In this article, we report on the results of a study of the diphthongs EAR and AIR, which are merging for many New Zealand speakers. Each stage of the study, which has been repeated at five-year intervals since 1983, involves the analysis of word lists and sentences read by over 100 14- and 15-year-old school pupils. The results show a clear trend toward a merger to the EAR diphthong. This is also confirmed in a study carried out in 1994 of 79 speakers (selected according to age, sex, and socioeconomic class) reading six word pairs. The results from both of these studies differ from those reported by Holmes and Bell (1992). We suggest the reasons for the different results could be methodological, and that the mechanism of the merger is merger by approximation after the AIR diphthong became involved in the New Zealand front vowel chain-shift raising.

In this study, we present the first agent-based simulation of vowel chain shifts across large communities, providing a parsimonious reinterpretation of Labov's (2007) notions of transmission, diffusion, and incrementation. Labov determined that parent-to-child transmission faithfully reproduces structural patterns such as the Northern Cities Shift (NCS), but adult-to-adult diffusion does not. NCS is transmitted faithfully to new generations of U.S. Inland North children. But St. Louis speakers, depending only on adult-adult contact, only attain an incomplete, unsystematic version. Labov (2007) attributed the difference to children's superior language-learning ability; transmission and diffusion are categorically different processes in that approach. By contrast, our multiagent simulation suggests that such transmission/diffusion effects can be derived by simple density of interactions and simple exemplar learning; we also find that incrementation is a natural outcome of this model. Unlike Labov (2007), this model does not require a dichotomy between transmission and diffusion. While dichotomous assumptions about child versus adult learning may be necessary in other contexts, our results suggest that the NCS effects in Labov (2007) may be explained economically in terms of simple density of interactions between speakers. Our results also provide an agent-based perspective supporting and explicating the notion of speech community.

In this paper I examine how a classic feature of variationist research, coronal stop deletion (CSD), operates at the end of the 20th century in one of the most renowned vernacular dialects in the United States, English in Appalachia. Through examination of CSD in a corpus of Appalachian speech, this paper also focuses on the methodological choices available. Several methodological questions are reviewed, such as the choices concerning voicing of the codas (wind vs. went vs. west). The corpus comprises interviews with 67 Appalachian speakers, yielding 17,694 tokens of potential CSD. These were analyzed using quantitative variationist methodology to reveal that morphological categories are less influential than even the preceding phonological environment. This finding is in stark contrast with some other vernacular varieties and suggests that apparent morphological influences are actually phonological influences that present themselves as morphological trends. Overall, the following phonological environment is overwhelmingly the most influential linguistic factor on the rate of CSD. These Appalachian speakers maintain relatively high rates, in effect constraining the social distinctions within Appalachia that could possibly be made using CSD, but marking them as vernacular speakers for those outside Appalachia.

Labov (1966, 1972b) described the variable production of coda /r/ in New York City English (NYCE) as a change in progress from above in the direction of rhoticity. Since then, scholars have commented on the slow rate of change toward rhoticity in NYCE and characterized (r) as a superposed feature restricted to formal speech (Fowler, 1987; Labov, 1994; Labov, Ash, & Boberg, 2006). This study's ethnically diverse sample of speakers from the Lower East Side of Manhattan (n = 65) shows a mean rate of /r/ production of 68%, with young people, women, and middle-class speakers leading in the production of /r/ in apparent time. Speakers from five ethnic backgrounds—African American, Chinese, Jewish, Puerto Rican, and white—show coherence for the internal constraints on variable nonrhoticity. However, only Chinese, Jewish, and white speakers participate in the change toward rhoticity. These findings highlight the role of ethnicity in patterns of variation and change and demonstrate that the change toward rhoticity in NYCE has accelerated and is no longer restricted to formal speech.

This study explores variability in particle placement across nine varieties of English around the globe, utilizing data from the International Corpus of English and the Global Corpus of Web-based English. We introduce a quantitative approach for comparative sociolinguistics that integrates linguistic distance metrics and predictive modeling, and use these methods to examine the development of regional patterns in grammatical constraints on particle placement in World Englishes. We find a high degree of uniformity among the conditioning factors influencing particle placement in native varieties (e.g., British, Canadian, and New Zealand English), while English as a second language varieties (e.g., Indian and Singaporean English) exhibit a high degree of dissimilarity with the native varieties and with each other. We attribute the greater heterogeneity among second language varieties to the interaction between general L2 acquisition processes and the varying sociolinguistic contexts of the individual regions. We argue that the similarities in constraint effects represent compelling evidence for the existence of a shared variable grammar and variation among grammatical systems is more appropriately analyzed and interpreted as a continuum rather than multiple distinct grammars.

