Orthographically mediated inhibition effects: evidence of activational feedback during visual word recognition.

Models of visual word recognition that have adopted an interactive activation framework (e.g., Coltheart, Curtis, Atkins, & Haller, 1993; Grainger & Jacobs, 1996) assume that activation can spread from semantic to orthographic representations via a feedback mechanism during visual word recognition. The present study used a mediated priming paradigm to test whether such feedback exists and, if so, under what conditions. Participants named aloud targets that were preceded either by a semantically related prime (e.g., dog-cat) or by a prime that was related to the target via a mediating word (e.g., cat-[dog]-bog). Direct evidence of activational feedback was obtained in the form of orthographically mediated inhibition effects. These mediated inhibition effects are consistent with activational feedback and support models of visual word recognition that have adopted an interactive activation framework.

Developmental differences in the ability to inhibit the initial misinterpretation of garden path passages.

The Hamm and Hasher (1992) procedure was used to examine whether children abandon their initial misinterpretation of garden path passages. Children (mean ages: 9 years-2 months and 12 years-4 months) and young adults (22 years-11 months) listened to either expected or unexpected (garden path) versions of passages that differed in the degree to which they provided contextual support for the final, correct interpretation. The first half of garden path passages initially led the listener to generate an incorrect interpretation. The second half of the passage provided information that clearly supported the correct inference and required the abandonment of the previous misinterpretation. Correct target inferences were formed and were equally available to all age groups. However, both groups of children were more likely than adults to accept competing inferences as being consistent with their understanding at the conclusion of garden path passages. The results suggest that developmental differences exist in the ability to inhibit thoughts that are no longer relevant in a listening comprehension and memory task.

Developmental changes in the inhibition of previously relevant information.

This study examined whether developmental differences exist in the ability to suppress irrelevant information that has entered working memory. Second graders, sixth graders, and young adults (M ages = 7. 5, 11.7, and 22.4 years, respectively) were asked to provide the ending for a series of sentences that highly constrained a terminal noun (e.g., "He mailed the letters without any stamps."). Responses to filler sentences were confirmed with the anticipated ending, whereas responses to critical sentences were disconfirmed with an unexpected ending (e.g., help). In either case, participants were instructed to remember the terminal noun (target) that was presented to them. Using another sentence-completion task, memory for both disconfirmed and target nouns was measured implicitly in terms of priming effects. Children, particularly second graders, showed priming effects with disconfirmed nouns, whereas young adults exhibited priming effects only with the target nouns. These results are consistent with the view that there are developmental differences in the ability to inhibit irrelevant information (Bjorklund & Harnishfeger, 1990; Hasher & Zacks, 1988).

The dissociation of repetition priming and recognition memory in language/learning-disabled children.

Recent research has found that the performance of learning-disabled and non-disabled children is dissociated on explicit and implicit tests of memory (Lorsbach & Worman, 1989). The current study further examined this phenomenon by comparing language/learning-disabled (L/LD) and nondisabled children (NLD) on tasks measuring primed picture-naming and item recognition. Included within the design of the experiment was the manipulation of both presentation format (pictures or words) and retention interval (immediate or 1 day). Children were initially presented with pictures and words. Performance was measured both immediately and following a 1-day retention interval on a picture naming task, an item recognition task, and a supplementary measure of memory for presentation format. The magnitude of facilitation associated with primed picture-naming was found to be independent of item recognition performance. In addition, the effects of population (L/LD and NLD) and retention interval (immediate test or 1 day) each produced dissociations between the magnitude of naming facilitation and item recognition performance. Results were discussed in terms of their implications for our understanding of the nature of memory difficulties in L/LD children.

Buildup of proactive inhibition as a function of temporal spacing and adult age.

Glenberg's theory of temporal distinctiveness (Glenberg, 1987; Glenberg & Swanson, 1986) was used to examine whether the buildup of proactive inhibition (PI) may be interpreted as an overloading of temporal retrieval cues. The Brown-Peterson task was selected on the basis of the assumption that successful performance on such a task requires the use of a recency rule where the subject must use temporal cues to retrieve the contents of the most recent study list. Two variables were manipulated that were presumed to affect the utilization of temporal cues: length of intertrial interval (ITI) (0 or 90 s) and adult age (young or elderly). Glenberg's theory suggests that temporal spacing improves memory by enhancing the distinctiveness of temporal retrieval cues. Thus, it was predicted that PI buildup should be less in the 90-s condition than in the 0-s condition. Assuming that older adults are more sensitive to the overloading of temporal cues, it was anticipated that older adults would experience a greater buildup of PI than younger adults at the 0-s condition. Results showed that although PI buildup varied with length of ITI, it did not vary with adult age. Analysis of delayed measures of recall, recognition, and temporal memory suggested that a retrieval interpretation based on Glenberg's theory is insufficient as an explanation of the current results; encoding difficulties also appear to contribute to the buildup of PI.

Dual-task performance as a function of adult age and task complexity.

A dual-task procedure was used to examine capacity demands of letter-matching in younger and older adults. Older subjects generally were slower on both tasks than were younger adults, but this difference was especially pronounced for the late stages of category matching, suggesting that retrieval and comparison of category information is particularly demanding for older adults.

Age differences in the rate of processing in short-term memory.

A probe-recognition task compared the speed and accuracy of visual, phonemic, and semantic recognition processes in younger and older adults. Each trial consisted of the sequential presentation of 10 words at a rate of either 350, 700, or 1400 msec per word. Following each study list, a probe word was given that could be identical to any word in the preceding study list or a homonym or a synonym. The two age groups were equivalent in their ability to identify recently presented identical words, homonyms , and synonyms. Age differences, however, were observed in the speed of access to information, and these differences were greatest on tasks requiring retrieval of semantic information from the secondary memory component of short-term memory.

Encoding and contextual components of word recognition in good and poor readers.

Good and poor readers from the third and sixth grades (9- and 12-year-olds, respectively), named visually presented words as rapidly as possible. Words were in clear or degraded form, and were preceded by related or unrelated words. Poor readers were hurt more by degradation than were good readers, and showed greater benefit from context. In general, the contextual benefit was greater with degraded words than with intact, and this interaction was especially pronounced in the poor readers. The results are consistent with an interactive-compensatory model of word recognition. Under conditions in which stimulus encoding is slow, contextual factors may compensate for this encoding deficit.