[Journal of the Simplified Spelling Society, 1994-1 pp7-16 later designated J16]
[See Journal, Newsletter, Anthology, SPB articles, Personal View 10 and website by Valerie Yule.]

Experiments in public response to surplus-cut spellings in texts.

Valerie Yule, Monash University, Australia.

This paper describes briefly three sets of experiments in which adults read text with letter deletions in words, plus a further experiment rating similar words from a list.

Series 1. Comparisons of spelling for reading and spelling for writing - the same text used for different tasks with different subjects.

Experiment 1.1. Readers' awareness of surplus-cut letter deletions in text
Experiment 1.2. Readers' awareness of non-SC letter deletions in text.
Experiment 1.3. Judgements of superfluous letters in words in text.
Experiment 2. Readers'. objections to spelling changes in text.

Series 3. The effects of practice in reading surplus-cut spelling in text on spelling preferences and perception of 'superfluous' letters in spelling.

Experiment 3.1 Preferences for spelling mode in reading.
Experiment 3.2. Detection of surplus spellings by readers.
Experiment 3.3. Spelling preferences of writers.
Experiment 3.4. Spelling preferences of writers following reading practice in surplus-cut spelling.
Experiment 4. Ratings of acceptability of listed surplus-cut spellings


Aims of the experiments.

The improvement of English spelling may be brought in thru official agency and education in schools, but the route to public acceptance must be thru increasing public familiarity with improved spellings and the principls behind them, in everyday life - the ej of the wej, so to speak.

It seems to me that this should begin in an experimental way before official moves, as part of the current trends of the 'living language'. Once the taste of spelling change becomes palatable the following changes may be easy and indeed welcome.

When the public can join in experimental reserch, this can help to ensure that reforms do indeed meet the needs of all categories of users and lerners, and amendments can be made as necessary, while there is time. And when people are themselvs involvd in experiments this makes a valuabl learning experience about the possibilities and the personal advantage of changes.

The series of experiments presented here investigates adult responses and preferences for various types of surplus-letter deletion in words, as a way to reduce the present clutter in English spelling. The concept of 'surplus' is that many letters in the spelling of English words serv no purpose in representation of meaning or pronunciation. Deletion of such surplus letters may be an essential step, and possibly the first step in reform. (See Yule, 1982, 1986, 1991, Upward & colleagues, 1992.)

These exploratory experiments observe the types of letter deletions that subjects do and do not notice in reading, and their preferences about them. The less that readers notice spelling modifications in text, since they 'fit their system', and the mor that these are changes that subjects themselves would like to make, the mor easily those changes can be implemented, and the mor helpful they may turn out to be.

Changes that are found to be intrusiv and disliked may require some rethinking of the principles that justify them, or judicious delay to avoid spoiling the good public relations produced by changes that are popular. As welcomed improvements are taken up informally as acceptable alternativ spellings and prove their value, appetite for full cleaning up of English spelling will increase rapidly. But first the gate must be opened.

Experiments in readers' and writers' responses to changing spelling can also contribute to basic understanding of the principls of spelling that they may be using, whether fonemic, morfofonemic, grammatical visual and orthografic - or unprincipled rote-memory. The experiments can also have practical consequences by improving participants' understanding of the underlying English spelling system.

These experiments are selected from a range of studies undertaken over some years in both Australia and the U.K as opportunity permitted. This strategy has been preferrd rather than one grand design, as so many factors are involvd in questions of attitudes and preferences in spelling. The volunteer subjects wer all adult citizens plus a few undergraduats and a group of 15-year-olds included in Experimental Series 3. Each brief report is selected to illustrate different aspects. To prevent tedious back-tracking for readers, comments are made during the course of presentation of each experiment, to be summd up and compared in the final discussion.


Series 1. Comparisons of spelling for reading and spelling for writing.

Materials. A seven-paragraf 900-word Sufi story about an intrepid girl was used as the text for three related experiments with different subjects and tasks. The first paragraf of approximatly 150 words provided a model. The last six paragrafs, of similar length, containd 308 word types, of which approximatly a third (104 types - around 11% of the total word-tokens in the text) containd letters that might be deleted on principls of extended 'Surplus-cut' spelling (SC, Yule 1991) then under the designation of Clipd spelling, similar but not identical to 'Cut Spelling' (CS, Upward, 1992), which has also been tested in some other experiments, e.g. Yule & Greentree (1986). The principles of extended SC include deletion of silent letters, doubld letters, and representations of vowels in unstressd syllabls (shwa vowels) that serv no purpose of representation of meaning or pronunciation.

