The past tense debate: is phonological complexity the key to the puzzle?

Theorists disagree over whether our language faculty is a single system or a dual one. Those supporting the latter position believe that English regular and irregular past tense verbs reflect this duality, with some proposing that each is processed by a rule mechanism and memorised lexicon respectively. Single system proponents believe instead that all verbs are processed by the same system, differing only in their degree of reliance on phonological and semantic representations. Regular past tense verbs involve greater phonological processing partly because they are phonologically more complex than irregulars. Early neuroimaging studies showing activation differences between the two have been taken as evidence for a dual system. However, it has been proposed recently that greater activation related to regular verb inflection was instead due to the failure to match regular and irregular verbs for phonological complexity (PC). Using a 2×3 ANOVA, the current event-related fMRI study tested this idea directly by manipulating regularity (regular, irregular) and PC (low, mid,and high) in 19 English-speaking monolingual participants. We found a main effect of PC, supporting the idea that phonological complexity cannot be ignored when considering differences between regular and irregular verbs. However we also found a main effect of regularity, demonstrating that differences over and above phonological complexity exist between the two types of verb. Even with phonological complexity matched, several regions including left inferior frontal gyrus and caudate were more activated for regular verb inflection. Temporal lobe regions and left hippocampus were among regions activated relatively more for irregular verb inflection. These latter findings suggest it may be premature to rule out a dual system account.