–Possessee’. (The basic word order in Kam is SV and AVO.) While two recent papers have featured this unique construction type for an array of other Southeast Asian languages, they fail to acknowledge its distinct semantic value in contrast to the related construction type ‘Possessee–Possessor–Verb’. The former construction type displays a so-called ‘zoom-effect’: the possessor is predicated IN, AT or THROUGH his/her/its possessee; the predication zooms from the possessor on his/her/its possessee. The latter construction, in contrast, views the possessee as an entity separated from its possessor, and the predicate as applying solely to the possessee. After illustrating the ‘zoom-effect’ for a representative sample of Kam constructions, I demonstrate that ‘zoom-effects’ do not merely exist when the possessee–possessor compound has the zero-role (=intransitive subject) as above, but also when it assumes other semantic roles (e.g. patient, force, etc.). A general definition of this construction type, called ‘zoom-on-possessee construction’, is proposed; it enables us to unify and account for an array of hitherto disparate construction types that run in the literature under labels such as ‘proprioceptive state expressions’, ‘body part locative constructions’, ‘dative of affect’, etc. Furthermore, I discuss in some detail whether zoom-on-possessee constructions are better accounted for within a multi-stratal or a mono-stratal framework and, finally, whether the concept of noun-incorporation has any relevance.

When the sizes of language families of the world, measured by the number of languages contained in each family, are plotted in descending order on a diagram where the x-axis represents the place of each family in the rank-order (the largest family having rank 1, the next-largest, rank 2, and so on) and the y-axis represents the number of languages in the family determining the rank-ordering, it is seen that the distribution closely approximates a curve defined by the formula y=ax−b. Such ‘power-law’ distributions are known to characterize a wide range of social, biological, and physical phenomena and are essentially of a stochastic nature. It is suggested that the apparent power-law distribution of language family sizes is of relevance when evaluating overall classifications of the world's languages, for the analysis of taxonomic structures, for developing hypotheses concerning the prehistory of the world's languages, and for modelling the future extinction of language families.

Korean has a very complex phonology, with many interacting alternations. In a coronal-/i/ sequence, depending on the type of phonological boundary present, alternations such as palatalization, nasal insertion, nasal assimilation, coda neutralization, and intervocalic voicing can apply. This paper investigates how the phonological patterns of Korean affect processing of morphemes and words. Past research on languages such as English, German, Dutch, and Finnish has shown that listeners exploit syllable structure constraints in processing speech and segmenting it into words. The current study shows that in parsing speech, listeners also use much more complex patterns that relate the surface phonological string to various boundaries.

In December 1999, as part of its tricentenary celebrations, the Berlin Academy of Sciences invited eleven speakers to discuss the Origin of Language (cf. the Trabant & Ward volume, henceforth T&W). In March 2000 a workshop (Crow 2002a) under the auspices of the British Academy and the UK Academy of Medical Sciences, ‘The speciation of modern Homo sapiens’, addressed the same problem. The speciation of Homo sapiens and the origins of language are surely two sides of the same coin. At about the same time, the Christiansen & Kirby volume on Language Evolution (henceforth C&K) was conceived at the Fifth Australasian Cognitive Science Conference. Together the contributions of these volumes constitute a substantial contemporary archive on the origin of language. Their publication provides an opportunity to review the status of attempts to account for the evolution of language. Do the contributions converge on a solution?

Robyn Carston,Thoughts and utterances: the pragmatics of explicit communication. Oxford: Blackwell Publishers, 2002. Pp. x+408.

Robyn Carston's Thoughts and utterances is, in every sense, a big book. The result of years of thinking about communication from a relevance-theoretic perspective, it is long, wide-ranging, rich, intricate, demanding and radical. And it is (or should be) controversial. Essential reading therefore for anyone with any interest in language and communication, philosophy of language, semantics and pragmatics.

Katalin É. Kiss,The syntax of Hungarian (Cambridge Syntax Guides). Cambridge: Cambridge University Press, 2002. Pp. xii+278.

László Varga,Intonation and stress: evidence from Hungarian. Houndmills: Palgrave MacMillan, 2002. Pp. xvii+229.

László Hunyadi,Hungarian sentence prosody and universal grammar. Frankfurt am Main: Peter Lang, 2002. Pp. 328.

It seems it's harvest time in Hungarian linguistics: a number of respectable publishers have in recent years put out books on various aspects of the Hungarian language by authors working in or outside Hungary. In addition to the monographs under review here, one can mention Szabolcsi (1997), Kenesei, Vago & Fenyvesi (1998), Koopman & Szabolcsi (2000) or Siptár & Törkenczy (2000).