Experiment 1.1. Readers' awareness of Surplus-cut letter deletions in words.

Aim. The aim was to observ the types of surplus-letter deletions that wer noticed or overlookd during reading directed to meaning. The assumptions wer that readers will be mor likely to notice letter deletions if the deleted letters are of use to them in reading, and that they will be mor likely to overlook deletions of letters that they do not need.

Materials. Letters that appeared to be surplus to readers' requirements wer deleted from 79 words in the last six paragrafs of the story. This limited extent of deletions was made on analogy with cloze principls of the proportion of omissions that need not impede reading for meaning. It was thought that reading for meaning might be impeded if the text wer crammd with deletions, and instructions wer to mark all those that wer noticed. Deletions wer also not so many as to change a readers' set from expectation of standard spelling.

Subjects and Procedure. Naive adult subjects (N=33) from a citizens' volunteer subject panel completed the task as an untimed filler in a memory experiment. They wer askd to read the story so that they would be able to answer questions about it afterwards, and while they wer reading thru the text, to slash any spelling mistakes that they noticed.

Results. A short multipl-choice comprehension test showed that subjects did fulfil the requirement of reading for meaning.

Subjects overlookd an average 11.3% of the deleted letters, with a range of 1-50 oversights and median of 8.5. Six re-spellings wer overlookd by 52-77% of subjects, suggesting that these deletions wer not particularly noticeabl: - ASKD, COLLECTD, EXPECTD, SEEMD, WOUD, COUD - that is, silent <e> within participls or between final <dt>, and silent <l> in the WOULD/COULD/SHOULD spelling pattern.

Ten re-spellings wer noticed by all subjects, indicating that these spelling changes stood out: FORIN, TRUBLS, SETL, SHOR, SEA-SHOR, WEL, HUMBL, HEVY, REVELED, RACKD, and a further seven wer noticed by all except one: BEFOR, COLAPSD, HAV, REDY, SEL, THRU, ULTIMAT - that is, words with mor than one letter deleted, novel three-letter consonant strings, final silent letters in mono- and disyllabls, reduction of a vowel digraf, and one infrequent word, ULTIMAT, which can be popularly pronounced with a final long vowel.

Experiment 1.2. Readers' recognition of non-SC letter deletions in words.

A control experiment tested readers' responses to 79 words with letter-deletions selected to make minimum change in their basic visual configuration. It was possibl to make such deletions from 20 words that wer also used for SC deletions in Experiment 1.1. In a similar 'filler' setting, 34 adult volunteer subjects wer given the same instructions as in Experiment 1.1.

Results. Subjects overlookd an average of 13.5% of misspellings, with a range of 0 to 34 oversights.

Five deletions wer overlookd by 52 to 68% of subjects, suggesting that these wer not particularly noticeably in CHILDEN, COST (coast), INTERPETER, PREDICTON, SUCESSIVE.

Fonology could be a factor, since the deletions minimally affect pronunciation in informal speech, except PREDICTON, where the omitted <i> is not visually distinctiv within that letter cluster.

All subjects noticed 13 misspelled words, indicating that the deletions in them stood out - HAPPNESS, BOUGT, ESSENTAL, HAPY, LEST (least), LIF, ROP, SPOK, SYMPATY, UNPLASANT, WAK (walk), ALON, GRATFUL.

17 words wer noticed by all subjects except one: ABL, ADVENTRES, ARRIV, CHOS, CULD, FORIGN, GREF, HAPPNESS, MAK, MISED, SINC, TROUGH, ULTMATE, WOMA, WULD, CARER, SLAV.

The missing letters in these prominent mispellings tended to make words ambiguous in meaning or strongly suggested another pronunciation. Four of the respellings produced unexpected 3-letter consonant strings.

