This paper proposes an underlying representation for passive sentences in Mojave and in Uto-Aztecan, and explores the broader issues that arise in extending the analysis to other languages and incorporating it in linguistic theory as a substantive language universal. In the introduction, this underlying representation is presented and discussed in relation to previous analyses of the passive. It is claimed specifically that passive sentences in Mojave and Uto-Aztecan are basically impersonal, and derive from structures in which a clause with unspecified subject is embedded as subject complement to the predicate be. Agentive phrases, when they occur, are said to derive from an external source, and are not considered an intrinsic part of the passive construction. The Uto-Aztecan evidence, primarily comparative and diachronic, is presented in §1. It is argued that reflexive constructions commonly assume passive function because both involve ‘non-distinct’ arguments, of which coreferential and unspecified arguments are special cases. The concept of non-distinct arguments receives considerable support from various syntactic changes that occurred in Uto-Aztecan. The Mojave evidence, primarily synchronic, is presented in §2. In §3, a variety of related issues are treated, including potential problems in extending the analysis to English, the source of by-phrases, the nature and status of unspecified arguments, the semantic claims implied by the proposed underlying representation, and the relationship between passive and perfective constructions.

Problems of number inConsistency arise in Italian when singular distributive quantifiers float rightward off of subjects, leaving these subjects plural. The inConsistency arises in the presence of adjectives, participles, and pronominal forms. Different proposals are discussed here for ordering adjective agreement and participle agreement with respect to Quantifier Floating, and different theories of pronominalization phenomena are also discussed. It is shown that significant generalizations about number Consistency can be captured only if a particular ordering for agreement rules and Quantifier Floating is assumed ALONG WITH a particular theory of pronominalization. Thus the data presented here suggest that agreement phenomena and pronominalization must be related.

The principal nominal deictic affixes of Bella Coola, a Salishan language of British Columbia, are examined. Semantically, a three-way distinction is found: Proximal, Middle, and Distal for deictic space; Present, Near Past, and Distant Past for deictic time. The notion of neighborhood and a definition of natural and unnatural circumstance are introduced to explain the usage of the affixes and the observed neutralization of the semantic components of space and time.

One sometimes finds gaps in the morphological systems of languages, such as the lack of a 1sg. non-past in certain Russian verbs. In some Arabic dialects, verbs with identical final radicals are defective: they lack forms where conflicts between rules would yield a phonetically uncomfortable configuration. For some Hungarian verbs in the ik-class which have a stem-final cluster, the grammar disallows epenthesis, so that these verbs lack the paradigms which would normally be formed by a consonantal suffix. However, though such defectiveness is often confined to individually specified words, Halle's extreme views of morphology (1973) cannot be accepted.

A number of syntactic constructions in Mandarin Chinese are analysed—which, synchronically, are unrelated and highly irregular. However, all reflect a diachronic drift from VO to OV typology which has been operating in Mandarin Chinese. In the light of this diachronic drift, the syntactic constructions can be viewed as structures in transition, and the synchronic irregularities as manifestations of the transitional process.

French phonology and morphology (Schane 1968, henceforth FPM) presents a tightly integrated set of arguments for an abstract morphophonemic stress rule in Modern French. This rule depends crucially on an underlying tense-lax distinction in the vowel system, and on various morphological restrictions. However, the FPM rule can be modified and restricted, within Schane's framework, to give a more accurate picture of stress assignment on a morphophonemic level. Such a reformulation leads to a synchronic rule in Modern French that is very similar to the Latin stress rule, thus providing a potential case where a particular generative phonological approach leads to a historically accurate synchronic grammar.

People frequently comment about wine when drinking it; but to what extent do they mean the same things by the words they use ? A structural analysis of the wine vocabulary used by wine experts is given. Experiments involving typical wine drinkers show that there is little consensus in how the words are applied to wines. Moreover, communication tasks show that the sender and receiver of messages about wine perform little better than chance.

A technique has been developed for observing the translation of cognitive input into language in a spontaneous, practiced speech event: descriptions of the lay-outs of apartments.

The great majority of such lay-outs are imaginary tours which transform spatial lay outs into temporally organized narratives. The set of discourse rules which govern these speech events is given, and the rules which determine a number of features of sentence grammar are- outlined. This study represents an initial description of the links between cognitive input, discourse rules, and the rules of sentence grammar.

The study of sign language reveals the effect of the modality of communication on the language system. This paper presents a comprehensive description and discussion of the manifestation of time, space, and person reference in American Sign Language. It is in this area that the effect of the modality on information transmission appears most clearly. Various aspects of space, time, and person are discussed: the manner in which visual language allows for deictic and anaphoric locative, temporal, and ‘pronominal’ reference, the surface manifestation of the conceptualization of time, the specialized use of the dominant and non-dominant articulators, the contrast between the ‘segmental’ nature of oral-language spatial terms and the continuous nature of locative expressions in ASL, and the manner in which verbs may incorporate agent and/or patient and manner adverbials.