[Journal of the Simplified Spelling Society, 26, 1999/2 pp30,31]
[See Journal and Newsletter articles, Pamflet 15, Cut Spelling and Papers by Chris Upward.]

Anglo(-Japnese Non-)Dyslexia.

Christopher Upward.

Chris Upward sumrizes and discusses th report 'A Case-Study of an English-Japanese Bilingual with Monolingual Dyslexia' by Taeko N Wydell and Brian Butterworth. JSSS previusly publishd th folloing article on japnese: Christopher Seeley 'The 20th Century Japanese Writing System: Reform and Change' in J19 1995/2, pp27-29. Th presnt revew is ritn in Cut Spelng.

1. Comparing litracy between languajs.

Evidnce for th harm don to litracy levls by th unpredictbl spelng of english has always been importnt to th Simplified Spelling Society in refuting th skepticism it ofn encountrs in making its case. One kind of evidnce that has been acumulating in recent years arises from comparisn between litracy standrds in difrnt languajs with difrnt riting systms, both alfabetic and non-alfabetic. A pioneerng work in this field was editd (1973) by former SSS President John Downing, tho its comparisns wer impressionistic rathr than statisticl. Mor recent studis with detaild statisticl analysis for pairs of individul languajs include: Thorstad (1991) for italian versus english lernrs; Upward for english students in english and jermn (1992); Landerl et al. (1997) for jermn versus english dyslexics.

2. Dyslexia and japnese, granularity and transparency.

Th reserch paper discusd here first outlines th developmnt of sientific thinkng about th natur of dyslexia, concluding that it now apears not as a singl, homojeneus disordr, but as a ranje of disabilitis that each sufrr is difrntly afectd by. Dificlty in recognizing and manipulating th sounds of words (fonlojicl deficit) is, howevr, a recurng featur, as ar poor short-term memry and jenetic predisposition to litracy problms.

Th japnese riting systm(s) is/ar then described, as functionng on two levls. One levl is th almost entirely predictbl fonic represntation of sylabls by singl symbls (kana). Th othr is th use of chinese caractrs (kanji) wich may be red in a variety of unpredictbl ways. Japnese riting thus difrs from alfabetic riting in not requiring fonemic analysis of sylabls, ie, consnnts ar representd jointly with th vowls that follo them.

Dyslexia as seen in alfabetic languajs is relativly rare in japnese, and seems to arise from dificltis of visuo-spatial perception (ie, of memrizing th complex structurs of th caractrs), rathr than of fonemic procesng. This sujests a 'hypothesis of granularity and transparency', by wich dyslexia wil be less comn 1) in riting systms wher th relationship between sound and symbl is mor predictbl (=mor transparent), and 2) wher th elemnts in th riting systm (letrs, caractrs) ar supra-fonemic, so they require less detaild analysis, and ther 'granularity' is corsr, ie, ther symbls represent at least two successiv fonemes rathr than just one.

This hypothesis cud explain wy dyslexia may be rarer in japnese than in english. First, th kana sylabris of japnese represent th sounds of sylabls mor predictbly than th letrs of ritn english represent english fonemes, ie, an importnt part of ritn japnese is mor transparent than english. Secnd, both th kana symbls and th kanji caractrs ar mor corsly graind (ie, hav corsr granularity) than th letrs of th roman alfabet used in english. In terms of transparency english is thus at a considrbl disadvantaj from th lernrs point of vew, and in terms of granularity al alfabetic riting systms ar at a disadvantaj compared with japnese. (This is not to say that japnese may not hav disadvantajs of its own, such as th long time required for lernng a suficient numbr of kanji caractrs, but they ar not at issu here.)

3. Th case histry.

That is th bakground to th analysis of th litracy problms experienced in english by AS, th teenaje son of hyly litrat anglo-australian parents livng in Japan. AS had al his scoolng in japnese scools, but spoke english at home, wher he receved a thoro groundng in ritn english. His scool performnce in jenrl was wel abov th avraj for japnese students of his aje, exept for his litracy in english. Here, quite severe dificltis had been noticed erly on by his parents, and at th aje of 13 he was diagnosed as dyslexic in english.

