Brazilian Portuguese prosody is organized in a rhythmic hierarchy which governs the phonology of the language, and which is demonstrated by instrumental and phonological evidence to be present diachronically and synchronically in different speech styles. Principles of stress, which yield two degrees, assign secondary stress to pretonic syllables and no stress to posttonic syllables. From this follows the relationship pretonic ⊃ posttonic—which means that, in a given style, any reductive process necessarily operates posttonically before it can operate pretonically. From this hierarchical relationship, vowel quality and syllable weight are predictable. However, the hierarchy may be disrupted slightly to accommodate the tendency toward stress-timing and an alternating strong/weak pattern.

It is argued that a wide range of extraction and coördination phenomena in English can be accounted for by a simple extension of Categorial Grammar. The same extension will account for a similar range of related phenomena in Dutch, including certain notorious cases of ‘intersecting’ dependency among discontinuous constituents of Dutch infinitival complements. Some universal implications of the theory are considered, and the relation of such grammars to processors which carry out incremental semantic interpretation is discussed.

This paper systematically considers the full range of topic structures in Chinese. Such structures are syntactically characterized by the rule schema S' → X {S, S‘}, where X, the topic, is any major category, and S or S’, the comment, is another topic structure or a sentence which is independently well-formed. Some constituent of the comment, or the comment as a whole, must be related to the topic. If the related constituent is an empty category, it is interpreted not as a variable bound by the topic, but as a pro-form whose antecedent is the topic.

An analysis of naturally occurring data reveals that the FALL–RISE intonational contour makes a context-independent contribution to utterance interpretation: the conveyance of speaker UNCERTAINTY. Two necessary conditions are proposed for the felicitous use of fall–rise: (a) some item referenced in an utterance must be perceived as linked by a scalar relationship with its context; and (b) speaker uncertainty with respect to the scale or scalar value evoked must be plausible. The conveyance of uncertainty constitutes a pragmatic contribution to utterance interpretation, which we analyse in Gricean terms as a case of conventional implicature. Other effects, such as irony or politeness, may be conveyed via this implicature.

The much-debated syntactic status of voici ‘here is’ and voilà ‘(t)here is’ in French relates principally to the theoretical status of syntactic categories and of grammatical subject. The first issue can be properly addressed only if one accepts a minimal theory of constituent structure. The surface distribution of voici/voilà then shows that they are two subjectless finite verbs with a single tense and mood: present indicative. Not surprisingly, they have lost some of these exceptional characteristics in certain dialects of French: they have acquired a personal conjugation in Québec French, and new tenses in Walloon and Franco-Provençal. These facts are relevant to syntactic theories which require all basic sentences to have grammatical subjects at some level of analysis, since no independent parameter is available to account for the systematic absence of surface subjects before a small set of verbs in a non-‘PRO-drop’ language like French.

Typological and theoretical speculations about clitics require that the CLITIC be adequately distinguished from the INFLECTIONAL AFFIX on the one hand and from the INDEPENDENT WORD on the other. The first of these tasks has been attended to, but the second has been slighted, with the result that many items labeled ‘particles’ have been treated as clitics.

After some remarks on the nature of ‘tests’ in linguistics, a series of tests is provided here for distinguishing clitics from independent words. On the basis of these, it is concluded that most of the ‘particles’ in the literature are simply words; from this conclusion, it is argued that treating words with idiosyncratic distributions as acategorial ‘particles’ is wrong.

We then consider the relevance of various cases of ‘particles‘—in German, Chrau, Hidatsa, and Welsh—to theoretical proposals about special clitics. These examples include some items that are independent words, some that are inflectional affixes, and others that are independent words with simple clitic variants.

