Overuse of Contractible Copula in Patients with Agrammatic Aphasia

• ¹ The Graduate Center, City University of New York, United States

In order to test Jakobson’s regression hypothesis, DeVilliers (1978) compared the 14 grammatical morphemes chronologically ordered by Brown (1973) in his longitudinal study of child language acquisition to the morphemes used by people with nonfluent aphasia. She reported that in the 5,000-word segments she studied from Howes’ corpus of transcriptions from 8 patients with non-fluent aphasia, fewer of Brown’s morphemes were used among the patients (8/14) than in the children and the order of morpheme use also did not reflect that of the children’s acquisition order. In particular, the patients used a high number of instances of both copula and contractible copula (e.g., It’s, They’re) which Brown reported to be relatively late-learned in children. Because we considered it unlikely that all the patients DeVilliers studied had agrammatism, and because multimorphemic contractions would seem compositionally complex for patients with agrammatism, we analyzed free narrative samples of people with agrammatic aphasia from the AphasiaBank (MacWhinney et al. 2011). Of the 111 speakers with aphasia in the AphasiaBank, we excluded data from all but the 19 participants whose databank label was Broca’s aphasia and classification by two SLPs classified as agrammatic (PWAA). Nineteen healthy controls (HC) were matched to the patients: Age: PWAA 53.7, HC 54 ; Education PWAA 14.4, HC 14.5. We employed the personal narratives participants produced, an average of 203 words per PWAA and 760 words for HC, segmenting the utterances according to criteria adapted from Saffran, Berndt, & Schwartz (1989). For each agrammatic utterance, we created a corresponding utterance to note the least complex full sentence for that utterance. Brown’s 14 grammatical morphemes were then tallied as correct, omitted, or erred on for each participant and frequency of use and percentage accuracy in obligatory contexts were calculated for each. We then followed DeVilliers’ (1974) method for rank ordering the morphemes by accuracy. Our findings indicate that only the first four morphemes acquired (present progressive, in, on, and plural) are used relatively accurately by both PWA and children in the early stages of acquiring language. However, one striking discrepancy is that reported but not discussed by DeVilliers (1974: the contractible copula, which was acquired relatively late in children (Brown, 1973; DeVilliers, 1978), was used with high accuracy in obligatory contexts among the participants with agrammatism. Indeed, it was used to provide over 20% of the verbs in our agrammatics’ discourse. By contrast, among 19 healthy controls in the data set matched for age and education the contractible copula was employed for only 6% of the verbs. However, among the possible forms of contractible copula the patients might have employed, only four forms were used more than once: It’s, That’s, I’m, There’s. Of the total number of copulas used the majority was the contractible copula “It’s”. The 19 healthy controls, by contrast, used contractible copulas with a fairly even distribution among the different forms. Thus, we conclude that PWAA appear to use these high-frequency items as single words rather than as contracted copula.