[Journal of the Simplified Spelling Society J22 1997/2 p26]
[Chris Upward: see Journals, Newsletters, Leaflets, Media, Pamflet 15, Book, Papers.]

Alarm Bels Ring for Fonics and/or Spelng Reform.

Christopher Upward anlyzs

'Aspects of Writing in 16+ English Examinations between 1980 & 1994'

by A L Massey & G L Elliott, University of Cambridge Local Examinations Syndicate (UCLES), 1996, 55pp + apendices, ISBN 1 873140 08 8.

Thanks ar du to Alastair Pollitt and Alf Massey of UCLES for ther helpful coments on a draft of this paper.
To ilustrate its final conclusion, th articl is ritn in Cut Spelng.

Note: Both th UCLES survey and th presnt analysis of it predate the Literacy Task Force (Barber, 1997) wich th new Labour govrnmnt in th UK has now elevated to th status of National Literacy Strategy. Th Strategy plans to takl th fonic deficit identifyd belo, but is not yet proposing to adress th undrlyng alfabetic deficit. It is th task of th SSS to persuade it to do so (se JSSS J21, pp27-32, and furthr documnts in this issu).

Abstract.

This paper first (§2) describes th politicl bakground to studis of litracy standrds and then revews (§3) som erlir atemts to identify chanjing standrds of spelng acuracy. In §4.1-4.3 it sets out th findngs of th UCLES survey that standrds of spelng acuracy in english scools hav declined dramaticly between 1980 and 1994. Th UCLES mispelng lists ar then anlyzd (§4.4-4.5) and a distinction made between errs that fail to represent pronunciation (ie, failur to mastr th riting systm itself), and errs that represent th pronunciation rongly (eg, failur to memrize anomlus spelngs). Furthr analysis leads to th conclusion (§4.6) that reduced fonic undrstandng is th major cause of th decline. Th final section (§5) then shos how redundnt letrs ar asociated with som two thirds of mispelngs, and argus th case for teachng Cut Spelng as a medium far mor conduciv to acurat spelng than TO.

1. Difrng aims of mispelng studis.

Th fenomnn of mispelngs has over th years been reportd on from varius angls. One angl has been th analysis of th mispelngs themselvs, to establish wat they tel us about th dificltis of th riting systm (an overvew of this aproach is givn in Upward, 1994). Anothr angl is th comparisn of spelng standrds acheved by difrnt categris of riters, wethr syncronicly between difrnt cuntris (eg, for english and italian, Thorstad, 1991; for english and jermn, Upward, 1992), or diacronicly (as linguists cal historicl comparisns) alias longitudinly (as sycolojists cal comparisns of peples spelngs at difrnt points in time). This revew discusses a new study of th latr kind.

2. Th politicl context.

Th perenial complaint of oldr jenrations that ther desendnts fal short of ther eldrs has ofn been aplyd to languaj, and, within languaj, to yung peples spelng in particulr. Monitrng standrds is necesry for developng educationl policis, but it is importnt to distinguish th curmujnly anecdotes of oldr critics from carefuly anlyzd statisticl evidnce. In th past decade this distinction has been complicated in Britn (perhaps elswher too) by a politicl and ideolojicl overlay. On th one hand a permissiv aproach, widely identifyd with th left, has regardd 'corect' spelng as secndry to encurajng childrens fre expression, and has therfor tryd to minmize its importnce. On th othr hand a C/conservativ reaction has hylytd poor spelng as symtmatic of a decline in pedagojicl disiplin encurajd by 'trendy progressivs' on th left. Meanwile, real proof of chanjing spelng standrds has been hard to find, partly because sampls of comprbl riting from difrnt jenrations hav not been availbl for analysis.

3. Previus studis.

Som erlir studis may, howevr, be noted.
At th beginng of th 1990s a booklet provocativly entitled Sponsored Reading Failure (Turner, 1990) arousd fierce controvrsy in Britn by lambastng syns of sharp decline in litracy standrds thru th 1980s. It based its claim on figrs from som 10 difrnt Local Education Authoritis in suthrn England, but its polemicl styl and inability to present its statistics clearly for th lay readr seriusly reduced its valu. Nevrthless, if one skims over th detail, one may be redily persuaded by its vivid pictur of a jenration of educators drivn to abandn fonics by a fashn for alternativ, far less efectiv teachng methods, and of litracy standrds consequently in a state of catastrofic decline. (Se endnote [1] for a small sampl of its argumnt.)