Comments on Experiments 1.1 and 1.2. Instructions alerted subjects to keep some sort of proofreading in mind, but the generally correct answering of multiple-choice questions following the story showed that subjects did indeed read for meaning.

i. There was no significant difference between CS deletions (88.7%) and non-CS deletions (86.5%), in total number of deletions noticed by subjects but mor non-CS words wer noticed by almost all subjects.

ii. Altho reading was silent, 20 words with deletions that markedly affect fonological representation wer noticed as spelling mistakes by mor subjects in Experiment 1.2 than wer the CS deletions in the same words - that did not affect pronunciation - noticed by the subjects in Experiment 1.1


Table 1. Comparison of SC and non-SC letter deletions in words in Experiments 1.1 and 1.2.

a) SC letter deletions overlookd mor often than non SC deletions
  N subjects
overlooking
mispelling
'Visual'
spelling
N subjects
overlooking
mispelling
COUD
WOUD
COLLECTD
SEEMD
UNPLESANT
PEPLE
ARIVE
ESENTIAL
VILAGES
HANDSOM
LOOKD
HAPPEND
MISSD
HAPINESS
26
20
19
18
13
10
9
9
9
9
8
6
3
1
CULD
WULD
COLECTED
SEMED
UNPLASANT
POPLE
ARRIV
ESSENTAL
VILLAGS
HANSOME
LOKED
HAPENED
MISED
HAPPNESS
1
2
3
7
0
3
1
0
8
2
3
5
1
0
No difference
TAUT
ULTIMAT
3
1
TAUGT
ULTMATE
3
1
 
b) SC deletions overlookd less often than 'visual-retention' deletions
SUCCESSIV
ANSER
TRUBLES
FORIN
3
3
0
0
SUCESSIVE
ANWER
TROBLES
FORIGN
17
10
2
5


Experiment 1.3. Subjects' judgments of 'surplus' letters in words.

Method and materials. Thirteen adult subjects in a filler experiment wer askd to slash all letters that they considerd surplus to representation of meaning or pronunciation of words, in the six final paragrafs of the same text presented in standard spelling. The first 150-word paragraf was set out as an exampl in which 'surplus' letters wer already radically slashd. This was presented as a cancellation task, and reading for meaning was not requested. It could thus be seen as bearing a closer relation to the act of spelling than of normal reading.

Subjects' responses wer compared with an A priori classification of 107 words containing letters possibl 'surplus to representation of meaning or pronunciation'.

Results: No subject left standard spelling unaltered. 'Surplus-cut spelling' type deletions wer made in a mean number of 39.3 words, 36.7% of the deletions presumed eligible Only three of the pre-classified 107 words wer not slashd by at least one of the 13 subjects. Final <e> was deleted in the recurring word WER by all subjects.

However, subjects also slashd letters in another 39 words, beyond the A priori classification of feasibl deletions. They made an average of 16 such deletions each - significantly fewer than their CS deletions. They deleted silent <e> and doubld consonants regardless of their function. A significant minority of subjects wer not aware of morfemic principls in spelling and mutilated stem morfemes, and so reducing resemblances for related words. All subjects altered 35 words (11.4% of word-types), but they did not agree on which letters wer surplus within them. This could be because improvements in the spelling of the words might have require change of letters as well as deletions, but could also be due to confusion when mor than one letter might be dispensable or to uncertainty about silent < e > in the participl <ed>. For exampl, MIHT, MIT and MIGT wer ways that subjects tried to cope with the obsolete spelling of MIGHT.

Of the possible types of surplus-cut deletions that could be made, these subjects wer most likely to delete doubld consonants, silent consonants as in COULD, FOREIGN and TAUGHT, and letters in words with more than one surplus letter, such as FOLLOWED and COLLAPSED. They wer least likely to delete silent functionless letters in -ED participles and in -EA, -OU, -TTLE, and -CK formations, as if they took these for granted as visual spelling patterns.

Fonological principls wer paramount, and the most commonly mutilated words remaind pronounceabl - apart from the vowel deletions of three idiosyncratic subjects whose cancellations produced words mor like those of young poor spellers, e.g. BOUGT, WOLD, SPOK suggesting that they had a rote visual memory for spelling that was not related to any underlying spelling system. Majority preferences wer to delete< gh > spellings and doubl consonants, but there wer differences and inconsistencies over visual or fonological solutions (e.g. ANSWR PEOPL), omitting the unstressd schwa vowel, deleting silent <e> and silent consonants, retaining stem morfemes, and using a singl letter to replace a digraf. Deletions made by subjects wer not related to the position of letters in a word.

