ИСТИНА

One of the goals of contact linguistics is to offer tools to reconstruct unknown historical patterns of interaction between ethnic groups based on the observed structural outcomes of language contact. In this reconstruction, a necessary intermediate stage is modelling the historical patterns of bilingualism. Based on a database of loanwords in minority languages of Daghestan, we show how simple counts of lexical borrowings may shed light on patterns of small-scale multilingualism in the past. Language contact is channelled through people, via their use of multiple languages (Milroy 1997: 311). Patterns of bilingualism vary considerably from society to society and to a great extent shape the outcomes of language contact. It is assumed that the amount of structural influence from one language to another is a function of the intensity of bilingualism (Thomason & Kaufman 1992: 74-76). When lexical influence is at work, the more intense the bilingualism, the higher the amount of loanwords from the donor to the recipient language is supposed to be (e.g. Scotton & Okeju 1973, Watson 2018). We cannot know the details of the individual loanwords’ journeys to the recipient languages. However, based on our counts, we conclude that not only the amount of Turkic loans is much lower in the north-west than it is in the south, but also, at least for some loans, the donor language could have been Avar, rather than Kumyk or Azerbaijani. The spread of Turkic bilingualism across Andic and Tsezic villages is likely to have been even lower than the counts in Table 1 suggest — if present at all. This conclusion is in full accordance with an independent reconstruction of the traditional small-scale bilingualism in (Dobrushina et al. 2017).