Som othr studis apear at first syt to sho no decline. One such was th NFER report Spelling it out (Brooks, et al., 1993), revewd in an erlir issu of JSSS (Upward, 1993). Its jenrl verdict was that th standrds of 11- and 15-year-olds mesurd between 1979 and 1988 wer "quite good"; but ther wer som statisticl uncertntis in th survey, and th pupils concernd may hav been too old to be afectd by any coruption of initial teachng methods during th 1980s such as was diagnosed by Turner. Simlrly not sujestng any decline (tho standrds wer found to be badly in need of improvemnt) wer two ALBSU reports (Hamilton, 1987; Elkinsmyth/ Bynner 1994) wich wer also discusd in erlir issus of JSSS (Upward, 1988, 1995a); but they too relate to jenrations ho aquired ther litracy skils wel befor 1980.

A mor recent study, howevr, dos giv evidnce for longterm decline. Publishd by ALBSU undr its new name, th Basic Skills Agency (1995), and revewd in JSSS J20 (Upward, 1996/1), it givs th results of a spelng test administrd simltaneusly to nearly 1,000 subjects spred over five jenrations of adlts (in that respect it difrs from th previusly mentiond surveys, wich compared standrds acheved at an intrvl of sevrl years by peple of th same jenration). In th 1995 study, th worst spelrs wer found to be th yungst: 16-24-year-olds avrajd 35.7% words mispelt; next worst wer th 55-60-year-olds, with 32.8% mispelngs; th midl groups performd betr, tho with a decline thru th jenrations, th 45-54-year-olds avrajng 27.0%, th 35-44-year-olds 27.5%, and th 25-34-year-olds 28.6% errs.

Th education of th oldst aje-group may wel hav sufrd from erlir scool-leving and th social turmoil of th 1940s (war, destruction of homes, sepration from and loss of parents, evacuation, etc), and ther may also be som efect of ajing. Th decline in standrds of th yungst aje-group is most striking: spelng-acuracy of 16-24-year-olds is over 7% worse than that of peple ten years oldr, over 8% worse than those twenty years oldr, and nearly 9% worse than those thirty years oldr. Or, difrntly expresd, todays 16-24-year-olds mispel over 25% mor words than do ther eldrs. It is concevebl that a reversd aje efect may be at work here, with adlts spelng acuracy improving with experience, and th Basic Skills Agency says (persnl comunication) that it is "wary about making firm judgements about standards declining on the basis of this research", altho they "do tend to think that there was a period in schools when rather less concentration was paid to 'secretarial' skills than was perhaps desirable". Wat is clear from these figrs, howevr, is that yungr jenrations spel worse than ther eldrs going bak sevrl decades.

4 Th 1996 UCLES study.

4.1 Scripts and examnations.


In 1996 th University of Cambridge Local Examinations Syndicate (UCLES) publishd a substantial comparativ study (Massey & Elliott, 1996) of vocablry, spelng, punctuation, sentnce structur and non-standrd english found in 1,199 16+ examnation scripts from th years 1980 (299 scripts), 1993 (420 scripts) and 1994 (480 scripts). Compared with th erlir surveys described in §3 abov, th UCLES survey stands out for its thoro and sutl analysis. Th sampl consistd of th fourth sentnce taken from th scripts of 30 boys and 30 girls awardd each of th grades A-E (or, for 1993-94) A-G in each of th years in question.

Th circmstnces of th examnations wer by no means identicl, and th study is duly cautius about its methodolojy and conclusions, saying (p5):
This paper does not pretend to solve the conceptual or methodological problems and cannot say conclusively if grading standards in English have risen or fallen in recent years. But it does present some rare comparative data concerning features of the writing of pupils awarded ostensibly 'equivalent' 16+ examination grades in 1980 and 1994, which are interesting and worth public consideration.
Th chief difrnces between th examnations wer that th 1980 GCE targetd th forml use of ritn english and was taken by a minority of pupils selectd for ability, wile th 1993 and 1994 GCSE examnations targetd wider use of languaj and wer taken by pupils of a wider ability ranje. Ther was som indication that th 1993 candidats myt also hav been of hyr avraj ability than those of 1994. In 1994 new National Curriculum criteria wer introduced, wich emfasized (p14) "knowledge about language ... and the importance of Standard English, together with presentation: explicitly including spelling, layout and neatness as integral parts of all writing tasks." Th examnation was ofrd at two levls, but thre quartrs of pupils wer entrd for th hyr 'tier'. In principl th grades awardd from 1980 to 1994 wer intendd to maintain standrds, but many factrs such as th abov, but also including chanjes in th ethos of english teachng, made it dificlt to be sure that like was always being compared with like.

