A final brief in the case against agrammatism: The role of theory in the selection of data