Abstract

We use the persistent homology method of topological data analysis anddimensional analysis techniques to study data of syntactic structures of worldlanguages. We analyze relations between syntactic parameters in terms ofdimensionality, of hierarchical clustering structures, and of non-trivialloops. We show there are relations that hold across language families andadditional relations that are family-specific. We then analyze the treesdescribing the merging structure of persistent connected components forlanguages in different language families and we show that they partly correlateto historical phylogenetic trees but with significant differences. We also showthe existence of interesting non-trivial persistent first homology groups invarious language families. We give examples where explicit generators for thepersistent first homology can be identified, some of which appear to correspondto homoplasy phenomena, while others may have an explanation in terms ofhistorical linguistics, corresponding to known cases of syntactic borrowingacross different language subfamilies.

 

Introduction (beta)

None

 

Conclusion (beta)

None