4.2 Mor errs: 1980:50 1993:93 1994:149.


Wen th results wer publishd in April 1996, they wer reportd in th press as revealng "a significant slide in teenagers' writing skills since 1980" (Charter, 1996), with figrs quoted of 50 mispelngs in 1980 deteriorating to 93 in 1993 deteriorating to 149 in 1994. In fact Table 5 (p23), from wich th figrs wer taken, marjnly undrstates th numbr of errs for 1994, since it excludes th scors of th new categry of 'star' performrs A*, hos riting paradoxicly containd mor errs than that of th plain A performrs; by avrajng th two categris (to ensure A/A* results wer not overepresentd in th total sampl), a total of 151 errs (rathr than 149) is obtaind. A furthr posbl slyt undrstatemnt may arise from th fact that th survey counts mispelngs repeatd in each year/grade only once, thereby not distinguishng between an err repeatd by one pupil (as was probbly th case with th one repetition from 1980, *hotle for hotel) and an err made by mor than one pupil (as was probbly th case with confusion of to/too by B-performrs in 1993): errs made by mor than one pupil surely do sujest loer overal spelng standrds. Tho less relevnt statisticly for th presnt survey, we may also considr that an err repeatd by th same riter is mor significnt, inasmuch as it mor likely reflects a firm misaprehension of th corect spelng, rathr than just momentry uncertnty or inatention. If we ad in subsequent ocurences of th same err, th numbrs of mispelngs in th survey (for Grades A-E only) rise to 51 for 1980, 101 for 1993 and 155 for 1994.

But we may also feel that th survey has been too severe in som of its jujmnts. Thus, a cupl of verbs spelt with th -IZE endng (eg, apologize) and a dozn instnces of th populr alternativ alright for all right wer countd as mispelngs (shud one not in fact distinguish between 'th ansrs wer alryt' [= satisfactry] and 'th ansrs wer al ryt' [= 100% corect]?). Ther wer also quite a few cases wher a slyt imprecision of handriting myt sujest a rong letr (eg, *pigedn for pigeon, *lake for take, *full for fall, *cares for cases, *but for put), and indeed (se belo) th survey aknolejs that th numbr of 'mispelngs' in a strict sense of th word may hav consequently been overestmated. It wud be intrestng to no wethr use of a word procesr (without spelchekr!) wud produce a loer mispelng count by elimnating such 'handriting' errs, or indeed wethr it wud produce fewr mispelngs jenrly.

Th figrs for word length and ranje of vocablry may also hav som berng on mesurs of spelng acuracy. It was found (p18) that mor able pupils tendd to use longr words than less able, or at least they used mor letrs in riting words (avraj of grade A performrs: 4.2 letrs; and of grade G: 3.6 letrs per word). This rases th question wethr mispelngs wich omitd letrs (eg, *no for know by a grade G performr) wer countd as 'shortr' words. Certnly, studis conectd with Cut Spelng (Upward, 1996a) hav repeatdly observd that many mispelngs involv omission of redundnt letrs. But th oposit efect may also hav arisn if alright was countd as one longr word compared with th two shortr words all right, altho it was also penalized as 'rong'. In adition to th decline in word-length from 1980 to 1994, ther was also a decline (p21) in ranje of vocablry, wich sujests that th decline in overal spelng proficiency was perhaps gretr than indicated by th numbr of mispelngs: th 1994 candidats mispelt far mor words, even tho ther vocablry was mor limitd, so if they had used as extensiv a vocablry as ther 1980 equivlnts, they myt hav mispelt even mor words. Th survey overlooks this posbility wen it says (p22; se belo for ful context) that girls "compensated for a narrower vocabulary with greater accuracy" - if boys had used a naroer vocablry, ther spelng myt hav been mor acurat, a hypothesis suportd by th findng (Moseley, 1989) that spelng dificltis inhibit th use of richr vocablry.

Altho a mor refined analysis myt therfor hav produced somwat difrnt figrs undr varius hedngs than those publishd, th gains and losses that wud hav arisn by one criterion or anothr wud tend to cancel each othr out, and ther is no reasn not to take th UCLES figrs as brodly indicativ of trends.

Th surveys acount of its aproach is worth quoting at length (pp 22, 23, 26):
The writing samples were checked for correct spelling by a research assistant, using word processing software for an initial screening and subsequently checking each word clerically. In a few cases it is possible that spelling errors were recorded where the real fault lay with poor handwriting; either way, communication was impeded!

Poor spelling is an emotive issue but the evidence here is reasonably objective. Table 5 shows the numbers of spelling errors for boys and girls awarded each grade in each year. It distinguishes between straightforward misspellings and wrong meaning errors (where wrong, but accurately spelled homophones of the word required were used) and also give (sic) the proportion of all spelling errors per 100 words [2], to assist fair comparisons between pupils awarded different grades and/or boys and girls, where some groups have tended to write longer or shorter sampled 'sentences'. ...

