Short-term memory processes in the deaf.

Abstract

Two experiments are reported. In the first, profoundly deaf schoolboys read consonant sequences for immediate written recall, reading, according to instruction, either silently or aloud. Detailed error analysis suggested a dichotomous classification of subjects into those primarily relying on artioulatory coding (A group) and those relying on some other mediating code which could depend on shape (non‐A group). This classification correlated significantly with teachers' ratings of speech quality.The A group showed no effect of reading mode on total number of errors. But the non‐A group were significantly worse when reading aloud. Errors were distributed across serial positions effectively identically, regardless both of coding behaviour and reading mode. A recency effect was found of an order somewhere between that for hearing subjects reading this kind of material silently and reading it aloud.A test of the reliability of the classification into A group and non‐A group was made in the second experiment. In this the same subjects silently read word lists for immediate written recall. Lists were drawn from one of two vocabularies: one consisting of five pairs of common homophones, the other of five pairs of common words assumed to look similar. The A group recalled significantly more words of the latter lists, while the non‐A group recalled both lists equally well.The results are discussed in relation to current models for short‐term memory and to their implications for the education of the deaf. In the former connexion, the presence of recency with subjects who have no auditory feedback would be an embarrassment to ‘echo‐box’‐type accounts of recall. With respect to the latter, two main points are made: (1) Deaf school children do not all memorize in the same code, and for some it may be different from that used by teachers. (2) If memorizing names of things is impaired by reciting them aloud, a hitherto concealed weakness of a common instructional procedure is exposed.