Relation of deletions made in Experiment 1.3 and spelling errors detected in Experiments 1.1 and 1.2 ie. the relation of what readers notice in reading, and how they may understand spelling for writing. Responses of subjects (tho not all subjects) in the three experimental groups followed fonological, morphemic and orthografic principls, and while important, visual configuration was not primary even for the two groups reading for meaning.

The insignificant correlation of .09 between subjects' cancellations in Experiment 1.1 and subjects' respellings of the same words in Experiment 3.1 was in part a matter of individual differences, accentuated by the small N of 13 subjects for Experiment 3.1. However, a four-way relationship with the awareness of deletions shown by subjects in Experiment 1.1 resulted from the principles the subjects in 1.3 followed for surplus-letter deletion, plus their lack of awareness of another spelling feature - that letters which have a function in some spelling patterns in words may have no function in the same spelling pattern in other words, as with a silent < E > in the participle -ED and in digrafs such as -OU-.

Thus, some deleted letters in words that wer overlookd by most subjects in Experiment 1.1 wer also overlookd as candidats for deletion in Experiment 1.3, as in words such as ASKD, SEEMD, WORKD, UNPLESANT. Some deleted letters in words that most subjects in Experiment 1.1 noticed, wer also deleted by subjects in Experiment 1.3, such as TAUT, FORIN, ANSER, ALTHO, THRU, FOLOWED, COLECTED, COLAPSD, WEL, SEL. And on the other hand, deleted letters from COUD, WOUD and WER wer most likely to be overlookd in Experiment 1.1, and removed in Experiment 1.2, while the deleted letters in HEVY, REDY, SETL, TRUBL and CORT wer most likely to be noticed in Experiment 1.1 and not taken out in Experiment 1.2.

The data overall indicate that there is in fact a relationship of spelling for reading and for writing, tho the two are not identical.

All groups generaly used fonological and morfemic principls and seemd to have some idea of a 'form of the word'. They did not show real understanding of the use of double letters and silent <e> - possibly because their principls are not reliabl in standard spelling. Visual features wer shown to play a part in the identity of some visually-distinctiv irregular words, e.g. ANSWER, PEOPLE, but wer not the prime factor even for readers.

Readers particulate tended to miss deletions in words that improved fonological relationships or shortend words while still retaining the visual appearance of stem morfemes, e.g. INTEPRETER, PREDICTON, REMEMBERD, and they tended to notice re-spellings if the pronunciation was ambiguous, e.g. HEVY, HAPY.

The orthografic legality of letter sequences in conventionl spelling did not seem important for either subjects' own responses or their awareness of spelling errors in reading, altho this has been considerd significant by some theorists (see Adams, 1981, and c.f. Baker, 1980.) For exampl, subjects tended to overlook some rarely found final letter-sequences, such as -OUD, -SKD, -EMD, -BERD, -SUR, -RKD, -AIND, -OKD, -REK but did notice other words with changes that still produced legal sequences e.g. -ORIN, -EVY, -ATY, -LON.


Experiment 2. Readers' objections to spelling changes.

Aim A replication study to observe adult readers' reactions to SC spellings in text, with a different subject group, different text and different task.

Materials and Method. A story about an adventure in a cave, which has also been used for reading time and comprehension experiments, both in print and on screen, consists of seven paragrafs of approximatly 100-150 words each. In this experiment, all paragrafs except paragraf 2 containd words in radical SC spellings, making a total of 105 word-types that wer modified by letter deletion. Some of these words recurrd up to four times (discounting TH for THE which recurrd 81 times), so that a total of 131 word-tokens could be rejectd if the spellings wer disliked. 33 naive adult subjects wer askd to mark all the changed spellings that they disliked.

Results. No subject disliked all the spelling changes, and all subjects objected to at least 14 changes. Disapproval ranged from 14 to 73 of the 105 re- spellings, i.e. all subjects accepted at least 30% of the respellings, and some as many as 86%.