The distinction between wrong meaning errors and other misspellings does not appear of any great significance but, overall, boys' writing included more spelling mistakes than girls' - who thus compensated for a narrower vocabulary with greater accuracy. ... 1980 GCE candidates made fewest spelling mistakes. Overall, 1993 GCSE candidates awarded grades in the A-C range achieved error rates much like those encountered in the 1980 writing samples but 1993's D and E candidates performed comparatively poorly. The proportion of misspellings in the 1994 writing sample had about two to three times the error rate of their 1980 equivalents. Spelling by 1994 candidates in grades F-G too compared badly, with more than twice the error rate of those obtaining equivalent grades in 1993.

4.3 Seekng causes for th decline.


Th survey natrly finds its results disquietng, and discusses posbl causes for th decline in th foloing terms:
There is no obvious explanation which might excuse such differences between the years. Any suggestion that 1993 and 1994 candidates were falling down as a result of trying to use a more adventurous vocabulary will not wash. The evidence above shows the opposite to be the case. Grade for grade, the 1994 candidates were using a more restricted vocabulary than those of 1980 but were less capable of spelling correctly.

The sudden and substantial turn for the worse in spelling observed in 1994, as compared with 1993, is of considerable interest. What can have brought about such a change between two successive years? ... Schools opting for an external examination in 1993 were by definition untypical, as the vast majority opted for the 100% coursework alternative. They included a relatively high proportion of selective and independent schools who might, arguably, have seen themselves as inheritors of the curricular traditions of 1980. In 1994, the introduction of new national curriculum based syllabuses forced all schools into the examination's net. Might schools which had formerly used the 100% coursework option (who formed the majority of those examined in 1994) have placed less emphasis on the necessity of accurate spelling? No other explanation comes readily to hand.

The shift between 1993 and 1994 is all the more surprising, and disappointing, in the light of the national curriculum's strictures concerning spelling, which had supposedly governed the preparation of 1994's candidates since they entered secondary scools five years beforehand.
Th survey thus givs no explnation for th worsnng spelng of successiv jenrations. Howevr, its acount of th sudn plunj from 1993 to 1994 implys that very likely ther has realy been a sloer but stedy decline from th 1980s onwrds, hos ful efects wer partialy maskd in th 1993 data. Yet th frase "there is no obvious explanation which might excuse such differences between the years" sujests a curius atitude. Wy shud we try to "excuse" a decline in standrds of litracy? Did th survey considr an explnation for th decline wich did not excuse it? Is not th decline a trend demandng to be reversd, rathr than excused? We ar here remindd of th Basic Skills Agencys coment on its own evidnce for a simlr decline, of being "wary about making firm judgements about standards declining", wile admitng to a suspicion that "there was a period in schools when rather less concentration was paid to 'secretarial' skills than was perhaps desirable". Both reports thus giv th impression of not quite liking to look ther findngs squarely in th face.

It is remarkbl that th othrwise capabl and sofisticated anlysts ho hav produced these reports do not even refer to th neglect of fonics in litracy teachng as a prime candidat for th cause, altho that is widely stated (se Turner, 1990; but mor recently Sanchez, 1996, for California, and OFSTED, 1996 for Londn) to undrlie th decline in standrds observebl thruout th english-speakng world. It has been comntd mor than once (eg, Upward, 1995b, p75) that a by-product of th caos of english spelng is confusion not merely about th spelng systm itself, but about th hole sycolojy of alfabetic litracy. It has been th fashn for som 20 years for theorists to considr litracy aquisition in english mor as a natrl, autonmus process of maturation (eg, Frith, 1980; Goodman, 1982; Smith, 1978) than as a specific skil to be mastrd first by being taut its basic elemnts and principls, and then by practisng ther aplication until autmaticity is acheved (Downing, 1987). At th hart of these elemnts and principls is that hy point of human invention, th alfabet, hos use is predicated on fonics.

We wil now anlyz th mispelngs listd in th survey to se if they confirm neglect of fonics as a likely cause of th decline.

4.4 Categris of mispelng.


One of th most valubl featurs of th UCLES survey is that al th mispelngs ar listd (Table 6, pp24-25), ie, 302 mispelt words from Grade A-E performrs across al thre years, plus anothr 431 from th weakst two grades, F and G, wich only arose in 1993 and 1994. Th survey itself scarcely atemts to anlyz th lists (Table 5 countd 'mispelngs' sepratly from 'rong meanngs', but th distinction was then found to be of litl significnce, with only around one sevnth of th total mispelngs constituting 'rong meanngs').

A initial scan of th UCLES mispelngs sujests two jenrl categris. One is wher riters sho a failur to mastr th principls of th riting systm (ie, riting errs); and th othr is wher riters trip over th failur of th riting systm to folo principls that can be redily mastrd (ie, spelng errs). Th reasn somtimes givn for many errs, carelesness, explains nothing, and certnly not wy pupils in 1994 shud be mor prone to err than those in 1980; in any case, close analysis of aparently careless slips sujests that somthing mor than unmotivated carelesness is usuly to blame (se th coments on *brough/*enoght belo).

