Split intransitivity has been an important issue in syntactic theory since the postulation of the Unaccusative Hypothesis in Relational Grammar in Perlmutter 1978; it was adapted into Government-Binding theory in Burzio 1981. In both theories, a purely syntactic approach to split intransitivity is taken. This paper argues that split-intransitive phenomena are better explained in semantic terms. The semantic analysis will be carried out in Role and Reference Grammar, which assumes the theory of verb classification proposed in Dowty 1979. The analysis will focus on Italian, Georgian, and Acehnese, languages which have been cited as providing support for the UH. It will be shown that two semantic parameters, inherent lexical aspect (Aktionsart) and agentivity, underlie split intransitivity crosslinguistically.

Nouns are predicates. In Southern Tiwa, nouns occur as serial predicates (‘incorporation’ is serialization). Besides their null pronoun argument, they can also license an argument in the role of possessor (‘Possessor Ascension’ is a superfluous notion). These conclusions emerge from a new morphosyntactic analysis, which also entails that the verb agrees with all and only final terms (multistratal reference is unnecessary). The paradigm gaps, constraints on incorporation, and limited conditions for optional passive, though seemingly intricate, all point to a single hierarchical principle. This hierarchy calls for a simple but previously unexplored formalism, a tiered model reminiscent of autosegmental phonology.

Texts in Classical Nahuatl of the year 1524, in the genre of formal oratory, reveal extensive use of lines showing parallel morphosyntactic and semantic structure; this parallelism expresses likeness rather than contrast, and is organized in binary sequences which display successive embedding. Analysis and translation of an extended passage point to (a) the applicability of structural analysis to ‘expressive’ as well as ‘referential’ texts; and (b) the importance of understanding oral literatures not just in terms imposed by translators, but in the light of their inherent esthetic principles.

Recent research has examined the early stages of language development in signed and spoken language. In this paper we discuss claims that, in the timing of early language milestones, there is an advantage for the acquisition of signed languages. In particular, we will review the evidence that the emergence of the first signs and of the first two-sign combinations is precocious in signing children as compared to the appearance of the first words and of the first two-word strings in speaking children. We conclude that the evidence for early emergence of the first signs deserves continued attention, but that the evidence for early syntax in signing children is not at all compelling. Finally, we discuss the import of a short-lived sign advantage in early language acquisition, in terms of understanding maturational timing mechanisms more generally. Specifically, we address the question of whether or not a single timing mechanism underlies early milestones in the acquisition of both vocabulary and syntax.

While descriptions and analyses of tonal downstep abound, considerably less attention has been devoted to cases involving tonal upstep. This is no doubt due to the relative scarcity of languages that exhibit upstep. The present study examines certain tonal phenomena in Krachi, a language with two phonologically contrastive pitch levels. I argue that the Krachi data are best analyzed as involving an upward shifting of the tonal register (upstep). I then argue that upstep in Krachi provides evidence for the inclusion of a register tier in tonal theory.

This study presents evidence for the psychological reality of the Subset Principle as a determinant of the acquisition of Governing Category Parameter settings for English anaphors and pronominals. We used prelingual deafness as a natural experiment in the acquisition of English in the presence of variably impaired access to English language data. Previous work has shown that deaf adults with low proficiency in English typically know the unmarked properties of English but have specifically less knowledge of the marked properties than more proficient deaf adults. In the present study we demonstrate that deaf adults with relatively low English language proficiency possess the correct governing category for English anaphors but an incorrect, larger governing category for English pronominals. This pattern respects the markedness predictions of the Subset Principle. These individuals appear to select grammars which generate a smaller set of sentences than the grammars of more proficient English language users. The results support the proposal that, when a parameter's values determine languages that are ordered as proper subsets, the learner selects the value which determines the smallest language consistent with the input data.

The Anatolian branch of Indo-European is characterized by a split-ergative case-marking system in which neuters inflect ergatively and common-gender nouns inflect accusatively; its ergative case originated via the reanalysis of an unproductive neuter instrumental marker in null-subject transitive clauses. A development from instrumental to ergative also occurred in the prehistory of the Gorokan languages of Papua New Guinea, and it is suggested that this process is a general mechanism for the development of split ergativity of this type. The well-known NP hierarchy discovered by Silverstein receives a natural interpretation as a hierarchy of instrumentality.

X-bar theory is widely regarded as a substantive theory of phrase structure properties in natural languages. In this paper we will demonstrate that a formalization of its content reveals very little substance in its claims. We state and discuss six conditions that encapsulate the claims of X-bar theory: Lexicality—each nonterminal is a projection of a preterminal; Succession—each Xn+1 dominates an Xn for all n = 0; Uniformity—all maximal projections have the same bar-level; Maximality—all nonheads are maximal projections; Centrality—the start symbol is a maximal projection; and Optionality—all and only nonheads are optional. We then consider recent proposals to ‘eliminate’ base components from transformational grammars and to reinterpret X-bar theory as a set of universal constraints holding for all languages at D-structure, arguing that this strategy fails. We show that, as constraints on phrase-structure rule systems, the X-bar conditions have hardly any effect on the descriptive power of grammars, and that the principles with the most chance of making some descriptive difference are the least adhered to in practice. Finally, we reconstruct X-bar theory in a way that makes no reference to the notion of bar-level but instead makes the notion ‘head of the central one.

