Cleary, A. M., & Greene, R. L. (2004). True and false memory in the absence of perceptual identification. Memory, 12, 231-236.
In 1990, Peynircioglu found that participants could discriminate between studied word fragments and unstudied fragments even when the fragments were unidentifiable. Cleary and Greene’s study further investigated these findings. They wanted to determine whether participants can discriminate between unidentifiable studied and unstudied words to show that this phenomenon is not dependent on the procedure Peynircioglu used. Also, discrimination without perceptual identification provides insight into how information might be processed through the cognitive system.
Experiment 1
Experiment 1 sought to determine if recognition without identification would occur using a perceptual identification task. The experimenters also wanted to investigate if it would occur even when the presentation modality was changed from study to test. They used both visual and auditory presentation, but testing was only visual.
Participants either saw words on a computer screen at a rate of 2 seconds each or heard the experimenter read words at the same rate. 60 words were presented in all. After hearing or seeing 15 words, participants took a 30 word test in which half of those words were the study words. During the test, participants saw alternating dollar signs and test words. After each word, they were asked to type in the word they saw or simply press “Return” if they couldn’t identify the word. They then rated how strongly they believed the word had been studied on a scale of 1 to 10, even if they didn’t identify the word. Finally, if the participant hadn’t previously typed in a word, he was forced to type in the word he thought he saw, even if it was just a guess. This continued for all 30 test words, then the next study phase began.
Table 1 shows recognition ratings. The proportions of studied and nonstudied words that were identified is represented in Table 2. In both conditions, unidentifiable words were given higher ratings if they were studied. Type of presentation did not matter. This means that recognition in the absence of perceptual identification occurs and does not depend on having the same perceptual features that were studied present at the time of test. In both conditions, higher ratings were given to identified items that were studied than those not studied. Participants were more confident in their recognition ratings when they could identify the words.
There were typical priming effects in that participants identified more studied than nonstudied words. Although it occurred in both conditions, these effects were larger in the visual condition.
Experiment 2
This experiment attempted to find false recognition without identification. The experimenters wanted to see whether participants would give higher recognition rates to critical distractor words when the associate words were studied (the DRM paradigm).
The test words came from the Stadler, Roediger, and McDermott norms (1999). The first five associates of the critical words were used and were presented in blocked format from strongest to weakest associative strength. The words were presented and tested the same way as in the visual condition in Experiment 1. The test list contained one associate from the four sets of studied words, the critical distractor from each of the studied sets, one associate from the four sets not studied, and the critical distractor from the nonstudied sets. An associate was always followed by its critical distractor.
Recognition with identification occurred, and participants gave higher ratings to unidentified studied associates than nonstudied associates. False recognition without identification also occurred, and higher ratings were given to unidentified critical distractor words when its associates were studied. An interaction between identification status and study status occurred; recognition and false recognition were greater for identified stimuli than for unidentified stimuli. Significant priming occurred for associates but not for critical distractors. This is interesting because priming did occur in stem and word fragment completion tasks (McDermott, 1997).
General Discussion
Participants can discriminate between studied and nonstudied words even when they are unidentifiable. This recognition does not depend on perceptual features. This indicates that information can travel directly from presentation to representation in memory. These experiments may resemble those investigating subliminal perception, but a striking difference is that subliminal studies deal with semantics, whereas this study only showed that participants can discriminate between studied and nonstudied words. This study does not determine whether meaning is extracted or used when participants make these recognition judgments. The researchers suggest that participants might be simply extracting letter information and matching it with stored letter memory. This can explain the results in Experiment 2 in that participants might think of the critical distractor while studying associates, then matching it with the stored orthographic features of that word at test.