Categry 1. Riting errs: failur to mastr th principls of th systm.
Riting errs ar those wich do not represent th pronunciation of th intendd word. These may be subdivided into sevrl typs, as folos:

1. Defectiv handriting (eg, *full for fall), wher th riter has inadequat control over letr formation. That this too may result from a failur of teachng methods is implyd in Gorman et al. (1996), wher no distinction is made between practisng letr-shapes and copying words.

2. Failur to identify th consnnt structur of a word and to represent it acordngly. Th errs *brough for brought and *enoght for enough sho that th presnce or absnce of a final T has not been related to its (non-)pronunciation; th riters wer here no dout misled by th plethra of confusing -OUGH(T) strings in english. A variant on this typ of err is seen in *shoudler for shoulder, wher th corect letrs hav been remembrd, but ther sequence has not been related to ther pronunciation. In these cases th riter has been unable to anlyz th fonic structur of th intendd word suficiently to represent it alfabeticly. (On th othr hand, if these words had been mispelt *braut, *sholder, they wud hav been countd in Categry 2, as those forms do represent th pronunciation and no failur of fonic analysis has ocurd.)

3. Vowls, wher english ofrs a bewildrng ranje of alternativ spelngs. But vowl errs in Categry 1 sho that th riter dos not even no wich spelngs ar posbl for a givn vowl sound, let alone wich is corect in a givn word. Thus th survey lists forms like *carfully for carefully, wher undrstandng th systm myt hav at least produced th mor plausbl Categry 2 mispelng *cairfully; and simlrly *fuhw for few, wher undrstandng th systm myt hav at least produced th mor plausbl Categry 2 mispelng *fue.

4. Unidentifyabl words: ther wer 16 spelngs (14 from Grade F-G performrs) wher th asesrs wer unable to identify th intendd word, eithr from th letrs used or from th context (eg, *clats).

Categry 2. Mispelngs: corect sound rongly representd.
Errs in this categry sho that th riter has mastrd th basic riting systm, but not its unpredictbl aplication to individul words. Errs in this categry may be subdivided into two typs, hos significnce is furthr discusd in th conclusion belo (§5).

1 Rong choice from alternativ posbilitis: exampls ar *shear, *bycicles, wich ar in themselvs no less plausbl represntations of th pronunciation than sheer, bicycles (th survey classifyd *shear/sheer as a 'rong word' mistake).

2 Redundncy errs: words containng letrs not required by th pronunciation at al, and wich th riter omitd, or misplaced, or insertd by false analojy. Exampls ar: *coctail (cf concoct), *freind, *possibal (cf Hannibal), *equippment (cf equipped). Th pronunciation of these forms wud again be no difrnt from th corect forms member, friend, possible, equipment, but th insertd, misplaced or omitd letr is redundnt to its represntation.

4.5 Distribution of errs.


A quik count was made of th eroneus forms listd in th survey, classifyng them rufly as belongng to Categry 1 or 2, with results as givn in Table 1. Th totals difr somwat from those publishd in th survey, partly because th survey countd mispelt words rathr than errs as such; thus *chaeous for chaos was countd as a singl mispelt word, but is here countd twice, as it contains two errs. It shud furthr be noted that th boundry between th two categris is somtimes blurd: *occaision, for instnce, myt argubly be classifyd in Categry 1 as a failur of fonic analysis (as explaind in §4.6 belo).

4.6 Interpretng th trends.


As we saw, th UCLES survey felt unable to sujest clear causes for th decline in standrds it identifyd. We hypothesized that chanjes in methods of litracy teachng (th abandnmnt of fonics) in th period concernd myt be an importnt factr, and we decided to anlyz th surveys findngs for evidnce. Without atemtng to determn th statisticl significnce of th figrs (is th sampl larj enuf?), we may now make certn deductions from them.

Table 1: Numbr of errs by year/grade/categry.