Disapproval ranged from 97% for CUD and 94% for AFTR, WITE, BELO, BETR, BOLDER (for boulder), DISLODG, FASND, LOOS, ROK and WAT to 54% for FRENDS, 51% for MAMOTH, 30% for FASINATN, 27% for EVRYWER, 6% for BOYS replacing BOY'S, and 9% for later occurrences of TH. One subject marked IRON as a disliked 'spelling change'.

Respelld words that wer not cancelld by subjects may have been left unchallenged for either of two reasons - because the subject was aware of the change but did not object to it, or because the subject did not notice the anomaly. For the purpose of the hypotheses behind this experiment, either interpretation is satisfactory. Further experiments could sort out the difference, or whether tedium became a factor (see below).

The passiv acceptance of TH in reading in this experiment contrasts with the lack of enthusiasm for the use of this shortening in the writing experiments reported here. Possibly its frequent recurrence relegated it to the status of background noise in this particular experiment.

Data from this experiment suggest some factors that may contribute to readers' perceptions of disliking or not disliking spelling changes, altho replication is required for verification.

1. Place in sentence. Readers of text have less difficulty in reading words that come later in a sentence, and there is some evidence that they pay less attention to their detail.

2. Place in text. Altho subjects wer askd to cancel every word they disliked when they saw it, the efect of repetition of a modified spelling during the reading session was to reduce objections to it, whether from modified attitude resulting from familiarity, or from acquired immunity to the novel appearance - or, of course, ignoring instructions thru tedium. This effect is significant on inspection. A changed spelling that was repeated, such as CASM WER, AFTR, and WITE, tended to be less noticed as an anomaly. Exceptions to this increased tolerance wer CUD, ENTRNCE and SEEMD.

3. Frequency of word.
i. The most common irregularly spelled words may appear mor objectionabl in an unfamiliar spelling.
ii. Subjects may be less certain of the correct spelling of a rare word.

4. Length of word. Subjects object less to or notice less a change in a long word.

5. Position of deletion. Subjects may object less to or notice less a change in medial position, or penultimate.

6. Silent letters are objected to less than omission of an unstressed syllabl.

7. Deletion from a doubld consonant arouses littl objection.

8. There may be mor objections to multipl deletions in a word.

Ratings of difficulty Following the task, subjects wer asked to rate the difficulty of the spelling on a 9 point scale. The second paragraf in TO provides a base-line.

These ratings can be regarded as low on a 9-point scale, altho subjects rated the radical version of 'surplus-cut' spelling as up to three times mor difficult than standard spelling. Individual subjects' mean ratings for the SC paragrafs ranged from 1.0 (the lowest possibl rating) to 7.8. Three of the 33 subjects rated standard spelling as mor difficult than the lowest possibl rating of 1.0, and their ratings for SC spelling wer not significantly different - possibly they are among the poor spellers who hope for improvements in English spelling.

However, in another experiment, not described here, that also used this text, the subjects' task was to read the story for meaning and answer comprehension questions. Ratings of spelling difficulty wer not significantly different from standard spelling after the initiating paragraf. It would appear that when the task has focussd on whether subjects object to the spelling, as in this experiment, ratings of spelling difficulty may be higher than when attention is directed to reading the content of the text.

Subjects wer invited to comment on the 'objections' experiment, and the four who responded show the extent and importance of individual differences - averages can obscure significant aspects of reading and spelling.

'Confused.' 'I did not like ordinary spelling either.' 'I have read a story in fonetic spelling before, so I was less irritated than I might have been'. 'It was easy to read and make sense of. I kept expecting the long and difficult words to be misspelt but they wern't always.' 'The abbreviations became easier but they annoy me. I'd rather see the whole word written out.'

TABLE 2. Mean ratings of difficulty of spelling within a text

Paragraphs
N words with letter
Ratings of difficulty
1
33
2.76
2(TO)
-
1.12
3
28
3.33
4
27
2.96
5
32
2.94


Series 3. The effects of practice in reading surplus-cut spelling on spelling preferences and perception of 'superfluous letters in spelling.

