This study discusses the results of the first dialectometric approach to geo-linguistic variation of Cappadocian, an Asia Minor Greek variety spoken for many centuries in a situation of bilingualism due to intense contact with Turkish. Cappadocian Greek exhibits remarkable variation among its different communities, reflecting different degrees of archaisms and/or innovations, some of which attributed to language-internal factors, while others to language contact (among others Dawkins 1916; Karatsareas 2011; Melissaropoulou 2016, 2019; Ralli 2019; Janse forthcoming).
To this end, we study the aggregate linguistic distances among twenty −based οn data availability− Greek-speaking communities of the Cappadocian plateau (including Pharasiot and Silliot for comparative purposes). Linguistic features of phonological and morphological nature −treated both separately and all together− are classed under two heads: language-internal vs. contact-induced phenomena. The characteristics of these layouts are quantified so that hidden patterns can be revealed and identified. Finally, the emergent patterns for each cluster of phenomena are being subject to comparative analysis, using methods of cluster analysis and multidimensional scaling (Nerbonne & Heeringa 2010; Wieling & Nerbonne 2015).
Our preliminary results suggest that the two discerning parameters are crucial for revealing and accounting for the various patterns of variation. Using multidimensional scaling (Black 1976; Embleton 1993, 2013), we plotted our results into two dimensions a. based on all attested phonological phenomena cumulatively (r = 0.96), b. isolating the contact-induced phenomena (r = 0.93) and c. isolating the language-internal phenomena (r = 0.97) (plots avalaible here: https://bit.ly/3mE3Tt9). Interestingly, contact-induced phenomena result to a dialect partition reflecting the social conditions and the degrees of bilingualism of the communities. Particularly, North Cappadocian and Pharasiot are supposed to be influenced the least, while South Cappadocian and Silliot the most (Dawkins 1916: 204, 208-209). However, language-internal phenomena reveal a reverse effect, compared to the contact parameter, and the distances among Cappadocian (and Pharasiot) subdialects are eliminated. These features mark the Greek substratum of each dialectal group, exhibiting enough uniformity to justify their treatment as separate dialects (Dawkins 1916: 211-212). Finally, when all phenomena are considered, both dialectal and sub-dialectal clusters (and continua) emerge, aligning yet only partially to the general assumptions of the relevant dialectological research (Bompolas & Melissaropoulou 2020, under review). Further research is on the way, expanding the set of examined phenomena to all levels of linguistic structure. We aspire that it will reveal additional (dis)similarities among the studied varieties, investigating whether the identified patterns pertain or are subject to change and, whether so, to what extent (Scherrer & Stoeckle 2016).
The implementation of both a qualitative and a quantitative dialectometric approach is advantageous in that it contributes to discerning the factors that are responsible for the occurring patterns of variation. Examining the interrelations of the linguistic features before they are aggregated provides a means of comparing patterns rather than merely accumulating them (Pickl & Rumpf 2012), leading to better interpretations for the pathways in which language variation develops in space (Wieling & Nerbonne 2015: 248-250).