Three experiments are reported in which Ss saw lists of letters for immediate recall, but had to say a neutral sound as they perceived each letter. In Experiment I this technique substantially reduced recall of letters, spoken as capitals, as compared with that for normally voiced lists, but no difference between "suppression" and voicing emerged if the letters were pronounced in altered fashion (e.g., BDF pronounced as "buh duh fuh. . ."). In Experiment II, having to say a sound which coincided with the dominant sound of the presented list led to slightly poorer recall than did the saying of non-coincident sounds at presentation. Experiment III showed that the main advantages for recall of articulation consisted in raising the recall particularly of lists of low acoustic similarity, reducing order errors, and raising the recall particularly of the early and middle items of lists (the latter only if articulation was normally intoned as opposed to monotonous). It is concluded that the advantage of articulation at presentation increases as the acoustic similarity of the items decreases. Some general implications concerning STM are drawn from this conclusion.

Ss either silently read, mouthed, whispered, soft-voiced, or loud-voiced lists of 8 consonants foT immediate free recall. White noise of an intensity sufficient to mask S's hearing of his own soft voice (as far as possible) was present during presentation and/or recall. The results showed that when there was no noise at presentation, recall increased as vocalization activity increased; when there was noise throughout presentation and recall, loud voicing was significantly superior to whispering; when the noise was switched off for recall, loud voicing gave insignificantly inferior recall to soft voicing and whispering. The results are interpreted as being consistent with the view that when lists are vocalized Ss try to attend particularly to the auditory feedback, but that if this is prevented, more attention is paid to the motor cue. Presentation rate was not found to play a significant role in this experiment. 1 This work was done while the author held a Studentship at Trinity College, Cambridge. He is particularly grateful to A. J. Watson for advice and discussion, and to Prof. O. L. Zangwill for his encouragement. The noise generator was kindly lent by M. E. Parlett.

No abstract

An experiment is reported in which patients with Korsakoff's psychosis were compared with alcoholic patients without Korsakoff's psychosis on a probe-recognition task. The evidence was consistent with the view that the Korsakoff patients had a greater susceptibility to proactive inhibition than did the control patients; a signal detection analysis of die results suggested that this might be reflected in lower initial registration strength for recent items. The results are discussed in the light of current views on •This work was supported by Ontario Mental Health Foundation Grant No. 211. We are extremely grateful to the superintendents of the Ontario Psychiatric Hospitals at Brockville, Hamilton, Kingston, Lakeshore, and Whitby for their generous assistance and provision of facilities for this work. In particular, we would thank Dr S.C. Laverty of the Kingston Hospital for his aid in helping us find the alcoholic control subjects. We are also grateful to Dr W.A. Wickelgren for his reading of an earlier draft. 414

Three experiments are reported in which items required for immediate recall were alternated with "interfering" items not required for recall, but demanding overt responses. Presentation rate was also varied. In Experiment I letters were alternated with "interfering" digits: having to calculate the digits mentally led to reduced recall of the letters. In Experiment II "required'' letters were repeated, three presentation rates also being used: lists with repetitions were more poorly recalled than lists without repetitions, but in die former condition, die slower the rate, the better was recall. In Experiment III "required" letters were alternated with digits or rhyming letters: letter interference produced poorer recall, but there was also an interaction with presentation rate. An "interfering" letter closely preceding a "required" letter was more likely to replace die latter in recall than was one preceding it at a longer interval. It is argued that both rehearsal opportunity and response competition determined recall, and that simple versions of decay theory or interference theory are inadequate to cover all these results.