In an unpublishd experiment (Yule & MacKay, 1986) 92 readers aged 15-50, classified as poor readers and average readers, undertook extended daily practice over three weeks in reading a series of 46 texts that had been transliterated into SC spelling, a total of up to 50,000 words for the faster readers, or read the same texts in standard spelling (traditional orthografy, TO) as controls. No training or explanation of Surplus-Cut spelling was given. Subjects wer only told to ignor any spelling changes. A series of paper and pencil and oral tests was included, and several of these tests, undertaken by some but not all subjects, showed how experience in reading SC tests significantly affected attitudes, understanding and application of SC principls.

Since this was an early exploratory study, texts varied in types of SC spellings - for exampl, EXPLAIND or EXPLANED; some texts retaind THE while others used TH; and some texts containd mor radical letter deletions than others.

Experiment 3.1. Preferences for spelling mode in reading

At the end of the first week of reading practice, 15 SC and 15 TO readers read a short passage about information tecnology which containd a choice of several spellings including TO for 40 words. Both groups of readers selected an average of eight words (20-30% of the options) that they would prefer to read in surplus-cut spelling rather than in TO. This indicates at least some degree of popular dissatisfaction with present English spelling.

Experiment 3.2. Detection of surplus spellings by readers

Early in the second week of the reading-practice experiment, thirty SC readers and thirty TO controls wer given the task of canceling letters that they thought wer not needed in words, in a story of 104 words about a magician's daughter in which 71 words had irregular TO spellings.

Both the average and poor readers in the SC groups showed effects of their experience in reading in SC by making significantly mor SC-type deletions (mean letter deletions of 24.8 and 13.8 - that is, 35% and 19.4%) than did the normal and poor reader TO groups (mean deletions of 21.2 and 11.7, that is, 30% and 16.5%) F (3,52) = 2.6 p<.05.

Average readers made mor SC-type deletions compared with the poor readers, F (3,52) = 9.9, p<.Ol, suggesting that reading ability is significantly related to understanding English spelling structure, as shown in greater competence in the task of judging what letters in words may be surplus to representation of meaning or pronunciation.

Experiment 3.3. Spelling preferences of writers.

Late in the second week of the experiment, the same 'Princess' story was presented to 43 subjects with instructions to write it out in the spelling that they personally would like to see establishd if they wer the masters to decide it.

The 22 average and poor readers who had been reading texts in SC tended to change the spellings of mor words (31.4 and 28.0 mean changes, i.e. 44.2% and 39.4% of the words open to changing) than did 21 TO readers (19.8 and 20.8, i.e. 27.9% and 29.3%). Tlc only subjects who made no changes at all to standard spelling wer four female average readers aged over 42 reading in SC and two in TO.

The SC group tended to make mor SC-type changes than the TO group (mean SC-type changes 27.7 and 17.4 words, i.e. 39.0% and 24.5%). Average readers tended to make SC changes that wer mor appropriat to the representation of pronunciation than did the poor readers, whose alterations wer mor likely to show a visual rather than fonological apprehension of word structure, e.g. ONC APON A TIM and IMAGATON.

Experiment 3.4. Spelling preferences of writers following reading practice in SC

Among the post-tests following the reading practice, 32 subjects who had been reading in SC and 23 subjects who had been reading in TO wrote out a paragraf of text in the spelling that they would like to hav if they wer the person who decided how the English language should be written down. The 80-word passage, about catalogs for tourists at a castl included 50 words in SC spelling, which subjects could change as they liked. (The TO-readers had experienced SC spelling in oral reading of two paragrafs two weeks previously.)

All target words wer transcribed in non-standard spelling by at least one subject. The average target word was written with shortend spelling by 66% of SC subjects and 25.9% of TO subjects, and with changed but not shortend spelling, by 0.6% of SC subjects and 0.3% of TO subjects. The shortend spellings wer not necessarily written according to the SC model provided, and often included letter changes. Longer respellings wer rare.

The importance of the experience of three weeks' daily reading in SC was shown in the high proportion of spelling shortenings that wer made by the SC subjects. This experience had given them a model of spelling change that influenced their own changes, and also influenced their thinking about spelling, as shown in the fact that they wer also mor radical than the TO subjects in devising their own spelling deletions and letter changes - individual subjects even shortend AND, WHO, FOR and BE.

