Items seen for the second time in an experiment (old items) can be perceived more readily (fluently) than items seen for the first time (new items) (e.g., Jacoby & Dallas, 1981). We hypothesized that perceptual fluency is used as a cue for discriminating old from new items. In the test phase of a recognition task, each item was gradually clarified until it was identified, at which time subjects made an old/new judgment. We expected that fluently perceived (quickly identified) items would tend to be judged old regardless of their actual old/new status. In Experiment 1, words were more likely to be judged old both if they were quickly identified and, independently of this, if they actually were old. The latter finding implicates a factor (e.g., directed memory search) other than perceptual fluency in recognition judgments. Experiment 2 succeeded in reducing the contribution of this additional factor by using nonwords rather than words. Recognition judgments for nonwords were much more dependent on speed of identification than they were on actual old/new status. We propose that perceptual fluency is the basis of the feeling of familiarity and is one of two important factors that make variable contributions to recognition judgments. We thank Teresa A. Weber for her assistance in conducting Experiment 2.

No abstract

Two experiments compared the working-memory performance of highly gifted 13-and 14-yearolds who showed (a) both mathematical and verbal precocity, (b) primarily mathematical precocity, or (c) primarily verbal precocity. Experiment 1 examined (a) working-memory representational capacity for digit, letter, word, and location stimuli and (b) manipulation in working memory of digit, letter, and location stimuli. Verbally precocious youths showed enhanced capacity for words, and mathematically precocious youths showed enhanced capacity for digit and location stimuli. In working-memory manipulation, the mathematically precocious outperformed the verbally precocious with digits and letters. Experiment 2 examined speed of encoding into working memory. The verbally precocious showed enhanced encoding speed. Different types of intellectual talent appear to be associated with different working-memory characteristics and with differences in how digit and word stimuli are represented in memory.1 Although the discussion is written as if the units are individually meaningful, like Morton's (1970) logogens, this is just a semantic convenience. The same points apply to representations comprising patterns of units, like the distributed memory of McClelland and Rumelhart (1985).

No abstract

Weber, Burt, and Noll (1986) estimated that the time needed to switch attention between memory and perception was around 300 msec. The first two experiments in the present paper estimated switching time using a variation of their task. Subjects reported aloud lists of six items. The items were read ofT a computer screen (perception), recited from memory, or reported alternately from the two sources. The data show that the switching-time estimate is influenced by input/output compatibility, response-initiation times, and memory load. When these factors were controlled, estimated switching time dropped to around 100-150 msec. The data suggest, however, that the switch from perception to memory might be slower than the switch from memory to perception, which would invalidate the formula used to compute switching time. Experiment 3 tested the time for a single switch from perception to memory and a single switch from memory to perception by restricting report to one pair of items in the list. When the to-be-reported pair was precued, estimated switching time dropped to zero. When the pair was not precued, the memory-to-perception switching time remained at zero, but the perception-to-memory time was more than 400 msec. The pattern of results forced a reconceptualization of the task in terms of memory retrieval rather than attention switching. The attention-switching times appear to reflect processes required to select items from memory.Weber, Burt, and Noll (1986) pointed out that many tasks require rapid switching of attention between perception and memory; that is, the tasks are structured so that information in the environment and information in memory must be alternately accessed. Using the workingmemory conceptualization of Baddeley (1982;Baddeley & Hitch, 1974), Weber et al. described a model in which attention switching is the passing of control between memory and perception subroutines. Attention was figuratively described as a switch that could be moved between perceptual information in a visual/spatial scratchpad and memorial information in an articulatory/phonetic scratchpad. Items to which the switch was directed could be accessed by the control processes of working memory and output as responses. When a task required that perceptual and memorial information be intermixed at output, the switch would have to alternate rapidly between the two sources. Weber et al. were interested in the time necessary to accomplish the switch between memory and perception and whether switching required any of the limitedcapacity control processes available to working memory. Weber et al. (1986, Experiment 1) estimated switching time by comparing oral production times for lists of letters three or six items long. All-memory lists of either three or six items were recited aloud after they were silently studied. All-perception lists of either three or six items were read off a computer screen. Mixed lists consisted of a three-item memory sublist and a three-item perception sublist. The subject alternated report of items from ...