The fullest account of English imperative sentences considered in the context of a transformational grammar is contained in Katz & Postal (1964: 74–9). Their account forms part of a discussion of the general thesis that transformations do not affect the meaning of sentences, and of the proposal that follows as a natural corollary of this thesis, that the set of optional singulary transformations be restricted to those usually referred to as ‘stylistic’ transformations. Instead of treating imperative sentences as derived from declarative kernels, therefore, they postulate the occurrence of an imperative morpheme (Imp) in the underlying phrase-markers of imperative sentences. This, they assume, marks these structures as the domain of the imperative transformations and imposes certain selectional restrictions upon them. The notion of the Imp morpheme, which is the most original and most important part of Katz and Postal's analysis of imperative sentences, is adopted in this paper and the occurrence of the morpheme in the deep-structure of all imperative sentences is assumed throughout it. In the rest of their analysis, however, Katz and Postal draw upon more traditional ideas, one of which is disputable.

The ideas of this paper spring from an attempt to grapple with certain phonological problems in a number of languages, and in particular with the special problems which arise when one desires to use a generative, ‘distinctive-feature’ approach. The facts of Tswana phonology summarized below (§5.1) proved especially awkward to handle; when the original twelve features were taken and assigned in a ‘classical’ manner (i.e. [+Vocalic, + Consonantal] for all Liquids, [—Vocalic, —Consonantal] for all Glides, etc.), there were found to be many different ways of characterizing the various segment-types. What was disturbing was that every one of these ways entailed writing phonological rules which failed to highlight the underlying structure, or even obscured it. Conversely, if assignments other than the ‘classical’ were permitted, any attempt to group the segment-types into an arrangement which faithfully reflected the relationships between them (such as Table 8 below) left one with insuperable phonetic problems of feature-assignment. Both these types of difficulty are exemplified in Appendix i.

In Biblical Hebrew, the tendency for rhythmic stress was primarily manifest in stress retraction, which moved word-final stress to the preceding syllable so as to avoid ‘stress clash’ with an immediately following word-initial stress, subject to certain restrictions. It was discussed in detail by the traditional grammarians of Hebrew; recent generative accounts include Prince (1975) and McCarthy (1979). In Modern Hebrew, the tendency for rhythmic stress can be realized in a variety of ways, of which stress retraction is by no means the commonest.

The first concern of this article is an analysis of locative sentences in the Iberian Romances. It is argued that both the existential (〈HAVE〉) and the stative (〈BE〉) construction derive from a single abstract verb. Their differences are based in the presence vs. the absence of an incorporation process over an otherwise identical lexical structure. The second topic of the paper is a study of the behavior of pronominal clitics within these sentences. It is observed that while Catalan has a rich paradigm of clitics (accusative, dative, locative, partitive), languages like Asturian, Galician and certain Spanish dialects resort to a ‘recycling’ strategy in order to palliate the deficiencies of their clitic paradigms. In this respect, we will show how accusative clitics are used as partitive, locative, and even subject clitics. We also propose some of the principles which constrain the application of this strategy. Finally, an Appendix is devoted to certain uses of the accusative clitics as modal markers, also within locative sentences. These uses are closely related with the behavior of certain clitics in Northern Italian dialects.

In an article in this Journal (Specification and English tenses) Crystal has pointed out the hitherto neglected role that adverbial determiners may play in giving unambiguous indication of time where the verbal tense-form itself is not specific enough: ‘It is [in such cases] not a question of tense-form alone giving the relevant distinguishing indication of time, as has been traditionally assumed, but of tense-form with or without adverbial specification which gives unambiguous indication’ (Crystal, 1966: 5). Adverbials may thus become ‘non-omissible determiners’ (a term borrowed by Crystal from Ivić, 1962).

The definite article in many European languages has its origin in a demonstrative or a pronoun. The development into a definite article is a typical case of grammaticalization. In this article I will demonstrate that this kind of grammaticalization, like all kinds of grammaticalization, can be explained as a case of reduction through reanalysis at acquisition. In addition to the prenominal definite article shared with other Germanic languages, the Scandinavian languages also have a postposed definite article. In Old Norse the postnominal definite article is a clitic merged as a head in D, while in its modern descendent Norwegian it is an inflectional suffix checking a grammatical feature in the Infl domain, expressing definiteness within the DP according to general principles of agreement. Thus, so-called ‘double definiteness’ (den gamle hesten ‘the old horse.def’) has become possible as an agreement phenomenon. In Old Norse, the clitic cannot trigger definiteness agreement. This change from a clitic to an inflectional suffix is obviously a case of grammaticalization, but it has wider implications than just the change of morphosyntactic status. ON is shown to have had two projections in the D domain (þau in stóru skip ‘those the large ships’). Later the independent definite article inn was lost and replaced by the demonstrative þann>den. As a result (or cause?) its projection was lost, and the postposed article was left without a free-word counterpart. This, combined with phonological reduction and semantic bleaching, reduced it to an inflectional suffix.

