Orthographic and lexical processing of visual letter strings.

Three experiments tested whether orthographic and lexical properties of a letter string influence the time to search for a component letter. 102 Canadian undergraduates served as Ss. Orthographic redundancy, defined by single-letter position-specific frequency, facilitated the search of targets specified prior to and simultaneously with the letter string. Words were searched faster than nonwords when the target followed the letter string. Neither orthographic nor lexicality had significant effects when the position of the target within the string was certain. Results are consistent with a hierarchical-levels model of word perception in which the activation of detectors at different levels is constrained by task demands. (French abstract) (31 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

More on semantic priming in a fully recurrent network: A response to Masson (1999).

The authors examine the account given by M. E. J. Masson (1999) of the processes underlying the performance of his (M. E. J. Masson, 1995) model of semantic priming of word naming. The account is shown to be incomplete at best and not necessarily inconsistent with the view that priming in such models reflects differences in interference and not facilitation. The interference interpretation also applies to performance when the network is not reset at the start of each simulated trial, which was proposed as an appropriate analogue to the start of human trials. Other issues raised by Masson (1999) are briefly examined. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Multiple processes in the automatic encoding of frequency information.

Conducted 2 studies in which a total of 96 undergraduates were unexpectedly asked to make frequency judgments about words used as distractors during a series of short-term memory tests for digit sequences. Frequency estimates of repeated items were influenced by the interaction of the duration of each vocalization and the number of trials intervening between the 2 occurrences of an item. As the separation between the occurrences increased, the frequency estimates increased only when the items were vocalized for long durations. This same result was not found for mere recognition of items, which was influenced independently by the frequency of presentation and duration of rehearsal. These findings are consistent with a multiple-process hypothesis of frequency representations, according to which distinct frequency traces vary in strength. Results also indicate that distinct representations for repeated events are encoded spontaneously during incidental processing. (French abstract) (15 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Savings in a recognition test.

Tested the hypothesis that maintaining a response set from original learning to relearning is necessary to produce savings in a recognition task (RCT). 32 Ss learned 28 number–noun paired associates and 6 wks later took a 4-alternative RCT. That test was followed by a relearning trial on items from the original pairing or in new pairings. The relearning list was tested with cued recall and a 4-alternative RCT. Significant savings were found on both the recall and recognition tests. Results are discussed in terms of the paradigm and research of C. M. MacLeod (see record 1989-07460-001). (French abstract) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Encoding, retrieval, main effects and interactions: Were Lockhart and Craik (1990) on the level?

Investigated whether a main effect of encoding emerges when retrieval tests are equated for difficulty. 132 introductory psychology students were tested in 2 experiments. Ss encoded words semantically or by rhyme, and then judged whether unstudied words were related to studied words. A list structure (L. Jacoby, 1975) was adapted to remove the overall difference in difficulty across semantic and rhyme retrieval tasks. Results indicate no main effect of retrieval task, and the main effect of encoding was not dependent on the retrieval task. Semantic encoding led to a greater likelihood that Ss would judge an unstudied word to be related to a studied word. It was concluded that test words shared either a semantic or a rhyme relation with studied words, the advantage of semantic encoding was not dependent on the retrieval criterion. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Limiting lexical activation limits perceptual priming in masked word identification

Four experiments tested the hypothesis that perceptual priming of a word depends on the prior lexical processing of the word (Rajaram & Roediger, 1993; Weldon, 1991). Experiment 1 showed that first-letter naming reduced priming relative to reading a whole word on two tests: word fragment completion and masked word identification. In Experiment 2, naming the first letter of a word took longer than naming the letter presented alone, and led to better masked word identification. Experiment 3 showed that masked word identification was enhanced by prior word processing more for low frequency words than for high frequency words, but only when words had been read aloud. Experiment 4 tested whether the auditory input accruing from reading a word aloud was the source of facilitation and frequency effects. Participants judged either the frequency of the whole display or the positional frequency of the first letter in the display, and indicated their decisions manually. The major findings from Experiment 3 were replicated, ruling out the cross-modal source of those effects. It was concluded that activation of a lexical unit, one component of word processing (Vriezen, Moscovitch, & Bellos, 1995), is a critical determinant of the perceptual priming of that word.     

Semantic priming in fully recurrent network models of lexical knowledge.

The claim that A E. Masson's (1995) model of lexical knowledge provides an account of semantic priming is evaluated in a series of simulations. It was found that although a semantically unrelated prime produces slower simulated naming of the target word than a related prime, the difference reflects interference by the unrelated prime and not facilitation by the related prime, which has little absolute effect. The facilitation reported by Masson was produced by the simulated "neutral" stimulus that intervened between prime and target, not by the semantically related prime. Simulations of a stochastic variant of Masson's model and a "brain-state-in-a-box" model of word naming also failed to produce semantic facilitation, raising the possibility that in fully connected recurrent networks, facilitation does not occur via indirect collateral pathways. Methodological implications for evaluating priming in simulated networks are raised. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Evidence against the computation of abstract letter identities in visual processing.

Conducted 2 studies conceptually similar to the study of D. Besner et al (see record 1985-05766-001) in which letter strings were reorganized into sets where the single letters that were different were phonologically similar (e.g., G vs C) or dissimilar (e.g., G vs K). Latencies for same–different decisions about mixed-case letter strings were faster when the different letters were phonologically dissimilar. Results suggested skilled readers access name codes of individual letters in making their speeded classifications. In the present studies with 16 undergraduates, phonological similarity was manipulated in either the 1st or 4th letter of a mixed-case 4-letter string. A similarity effect was found when the 1st letter was varied. Results are not consistent with the view that abstract letter identities are computed in the course of same–different decisions about simultaneously presented mixed-case displays. (French abstract) (12 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)