Abstract
This paper expands on recent studies of very large treebank collections aiming to find empiricalevidence for language universals, specifically for the functionally motivated Dependency Length
Minimization (DLM) hypothesis. According to DLM grammars are set up to support the expressionof utterances in a way that minimizes the distance between heads and dependents. Weconstruct several incremental baselines that lead from the random free order linearization to thereal language by adding various word order constraints. We conduct detailed analyses on 55 treebanksand find that all of the constraints contribute to DLM.We show that DLM on the one handshapes the regularity and on the other motivates the attested exceptions from canonical word order.
The findings contribute to a more fine-grained, differentiated picture of the role of DLM inthe interaction of competing constraints on grammar and language use.