1. The development of transformational theory has been marked by a considerable shifting about of powers allotted to the various constructs. As a theory of lexicon arose, for example, some of the function of the context-sensitivity of the phrase-structure rules was rendered unnecessary; correspondingly, when the rewrite system was tightly confined to a context-free nature, the transformational component took up part of the burden of providing for certain dependencies and concord phenomena. A fair estimate of the history of that component would have to concede that its powers – in spite of the cycle, and some scattered efforts (Emonds, 1969; Postal, 1971; Ross, 1969; Sanders, 1970)–have grown significantly. For example, Emonds's work aside, no general principles of derived constituent structure have developed (as originally anticipated, say, in Lees, 1957b: 400–401); quite the contrary, the particular elementary operations at the root of transformational relationships have been extended, so that at the moment, in addition to sister-adjunction and daughter-adjunction, we have at our disposal Chomsky-adjunction – a range of moves allowing just about any sort of bracketing relation to be developed. Similarly, the recent suggestion that a transformation be considered, most generally, a relation holding between P-markers (not necessarily ‘adjacent’ and not necessarily paired) engenders an enormous increase in power – all the more so since the so-called local and global constraints that are intended to offset the magnification of power are only promissory (Lakoff, 1970).

This paper investigates two ways goals of motion events can be expressed in so-called ‘verb-framed’ languages (Talmy 2000), focusing on the Japanese postpositions -made and -ni. It is typically assumed that these postpositions are both goal-markers, but differ in the exact goal semantics they encode, giving rise to non-overlapping distributions. Based on a range of distributional differences, I argue instead that they are more radically distinct than this: -made marks the endpoint of event participants (including but not limited to paths of motion), while -ni is a dative case that marks the goal argument of motion verbs. This suggests that it is possible for two functionally distinct participant markers to converge and give the appearance of being alternate ways of realizing the ‘same’ participant. Furthermore, adpositions such as -made, an inherently non-motion-encoding resource, represent an understudied strategy for marking goals across languages, something that has ramifications for how motion typologies are constructed.

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.

Over the course of the last twenty-five years, linguistic theory has established itself as providing one of the major routes towards an understanding of the human mind. With increasing precision we have been able to address the problem of articulating in detail the structured capacities the human mind brings to the problem of language acquisition. Along the way there have been doubters, much of the doubt having arisen because of the apparently unbridgeable gap the theory demands between the language user's capacity and the interaction of this capacity with more general cognitive skills. Such doubters as there were received little reassurance from looking at work on performance or language use, for there has been little more than speculative philosophy on the one hand (most notably by Paul Grice, 1975) and articulation of detailed processing mechanisms on the other (e.g. Fodor, 1978; Frazier, 1979; Frazier & Fodor, 1978), without any overall theory. J. A. Fodor (1982, 1983) has led the field in cognitive psychology with his representational theory of mind. But we have till now had no theory of the central cognitive mechanism. Relevance Theory (Sperber & Wilson, 1986) provides us with a contender. It makes specific claims about the central cognitive mechanism and about the relation of natural language to that central mechanism. Against this framework, we are at last able to formulate precise proposals about the grammar-cognition interface, and in so doing provide answers to the psychological reality questions which have gone unanswered for so long: What is the relation of grammars to utterance–interpretation? What is the nature of the link between grammars and the central cognitive mechanism? What is the relation between a speaker's knowledge of his language and his general knowledge?

This is a response to Carston's critique of my account of presupposition-cancellation. While accepting her demonstration that (contra ‘On Horn's dilemma’) presupposition-cancellation does not involve a linguistically encoded contradiction, I show that this is nevertheless consistent with the account of presupposition proposed in Burton-Roberts 1993/7. In fact, Carston's and my accounts of presupposition-cancellation both treat it as involving a pragmatically derived contradiction. I also reconsider the nature of so-called ‘metalinguistic negation’, arguing against Carston that there is a special use of negation (!MN) which involves a use-mention mix and a pragmatically derived contradiction, and is non-truth-functional. I show that, although !MN is echoic, not all echoic negations are examples of !MN.

This paper focuses on the role of modality in determining certain properties of grammaticalization processes in signed vs. spoken languages. The process examined here is the evolution of a case-marked pronoun in the pronominal system of Israeli Sign Language. This pronoun is shown to have evolved from the homophonous noun PERSON. The grammaticalization path leading to the evolution of a case distinction is compared to the evolution of case markers in spoken languages. This comparison reveals that languages in different modalities target different words as sources for grammaticalization. Case markers in spoken languages usually evolve from certain nouns or verbs denoting spatial relations, while in sign languages this is not the case. It is suggested that this difference might be attributed to the scarcity of prepositions in sign languages, and to the iconicity of spatial predicates, which may restrict the possible grammaticalization processes in which they may participate.

