Commentary Commentary on Ullman et al

Ullman and colleagues should be congratulated for an energetic assault on some of the core issues in neurolinguistics. This work attempts to detect an impact of productivity differences in morphology upon fluency differences in aphasia. This basic goal is well-motivated empirically and should eventually succeed. However, I believe that the links currently being proposed between language and the brain are not well structured conceptually. If these conceptual problems can be successfully addressed, then this work could make a contribution. Without these clarifications, I believe that the current formulation represents a conceptual step backward. The idea that word formation involves both productive and non-productive processes is an old one, dating back to Panini and the Greeks. This contrast has served as a cornerstone of diachronic morphology for nearly two centuries. In developmental psycholinguistics, the decision to analyze words as based on rote or rule crops up at virtually every stage in the analysis of emerging grammars (MacWhinney, 1978). From the viewpoint of neurology and aphasiology, the contrast between fluent and non-fluent aphasia is also fundamental. The traditional interpretation of this contrast has been one that focuses on the role of anterior areas in motor processing and posterior areas in sensory processing. In the 1970s, evidence that nonfluent aphasics also had disabilities in comprehending complex syntax (Zurif, Caramazza, & Myerson, 1972) led researchers to challenge this analysis. Following Jakobson (1955), neurolinguistics began to imagine that frontal areas might control grammatical processing, whereas posterior areas control lexical insertion. The dual-route analysis of Ullman and colleagues represents an attempt to advance the Jakobsonian view