Introduction
While semantic and phonological deficits in aphasia are relatively well-studied, less attention has been given to morphological processing. Nevertheless, there are active debates about morphological processing that may be informed by investigating morphological deficits, including the extent to which regular and irregular forms are computed similarly. Theories of processing vary with respect to this point: dual-mechanism accounts propose that regular verbs are computed via rule-based processes combining stems and affixes, while irregular verbs are stored and retrieved separately from their stems. In contrast, full-decomposition accounts posit that both regular and irregular verbs are stored in a decomposed fashion and computed using combinatorial processes. The present study compares the predictions of these two accounts using the single-word reading performance of an aphasic individual with a morphological deficit. We designed two tasks to decouple effects of morphology and phonology and evaluate regular and irregular verb production. In particular, we first compared error patterns of regularly-inflected forms to uninflected homophones (e.g., praise vs. prays) to establish the presence of a morphological deficit. We then compared error rates and types for regularly-inflected (sin-sinned), irregularly-inflected (win-won), and phonologically-matched word pairs (tin-ton). Dual mechanism accounts predict that error rates and types of the irregularly-inflected forms will match the phonological word pairs as all are listed separately in the lexicon. In contrast, full-decomposition accounts predict regular and irregular verbs would demonstrate similar rates and types of morphological errors, distinct from monomorphemic words.
Case Report
RMI, 39, right-handed male presented with aphasia secondary to L-MCA CVA. His production in spontaneous speech, reading, and writing, includes frequent morphological errors, with semantic and phonological errors also occurring.
Experiment 1
RMI was administered a single-word reading task containing homophone pairs that orthogonally varied morphological and phonological complexity (e.g. prays-praise, locks-lox). The list contained 53 homophone pairs and was administered 4 times (N=424 words total). Deletion of final consonants occurred significantly more often for morphologically-complex words (locks→lock; 94/212, 44%) compared to homophones (lox→[lak]; 20/212, 9%; X2=63.94, p<.05). This difference reveals a morphological deficit.
Experiment 2
RMI was administered a single-word reading task containing 40 irregular verb pairs (win-won), 40 regular verb pairs phonologically-matched to the stem (sin-sinned) and 40 monomorphemic word pairs matched to the irregular verb pairs (tin-ton). He produced morphological deletion errors on 30.0% (12/40) of inflected regular verbs (sinned→sin) and on 42.5% (17/40) of inflected irregular verbs (won→win), whereas the analogous error (e.g. ton>tin) never occurred on the monomorphemic pairs (see Table 1).
Discussion
Taken together, the results of Experiments 1 and 2 indicate a combinatorial morphological deficit affecting both regular and irregular verb production. The errors in producing irregularly-inflected verbs are similar to the production of regular verbs, and not similar to phonologically-matched monomorphemic words. These findings are consistent with full-decomposition accounts of morphological processing and inconsistent with accounts that posit differential processing of regular and irregular verbs. We will present additional analyses from these tasks and related tasks involving morphological comprehension and production across modalities.