New directions in the study of semantic deficits: a comment on Storms et al. (2003).