The significance of an availabl model for spelling change is shown in the greater proportion of spellings that wer shortend by the TO subjects, compared to the shortenings made by TO subjects who wer given no model in Experiment 1.1.

For SC subjects, the most important factor influencing letter deletions in their transcriptions was the length of the word. On average, one syllabl words wer shortend by 57% of SC subjects, words of two syllabls wer shortend by 62%, three syllabl words by 70.9% and if words wer four or mor syllabls, such as RECOMMENDATION, INTERROGATE or EXCEPTIONALLY, 90.6% of SC subjects shortend them. 75% of SC subjects changed the spellings of words which wer open to two or mor changes, e.g. LITL, MARVELUS.

For subjects who had read thruout in TO, the trends for spelling changes of most words tended to follow the same direction as SC subjects but at a modest level with only 4 to 35% of TO subjects changing these words from standard spelling. However, one set of words was dramatically different, and closer to SC proportions of change. 40-60% of TO readers shortend the spellings of FASHIONED, EXCITING, RECEIVE, RECOMMENDATION, DISCOURSE, TRAVELLING, CHARACTER, MARVELLOUS, PROGRAMMES, CATALOGUES, EXCEPTIONALLY and INTERROGATE. The likely explanation is that these are all words that writers find difficult to spell and write out correctly in any case. Long words that are difficult to spell therefor seem obvious candidats for immediat public adoption of SC changes as alternativ spellings.

All types of SC deletions wer made by subjects - deletions of doubld letters, superfluous letters in vowel diagrafs, and silent vowels and consonants, including final silent <e>, altho deletion of schwa vowels, as in VISITR, was less popular.

It may be noteworthy for public relations in introducing surplus-cut spellings that no TO subjects and only 18% of SC subjects copied the model of TH for THE, altho this would be a major time-saver.

In subjects' post-test comments on the total reading-practice project, reactions to surplus-cut spelling varied from hearty welcome to increased irritation. Some subjects gave instances of difficult TO spellings as the spellings they had disliked the most, while others spontaneously instanced irritated responses to recurring TH. One SC subject claimd to have noticed no difference from normal spelling. Others reported rapid adjustment. Overall, the responses wer positive


Experiment 4. Ratings of acceptability of listed surplus-cut spellings.

Twenty adult subjects rated their approval of 60 sixty SC-type respellings of words on a five-point scale ranging upward from strongly disliked to greatly liked. (1) Tables 3 and 4 show the findings.

Deletions in less familiar words listed also appeared to increase disapproval, perhaps thru lack of context to reinforce semantic access. Length of word or position of deletion in the word was irrelevant.

TABLE 3. Mean ratings of acceptability of SC-type respellings of words.

WordsRatings
MAGNAT LETRHED1 .7
 
MAMAL CATRPILR KNOLEGE SUDNLY NIBL
AQITL SOVREN SIV CARACTR FORIN PEPLE
INOCUUS LITL WER PERLY ZELUS ENDEVR
2.0-2.5
 
MISCHIVUS UNPARALELD THERFOR ILITERAT
ALTHO PSYCOLOGY IMEDIATLY PEKS WHER
EXESSIV EFECTIV PROFESSR WOUD POSIBL
2.6-3.0
 
GARDIAN ASASSINS BALERINA DIFERENCE
DISIPLIN HORD MOLD MILIONAIR OBSTINAT
ACOMODATE MOSQITO FESANT THEMSELVS
OCASION FRENDS
3.1-3.5
 
COCO ESPECIALY RECOMEND DONKY
DISAPOINT MATRESSES
3.6-4.0
 
COLOR MINITUR.4.0-4.3


TABLE 4. Mean ratings for categories of letter deletions.

Generally approved:
Double letters
Unstressed vowels
Final silent vowel
Neutral:
Consonant
Disapproval:
3 letters deleted
2 letters deleted
Words
10
11
9

2

6
24
Mean Rating
3.4
3.4
3.0

2.9

2.2
1.6

Discussion

This series of experiments has focussed on public response to the deletion of surplus letters in English spelling for both reading and writing on the grounds that -

1. Reforms that have the support of the literat population have a better chance of early adoption.

2. Informal explorations indicate that spelling changes by deletion rather than by substitution are more acceptabl to readers because they make minimal disturbance to visual configurations of words, and are more acceptabl to writers because they are more economical and make minimal disturbance to grafo-motor habits. Yule & Greentree (1986) found that readers adjusted rapidly to 'surplus cut' spellings in text, but spelling reforms involving letter changes wer more disruptive.

