Why Chinese is not so foreign

Not so long ago, psycholinguistic studies of the reading process and literacy acquisition focused almost exclusively on English. But writing systems, and the structures of the spoken languages for which they are used, are so diverse that no one language or script can give a fair picture.

This book assembles 12 chapters that give an up-to-date account of what is known about literacy acquisition in the languages and scripts of the world, by means of in-depth studies of distinctive cases. The cases span the typology of scripts. It is well known that Italian contrasts with English in having very regular sound-letter correspondences, whereas Chinese is non-alphabetic, and Japanese mixes Chinese morphemic writing with syllabic kana symbols; so it is no surprise to find chapters on each of these writing systems.

Less obviously, the concept of phonetic regularity is ambiguous: a script can be perfectly regular in the way that its letters translate into sounds,while containing much unpredictability in the mapping of sounds into letters, and one chapter examines Greek as an example of this.

Another chapter looks at English-language literacy acquisition in Singapore, where history happens to have created a natural experiment allowing comparison of reading transfer skills between children whose mother tongue is Malay and those whose mother tongue is Chinese, written respectively in a very regular alphabetic script and an entirely non-alphabetic script, but who learn English in identical schooling conditions.

The book will be of interest to literacy researchers. Some insights, for instance Usha Goswami's finding that phonological awareness in children proceeds universally from syllables down to phonemes and smaller units, irrespective of script, are by now not surprising; and some are relevant more to academic psycholinguists than to people who teach children to read and write. That would apply to the claim by J. Richard Hanley, Ovid Tzeng, and H. S. Huang that "there is now overwhelming evidence that phonological processing strategies play a crucially important role in the way that children read Chinese", so that "the underlying cognitive skills and strategies involved in learning Chinese and English are probably not as different as was once imagined". It would also apply to the finding quoted by Peter Bryant, Terezinha Nunes and Athanasios Aidinis that French children learn word endings as markers of plurality independent of the verb/noun distinction, and thus write forms such as les timbres for "the stamps".

But many points have big implications for the primary-teaching profession.Thus, Ingvar Lundberg of Gothenburg University, examining the results of a 30-country survey of reading among nine-year-olds and 14-year-olds, argues powerfully that neither language or script differences nor different school systems nor teaching strategies explain variations, which are determined almost wholly by non-school cultural differences: "Reading is a skill that to a large degree is developed in arenas outside school."

Heinz Wimmer, Karen Landerl and Uta Frith find that dyslexia manifests itself differently among speakers of German, with a relatively regular script, than among English speakers: the reading accuracy of English dyslexics depends heavily on word frequency, which is not important for German dyslexics. Even with a script as regular as Italian, Giuseppe Cossu shows, children cannot spell every word that they can read.

David Share and Iris Levin, in a chapter on Hebrew, show that differences in the rate that children learn to write consonants and vowels are language dependent: not only Israeli children, using a script in which vowel indications are peripheral, but also English-speaking children are better at writing consonants than vowels in the early years, but for Spanish and Italian children, the reverse is true.

The book is written in a compact style, densely larded with references to the experimental psycholinguistic literature. This will inevitably limit its accessibility to a non-academic readership, but makes it a good starting point for exploration of the fields it covers. There cannot be many significant publications that are not cited here, provided the research has involved one of the languages used as a case study. One weakness, though, is that topics that have been discussed using other languages as test-beds fall through the net: the "phonic mediation" controversy happens to have been researched largely via Serbo-Croat and is not discussed here.

Geoffrey Sampson is professor of computer science and artificial intelligence, University of Sussex.