Degrees of severity and recovery in agrammatism: Climbing up the syntactic tree

Background: Agrammatic aphasia impairs syntactic abilities in production and comprehension. The Tree Pruning Hypothesis (TPH) suggests that the syntactic deficit in production can be described in terms of inability to access the high nodes of the syntactic tree. Aims: The current study explored patterns of individual differences between individuals with agrammatic aphasia, and suggested a characterisation for different degrees of agrammatic severity using the syntactic tree. A second aim was to test the path of spontaneous recovery in agrammatic aphasia. Methods & procedures: The first experiment tested 18 individuals with agrammatism: 16 were Hebrew speakers, and 2 were speakers of Palestinian Arabic. The syntactic ability of the participants was assessed with respect to three levels of the syntactic tree. To test the ability at the Agreement Phrase (AgrP) level, a task of agreement completion was used. To test the ability at the Tense Phrase (TP) level, a task of tense inflection completion was used. The ability at the highest level of the tree, the Complementiser Phrase (CP), was tested using elicitation tasks for two structures: Wh-questions and relative clauses. The second experiment tested the recovery of these four abilities over time in SB, an individual with agrammatism, starting 4.5 months post her brain injury until 18 months post-onset. Outcomes & results: The main findings were that the variation between the performance of different individuals with agrammatism and degrees of agrammatic severity could be accounted for by different sites of pruning on the syntactic tree. Severe agrammatism results from inability to access TP and the nodes above it, which impairs both tense inflection and CP-related abilities like the production of embedded sentences and Wh-questions. Milder agrammatism results from the inaccessibility of a higher node, CP, which causes a deficit to embedded sentences and Wh-questions, but leaves tense unimpaired. For both degrees of severity, agreement inflection was unimpaired. The second experiment showed that the spontaneous recovery of SB proceeded on the syntactic tree: the starting point was impairment in AgrP, TP, and CP, at the next stage AgrP recovered, and at the following stage TP recovered too. Conclusions: The results show that the syntactic tree is not only a useful tool for the characterisation of agrammatic aphasia at one point in time; it can also account for individual differences as well as for degrees of agrammatic severity, and can describe stages of spontaneous recovery. A milder impairment, or improvement in agrammatism, manifests itself in the ability to access higher nodes of the syntactic tree.