The following online article has been derived mechanically from an MS produced on the way towards conventional print publication. Many details are likely to deviate from the print version; figures and footnotes may even be missing altogether, and where negotiation with journal editors has led to improvements in the published wording, these will not be reflected in this online version. Shortage of time makes it impossible for me to offer a more careful rendering. I hope that placing this imperfect version online may be useful to some readers, but they should note that the print version is definitive. I shall not let myself be held to the precise wording of an online version, where this differs from the print version.

— In fact I think I am right in saying that this review was never published; something strange happened at the journal which had asked me to write it.

Geoffrey Sampson, Sussex University


Writing Systems: a linguistic approach.  Henry Rogers.  Oxford: Blackwell, 2005.  xvii + 322 pp.


Writing Systems: an introduction to their linguistic analysis.  Florian Coulmas.  Cambridge: Cambridge University Press, 2003.  xix + 270 pp.



Both of these books are student textbooks about writing systems, and I should probably begin by declaring an interest:  twenty years ago, I myself wrote an earlier book in this genre.  Rogers’s and Coulmas’s volumes are rather different from one another (and from mine).


Rogers aims to give an account of all scripts used for the world’s languages now and in the past.  After two ground-clearing chapters he devotes successive chapters to different scripts or script-families.  Rogers’s final chapter discusses how writing systems should best be classified; and he includes a series of useful appendices on matters such as linguistic technical terms and phonetic transcription.


Rogers’s exposition is remarkably well-informed, up to date, and comprehensive; I learned a great deal from it.  The last script Rogers discusses, for instance, is one of which I knew nothing previously, but which apparently settles a question that I had thought could be answered only conjecturally.  My book defined a category “semasiography” for systems of visible symbols (such as international road signs) which convey meaning independently of any particular spoken language, and I speculated whether in principle there could be a semasiographic system complete enough to allow us to express anything we want to express, as spoken languages do.  I argued that this was quite possible, though I did not think it likely in practice; to my surprise some critics objected vehemently that the concept was unreasonable.  But Rogers tells us that such a system already exists:  it was published in the mid-20th century by an immigrant to Australia, Charles Bliss, and has been extensively used in a few countries, including Rogers’s Canada, in education of people with disabilities such as cerebral palsy.


Coulmas is more concerned with abstract conceptual issues; his chapters deal with themes, such as “consonants and vowels”, or “analysis and interpretation”, rather than with particular script families.  For a student readership in particular, one problem with Coulmas’s approach is that the questions which most interest him are ones that many readers will not see the need to ask, as when he raises “the non-trivial question of the c-ness of <c>”.  It may be reasonable to approach the relationship between written and spoken language by surveying the history of ideas on the topic from Plato and Aristotle onwards, as Coulmas does in his opening chapter, but such a survey needs to be selective.  It is not obviously worthwhile to ask what the Zen school of Buddhim thought about writing, particularly since (as Coulmas says) “there is no consistent Zen view on writing”; yet this is the longest section of the introductory chapter before Coulmas reaches the present day.  He draws a Zen analogy, obscure to me, between a calligraphic rendering of a Chinese character and Magritte’s painting of a pipe captioned “Ceci n’est pas une pipe”.


Underlying this chapter are real and significant questions about whether the development of writing leads to fuller articulation of cognitive processes that remain relatively simple and vague while a language is only spoken.  But that strikes me as an issue to raise towards the end (if at all) rather than in the first chapter of a textbook, and I am not sure how many of Coulmas’s readers will grasp what the problems are which his references to past thinkers are intended to illuminate.


One desirable property for a textbook on an intricate subject about which different cultures have developed separate analytic traditions is a willingness to cut through those traditions in order to explain the facts in terms accessible to members of our own culture.  Rogers is good at this, finding neat English analogies to clarify properties of alien scripts.  He cites the abbreviation lbs for “pounds”, where a root which is a translation equivalent in another language (Latin) takes an English grammatical suffix, in order to clarify how speakers of Middle Persian used Aramaic “morphograms” in writing their language.  Coulmas repeatedly allows the non-English traditions to get in the way of explanations that could have been straightforward.  Discussing how the Greeks used Semitic consonant letters to represent Greek vowels, he transcribes some Greek words phonetically into modern Greek pronunciation – which for a Greek would be the natural thing to do, but is not very relevant to the way the Greeks adapted the Semitic alphabet almost 3000 years ago.  Coulmas transcribes a word for “favourable” as [evnus], and then says “Take away the vowels, and almost nothing is left:  … ns”.  This will be lost on most readers, who will hardly realize that modern Greek [ev] derives from ancient Greek [eu].


