Studies in Language. International Journal sponsored by the Foundation “Foundations of Language” - Volume 22, Issue 2, 1998

This study reports on a significant negative association found in a cross-linguistic sample between the degree of predictability of word stress from a word boundary and the extent to which stress has segmental effects. In other words, in a given language the less predictable stress is from the word boundary, the more likely that the language will have vowel reduction in unstressed syllables, vowel lengthening in stressed syllables, and consonantal changes restricted to stressed or unstressed syllables. These findings are interpreted as part of a major diachronic tendency for stressed and unstressed syllables to become more differentiated in terms of duration as a cumulative effect of phonetic change, which in turn leads to the deletion of unstressed syllables, which renders stress unpredictable in some cases. A model of phonological representation that best accounts for the unidirectionality of this strong tendency is one in which stress, even while it is still predictable, is considered an inherent part of the word, and phonetic changes have a permanent and cumulative effect on lexical representation.

The article reports an oral corpus-based study of epicene pronominal constructions (i.e., pronouns coreferent with singular antecedents and referring to referents of indeterminate sex) in English. They is used in 60% of the tokens, he in 25%, and other forms were used minimally. That variation corresponds to three semantic factors: perceived sex stereotypes associated with the referent, notional number, and, surprisingly, degree of individuation. These findings support accounts of the importance of agreement as a discourse-level phenomenon (e.g. Barlow's Discourse-Linking Theory), and of pronouns as elements whose informational content goes beyond mere denotation.

This paper reports the results of a detailed text-based study of the use of Imperative constructions in Biblical Hebrew and English, and shows that the two languages differ significantly in this regard. The use of the English Imperative in the database is conditioned largely by social and interactive factors, e.g. the relationship between the speaker and the listener, their relative social status, the sensitivity of the action of giving the command, the setting of the interaction, who will benefit from the action, etc.; on the other hand, the usage of the Imperative in the Hebrew database is mainly determined by semantic and structural factors, e.g. the point in time when the commanded action is to take place, the linguistic form of the preceding clause, whether the command is the first in a conversation, etc. The clear differences here show that there cannot be any uniform explanation about why Imperatives in general are used, as have been proposed in speech act theory (e.g. Searle 1975); on the other hand, these differences are sufficiently complex that they also cannot be accounted for with simple statements regarding cultural differences (e.g. Blum-Kulka 1991). The results of this study suggest that theories about speech acts should be based not upon philosophical speculations using data from a single language, or upon limited linguistic and cultural data carefully selected to support a particular theory, but upon extensive, detailed, and exhaustive linguistic analysis which will clearly establish the descriptive facts of speech act usage in a variety of languages.

The Tungusic language Oroqen has a number of qualitative suffixes used to express intensification, augmentation, approximation, diminution, and the like. Many of these suffixes have been previously identified, though the descriptions of their meanings are incomplete and are often in conflict. In this paper, we provide an overview of these suffixes which includes a more thorough description of their meaning and use. We then examine one of the suffixes, -Jean, in more detail. The suffix has been identified both as a diminutive and as an intensifier. We offer a model of the semantic range of -kan (based on Jurafsky 1996) which accounts for the apparent conflict in its meanings.