Reinstating study context produces unconscious influences of memory

Abstract

Having read a word does more to benefit its later perceptual identification when many, rather than few, of the words in the test list have been previously read. Some have suggested that this proportion overlap effect is produced by an intentional use of recognition memory or recall in the perceptual identification task. Contrary to this account, we found that words that are easily recognized (words generated from an anagram at study) do not gain more from increasing overlap than do words that are poorly recognized (words read at study). These findings are problematic for claims that word perception relies on a module, such as a logogen system, that is separate from the rest of memory.Evidence of memory is often shown by performance on an indirect test, although a direct test reveals no evidence of memory (for reviews, see Johnson & Hasher, 1987;Richardson-Klavehn & Bjork, 1988). For example, reading a word makes it more likely that a person will be able to identify the word later when it is flashed for a test of perceptual identification, and that effect on perception is observed even when the person does not recognize the word as previously read (e.g., Jacoby & Dallas, 1981). The test of perceptual identification is an indirect test of memory because the instructions refer only to the task at hand and do not refer back to a particular prior event, although the subject's performance on the task might be influenced by memory for that prior event. Recognition and recall are direct tests of memory because the instructions refer to a target event in the personal history of the subject and ask the subject to consciously recollect that earlier event. We use the phrase unconscious influences ofmemory (e.g., Jacoby & Kelley, 1987) to refer to effects of a prior experience that arise on an indirect test of memory although a person is unable to recall or recognize the relevant prior experience. Others (e.g., Graf & Schacter, 1985) have referred to such effects as effects on implicit memory.A major problem for using indirect tests as a means of investigating unconscious influences of memory is that performance on these tests might sometimes reflect aware influences of memory that are undetected by the experimenter. Although not instructed to use memory for a particular prior event, subjects given an indirect test of memory might sometimes discover the relation between study and test and intentionally use memory much as they would for a direct test of memory. The problem is similar to the one that has plagued attempts to show unconscious perception. Indeed, because of the possibility of awareness undetected by the experimenter, Holender (1986) argues that there is no convincing evidence for the existence of unconscious perception. Similarly, in studies of memory, Richardson-Klavehn and Bjork (1988) note that many effects ascribed to unconscious influences of memory might be contaminated by conscious recollection. For example, the enhanced perceptual identification performance of old, relative to new, words might sometimes r...

In defense of abstractionist theories of repetition priming and word identification

“…Allen and Jacoby (1990) replicated this pattern of results using the same task, but included a manipulation intended to assess whether the increased priming in the high-overlap condition was the product of participants "catching on" and using explicit memory to enhance their performance. Specifically, two study conditions were included.…”

“…Bodner and Masson (1998) found that masked priming was enhanced when a high rather than a low proportion of the prime-target pairs were the same, particularly for low-frequency words. Unlike the Jacoby (1983a) and Allen and Jacoby (1990) studies, these results cannot be the product ofexplicit contamination, because the primes were presented for 60 msec and masked, making the primes unavailable to consciousness. On the basis of these results, the authors supported an episodic account of short-term priming.…”

“…First, in Experiment I of Allen and Jacoby (1990), recognition memory was actually more sensitive to the proportional context manipulation when items were read than when they were generated, and there was no significant difference in Experiment 2 (although there was a trend for larger effects in the generate condition). Accordingly, the observed tendency to obtain larger context effects for priming in the read condition in the two studies does not provide strong evidence for a dissociation between priming and recognition memory.…”

Monitoring source in an unconscious plagiarism paradigm

“…We have argued that there may not be as much source monitoring going on in these generative tasks as in the recall-own task. This does not exclude the possibility that we have raised elsewhere (Marsh, Landau, & Hicks, 1996, in press) that people who are busily engaged in generating a new solution may not adequately monitor the source of each candidate (e.g., Allen & Jacoby. 1990;Jacoby & Kelley, 1987).…”

Context effects in repetition priming are sense effects

“…Consider experiments in which context was provided at study but not at test (see, e.g., Allen & Jacoby, 1990;Blaxton, 1989;Jacoby, 1983;Levy & Kirsner, 1989;MacLeod, 1989;Oliphant, 1983). When single words formed the context manipulation, as in Jacoby's (1983) experiment, study items were accompanied by various cues-for example, an antonym (HOT-COLD) or a related item (SNOW-COLD) versus a neutral letter pattern (XXX-COLD).…”