YearGradeCategry 1: riting errCategry 2: mispelng
1980A
B
C
D
E
0
2
5
3
7
5
6
6
9
9
1980totals17 35
1993A
B
C
D
E
0
2
0
3
14
3
13
9
15
35
A-Etotals1975
 F
G
27
30
31
28
F-Gtotals5759
1993totals al grades76 134
1994A
B
C
D
E
8
0
11
15
24
8
12
18
24
46
A-Etotals58108
 F
G
86
86
84
78
F-Gtotals172 162
1994totals al grades 230270
GRANDTOTALS323 439

Th survey itself pointd to a posbl jenrl reasn for th dramatic kink in th curv aftr 1993 wen it said that th 1993 candidats came from scools including "a relatively high proportion of selective and independent schools who might, arguably, have seen themselves as inheritors of the curricular traditions of 1980". Now such scools (eg, Montessori scools) hav been less inclined to abandn fonics in ther initial litracy teachng than hav scools in th state sectr, and if pupils ho had receved trainng in fonics ten years erlir ar therfor overepresentd in th 1993 results, we hav a plausbl explnation for th relativly smal increse in ther Categry 1 errs. In 1994, by contrast, we shud be seing th impact of th abandnmnt of fonics across th state primary sectr a decade befor: far mor of th 1994 candidats wil nevr hav had systmatic trainng in how letrs ar aranjed to represent th sound-structur of words. Th figrs suport such an interpretation.

Table 2 belo hylyts th variations in rate of decline by candidats acheving difrnt grades, and permits furthr interpretation of th results.

Table 2: Percentaj increses in numbrs of errs.

Years 1980 1980-1993(1980-94) 1993-94
Grade A-ECat.1: riting err 100%112%(341%) 305%
A-ECat.2: mispelng100% 214%(309%) 144%
F-GCat.1: riting err  1993 100%325%
F-GCat.2: mispelng  1993 100%275%

Th triplng (305%) of errs in th 'riting err' categry by Grades A-E from 1993 to 1994 is replicated in even mor severe terms by th weakr candidats (Grades F-G, 325%). This myt be expectd, since brytr candidats wil hav workd out mor of th principls of fonics for themselvs over th years than weakr candidats cud.

But we may also considr that reduced proficiency in fonics very likely has som efect on th numbr of mispelngs in Categry 2 as wel. Ignrnce of fonics can make spelng patrns jenrly apear mor randm than they realy ar, even in english. Thus a fonicly aware riter myt apreciate that ther is a set of (4) words endng in -ASION, but that ther ar no words endng in -AISION. So altho plaice/place with AI/A ar jenuin alternativ spelngs for a word pronounced /ple:s/, and *planely is therfor a concevebl represntation of plainly, ther is no such empiricl basis for th form *occaision as a represntation of occasion, wich myt almost be ruled out as categoricly as *carfully must be ruled out for carefully (despite th valu of Car in scarce). Therfor, altho we hav positd two categris of mispelngs, one directly related to fonics and th othr not, th distinction between th two is not always clearcut: *occaision myt be countd as a 'hyr levl' Categry 1 err.

Poor fonic awareness may also encuraj th feelng that, since ther is so litl useful systm in th spelng of english anyway, ther is litl advantaj to be gaind from taking care over spelng. This atitude ("spelng dosnt matr") has been widespred among yung peple, including yungr primary scool teachrs, many of hom themselvs aquired litl undrstandng of fonics from ther own education or trainng. Teachr trainrs hav in recent decades encurajd teachrs to encuraj children to 'invent' ther own spelngs in th erly stajes, in th fond belief that corect spelng is a skil that develops later (Gorman et al. 1996, p70). This cavlir aproach to acuracy in ritn english probbly also contributed to th decline demnstrated by th UCLES survey.

Altho ther may be a strong link between poor fonic awareness and 'rong choice' (Categry 2) errs, th latr display a somwat difrnt groth patrn. In evry group but one (F-G performrs, 1994) ther wer mor Categry 2 errs than errs from direct fonic failur as such, tho th predomnnce of pure fonic errs by 1994 F-G performrs implys that, for th weakst candidats, poor fonic awareness may hav disproportionatly damajng consequences. Th A-E candidats wer alredy making a litl over twice as many Categry 2 errs in 1993, but such errs incresed less steeply in 1994 - as tho a jenrl cultur of orthograficl insuciance had alredy infectd 1993 candidats hos fonic skils wer not yet in serius decline. For F-G performrs th 1993-94 slope is much steepr (59 rising to 162). But despite these variations, Table 3 shos a striking consistncy in th overal decline in acuracy from 1980 to 1994: we can say that th 1994 candidats wer wel over thre times mor prone to errs in ritn english than ther predecesrs in 1980.

5. Conclusions for english spelng.

Th UCLES survey presents som alarmng data, but is puzld as to wat they mean. Observrs with a perspectiv of th politics of litracy teachng in recent decades wil imediatly grasp one importnt part of th messaj: th 1994 candidats wer substantialy less proficient in fonic undrstandng (a deficiency that was particulrly markd toward th loer end of th ability ranje), and th decline at least coincided with and was quite posbly causd by reduced atention to fonics in initial litracy teachng. For this part of th diagnosis ther is a relativly simpl (if not necesrly cheap) remedy: teachrs need to undrstand th sycolojy of litracy aquisition in an alfabetic riting systm and to be traind in th principls and practis of fonics (and this is indeed required by th 1997 National Literacy Strategy).