In vew of this disparity between his problms with english and his abov-avraj performnce in japnese, AS undrwent detaild asesmnt of his litracy standrds in both languajs. For both japnese and english, th reserch took care to compare his performnce with aje-machd monolingual subjects. Th paper describes his impressiv proficiency in handlng th profound ambiguitis of th sino-japnese kanji caractrs, wich gave him a readng aje in japnese wel abov his cronlojicl aje; and th sylabic kana symbls causd him no trubl, wethr they representd real words or nonwords, th latr being especialy significnt since dyslexics in english typicly find nonwords hard to decode.

In english AS had no dificlty with th alfabet as such, and his ability to anlyz fonemes was norml. Howevr, wen it came to identifyng hole ritn words in english his performnce was poor. His 'dijit span', ie, his ability to take in letrs preceding and folloing th one on wich he was focusd, was only 5, wel belo avraj (this is an efect of poor short-term memry); a consequence of this is that, since japnese words require fewr 'dijits' than english, his ability to absorb hole words in ritn english was autmaticly less than in japnese, wher for instnce a singl kanji caractr may represent a hole word. AS cud repeat polysylabic english nonwords quite succesfuly, but found som consnnt manipulation tasks hard in english. In singl-word readng tests, he wud al too ofn substitute a word of simlr apearnce for th corect one, evidnce that his fonic decoding was weak. This showd very strongly in his atemts at readng nonwords.

Varius tests mesurd ASs performnce in ritn english against that of both english and japnese students of th same aje. These tests found him always to perform significntly worse than th english control group, and jenrly worse even than th japnese controls, tho his comand of spoken english and his lifetime exposur to english wer far superir to thers. In particulr, wher th japnese students comnly mispelt english words by using alternativ permisbl spelngs for th sounds concernd, ASs spelngs wud mor ofn be randm gesses.

4. Discussion points.

Wydell & Butterworths reserch represents a valubl adition to th evidnce for th dificltis that th traditionl orthografy of english causes by comparisn with that of othr languajs. Spelng reformrs wil find th acount of curent litracy testng tecniqes intrestng, and wil welcm th multilingual perspectiv adopted, wich includes refrnces to chinese and danish (problmatic riting systms), italian, jermn, malay and spanish (al relativly straitforwrd riting systms, with jenrly predictbl sound-symbl corespondnces) beside, of corse, english and japnese. Th acount givn of th simplicitis and complexitis of japnese, with its two kana sylabris and ambiguus use of chinese caractrs (kanji) is a useful introduction to its uniqe riting systm.

A numbr of ideas developd in th paper ar worth reflectng on.

Friths (1985) concept of stajes in litracy aquisition (an initial 'logografic' staje, then an 'alfabetic', and a final 'orthografic' staje) is at one point aplyd as a benchmark for ASs developmnt, altho todays undrstandng of th importnce of fonics for initial litracy must cal it into question (as is indirectly hintd elsewher in th paper). Recent experience of oposing initial teachng methods surely sujests that they, rathr than any predetermnd 'staje', ar wat decide how lernrs first com to grips with alfabetic riting.

