Introduction

Patients suffering from an acquired deficit of verbal short-term memory (phonological auditory–verbal short-term input store, PSTS: see Warrington & Shallice, 1969; Vallar & Baddeley, 1984a; Shallice & Vallar, this volume, chapter 1) typically show a cooccurring impairment of sentence comprehension (see Vallar & Baddeley, 1984b). This association had led a number of investigators to suggest that the PSTS contributes to certain aspects of sentence comprehension (see, e.g., Saffran & Marin, 1975; Shallice, 1979; Caramazza, Basili, Koller, & Berndt, 1981; Vallar & Baddeley, 1984b, for further details).

If one takes a closer look at the empirical data base from which these conclusions are drawn, it is, however, apparent that the observed patterns of comprehension impairment are, at least prima facie, rather heterogeneous. Most reported cases show a defective performance on the Token Test (e.g., Warrington, Logue, & Pratt, 1971; Caramazza et al., 1981; Basso, Spinnler, Vallar, & Zanobio, 1982; McCarthy & Warrington, 1987a). When more specific aspects of this sentence comprehension deficit, putatively produced by the phonological memory disorder, are considered, a comparative evaluation is more difficult, since variable methods and materials have been used in different patients. In a sentence repetition task in which the ability to paraphrase is regarded as an index of comprehension, case IL is, in most instances, able to offer an “acceptable” paraphrase. However, in the presence of an introductory clause passive constructions are frequently transformed into active, and when the sentence is semantically reversible, the meaning is also reversed (Saffran & Marin, 1975).