Motion Verbs Corpus Study

This corpus was a control study which tested the telicity of verbs from the self-paced reading study.

200 occurrences of each experimental verb were extracted from the Deutsches Referenzkorpus using Cosmas II. Every occurrence was accompanied by the last two sentences of the preceding context. Only examples containing finite forms of intransitive, literal uses of verbs were chosen for the analysis. Two annotators classified the verbs' occurrences as "telic", "atelic" or "ambiguous", based on standard diagnostics for (a)telicity. Ambiguous cases were discarded; occurrences were labeled "ambiguous" if the two annotators classified them oppositely or agreed on an ambiguous interpretation.

The annotation procedure led to a corpus of 2392 occurrences (113–184 per verb) of which 1470 (61%) were telic ones.