An analysis of a corpus of naturally-occurring data reveals that verb phrase preposing serves two functions in discourse: to affirm a speaker's belief in a salient proposition explicitly evoked in the prior discourse, or to suspend a speaker's belief in such a proposition. Various types of VP preposing are identified, based on the type of discourse connective that introduces the preposing and on the epistemic status, relative to the speaker, of the affirmed or suspended proposition. The felicity conditions on VP preposing are shown to be related to those of other OSV constructions in that (1) a particular anaphoric relation must obtain between the preposed element and the preceding discourse, and (2) the preposing must effect the instantiation of a salient open proposition presupposed in the discourse.

In the process of communicating, conversationalists constantly make decisions about their interlocutors' state of knowledge, and on the basis of these decisions make lexical, grammatical, and intonational choices about how to manage the ‘flow’ of information. This paper focuses on how such decision-making affects choices in relative clause constructions in American English conversations. On the basis of a quantitative analysis of a corpus of natural conversations, we show that the structural choices in relative clause constructions are best explained as symptoms of interactants' attention to information flow.

Two curious phenomena which have to this point been treated separately in the literature are seen to be consequences of the same general interpretive mechanism. We propose that the theory of focus not only accounts for the ‘definiteness restriction’ with respect to material extraposed from NP, but also contributes crucially to an explanation for the variable acceptability of sentences containing extractions from extraposed PPs.

The oft-noted genetic and typological diversity of the languages of the New World is due not to accidents of history but to the operation of regular principles of linguistic geography. Four building blocks of a theory of linguistic diversity are presented here: a classification and measures of diversity, an account of the causes of diversity, measures of rates of differentiation, and means of estimating the age of a linguistic population. On the most conservative archeological view, the New World was first colonized some 12,000 years ago; on the received view, up to 20,000 years ago. But no high-latitude area like Siberia, Beringia, Alaska, or northern Canada could have contributed or transmitted the linguistic diversity to seed the New World in these time frames, and no single dispersing family could have given rise to the attested diversity in these time frames. The unmistakable testimony of the linguistic evidence is that the New World has been inhabited nearly as long as Australia or New Guinea, perhaps some 35,000 years. Genetic unity for ‘Amerind’ is incompatible with the chronology demanded by the linguistic facts.

The theory developed here is that quotations are demonstrations that are component parts of language use. Demonstrations are unlike descriptions in two main ways. They are nonserious rather than serious actions. A person demonstrating a limp isn't actually or really limping. And they depict rather than describe their referents, though they depict only selected aspects of the referents. The demonstrator of the limp depicts some but not all of its aspects. Quotations, we argue, have all the properties of genuine demonstrations. They too are nonserious actions and selective depictions. For evidence we appeal to a wide range of phenomena in spontaneous spoken and written quotations.

Verb agreement in the K'iche' agentive voice appears to deviate from the general ergative/absolutive system that K'iche' shares with other Mayan languages. This has led some to treat agreement in the agentive construction as falling outside the scope of the regular agreement system. It has also led to differing views regarding the appropriate syntactic representation of the agentive construction with respect to final transitivity vs. intransitivity. In this article we examine facts relevant to these issues and propose that the agentive voice is a detransitivizing antipassive construction. Recognition of this has the following consequences: (i) we can state a unified condition for the antipassive morphology, and (ii) we demonstrate that this nonregular agreement can be treated as a special case of the general K'iche' agreement system by adopting a lateral feature-passing analysis of the type proposed in Aissen 1987a. The proposed analysis also provides empirical justification for a heretofore somewhat weakly motivated claim of Relational Grammar, Postal's (1977) demotion analysis of antipassive.

Proposals for foot structure in Japanese have hitherto been limited to versification and to mechanisms for assigning tone that bear little resemblance to stress feet. This is not surprising, since the metrical foot has thus far been associated with stress, which Japanese appears to lack. I argue here that a bimoraic foot whose properties are similar to those of stress feet in other languages plays a significant role in Japanese morphophonology. The existence of such feet in Japanese has a number of implications, the most important of which is that it suggests that in Japanese a rhythmic system like that underlying many stress systems may coexist with the independent pitch accent system.

The child's creative contribution to the language-acquisition process is potentially most apparent in situations where the linguistic input available to the child is degraded, providing the child with ample opportunity to elaborate upon that input. The children described in this paper are deaf, with hearing losses so severe that they cannot naturally acquire spoken language, and their hearing parents have chosen not to expose them to sign language. Despite their lack of usable linguistic input, these children develop gestural communication systems which share many structural properties with early linguistic systems of young children learning from established language models. This paper reviews our findings on the structural properties of the deaf children's gesture systems and evaluates those properties in the context of data gained from other approaches to the question of the young child's language-making capacity.

1. Joseph H. Greenberg's Language in the Americas (LIA; 1987) has been greeted with dismay by many specialists in Amerindian linguistics1 (cf. Chafe 1987, Campbell 1988, Adelaar 1989), and defended by the author in a reply to Campbell in Language 65.1 (1989).