Minimality and Linguistic Variation
The goal of this project is to show that the key notion of minimal difference between languages can receive a precise definition as minimal variation of core grammatical principles.
In historical language change as well as in dialectal variation languages often vary minimally. For instance, the historical process of s-retraction in German, changing [s] to [ʃ] word-initially before consonants (e.g. [ʃ]tein, 'stone'), has taken place in a stepwise fashion, involving always larger sets of preconsonantal contexts:
(1) t1: --> ʃ before k
We hypothesize that minimal differences between stages of language change correspond to minimal differences of typological properties in the sense of Alber&Prince 2015.
This hypothesis is verified by creating a database of relevant microvariation patterns, modelling them computationally and performing a typological analysis yielding patterns of minimal difference:
(i) The database of microvariation patterns (found in the literature and databases we have created for previous projects) gives us a solid empirical base on which to test our hypothesis.
(ii) The computational model of microvariation patterns is created with the free software component OTWorkplace. The formal typology thus generated is compared to the empirically attested patterns.
(iii) We then submit the formal typology to a typological analysis, extracting its typological properties, in the form of ranking conditions in an Optimality Theoretic Grammar. In s-retraction, for instance, the typological properties distinguish between the various degrees to which languages exhibit the phenomenon. Our hypothesis (to be verified on the microvariation database) is that minimal variation between languages can be defined as minimal variation in property values.
We expect this research project to provide a solution to a long-standing problem, the definition of minimal grammatical difference. Its originality and innovativeness lies in the synergy of traditional, computational and theoretical methodologies.