Lexical restrictions on passive uses in English: a large-scale corpus investigation [Poster]
A. Myachykov, Dominic Thompson, Fernanda Ferreira, Christoph Scheepers · 2012
Using a combination of Principal Component Analysis and machine-learning techniques on large-scale corpus data, we demonstrate (a) that passive verb-uses are consistent across regional variants (American vs. British English) and modalities (spoken vs. written), (b) that the use of different passive-forms (get or be, full or truncated) is constrained by the notional verb, and (c) that there is a clear relationship between syntactic and semantic similarity in the clustering of notional verbs. The results support lexical/functional accounts of the English passive.