Variable productivity: evidence from the English lax vowel constraint

Abstract

The productivity of the English Surface Phonetic Constraint which forbids the occurrence of a lax vowel word finally is tested by means of a number of experimental techniques, in order to ascertain whether apparent rule productivity varies in accordance with the particular criteria adopted: borrowing; memory for nonsense forms; a syllable division language game; and pluralization tasks including both recognition and production. The results of the experiments reported in section two demonstrate that there exists a methodological difficulty for claims about rule productivity: productivity tests do not all give the same results. For example, speakers make very few errors in repeating nonsense words ending in the disallowed lax vowel, yet no borrowings into English retain a final lax vowel from the source language. In addition, this research has some bearing on the usefulness of the tense/lax distinction for English vowels in such producitivity testing.