Language& Communicarion, Vol. 2, No. 1, pp. 91-99, 1982. Printed in Great Britain. 0271-5309/82/01009-09$02.00/0 Pergamon Press Ltd. zyxwvutsrqpo WHAT WE MIGHT LEARN FROM ACQUIRED DISORDERS OF READING RHONDA B. FRIEDMAN M. Coltheart, K. Patterson and .l. Marshall (Eds), Deep Dyslexia. Routledge & Kegan Paul, London, 1980.444 pages, price $45. You are currently engaged in a complex perceptual and cognitive process; you are reading. The study of reading has a long history, and many of the original issues remain today. For example, when learning to read you may have been taught via the whole-word or ‘look-say’ method-that is, words were learned as whole units. Alternatively, you may have learned to read via the ‘phonics’ method. Each letter of a word represented a sound; when these sounds were strung together, a word was ‘sounded out’. These different methods of reading instruction reflect different convictions about what it is that we are doing when we are reading. Scientists have been trying to resolve such issues for nearly a century (see, for example, Cattell 1886, Huey 1906). Yet debate concerning the best method of reading instruction continues today (see, for example, Resnick and Weaver 1979), underscoring the lack of a universally-accepted theory of what we are doing when we read. A major question centers around the role of speech recoding in reading. According to one view, written language is simply a graphic representation of spoken language. There- fore reading is the process of transcoding graphic symbols into speech sounds, and then comprehending the covert ‘speech’ which is produced. An opposing point of view states that reading does not involve speech, that the reader derives meaning directly from the printed word. Proponents of the first point of view advocate the phonics method of reading instruction; proponents of the second view believe that the whole-word method is more beneficial. In the past few decades the debate has been taken into the psychology laboratory as psychologists have designed experiments to determine the extent to which speech recoding occurs during normal reading. This is not an easy task, since reading is an extremely complex skill composed of many subprocesses which occur automatically and simultaneously. It is difficult, then, to focus on any one aspect of the total process. One approach has been to present words or phrases very briefly and to measure the time it takes a normal reader to perform a particular task. For example, Rubenstein et al. (1971) asked readers to decide whether strings of letters were real words or nonwords. They found that it took longer to reject nonwords which are homophonous with real words (e.g. hoam) than it did to reject nonwords which are not homophonous with real words (e.g. hame). They concluded that such an interaction indicates that the letter strings are being recoded into speech sounds via grapheme-to-phoneme (spelling-to-sound) correspondence rules. Baron (1973) provided evidence for the opposing point of view when he showed that phrases such as ‘my knew car’, which make sense when pronounced, are judged to be nonsensical in meaning just as rapidly as phrases like ‘my no car’, which do not make sense even when pronounced. 91