Most phonological textbooks and treatises do not define lenition or weakening, as it is also called. Instead they provide a list of examples of processes which they wish to term ‘lenitions’ or ‘weakenings’. It is then hoped that the reader will deduce a correct definition. This is what is done in, for example, Lass & Anderson (1975: 150) and Sommerstein (1977: 228). One of the few available definitions of lenition is to be derived from Vennemann's definition of relative weakness (cited in Hyman, 1975: 165): ‘A segment X is said to be weaker than a segment Y if Y goes through an X stage on its way to zero.’ According to this definition, lenition is primarily a diachronic process affecting the sound structure of languages. It may also have synchronie reflexes (such as initial consonant mutations in the Celtic languages), but these are called ‘lenitions’ by an extension of the term from diachrony.

Taking a cognitive perspective, and concentrating on instances of exophora (or so-called ‘antecedentless’ anaphora), where by definition there is no co-occurring expression in terms of which a given anaphor might be interpreted (i.e. a potential ‘antecedent’), I aim to show, firstly, that so-called exophora falls within the category of anaphora proper and not deixis; secondly, that it is in terms of a conceptual representation of the situation being evoked, and not in terms of the physical situation itself, that the anaphor is interpreted; and finally, that exophora is in reality a more central manifestation of anaphora than the ‘endophoric’ type, where the ‘antecedent’ expression co-occurs with the anaphor.

I will base the discussion on naturally occurring data from French and English, and will consider the contributions of gender- and number-marking within pronominal anaphors, as well as of such features of the anaphoric segment as the argument and referent-order statuses assigned to an anaphor by the governing predicator and its modifiers, and the stress and pitch characteristics of the anaphor. All these features play an important role in the assignment of a full interpretation to so-called ‘endophoric’ anaphors just as much as ‘exophoric’ ones, thereby weakening the theoretical basis for the distinction between the two types.

Much recent work in phonology concentrates on the role of sonority in the phenomenon of vowel reduction, capitalizing on two facts: that reduction involves raising and/or shortening and that higher vowels and schwa are normally interpreted as having low sonority. This paper presents a different approach to vowel reduction in Standard Russian. It is proposed that the apparent sonority-driven effects in Russian are epiphenomenal. In particular, reduction to schwa is outside the domain of phonological computation in Russian, being an artifact of reduced duration. Other types of neutralization arising in vowel reduction are potentially amenable to a sonority-based analysis, but I argue that current approaches to sonority-driven reduction suffer from representational shortcomings. When these shortcomings are rectified, however, sonority is unnecessary as an explicit factor in vowel reduction: standard markedness mechanisms suffice to explain the data.

I. …any noticeable reiteration of the same grammatical concept becomes an effective poetic device. Any unbiased, attentive, exhaustive, total description of the selection, distribution and interrelation of diverse morphological classes and syntactic constructions in a given form surprises the examiner himself by unexpected, striking symmetries and antisymmetries, balanced structures, efficient accumulation of equivalent forms and salient contrasts,… by rigid restrictions in the repertory of morphological and syntactic constituents used in the poem, eliminations which… permit us to follow the masterly interplay of the actualized constituents. Let us insist on the strikingness of these devices (Jakobson, 1968: 602–3).

1. The paraphrase relation or the cognitive synonymy between sentences has played a key rôle in the development of the theory of transformational grammar. It was a remarkable feat of early transformational grammarians to discover that the cognitive synonymy as well as the similarity of certain selectional restrictions between pairs of active and passive sentences can be accounted for by positing single deep structures, and by relating the two types of surface sentences by transformations: I have in mind the restrictions which apply to sentences such as John frightens sincerity, Sincerity is frightened by John.

The publication of Syntactic Structures in 1957 stimulated a much-needed re-evaluation among linguists as to the goals of linguistic theory and the nature of language. Part of the discussion which has ensued has centred around the question of linguistic competence versus performance. Competence has been related to performance as ‘langue’ is to ‘parole’. ‘Competence’ thus refers to the ‘underlying system of rules that has been mastered by the speaker-hearer’ (Chomsky, 1965) and ‘performance’ to the way the speaker-hearer utilizes this ‘internalized grammar’ when he actually produces and understands utterances. Despite the continued controversy about this distinction, little can be added to the justifications for it put forth over many decades (cf. Chomsky, 1957, 1964, 1965; Katz, 1964, 1966; Postal, 1966; Sapir, 1933; Levin, 1965; de Saussure, 1916; etc.). Yet there remains much vagueness as to the limits of each and the relationship between the two. For many years the confusion was due to the influence of Bloomfield who centred his attention on the speech act; his aim was the classification of the OUTPUT of performance, i.e. the utterances, and led to no theory about the dynamic process of performance itself (Bloomfield, 1924, 1926, 1927, 1933). While giving lip service to a concern for ‘langue’, his own mechanistic approach negated any possibility for the rules of ‘langue’ to be anything more than lists of recurrent patterns found in ‘parole’. And since he was of the opinion that ‘the physiologic and acoustic description of acts of speech belongs to other sciences than ours’ (Bloomfield, 1926: 153) he did not direct himself to those aspects of ‘parole’ which could explain speech performance.