Semantic role clustering: An empirical assessment of semantic role types in non-default case assignment

This paper seeks to determine to what extent there is cross-linguistic evidence for postulating clusters of predicate-specific semantic roles such as experiencer, cognizer, possessor, etc. For this, we survey non-default case assignments in a sample of 141 languages and annotate the associated predicates for cross-linguistically recurrent semantic roles, such as 'the one who feels cold', 'the one who eats sth.', 'the thing that is being eaten'. We then determine to what extent these roles are treated alike across languages, i.e. repeatedly grouped together under the same non-default case marker or under the same specific alternation with a non-default marker. Applying fuzzy cluster and NeighborNet algorithms to these data reveals cross-linguistic evidence for role clusters around experiencers, undergoers of body processes, and cognizers/perceivers in one- and two-place predicates; and around sources and transmitted speech in three-place predicates. No support emerges from non-default case assignment for any other role clusters that are traditionally assumed (e.g. for any distinctions among objects of twoargument predicates, or for distinctions between themes and instruments).