Priming of Old and New Knowledge in Amnesic Patients and Normal Subjects a

Neuropsychological studies of memory pathology have demonstrated that amnesic patients who are characterized by diverse forms of neurological dysfunction are severely impaired on tasks that tap recall and recognition of recently studied information.'-' One feature of virtually all memory tasks on which amnesic patients are impaired is that they demand explicit recollection of the context and content of recent learning episodes. During the past decade, however, two kinds of evidence have established that amnesic patients can demonstrate relatively normal memory performance on implicit tests, which do not demand explicit recollection of a learning episode. One is that amnesic patients can acquire motor, perceptual, and cognitive skills in a normal or near-normal manner, even though they recall little or nothing of the learning episodes in which they acquired the skills." A second kind of evidence that amnesic patients perform normally on implicit tests is provided by the phenomenon of repetition priming: A single exposure to an item on a list facilitates amnesics' processing of that item on a variety of retention tests that do not require explicit recollection of the study episode. Warrington and Weiskrantz?.' for example, demonstrated that amnesics and controls showed a similar tendency to complete three-letter fragments of familiar words with items from a recently studied list. However, the amnesics' performance on a Yes/No recognition task was seriously impaired. Similar data have been reported by others.'"*'' Amnesics also exhibit intact repetition-priming effects, in the face of poor recall and recognition, on implicit tests such as lexical decision' and homophone spelling.'2 The dissociation between priming and recollection in amnesia is paralleled by demonstrations in normal subjects that experimental variables that have large effects on recall and recognition have little or no effect on the magnitude of repetition priming in word c ~ m p l e t i o n , ' ~ ~ ' ~ word identifi~ation,"~'~ and lexical decision.".'' Moreover, the magnitude of priming effects on these tasks can be statistically independent of recognition memory p e r f ~ r m a n c e . ' ~ ~ ' ~ ~ ' ' The observed dissociations have led a number of investigators to propose that repetition-priming effects are mediated by a "memory system," "memory process," or "form of memory" that is relatively spared in amnesia and that can function independently of the damaged memory process or system that underlies conscious and explicit recolle~tion.~~~~'~~''~'~~'~~~~ A fundamental question regarding the process or system that underlies repetition priming is whether it can support the acquisition of new knowledge or it is restricted to the activation of old, existing knowledge. One of the hallmarks of organic amnesia is the virtual absence of any capacity to retain new information. The possibility that the