Can we separate verbs from their argument structure? A group study in aphasia

Background: Given the integral role that verbs play in sentence production, understanding verb deficits is critical to clinical practice. Difficulties in sentence production are often directly related to an inability to retrieve argument structure information which, according to most theoretical accounts, is specified at a lexical level as part of the semantic representation of the verb. The presence of an argument complexity effect when retrieving verbs in isolation, i.e. increased difficulty with increasing number of arguments required by the verb, is considered evidence of the integral link between verbs and the ensuing structure. Recent reports, however, of relatively intact verb retrieval and impaired argument structure suggest that difficulties with verbs and argument structure may not always co-occur. One explanation for this is that lexical-syntactic information may be stored separately to lexical-semantic information at the lemma level and then differentially impaired. Aims: To determine whether the presence of difficulties creating argument structure in a group of people with aphasia with verb-retrieval deficits consistently aligned with the underlying nature of their impairment, where semantic involvement resulted in argument structure difficulties and post-semantic impairment resulted in intact argument structure. This would provide evidence either to support or to challenge the view that argument structure is lexically specified. The presence of an argument complexity effect within the participants' single verb naming was also investigated to examine the relationship between verb argument information and any subsequent difficulties creating argument structure for sentence production. Methods & Procedures: Verb retrieval and argument structure production were investigated in 12 people with aphasia with verb-retrieval deficits on single-word and sentence-production measures. Outcomes & Results: Eight of the 12 participants presented with a semantic verb deficit (five semantic and three mixed semantic and phonological) and had difficulties in creating argument structure. Four participants presented with a phonological verb deficit. Of these, two had intact argument structure while two had impaired argument structure. An argument complexity effect in verb naming was seen in both groups. Conclusions & Implications: Semantic verb-retrieval deficits consistently resulted in argument structure production difficulties, supporting the view that syntactic information is stored within the lemma. The presence of argument structure difficulties in two participants with phonological deficits; however, lends support for the separate coding of lexical-syntactic from lexical-semantic information at the lemma level, a factor that has both theoretical and clinical implications with respect to diagnosis and management.