But th rest of th messaj is even mor importnt, since it explains th mor numerus Categry 2 errs. These ar th product of th sheer perversity of english spelng, wich for exampl has th consequence that, wile end, bend, lend pose no dificltis, friend is perenialy mispelt as *freind or *frend. We can now usefuly take our analysis of th surveys mispelng list a staje furthr, by comparing th two err-typs distinguishd in §4.4 abov within Categry 2.

It has been previusly found (Upward, 1987) that around two thirds of mispelngs ar asociated with th widespred fenomnn of redundnt letrs in ritn english. These fal into thre classes: i) letrs irelevnt to pronunciation, ii) letrs representng unstresd centrl vowls (typicly shwa) befor L, M, N, R, and iii) dubld consnnts. Mispelngs involvng redundnt letrs can present as omissions, insertions or substitutions. Exampls of each class ocurng in th UCLES survey ar: from Class i) *were (H omitd from where), *where (H insertd in were) and *freind (omission of I befor E, insertion of I aftr E, that is, reversl of IE); from Class ii) *possibal (A insertd in possible), *membrs (E omitd from members),*vandilism(I substituted for A in vandalism); and from Class iii) *equippment (extra P insertd in equipment, cf equipped) and *setee (TT simplifyd from settee). Th othr, non-redundnt, typ of mispelng may be classifyd as 'rong letrs', such as in th UCLES list A for E in *shear (for sheer) and C for Q in *incuired (for inquired); these rong letrs ar not fonicly redundnt.

Table 3 belo demnstrates how, as in previus studis, almost exactly two thirds of al th mispelngs listd in th survey ar in one way or anothr conectd with th fenomnn of redundncy. Wy shud redundnt letrs prove so trublsm? Th reasn is simpl: with straitforwrd anomlus letrs, we merely hav to remembr wich letr is in fact required (ie, U, not I, in busy); but with redundnt letrs we hav to remembr thre things: i) that a redundnt letr is required, ii) wich letr it is, and iii) wher to place it. This massiv demand on memry ofn defeats riters, and wen one or othr of th thre redundncy conditions is not corectly remembrd, errs result. That is wy a word like business is comnly mispelt as *busness (th need for th redundnt I has been forgotn) or *buisness (th redundnt I has been remembrd, but not its corect position); on th othr hand rong-letr mispelngs such as bisness ar relativly rare, because remembrng th U in business is quite esy.

Table 3: Categry 2 errs by typ.

Year/GradeCategry 2: mispelngs
1980Rong letr Redundncy errTotal
A
B
C
D
E
1
1
1
3
4
4
5
5
6
5
5
6
6
9
9
1980 totals1025 35
1993   
A
B
C
D
E
0
3
2
6
12
3
10
7
9
23
3
13
9
15
35
A-E totals235275
F
G
8
6
23
22
31
28
F-G totals144559
1993 totals3797 134
1994   
A
B
C
D
E
3
5
8
5
17
5
7
10
19
29
8
12
18
24
46
A-E totals3870108
F
G
25
36
59
42
84
78
F-G totals61101 162
1994 totals99171 270
GRAND TOTALS146293 439
% of mispelngs33.3%66.7% 100%

Wen we then com to considr how this major problm of redundncy in english spelng cud be delt with, we find th ansr larjly coincides with th rules of Cut Spelng. In othr words, if english spelng wer releved of its 10% redundnt letrs (perhaps th least disruptiv kind of spelng reform), few of this most comn typ of mispelng wud ocur.

Our analysis of th UCLES survey thus concludes with two remedis for th serius decline in litracy standrds reveald. One is to ensure teachrs ar traind in th theory and practis of fonics, and th othr is to teach a fonicly simplifyd form of ritn english, stripd of its most confusing redundnt letrs, along th lines described in th Cut Spelling Handbook (Upward, 1996a).

Refrnces.

Note: th extensiv bibliografy givn in th UCLES survey (pp53-54) dos not include Turners (1990) pamflet, nor Joyce Morrises (1994) Phonicsphobia, nor th work of th reserchrs (eg, Chall, Flesch, Ehri, Treiman) ho time and again over th past 40 years hav demnstrated that fonics is fundmentl to th achevemnt of optml standrds of litracy.

Barber, Michael (Feb.1997) A Reading Revolution: how we can teach every child to read well (Preliminary Report of the Literacy Task Force), London: Institute of Education.

Basic Skills Agency (1995) Writing Skills: a survey of how well people can spell and punctuate, London.

Brooks, Greg, Gorman, Tom, & Kendall, Lesley (1993) Spelling it out: the spelling abilities of 11- and 15-year-olds, National Foundation for Educational Research.