The distribution of variants of the face and goat vowels in Tyneside English (TE) is assessed with reference to the age, sex, and social class of 32 adult TE speakers. The effects of phonological context and speaking style are also examined. Patterns in the data are suggestive of dialect leveling, whereby localized speech variants become recessive and pronunciations typical of a wider geographical area are adopted. Within this broad pattern, however, there is evidence of parallelism between the vowels in terms of the relative proportions of their variants across speaker groups. It is suggested that pressure to maintain the symmetrical structure of the underlying phonological system is guiding this process. Labov's (1991, 1994) principles of chain shift are discussed in this connection. However, it is argued that the patterns in the data are more plausibly explained by considering the social significance of each variant instead of making reference to variants as socially neutral expressions of abstract phonological categories.

This paper examines the acquisition of the variable rules constraining Spanish syllable-final /s/-lenition in Chilean Spanish–speaking children, and whether adult-to-adult speech differs from child-directed speech in the production of s-lenition. The data of 10 children (ages 2;04–5;09) and their caregivers is presented. Tokens of syllable final /s/ were coded for pronunciation and a variable-rule analysis examining the effect of various linguistic and extralinguistic constraints was carried out. The results show that child and caregiver use of s-lenition is similar to that found in adult-to-adult speech; however, two of the youngest children showed near-categorical behavior. This paper highlights the important role of caregiver input on acquisition of s-lenition and proposes that production of s-lenition in utterance-final position in the input (caregiver's speech) is linked to young children's acquisition of s-lenition; young children exposed to an input with high rates of [s] in utterance-final position acquire s-lenition earlier.

Linguistic parallelism (i.e., the tendency of similar forms to occur together within a stretch of discourse) has been shown to be very strong in many linguistic phenomena and in a vast number of languages. Examining the role of phrase-level parallelism on noun phrase number agreement, this article demonstrates that Puerto Rican Spanish and Brazilian Portuguese exhibit more similarities than differences with regard to this constraint. A detailed analysis of Brazilian Portuguese data is presented, and the results are compared with those found by Poplack (1980a) for Puerto Rican Spanish. Guy's (1981a) and Labov's (1994) hypothesis about missing zeroes in Brazilian Portuguese is also discussed. In conclusion, it is claimed that the phrase-level parallelism effect on noun phrase number agreement is embedded in a universal principle of linguistic use: parallel processing.

Whereas the alternation between the s-genitive (the New Year's message) and the of-genitive (the message of the New Year) is well documented, our study offers a more accountable analysis of genitive variation by including noun-noun (NN)-genitives (the New Year message). We consider four different variable contexts (s versus of, NN versus of, NN versus s, and NN versus s versus of ), which we analyze using regression analysis. The dataset consists of 10,054 variable genitives drawn from the Representative Corpus of Historical English Registers. The material covers the period between 1650 and 2000, thus enabling us to track the evolution of variable genitive grammar in real time. We report that there is an overall drift toward the NN-genitive, which is preferred over other variants when constituent noun phrases are short, possessor constituents are inanimate, and possessum constituents are thematic. In addition to these substantive contributions, we showcase methods of dealing with a complex dataset covering nonbinary grammatical variation.

Although the order of major sentence constituents at all stages of Ancient Greek is quite free, the distribution of clause types is not random over time, but changes from predominantly verb-final to predominantly verb-medial, suggesting a change in progress. Using the paradigm of Kroch (1989), in which it is assumed that syntactic change involves competition between grammatical systems, two models are constructed, one for the verb-final grammar of the Homeric period (pre-800 B.C.) and one for the verb-medial grammar of the Hellenistic Koiné (c. 100 A.D.). The language of an intermediate stage (Herodotus, c. 450 B.C.) is shown to pattern in part like Homer and in part like the Koiné. More strikingly, the ratio of verb-final to verb-medial structure that best fits the distribution of clause types found in Herodotus is extremely close to an independent measure of this ratio obtained from the distribution of weak pronouns and clitics.