The concept of difrng granularity of riting systms afectng ese of litracy aquisition is thot-provoking. Is th notion convincing that japnese kana symbls and kanji caractrs (once lernt) ar esir to read (especialy by dyslexics) than alfabetic script because they stand for longr segmnts of speech, fal within a shortr dijit span and do not require fonemic analysis? Let us compare th consnnt + vowl valus of japnese kana with othr sylabic riting systms, such as ethiopic (Bloor, 1995) or indian devnagri (Coulmas, 1989) or, despite gretr structrl variability, korean hangul (Sampson, 1985): in these a symbl for a givn consnnt is distinguishd by a predictbl aditionl markr for th folloing vowl, but they hav a simlr rufly square shape to kana/kanji; shud they then be hardr to read than japnese because they mark consnnts sepratly from vowls? Being mor predictbly structurd, surely they shud be esir to lern. Furthrmor, may we not considr alfabetic riting with regulr sound-symbl corespondnces to be simlrly sylabic, only with a horizontl sequence of consnnt markrs insted of havng them integrated into th vowl symbl itself? Wy shud can, fan, man, pan or kid, kill, kin, kiss be hardr to read than th same sylabls representd by a singl (perhaps compound) symbl? In th case of AS, it wud seem that his limitd dijit span may hav hindrd his alfabetic readng, but conversly lernrs with impaird visuo-spatial awareness ho hav trubl lernng japnese caractrs myt find alfabetic readng esir. Ther is a gret deal mor that one wud like to no about readng processes in japnese befor comng to firm conclusions on such questions; for instnce, is silent readng jenrly fastr in japnese than in alfabetic languajs? and how do japnese readrs react wen they meet unfamilir kanji?

Finaly, we do not no how severe ASs dificltis wud hav been if 1) he had been introduced to ritn english thru systmatic fonics, and 2) if his alfabetic languaj had been far mor transparently spelt than english, eg, if he had been a spanish-japnese bilingual.

Insofar as limitd dijit span agravated ASs dificltis with ritn english (perhaps it dos for many dyslexics), we may reflect how much help they wud derive from simplifyd spelng in english that significntly shortnd th avraj length of words. Cut Spelng for instnce uses som 10% fewr letrs than dos traditionl orthografy - thus wen th 6 letrs of freight ar cut to 4 in freit, th reduced word length fals within a 5-letr dijit span.

5. Conclusion.

Tho larjly a singl case study, th implications of this reserch paper ar far-reachng for our undrstandng both of dyslexia and of th difrnt impact of difrnt riting systms on th process of litracy aquisition. Without concluding that english spelng needs reformng, it puls no punchs in pinng a major part of th blame for ASs problms on its lak of transparency, and increses th quotebl evidnce availbl to spelng reformrs in making ther case.

Abundnt fresh evidnce of a simlr kind comparing litracy in sevrl othr languajs to that in english has apeard since th Wydell-Butterworth paper, in a colection of papers editd by Harris & Hatano (se belo). This wil be revewd in JSSS J27 - 2000/1.


Bloor, T (1995) The Ethiopic Writing System: a Profile. JSSS J19 1995/2. 30-36.

Coulmas, F (1989) The Writing Systems of the World. Oxford: Blackwell. 181-190.

ed. Downing, J (1973) Comparative Reading: Cross-National Studies of Behavior and Processes in Reading and Writing, New York: The Macmillan Company.

Frith, U (1985) Beneath the surface of developmental dyslexia. Eds. Patterson, K E, Coltheart, M, & Marshall J C Surface dyslexia: Neuropsychological and cognitive studies of phonological reading, London: Lawrence Erlbaum Associates Ltd.

Landerl, K, Wimmer, H, & Frith, U (1997). The impact of orthographic consistency on dyslexia: a German-English comparison. Cognition, 63, 315- 334.

Harris, M & Hatano, G (eds, 1999) Learning to Read ad Write: A Cross-Linguistic Perspective, Cambridge University Press: Cambridge Studies in Cognitive and Perceptual Development.

Sampson, G (1985) Writing Systems. London: Hutchinson. 120- 144.

Thorstad, G (1991). The effect of orthography on the acquisition of literacy skills. British Journal of Psychology, 82: 527- 37.

Upward, C (1992). Is traditionl english spelng mor dificlt than jermn? Journal of Research in Reading, 15(2): 82-94.

Wydell, T & Butterworth, B (1999) A Case Study of an English-Japanese Bilingual with Monolingual Dyslexia. Cognition, 70, 273- 305.

Back to the top.