AUDITORY RECENCY IN IMMEDIATE MEMORY

Six experiments investigated the locus of the recency effect in immedi ate serial recall. Previous research has shown much larger recency for speech as compared to non-speech sounds. We compared two hypotheses: (1) speech sounds are processed differently from non-speech sounds (e. g. Liberman & Mattingly, 1985); and (2) speech sounds are more familia r and more discriminable than non-speech sounds (e.g. Nairne, 1988,199 0). In Experiments 1 and 2 we determined that merely varying the label given to the sets of stimuli (speech or non-speech) had no effect on recency or overall recall. We varied the familiarity of the stimuli by using highly trained musicians as subjects (Experiments 3 and 4) and by instructing subjects to attend to an unpracticed dimension of speec h (Experiment 6). Discriminability was manipulated by varying the acou stic complexity of the stimuli (Experiments 3, 5, and 6) or the pitch distance between the stimuli (Experiment 4). Although manipulations of discriminability and familiarity affected overall level of recall gre atly, in no case did discriminability or familiarity alone significant ly enhance recency. What seems to make a difference in the occurrence of convincing recency is whether the items being remembered are undegr aded speech sounds.