UC Berkeley

Background: Effects of word frequency on language comprehension and production are pervasive in speakers with and without aphasia. Frequency effects at the sentence level are likewise pervasive in neurotypical speakers, but have received relatively little attention in the aphasia literature. When discussed, frequency-based explanations have typically been dismissed as explanations of sentence-level deficits. Usage-based approaches to understanding the accessibility of clause and phrase structures are not widely used in aphasiology, in spite of their psycholinguistic plausibility. Aims: One aim is to show that evidence in support of probabilistic models of aphasic sentence comprehension and production has been available for a long time, but has lacked visibility. Our second aim is to respond to one of the few papers [Bastiaanse, R. Bouma, G., & Post, W. (2009). Linguistic complexity and frequency in agrammatic speech production. Brain and Language, 109, 18–28.] that directly challenge frequency-based explanations. The third aim is to discuss some of the reasons why probabilistic models in aphasia have been far slower to gain traction in aphasiology than in psycholinguistic research on neurotypical speakers. Methods & Procedures: We review available studies of probabilistic sentence-level deficits in aphasia. We revisit and reanalyse some of the data presented in Bastiaanse et al. [Bastiaanse, R. Bouma, G., & Post, W. (2009). Linguistic complexity and frequency in agrammatic speech production. Brain and Language, 109, 18–28.] Outcomes & Results: We find the data compatible in Bastiaanse et al. to be consistent with the predictions of a usage-based account. We argue that the paucity of studies of probabilistic and other usage-based effects in aphasia has its roots in the dichotomy drawn between “algorithmic” versus “heuristic” factors in sentence processing. We argue that this dichotomy is artificial. “Heuristics” are part and parcel of sentence processing in all speakers. Conclusions: The effect of usage—that is, of the individual’s past linguistic experience—is evident in the language of people with aphasia, as it is in speakers generally. We argue that usage-based approaches are promising because of the gradient and “noisy, but not random” nature of language processing in aphasia, and that they are natural companions to life-participation approaches to aphasia rehabilitation.