This article considers language variation within one “ethnic” group: Israelis of Middle Eastern origins. Earlier studies (Yaeger-Dror, 1988, 1991) found that singers from the dominant “koiné” -speaking social group (Blanc, 1968) use [r] in pop songs and [R] in casual interviews. This can be defined as a register distinction. On the other hand, singers from a MidEastern ethnolinguistic background, whose underlying dialect includes [r], use [R] even in songs. Given that singers whose vernacular consonant invetory does not even include [r], and who should find it easire to use it categorically, have such a difficult time maintaing [r] consistently (and appropriately) in the song register? One of the recorded variants for these singers “merges” the [r] and [R] into coarticulated [rR]. Why does this previously unattested sound arise, and what does it tell us about the linguistic and sociolinguistic situation? Data from various registers are analyzed in order to discover the answers to these questions. This analysis is concerned with the quantifiable evidence of systematic patterns in the use of these three pronunciations for [r] and uses this evidence to demonstrate that subconscious sociolinguistic pressures on members of the minority community influence them to assimilate to the dominant social group while still retaining ethnolinguistic proof of a narrower ethnic identity. For example, the use of [rR] is found to be correlated with a wish to affiliate with both an [R]-using group and an [r]-using group, showing that sociolinguistic techniques can reveal social psychological ethinc affiliation. Like Trudgill's (1986) discussions of dialects in contact, the present theoretical discussion takes advantage of proposals advanced by Giles, to explain why the data reveal both convergence (toward the dominant out-group) and divergence (toward the in-group) (Giles & Coupland, 1991). Sociolinguistic methods permit a quantitative analysis of the strength of these conflicting tendencies, both of which are subsumed under the technical term “accommodation.” Methods are proposed to determine if choice of the

[R] or [rR] variant is conscious or not, and variable rule analysis reveals that for most of the singers the less cognitive salience, the greater the degree of convergence to the Koiné norm [R]. The linguistic factors that are correlated with the relative degree of salience can be used in future studies when the relationship between convergence toward another dialect or language and relative cognitive salience is also at issue.

The Portuguese NP a gente, meaning “the people,” is undergoing grammaticalization and is acquiring characteristics of a personal pronoun, increasingly replacing first-person plural nós, meaning “we,” in speech. In Brazilian Portuguese, this process seems to be correlated with a number of other ongoing morphosyntactic changes. In this study I compare data from Southern Brazil on the use of a gente in the 1970s and the 1990s. Quantitative analyses are conducted in terms of two methodological approaches: apparent-time and real-time studies. In the real-time analysis, two kinds of studies are discussed: a trend study, with two comparable groups of speakers, and a panel study, with the same speakers compared longitudinally. The linguistic and social embedding of this process is discussed in terms of the Labovian classification of changes as being “from above” or “from below.”I am very grateful to Gregory R. Guy for supervising this research project while I was a visiting scholar at New York University (2001–2002) and for his kind and wise assistance in the preparation of the lecture (presented at NYU on September 20, 2002), on which this article is based. I also acknowledge the valuable work of my research assistants at Universidade Federal do Rio Grande do Sul, Brazil: Kátia M. L. Aires, Greice L. de Souza, Karine Q. da Silva, Patrícia da R. Mazzoca, Leonardo Z. Maya, and Melissa Schossler. This research was conducted with the support of Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), an agency of the Brazilian government dedicated to scientific and technological development, grant 200740/01-6(NV); Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS), grant 00514482; and Pró-Reitoria de Pesquisa da Universidade Federal do Rio Grande do Sul.

Data from fieldwork and a telephone survey is used to demonstrate that the merger of the vowels in words such as cot and caught, traditionally considered a defining characteristic of the speech of western Pennsylvania, is well established in the former anthracite mining towns of eastern Pennsylvania. Demographic and linguistic data indicate that the merger arose independently in the anthracite towns as a result of massive in-migration of speakers of Polish and other non-Germanic languages around the turn of the century. Having identified the proximate cause of the merger makes it possible to predict the presence or absence of the merger in various towns using demographic data alone.

Optimality theory (OT) (McCarthy & Prince, 1993; Prince & Smolensky, 1993) has been proposed as a constraint-based theory of phonology in which the phonological facts of each language are accounted for by a language-specific ordering of a universal inventory of constraints. The constraints, expressing desirable (i.e., optimal) phonological states, evaluate possible candidate forms, selecting the optimal output. Any constraint may be violated by a surface form if it is overridden by a higherranked constraint; the ordinal sequence of constraints provides a weak quantification of constraint effects. Variability has been treated within OT by varying constraint orders. This model is analogous in several important respects to the variable rule model (VR) of Labov (1969) and Cedergren and Sankoff (1974). In VR, variable constraints express desirable phonological states which are variably realized on the surface, when not overridden by other constraints; constraints are probabilistically quantified. This article compares the OT and VR models, arguing that the VR model is superior on theoretical and empirical grounds: constraint effects in VR are stable, transparent, and learnable. Moreover, the probabilistic treatment of constraint effects allows VR to model successfully cases in which multiple violations of a single constraint lead to a cumulative reduction in likelihood of a form; such cases cannot be efficiently treated in OT.

