Sociolinguists assume that standard varieties are rapidly gaining influence in the repertoires of post-industrialized societies [1, 2]. This is typically due to the prestige associated with standard forms, their presence in media and education, and mobility within mutually intelligible dialect areas [3, 4, 5]. Under such circumstances, speakers are more likely to engage in inter-personal accommodation [6]. The results of this are generally called dialect levelling, characterised by two tendencies [7]: language internally, levelling out of marked forms [6], and language-externally, diffusion of variants with the widest geographical currency. Similar contact-induced phenomena are now being observed in Basque, with young speakers progressively converging with supra-local norms, especially in morpho-syntax and lexis [8, 9].
Therefore, this paper investigates the social distribution of one phonological variable in a Western Basque rural town, located in a dialect area considerably divergent from Standard Basque. 20 participants were studied, equally divided by gender (male/female) and generation (17-21-year-olds and 47-55-year-olds). They were selected with specific criteria in mind: all speakers have Basque as their L1 and language of the home, they have lived in the same town since birth, they usually use Basque outside the home for socialization, and their parents are from town or a neighbouring town. Young speakers were alphabetised in Basque, whereas adults were first alphabetised in Spanish and Basque later. To gather data, semi-structured dyadic sociolinguistic interviews were conducted with same-gender and -age pairs – one speaker in each group was interviewed twice with a different partner to control for intra-speaker variation. Post-interview questionnaires were also administered where participants had to answer 5-point Likert scales on language attitudes based on the BLP [10].
Through auditory transcriptions using ELAN [11] and conducted twice to avoid intra-rater variability, a corpus of 1513 tokens of the variable was analyzed, where two variants (i.e. local [e̞] and supra-local [a]) were identified. Group uses of each variant were tested for statistical significance, and the attitudes towards the town emerged as the only significant factor. Moreover, using a linear model analysis, a strong correlation was found between local variant incidence and positive attitudes towards the town and local vernacular. This is in line with recent trends in the literature on variation in the Basque-speaking area [12, 13]. Post hoc factor contrasts revealed a great, though non-significant, gender difference between adults, which may be the result of historically substantiated patterns: men have traditionally tended towards socializing environments that favoured dialect contact with other varieties while women stayed in more local circles [14, 15]. In the younger generation, females show a retreat from local forms, to parallel intergenerationally stable male values.
In conclusion, this paper reports on the importance of investigating subjective factors to explain language variation in cases of rapid language change, and shows that gender-correlated patterns of language use are to be considered within the context of speakers' access to incoming forms.