Analogical leveling in progress of a potential suffix in Tôkyô Japanese is analyzed within a quantitative paradigm. The phenomenon, whereby an innovative potential and a conservative potential alternate, is shown through a multivariate analysis to be controlled by five factors: sociological variable complex, length of the verb stem, conjugation pattern of the verb, the following inflectional form, and embeddedness of the clause containing the suffix. Most of the linguistic constraints are observed crosslinguistically in language change or variation, giving further credibility to the analysis. Although traditional frequency-based theory of analogical leveling would predict stem frequency to be a possible factor, I demonstrate that it is not in this case. As a principled explanation for this apparent lack of contribution from frequency, the Revised Frequency Hypothesis is proposed.

In Portuguese, subject pronouns may be omitted or expressed explicitly. This variable usage is often attributed to the richness of the system of verbal desinences. However, evidence drawn from a sample of informal written language—personal correspondence of people from Rio de Janeiro—shows that the most important constraints on pronoun usage in subject position are discouse-based. The pronoun tends to be used whenever it is needed to clarify or highlight the correct referent from among a set of possible candidates. Furthermore, the pronoun is less likely to occur when its clasue is tightly connected to the preceding discourse. The quantitative analysis presented here supports the hypothesis that pronoun usage is “functional” in the sense that semantically relevant information tends to be preserved in surface structure.

The development of language conditions in Helsinki makes it possible to observe many processes of sociolinguistic interest. By European standards, the beginning itself is exceptional. Helsinki was originally founded in the middle of a Swedish-speaking area, which means that it lacked all natural connections with any basis of Finnish dialect from which urban colloquial language and standard spoken Finnish might have emerged. However, throughout the 20th century Helsinki has been the most important Finnish-speaking city, and its colloquial Finnish has served as a model for the evolution of colloquial Finnish throughout the country. One may well wonder how colloquial Helsinki Finnish came into existence in the course of only a few decades and consolidated its position as the model for everyday Finnish.

This article uses the quantitative methodology of goldvarb to examine the variable distribution of lexical noun phrases representing core arguments of the verb in a corpus of spoken French and a corpus of spoken Spanish. It is shown that this distribution is not random, but instead conforms to a grammatically and pragmatically motivated pattern known as Preferred Argument Structure.

This article examines the type of relatively rapid speech modification sometimes called “accommodation” (Trudgill, 1986). In this case, we track change on the part of speakers of Honduran Spanish newly resident among speakers of Northern Mexican Spanish. The two variables investigated are final /n/ velarization and spirantization, both of which differ in the two dialects. (Honduran Spanish velarizes final /n/, while Northern Mexican Spanish does not; Honduran Spanish conserves stops l, r, s and glides, whereas Northern Mexican Spanish produces fricatives.) It is found that, although there is change in the direction of local norms in both variables, velarization shows greater change. In addition, the degree of contact with local speakers also leads to greater change for velarization only. Reasons for the different behavior are considered.

We present results on the grammar of spoken Brazilian Portuguese, focusing on the syntactic boundaries of the main constituents of an utterance. The purpose of examining such sites is twofold: (1) to provide evidence as to the relationship (or mutual exclusion) between boundary and type of filler, and (2) to determine the relationship between boundary and frequency of filling. We find major differences both in the distribution of filling by syntactic adjuncts versus discourse elements and in the susceptibility of different syntactic boundaries to be filled. We discuss the lack of demographic conditioning on the rate of filling.