A telling example of differential resistance to other cultures’ intellectual blinkers relates to the authors’ accounts of Chinese script.  Rogers, though not apparently an Oriental specialist, demonstrates a thorough grasp of the evolution of Chinese writing.  He understands that the script went through a stage when a character invented as a picture for one word was used also for a wide range of other words pronounced similarly, until, later, most words in such a set would be differentiated by adding extra elements related to their respective meanings.  Chinese discussions of the structure of Chinese characters almost invariably set out from the six-way classification of character types laid down by the Han-dynasty lexicographer Hsü Shen.  By the Han dynasty, Chinese writing was already a fully evolved system; Hsü Shen’s classification of character structures was necessarily aprioristic, and for someone trying to grasp the workings of the script his categories are more misleading than enlightening.  Rogers mentions Hsü Shen’s categories, after giving his own account of the script and with a comment on their limited value; Coulmas treats them as gospel.


People are naturally interested in the origins of the oldest script of all, which is normally taken to be Sumerian writing (though, incidentally, Rogers quotes a 2001 article which apparently gives fresh evidence suggesting that Egyptian writing may be older).  Since the 1970s, Denise Schmandt-Besserat has achieved considerable publicity for a theory, originally proposed by Pierre Amiet, which claims that Sumerian writing evolved from a system of clay tokens used in keeping agricultural accounts.  According to Coulmas, Schmandt-Besserat has “convincingly demonstrated” this.  But the theory is highly debatable; some scholars who are versed in the relevant area of archaeology have voiced strong objections, and if other authors who discuss writing systems accept Schmandt-Besserat’s conclusions this could be partly because many of us know too little about the facts.  Rogers’s comment is that Schmandt-Besserat’s theory “is not without controversy, but it seems to be gradually gaining acceptance”; that seems exactly right.


Too often, Coulmas’s own theoretical conclusions are questionable.  He repeatedly stresses that “no writing system encodes every distinction relevant in its language”:  alphabetic scripts never embody a one-to-one correspondence between letters and phonemes, and syllabic scripts “hav[e] fewer signs by a large measure than the languages they are used to write have speech syllables”.  Is that so?  As a Professor of Japanese Studies, Coulmas certainly knows much better than I, but in the latter case I wonder why the Japanese kana syllabaries should not count as complete. As for alphabetic scripts, Finnish is almost perfectly phonemic; Rogers lists the very few trivial ways in which it could be said to deviate from the ideal.  I do not believe, as Coulmas does, that there is something about language or about the nature of writing which ensures that no language’s script could ever approach scientific perfection – though it is true that scripts normally emerge from series of historical accidents which sometimes interfere with exact representation of speech, and that goal is less important to users of a writing system than a linguistic scientist might expect.


There are a handful of errors in Rogers’s book; but they concern side-issues.  A passage on p. 24 implies that the Chinese capital has sometimes been located in non-Mandarin-speaking areas; I believe that is not so.  On p. 76 the Chinese character for “garment” is wrongly glossed “sky”.  The Roman alphabet acquired the letter Y for the same reason (to spell Greek loanwords) and at the same period as it acquired Z; Y is not comparable to the much later innovations J, U, and W (p. 174).  Britain was never “the westernmost region of the ancient Roman Empire” (p. 185), though it was the northernmost.  I question whether English handwriting ever had a sign derived from long-s + z (p. 189), though German script does.  And Rogers confusingly insists on using the term “moraic” for most of the scripts which are usually called syllabic, because he uses “mora” for the CV part of a CVC syllable, e.g. the cu- of English cut.  Surely, in standard linguistic parlance, moras are irrelevant to English? – the ‑n of a Japanese syllable like kan is a separate mora because, in that language, kan lasts for two beats rather than one in poetic metre.


But these are forgiveable minor blemishes.  Rogers’s book could scarcely be bettered as a textbook on its subject, and it has much to offer to advanced scholars.  Coulmas, I fear, has missed his target.



Department of Informatics

Sussex University