The present study constitutes a study of the different realizations of the morphophonological personal pronoun variable (tin),of the dialect of Agrinio (Southern Greece),taking into the account the dialect contact between different speech communities of Agrinio.
Ηigh Vowel Loss –i.e. the deletion of unstressed [i] and [u]- has been defined as one of the majorphonological isoglosses that splits Greece into two main areas, those of the Northern Greek dialects which delete the unstressed [i]s and [u]s, and the Southern dialects that they do not delete the unstressed [i]s and [u]s(Newton 1972, Trudgill 2003 among others). Nevertheless, resent studies (Topintzi & Baltazani 2012, Kainada & Baltazani 2015, Lengeris et. al. 2016, Παπαζαχαρίου 2019, among others) have shown that the above phonological phenomenon is not categorical but a gradual one, with the percentage of realization to be lesser than 35% in most of the Northern dialects under study.
The dialect of Agrinio, although it is geographically located in the South of Greece, it is considered a Northern variety(Kontosopoulos 1978) and it shows the existence of High Vowel Loss as well. Our study, which is based on the acoustic analysis of casual speech of 90 informants (analysis of 10 minutes’ speech from each informant, or approximately 64.000 [i] and [u] vowels), fund similar results to the above mentioned resent studies, i.e. a small overall percentage of phonological phenomenon (34% of the overall number of the unstressed [i]s & [u]s).
Nevertheless, during the analysis of our data, it became obvious that the female personal pronoun [tin], in which an unstressed [i] exists, shows a clearly different pattern of high Vowel Loss, which affects the different realizations of the personal pronoun. Moreover, these different realizations are significantly correlated with social parameters.
In particular, we will present three different realizations of the personal pronoun [tin], i.e. [tin], [tn], and [n] or [ne] -according to the next sound, vowel or consonant respectively. We will further showthat highly educated young people from new and mixed neighborhoods produce significantly smaller percentages of the [tin], [tn], [n], [ne], [sn] realizations, in contrast to middle-aged/old speakers from old neighborhoods.
Finally, we will argue that the sociolinguistic behavior of the variable (tin) can be explained as part of the process of dialect levelling (Trudgill 1986, Hinskens 1992), which has taken place in the new and mixed neighborhoods since the 60s, when a significant number of new inhabitantsstarted settling in the city.