Semantic and phonetic memory codes in beginning readers

Abstract

In two experiments a group (N = 15) of poor beginning readers and of good readers (N = 15) were auditorily presented with continuous item lists. The children were asked to indicate whether each item had occurred previously in the list. In Experiment 1, using real words, later items were related either semantically or as rhymes to earlier ones. False positives to each item type were taken as indices of memory coding and showed that good readers encoded both semantic and surface aspects of items. In contrast, poor readers made a large number of meaning-based confusions (saying "old" to house when home had been presented earlier) but almost none based on rhyme (home/comb confusions). In a second experiment, with phonetically legal nonwords as items, poor readers made a significant number of phonetic false positives, although the good reader controls made more. The results are interpreted as confirming that poor readers are relatively insensitive to surface features of language but that this weakness is most marked when sound and meaning are both available as memory codes. 333The research reported in this paper addresses the question of how young retarded readers code words and pseudowords in a memory task. The work was motivated by two sets of considerations: first, the accumulating evidence that poor readers show only a weak active awareness of the structure of the speech code and, second, arguments concerning the primacy of meaning over form in preliterate children's reaction to language. These latter ideas will be explored more fully following Experiment 1.There is now considerable evidence which indicates that preliterate children, and those who are experiencing difficulty in learning to read, perform poorly on word games that require them to manipulate aspects of the sounds of words. For example, Firth (1972) discovered that 8-year-old backward readers were much worse than an intelligence-matched group of average readers in eliding segments from orally presented words-in saying,for instance, what is left when IfI is taken from If ae nl (see also Bruce, 1964). In an unpublished study from our laboratory, Arnold (Note 1) confirmed Firth's findings with phoneme elision and extended the investigation to elision of syllabic segments from words (e.g., /lImItI minus /lIm/). Poor readers were also weaker at this task. Rozin and Gleitman (1977) have reviewed much of the available data on word segmentation and combination skills in retarded readers; they found consistently that these subjects were poorer than comparable coping readers.Difficulties in operating with the speech code are not (1976) showed that introducing acoustic confusability into to-be-memorized letter strings adversely affected performance of superior readers more than that of marginal and poor ones. This was true with both immediate and delayed recall and with visually or auditorily presented strings. In a similar vein, Mark, Shankweiler, Liberman, and Fowler (1977) found that the false-positive errors of good readers in a word recognition test ...

Development of phonetic memory in disabled and normal readers

“…They cited an earlier study by Shankweiler and Liberman (1976) who showed that auditory presentation of rhyming and nonrhyming letter strings gave the same results as a visual presentation. This was later confirmed in an auditory version of the Mark et al task (Byrne & Shea, 1979). Thus, the insensitivity to phonetic similarity of memory items found in young disabled readers indicates a general deficit in phonetic memory rather than a reading-specific deficit.…”

Creating proactive interference in immediate recall: Building a DOG from a DART, a mop, and a FIG

“…As each item is generated, it is matched to each item in memory, and if a match exceeds threshold, that item is produced as the response. In long-term recognition studies, it has been documented that distractors (new items) that are rhymes of targets are more likely to be falsely recognized than nonrhyming distractors (Byrne & Shea, 1979;Runquist & Blackmore, 1973), although this effect is not large. Ifwe assume that phonemic information in the short-term domain has a privileged role, such false recognition might occur more frequently in the short-term domain.…”

Errors in short-term memory for good and poor readers

“…Using a paradigm that tested recall for time ERRORS IN SHORT-TERM MEMORY 445 periods longer than the assumed limits of STM, Byrne and Shea (1979) obtained evidence that children who were poor readers were able to use a phonetic code (albeit poorly) when forced to do so by the use of pseudoword stimuli, but otherwise tended to favor a semantic code (although at least one study failed to replicate this finding [Winbury, 1984]). However, when some of us examined the errors that good and poor readers made on a standard word string repetition task, we found no indication that either group of children was using a semantic strategy (Brady et aI., 1983).…”