Chall, Jeanne, Jacobs, Vicki A, Baldwin, Luke E (1990) The Reading Crisis, Harvard University Press.

Charter, David (1996) 'Words fail the GCSE students who can't spell or punctuate' in The Times, 8.4.96.

Downing, J.(1987) 'The Transfer of Skills in Language Functions' in JSSS J5 87/2, pp5-12, reprinted J28 2000/2 pp3-11.

Ehri, Linnea (1995) 'The Emergence of Word Reading in Beginning Reading' in eds. Owen, P. & Pumfrey, P. D Children Learning to Read, London: Falmer Press.

Elkinsmyth, Carol & Bynner, John (1994) The Basic Skills of Young Adults, London: Adult Literacy and Basic Skills Unit.

Flesch, Rudolf (1955, 1985) Why Johnny can't read?, New York: Harper and Row.

Frith, Uta (1980) Cognitive Processes in Spelling, London: Academic Press.

Goodman, Kenneth (1982) Language & Literacy, Boston, Mass.: Routledge and Kegan Paul, Ltd.

Gorman, T.& Brooks, G. (1996) Assessing Young Children's Writing, London: Basic Skills Agency, pp27-33.

Hamilton, Mary (1987) Literacy, Numeracy and Adults: evidence from the national child development study, London: Adult Literacy and Basic Skills Unit.

Massey, A J & Elliott, G L (1996) Aspects of Writing in 16+ English Examinations between 1980 and 1994, Univers. of Cambridge Local Examinations Syndicate.

Morris, Joyce M (1994) 'Phonicsphobia' in JSSS J17 1994/2, pp3-12.

Moseley, David (1989) 'How Lack of Confidence in Spelling Affects Children's Written Expression' in Educational Psychology in Practice, April 1989.

Office for Standards in Education (OFSTED, 1996) The Teaching of Reading in 45 Inner London Schools.

Sanchez, Rene (1996) 'What's Old is New: Phonics Begins to Push Out 'Whole Language'' in The Washington Post, 17 March.

Smith, F.(1978) Reading, Cambridge University Press.

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

Treiman, Rebecca (1993) Beginning to Spell, Oxford University Press.

Turner, Martin (1990) Sponsored Reading Failure, Warlingham: IPSET Education Unit, Warlingham Park School, July 1990, reprinted October 1990.

Upward, Christopher (1987) 'Can Cut Spelng Cut Mispelng?' in JSSS J6 1987/3, pp21-24.

- (1988) 'ALBSU: Literacy, Numeracy and Adults' in JSSS 1988/2, J8 p33.

- (1992) 'Is traditionl english spelng mor dificlt than jermn?' in Journal of Research in Reading, 15(2), pp82-94.

- (1993) '"Quite good" or "totaly unacceptbl"' in JSSS J15 1993/2, pp9-11.

- (1994) 'Err analysis: som reflections on aims, methods, limitations and importnce...', Pt.I, in JSSS J16 1994/1, pp29-33; Pt.II, in JSSS J17 1994/2, pp21-24.

- (1995a) 'No ansrs here yet' in JSSS J18 1995/1, pp37-41.

- (1995b) 'Writing Systems in Different Languages: A Factor Affecting Literacy Standards', Chapter 5 in Vol. II ('Curriculum, and Assessment Issues: Messages for Teachers') of eds. Owen, Pamela, &Pumfrey, Peter Children Learning to Read: International Concerns, London: The Falmer Press, pp70-87.

- (1996a) Cut Spelling: a handbook to the simplification of written English by omission of redundant letters, Birmingham: SSS, secnd (revised and expandd) edition.

- (1996b) 'Wat's th Problm - Spelrs or Spelngs?', revew of Writing Skills: a survey of how well people can spell and punctuate in JSSS, J20 1996/1, pp30-33.

Notes.

[1] From Turner (1990), p33: "The language in children's books ... is becoming impoverished. Often placed at the edge of a large, conceptually challenging double-page spread, the text does not draw the child's attention; indeed in the earlier stages of learning to read illustrations are a great debaucher of infant attention. Picture-cues are poor indicators of what is contained in the text. Some nouns can be shown; verbs to a lesser extent; narrative structures require the wholesale adoption of highly schematic, stylised conventions (as in children's comics) to mimic language; and thoughts cannot be represented at all...

But the great difference between stories and words from an educational point of view is that we come upon written language in the course of some kind of search, actively intended and initiated; but most pictures are encountered incidentally. ... This may explain why our memory for information read is so much better than for scenes watched on television..."

[2] Part of th statisticl uncertnty of th NFER Spelling it out survey (Brooks, et al., 1993) arose from failur to take this mesur.


Back to the top.