Areal pressure in grammatical evolution: An Indo-European case study

Typological features can be expected to exhibit the same diachronic development multiple times during the history of a language family, exhibiting what is known variously as recurrence, homoplasy, or drift. This paper seeks to quantify the areal signal shown by recurrent mor- phosyntactic changes across the Indo-European language family. We use 108 morphosyntactic feature variants taken from the DiACL Typology Eurasia database, which contains data from 117 Indo-European languages (including ancient and medieval doculects). We partition types of change (i.e., gain or loss) for each variant according to whether they roughly bring about a simplification in morphosyntactic patterns that must be learned, whether they are neutral (i.e., neither simplifying nor introducing complexity), or whether they introduce a more complex pattern. We hypothesize that the first group will exhibit the greatest areal sensitivity, followed by the subsequent groups. We combine phylogenetic information regarding the likely evolutionary histories of the features under study with a phylogeographic model of linguistic dispersion, allowing us to simulate gain and loss locations in time and space for each feature variant. We calculate the average of the spatiotemporal distance between each simulated change event and the most proximate change event, grouped by feature variant and type of change (i.e., gain or loss). Using principal component analysis (PCA) to account for the significant correlation between spatiotemporal distance and cophenetic (i.e., genetic) distance between locations in a phylogeny, we take the second component (PC2) as a variable which represents areality that is orthogonal to genetic relatedness. We fit a linear model in which ranks for PC2 serve as a response, and employ a three-level categorical predictor representing the cline in complexity mentioned above (simplifying > neutral > complexifying). We find that changes which in- troduce complexity are show significantly less areal signal than changes which simplify and neutral changes, but find no significant differences between the latter two groups. This result is compatible with a scenario where certain types of parallel change are more likely to be mediated by advergence and contact between proximate speech communities, while other developments are due purely to drift, and are largely independent of intercultural contact.