Finally, a class of DISCOURSE MARKERS is delineated: a grammatical category of items which are often classified as ‘particles’ but which turn out, again, to be independent words rather than clitics of any sort.*

This paper offers a reply to Coopmans' 1984 critique of Hawkins 1979, 1980, 1982. Hawkins had attempted, inter alia, to make typological word order universals relevant to concerns of generative grammar. Coopmans denies this relevance. His critique raises fundamental issues about the nature of language universals and their explanation, and about research methodology. Some of these issues are taken up in this reply. It is argued that Coopmans' dismissal of the typological approach is neither well-founded nor very helpful in the current state of the art. The different methodologies have complementary strengths, and can be of considerable mutual benefit.

The punctuation (accent) system of the Masoretic Hebrew Bible contains a complete unlabeled binary phrase-structure analysis of every verse, based on a single parsing principle. The systems of punctuation, phrase structure, and parsing are each presented here in detail and contrasted with their counterparts in modern linguistics. The entire system is considered as the product of linguistic analysis, rather than as a linguistic system per se; and implications are drawn for the study of written language and writing systems.

This article offers a holistic approach to the understanding of Arizona Tewa passives by examining them from typological, genetic-historical, and areal-historical perspectives. Two types of functional passives, IMPERSONAL and SEMANTIC, are contrastively analysed in terms of their grammatical properties and discourse functions. Genetic-historical comparison reveals the Arizona Tewa semantic passive to be rather anomalous in terms of other Kiowa-Tanoan languages. Areal-historical comparison suggests the influence of Apachean languages. A strategy of combining these distinct perspectives for historical analysis is shown to provide a more precise tool for the anthropological linguistic study of the past.

This paper describes a fairly rich area of complementation in Caribbean English Creole (CEC) within the broad framework of Chomsky's Revised Extended Standard Theory, and compares the description with other recent treatments. The area covered includes a number of complement types in which the CEC particle fi functions. It is argued that the key to understanding the syntax of the complement types lies partly in the variety, and sometimes ambivalence, of fi's functions. Earlier treatments are criticized for failing to clarify these functions, and for assuming too close a parallel between CEC and English complementation. In addition, it is suggested that the present analysis may provide a basis for explaining changes in the grammar of CEC complementation that result from the process of decreolization in continuum situations.

It is well known that, in languages of the world, the passive expression is related to constructions such as reflexives and reciprocals. This paper pays special attention to the less familiar correlations of passives to spontaneous, potential, honorific, and plural formations; and it is argued that presently available characterizations of the passive are unable to offer explanations for these correlations. A satisfactory characterization must include a framework in which one can determine the nature of ‘pseudo-passives’, ‘impersonal passives’ etc. To achieve these goals, a prototype approach is adopted, and the passive prototype is defined. It is also shown that current controversies over whether passives should be analysed as promotional or demotional phenomena find a natural answer when a wider range of passive-related phenomena is examined.

Italian sentences which appear to lack propositional arguments of verbs are shown to be base-generated without these verb complements. This analysis accounts for a wide range of syntactic and semantic data, including the fact that movement rules never appear to have operated out of missing complements. The possibility for a given lexical item to lack such a complement is arbitrary (i.e., it is a subcategorization fact); the so-called missing complement need not be associated with a specific interpretation. The polarity, tense, and person of a verb affect the possibility of its allowing a missing complement.

The conceptual domains REALIS and IRREALIS form an epistemic scale representing a speaker's subjective evaluation of the reality of a given situation. This scale accounts for grammatical phenomena such as conditionals, Japanese complementizer choice, and ‘evidential’ in Turkish, Bulgarian, and Macedonian. Conditionals are identifiable not by their syntactic forms, but by speaker attitudes within the IRREALIS division of the scale. Surprise is also an IRREALIS phenomenon, since it takes time for newly-learned information to be assimilated into one's established body of knowledge.

