ABSTRACT
This chapter presents two single-case studies that illustrate rehabilitation programs designed for patients with impaired numerical transcoding abilities. The two treatment programs were framed according to the asemantic model: The patients were simply taught small sets of rewriting rules. The two models of number-numeral transcoding operations resort to a multicomponent information processing analysis. The data-driven models are tailored to the formal structural features that we assume to be relevant for the processing of number and numeral lexical primitives. In such a precise theoretical framework, the transcoding mechanisms may be viewed as consisting of psycholinguistic procedures operating essentially in stacklike environments. The multicomponent structure of the transcoding models distinguishes four main components, the impairment of which should induce identifiable errors. The transcoding process per se is a set of rewriting rules triggered by the lexical information received from the preceding component. The result consists in loading an output buffer with the information specifying the target code lexical primitives to be produced.