Formal and usage-based approaches to phonology make competing predictions that can be tested with variationist methodology. This paper investigates formal, functional, and frequency effects on (t/d)-deletion in Canadian English. Although initial results suggest a correlation between lexical frequency and deletion, once interaction and lexical effects are taken into account, only phonological and morphological factor groups are significant. Previous reports of frequency effects may result from different measurements of frequency and the contribution of overlapping factor groups. These results suggest that frequency does not operate monotonically but interacts dynamically with the lexicon.

British Sign Language is a language exhibiting extensive regional variation and undergoing rapid change, in a period of changing attitudes to the language and the community of signers. Fingerspelling (a representation of written English) is an aspect of BSL that reflects these variations strongly. Analysis of interviews with Deaf signers appearing on BBC television shows that use of fingerspelling may vary according to various demographic factors and the signers' use of voice. There has also been a decrease in the overall use of fingerspelling in BSL over the last 10 years. Strong evidence is presented to support the claim that fingerspelling in BSL is of two distinct linguistic types. Some fingerspelling involves code-mixing with English, whereas some has become incorporated into the language itself. Further evidence for such a “discontinuity” between the code-mixing and borrowing is presented and discussed.

Korean has a typologically unusual three-way laryngeal contrast in voiceless stops among aspirated, lenis, and fortis stops. Seoul Korean is undergoing a female-led sound change in which aspirated stops and lenis stops are merging in voice onset time (VOT) and are better distinguished by the F0 (fundamental frequency) of the following vowel than by their VOT, in younger speakers' speech. This paper compares the VOT pattern of Homeland (Seoul) and Heritage (Toronto) Korean speakers and finds that the same change is in progress in both. However, in the heritage variety, younger speakers do not advance the change, unlike their Seoul counterparts. Rather they have leveled off or are perhaps reversing the change, and there is very little sex difference among the younger heritage speakers' patterns. We consider possible accounts of the differences between the Seoul and Toronto patterns, building our understanding of how language-internal variation operates in bilingual speakers, a topic that has received relatively less attention in the variationist literature.

Recipes exhibit a phenomenon that does not exist in other varieties that are commonaly studied (e.g., narrative, sentences in isolation, conversational discourse, etc.)—namely, zero anaphors as direct objects, as in Beat [ø] until stiff. The two goals of this article are to examine this phenomenon in both contemporary recipes and across time and to explore its consequences for linguistic theroy. It is found that stylistic, semantic, and discourse factors are the most important in the phenomenon, with syntax and morphology playing relatively minor roles. Furthermore, these zero anaphors pattern like overt pronouns and have in fact replaced overt pronouns over time. A preliminary model of register variation and its consequences for the synactic analysis of recipes are also presented.

The unconditioned merger of the low back vowels and the variety of realizations found for the low front vowel have been noted as leading to greater distinctiveness across U.S. English regional dialects. The extent to which the movements of these vowels are related has repeatedly been of interest to dialectology as well as phonological theory. Here, examining production and perception data from speaker-listeners across three major regions of the United States, the relationships among these low vowels within and across regions are investigated. Participants provided speech samples and took part in a vowel identification task, judging vowels along a continuum from /æ/ to /ɑ/. Results of acoustic analysis and statistical analysis of the perception results indicate that a structural relationship between /æ/ and /ɑ/ is maintained across regions and that listeners’ own degree of low back vowel merger predicts their perception of the boundary between /æ/ and /ɑ/.

This study examines the internal and external constraints on glottal replacement of /t/ among adolescents in London and Edinburgh. Results show that phonological and stylistic constraints play an important role in determining the realizational variation of /t/, as many similar studies have shown. However, there is also clear evidence that our understanding of this phenomenon has been restricted by the limited set of factors that have been investigated previously, as results show that this feature is also constrained by word frequency and morphophonological factors. These findings raise important questions concerning the role of morphological compositionality in language change and the nature of lexical diffusion.