Our study presents a variationist analysis of subject doubling in the French of Ontario, Canada. Two principal variants are distinguished: a non-doubled variant and a doubled variant containing a clitic agreement marker. In our analyses, both linguistic and social factors are taken into account and analyzed using goldvarb2. It is proposed that subject clitics are marked for default features, and that the doubled variant is favored when the clitic's default features match those of the subject NP; lack of matching favors the non-doubled variant. Discussion of linguistic factors for the present study, therefore, is limited to those factors which can be explained in terms matching. The principal social factor studied is restricted language use (cf. Mougeon & Beniak, 1991). Our results show that the greater the restriction, the fewer doubled subjects one finds.

This article investigates the constraints on variable deletion of short vowels in word-final unstressed syllables found in the variety of Latvian spoken in Riga. The affected vowels are almost always inflectional endings. Results from a variable rule analysis of 8 native speakers from Riga indicate that internal phonological and prosodic factors (especially distance from the main word stress) act as the strongest constraints on vowel deletion, along with the educational level of the speaker. The functional constraint of the recoverability of the deleted vowel is not significant.

In a near-merger, speakers produce two contrasting words differently, without reliably being able to discern the contrast in their own speech or in the speech of others. Acoustic measurements typically reveal small differences between the elements of near-merged minimal pairs along several acoustic dimensions. We argue that statistical evaluation of the potential distinctiveness of these near-merged elements must simultaneously take into account all of these dimensions. For this reason, discriminant analysis is used to assess the differences between near-merged/il–Il/, /el–εl/, and /ul–υl/ for five Utah speakers. In contrast with independent univatiate analyses of variance of F1, F2, f0, and spectral slope, the multivariate discriminant analyses suggest that all three contrasts are preserved by all five speakers. However, hompohones like heel and heal are not distinguished by the discriminant analyses. Discriminant analysis is thus a powerful technique for assessing whether a reliable basis exists for the claim that two potentially contrastive items are in fact distinctive.

Gender distinctions constitute one of the central grammatical categories in Hebrew grammar and are closely related to number distinctions. Both are acquired early on, since they apply to almost every morphological category: three major classes of content words, as well as two classes of function words. Feminine words are marked by Suffixal -α and -t, while masculine words carry either a zero or -e suffix. A small class of what might be termed “numeral nounclassifiers” are supposed to agree in gender with their head nouns, but carry a mirror-image gender marking, resulting in rule opacity. The subjects, 40 children (20 fourth graders, 20 seventh graders) from a lower middle-class background, were tested on gender markings of numerals in two situations involving monitored and unmonitored situations. The results indicate the disappearance of gender agreement in Modern Hebrew numerals and a reanalysis of numeral suffixes by speakers.

Recent work in the acquisition of variation has shown that children begin to learn patterns of stable variation at a very early age. In fact, it appears that they acquire variable rules at about the same time as they are acquiring related categorical rules. Little is known, however, about the transmission from generation to generation of features undergoing sound change in progress. Therefore, this study examines the acquisition of the Philadelphia short a pattern by 18 3- and 4-year-old children. Even though this pattern of the raising and tensing of short a is a complex one, the children had, for the most part, acquired it. In almost all cases, the children matched the short a distribution both of their parents and a group of adult Philadelphians who were interviewed in the mid 1970s and described in Labov (1989b). These results indicate that even the youngest members of the speech community are actively participating in ongoing sound change.

Enumeration is a figure of speech which evokes a thematic set through the concatenation of several of its elements. Its linguistic structure is variable with respect to the number of components, their syntactic nature, the use of coordinating markers and of the processes of repetition, reduction, and expansion. In studying the variation among 3,464 tokens of enumeration in Montreal French, we explain how the constitutive processes interact in terms of Slobin's (1977) charges to language: be clear, processible in real time, quick and easy, and expressive. We suggest how the six structural factors or processes contribute to the fulfillment of these charges. The data analysis evaluates the interaction of the processes and, indirectly, the pertinence of each of Slobin's four functional criteria. This analysis is a new method for attributing a pattern of process co-occurrences to a number of functions, focusing on how these processes compensate for, or cooperate with, each other.