ABSTRACT
These assumptions and goals in the study of language pathology favour a single-case methodology. Dissociations between intact and impaired processes can only be identified within a single system; what matters is that there is a discrepancy for a single patient. On the assumption of a universal processing system it is enough to show that one system exists where a dissociation between functions occurs to infer the discreteness of those functions within the universal processing system. If even one patient can read words but not non-words, the system must allow distinct processing for these. Where a number of patients exhibit the same dissociation, the evidence for discrete functions is reinforced. But such reinforcement is provided by a series of single case studies rather than by traditional group studies. Group studies seek statistically valid generalisations about groups of patients, and these cannot throw light on possible dissociations and hence on discrete components of language processmg.
