Abstrakt
Characterizing the distribution of crossing dependencies in natural language dependency trees isa crucial task for building parsers and understanding the formal properties of human language. Anumber of formal restrictions on crossing dependencies have been proposed, including bounds ongap degree, edge degree, and end-point crossings. Here we ask whether the empirical distributionof crossing dependencies in dependency treebanks offers evidence for these formal restrictions astrue, independent constraints on dependency trees, or whether the distribution can be explainedusing other, more generic constraints affecting dependency trees. Specifically, we explore thenull hypothesis that crossing dependencies are formally unrestricted, but occur at a low rate.
We implement the null hypothesis using random trees where crossing dependencies occur at thesame rate as in natural language trees, but without any formal restrictions. We find that thisbaseline generally does not reproduce the same distribution of gap degree, edge degree, endpoint-crossing, and heads’ depth difference as real trees, suggesting that these formal constraintsare a consequence of factors beyond the rate of crossing dependencies alone.