Revisiting the classification of Gallo-Italic: a dialectometric approach

Abstract

While Gallo-Italic varieties clearly belong to the Romance language family, their subgrouping as either Gallo-Romance or Italo-Romance has been the source of disagreement in the classificatory literature. While earlier analyses tended to classify Gallo-Italic as Gallo-Romance (notably Schmid, 1956; Bec, 1970-1971), later work has either argued for or tacitly assumed a classification of Gallo-Italic as part of the Italo-Romance branch, a view that is both different from as well as irreconcilable with the earlier Gallo-Romance classifications. In this paper we aim to contribute to the development of an empirically-based classification of Gallo-Italic through the use of dialectometry applied to atlas corpora, and specifically through the measurement of Levenshtein distance. Using three wordlists (Swadesh 100, Swadesh 200, Leipzig-Jakarta) and comparing twenty-six linguistic varieties across Italy and southeastern France, we show that Gallo-Italic is best classified as a third subgroup within the Gallo-Romance branch. Our results also clearly identify all the major bundles of isoglosses established through traditional dialectological methods and confirm Gallo-Italic as a relatively homogenous group distinct from Italo-Romance.

Multilingual Computational Linguistics (Lecture at the University of Passau in the Winter Term 2022/2023)

“…The lack of availability is very annoying, not in the sense that we have no access to a given article in the form of a scan or a book, but rather because many authors collect data, write articles about them, but then do not share their data officially. It is still not surprising that articles are being published in which new ideas are postulated or new conclusions are being made, but in which scholars do not share the data upon which they base their conclusions openly (Tamburelli and Brasca 2017). The same holds for many grammatical descriptions, in which scholars extract individual sentences from their personally collected private corpus but never reference them sufficiently, nor offer the full corpus.…”

Proto-Austronesian historical relations on the development of the proto-melayic aspect of Malay with Lampung language

“…(Mahsun, 1995;Onishi, 2019) Furthermore, the results of the calculation via the dialectology formula need to be consulted with the isolect status rules based on the following Comparative Historical Linguistics criteria/rules. Tamburelli & Brasca (2018) If the historical relationship between ML and LL in figure 2 is the relationship between dialects of a language, then in Table 2, this determination has undergone an innovation, namely the average number of kinship relations is the result of the number 48.5%, this percentage is in the range of status criteria isolect of a language, which is 36-81%. So the isolect status, shows the kinship relationship of the language of 'Language of Families'.…”

Cross-Linguistic Data Formats, advancing data sharing and re-use in comparative linguistics

“…It is still very difficult to find particular datasets, since linguistic journals often do not have a policy on supplementary data and may lack resources for hosting data on their servers. It is also often difficult to access data, and many papers which are based on original data are still being published without the data 1 and having to request the data from the authors is sometimes a more serious obstacle than it should be 22 , 23 . Due to idiosyncratic formats, linguistic datasets also often lack interoperability and are therefore not reusable .…”