Two theses concerning word order in Mandarin Chinese have been investigated through a quantified study of written and spoken contemporary Mandarin. It is found, first, that Mandarin is synchronically a typical VO language, in terms of text distribution of VO and OV orders. OV appears at the level of 10% or lower in text, and this is true for both definite and indefinite objects. Further, the functional distribution of OV in both texts suggests that it is an emphatic/contrastive discourse device, having little to do with the contrast between definite and indefinite object. Finally, neither the evidence from our text distribution data nor a comparison with a recent study of the acquisition of Mandarin by native children suggests the existence of a diachronic drift toward SOV order.

In previous work on cliticization, it has been assumed that two and only two strategies are available for clitic positioning: sentential second-position cliticization (Wackernagel's Law) and cliticization to a specified lexical class, most commonly V. Moreover, it has been assumed that the host in terms of which the clitic is positioned, i.e. the structural or lexical host, must also necessarily be the phonological host. This paper shows that the structural and phonological hosts need not be the same; i.e., clitics can simultaneously attach syntactically to a structural host, while attaching independently to a different phonological host. Thus the two-strategy assumption of clitic positioning and attachment is inadequate. Instead, three independent parameters are required. These binary parameters encode two structural notions—dominance and precedence—and one phonological notion—liaison. The values of the parameters constrain possible clitics to eight types, each of which is illustrated in this paper. These parameters are encoded in the lexical subcategorization frame for clitics. It is shown that clitics are phrasal affixes. Although languages appear to differ widely in types of clitics and cliticization, this paper shows how a unitary analysis of apparently diverse clitic types is possible in terms of the three-parameter system. Languages analysed include Classical Greek, Spanish, French, Ngiyambaa, Nganhcara, and some Uto-Aztecan languages.

When ‘marked’ or ‘external’ negation has not been treated as an additional semantic operator alongside the straightforward truth-functional, presupposition-preserving ordinary (‘internal‘) negation, it has been collapsed with internal negation into a unified general logical operator on propositions. Neither of these approaches does justice to the differences and kinships between and within the two principal varieties of negation in natural language. Marked negation is not reducible to a truth-functional one-place connective with the familiar truth-table for negation, nor is it definable as a separate logical operator; it represents, rather, a metalinguistic device for registering objection to a previous utterance (not proposition) on any grounds whatever, including the way it was pronounced.

Conversational coherence is a coöperative enterprise in which speaker and hearer jointly negotiate (a) a focus of attention—a referent—and (b) a response which further selects what aspect(s) of that referent will be attended to. Because not all potential referents can be attended to simultaneously, discourse markers like well help speakers locate themselves and their utterances in the on-going construction of discourse. Analysis of everyday talk shows that well anchors a speaker in a system of conversational exchange when the options which a prior referent has opened for upcoming coherence are not fully met. Thus well is sensitive to the information structure of questions, answers, the underlying conditions of requests, and various participation shifts in talk.

Taiwan Mandarin (TM) is the variety which is learned and used, primarily as a second language, by the people of Taiwan, 80% of whom speak Taiwanese (Tw), a variety of Minnan, as a native language. Taking Peking Mandarin (PM) as the designated standard, this paper identifies those PM features which (a) are not common in TM, e.g. zero markers for future action, lack of contrast between simple past and perfect, the use of verbs or adjectives instead of auxiliary verbs in short answers; and (b) those that are shared by Tw and PM, e.g. the use of an inclusive 1pl. pronoun and the discontinuous A-not-A question forms. In addition to the common phenomena of transference of features from a first language (Tw) to a second language (TM), and the adopting of features that are simple and regular—and therefore easy to process—there is a third important difference between TM and PM. TM displays an obvious drift toward those characteristics which are common to the southern Chinese dialects (Verb-Object languages), and away from those features of Altaic (Object-Verb languages) which are characteristic of PM. The structural and extralinguistic aspects of this trend are examined.

On the basis of recent linguistic changes in TM and Tw, it is claimed that a significant correlation exists between sociolinguistic conditions and the way in which TM has changed. This correlation may be useful in reconstructing the sociolinguistic situation under which Altaic features were brought into PM and other northern Chinese dialects.