L. Weiskrantz University of Oxford
One of the truly gratifying developments over the past decade, and one to which Endel Tulving and his clinical colleagues have made particular contributions, has been the marriage between theory construction in human experimental psychology and research findings in neuropsychology. For the latter to be of any use for the former, conclusions and inferences must be capable of being cast in terms of independent or interacting parallel pathways, or in ordered sequences of hierarchies of processes along given pathways, including feedback loops upon that pathway. Needless to say, this is so only if the theoretical flow diagrams are themselves cast in such terms. But, in neuropsychology, the raw data are always cast in the form: Lesion X produces Deficit on Task A. If one substitutes "Treatment" (e.g., a drug, sensory deprivation, stress, etc.) for "Lesions" and substitutes "Behavioral Outcome" for "Deficit", then precisely the same logic applies, and so the issues have much wider application (cf. Weiskrantz, 1968a). What I explore here is how we go from those simple conjunctions in the raw data to inferences that are in a useful form. Having considered some of these general principles in the first part of the chapter, I then follow up with some evidence about dissociations that lead one to inferences about multiple memory systems.
In neuropsychology, as it is practiced today, there are two approaches, sometimes it seems, two cultures, that derive naturally from neuropsychologists collectively, but not necessarily individually, having quite practical responsibilities as well as a deep curiosity about the mind-brain question. There are those who have an abiding and important practical concern, often driven by the everyday needs of the clinical scene, who wish to describe and treat patients with similar clusters of deficits, typically drawn from common aetiological sources. There are those, on the other hand, who focus on particular pieces of neuropsychological