Findings in the present experiments show that there would be considerabl public support for moves in the direction of omitting surplus letters. Among the hundreds of subjects participating in these and similar experiments and so gaining hands-on experience, there have been hardly more than a dozen who rigidly opposed any change, and 'if you wer the master of spelling' experiments are always popular. (Opinions in any surveys about changes are mor likely to be conservative of course.)

Findings are of interest for theories of reading and writing, which in turn must provide the foundations for any reform that will really be of benefit to users and learners. For example, the data support claims that spelling for reading and for writing are not handled in the same way, (see e.g. Frith 1979, 1982) but the similarities between them do not support sweeping generalisations of 'reading by eye, spelling by ear' or claims that what benefits writers will handicap readers. A full discussion is postponed for reasons of space, but some start-off references are included in the bibliography.

Comments here focus on practical points, including features of surplus-cut spellings that appear of primary acceptability and comprehension, and features which may initially elicit rejection of the whole enterprise and might advisedly be postponed pending public education. However, replication and extension of these experiments is required to establish these findings, particularly to explicate systematic relations between the spelling deletions accepted by readers and those applied by spellers.

i. Readers of text tended to overlook (accept) letter deletions, whether SC, CS or not, that improve fonological relationships, or occur in long words that retaind the visual appearance of stem morfemes, e.g INTEPRETER, PREDICTON, REMEMBERD.

ii. Readers wer more likely to notice deletions when pronunciation became ambiguous or distorted (CUD and WUD wer unpopular, probably for this reason) two or more letters wer deleted, or novel three-letter consonant strings resulted. Deletion of final silent letters in monosyllabls (apart from WER) and reduction of vowel digrafs also tended to be intrusive, even altho improving fonology.

iii. Some of the most popular spelling deletions made by writers in Experiment 3 (WOUD COUD WER ASKD) wer among those most likely to be overlookd in reading too, a double support for the proposition that the letters deleted in these words wer in fact surplus. However, the more often a misspelld common word appeared in the same text, the more likely it became to be passd over, despite instructions to miss no examples, suggesting that either familiarity or tedium reduces continued canceldation.

iv. The SC principles that writers picked up and applied immediately wer deletions of silent <e>, silent consonants, doubld letters and in vowel digrafs. Vowels in unstressd syllabls wer less commonly deleted.

v. Legal and illegal letter sequences in TO do not seem relevant for either subjects' own spellings nor in noting changes in reading modified spellings, contrary to the literature, often based on single-word studies, that has assumed that orthografic legality could be more important to users than spelling structure. (See e.g. Adams, 1981, Baker, 1980)

vi. Morfeme representation - the -ED participle. Results confirm the findings of Smith & Pattisonn (1982) who found that omission of errors in letter cancellation are more common in affixes or pseudo-affixes than in the same position in a word without affixes. (See Henderson 36, 1985.) Henderson (63) also cites studies showing that subjects asked to cancel out the letter <e> are more likely to miss targets that are part of an inflectional suffix. Smith & Sterling (1982) found that the letter <e> was overlookd most in the -ED inflection, but not so missd when nonaffixd as in hundred, with a roughly equivalent rate for <e> in comparativs such as CLOSER, agentivs such as DRIVER and pseudo-affixd simple words such as RIVER.

It seems to me that both fonology and unstressd schwa are involvd here. Such findings indicate that grammatical features shown in orthografy may be relevant in reading, and that affix-like orthografic patterns have a distinct perceptual status. This is supported by the general failure of naive re-spellers (Experiment 1.3) to recognise when <c> in such affixes has no fonemic function, and of readers to notice their omission less in monosyllabls than in polysyllabls (e.g. overlooking ASKD and SEEMD but not REVEALD or COLLAPSD).

It could be possibl that the spelling WER is generally accepted by both readers and writers because of its relationship with ARE (which is not accepted easily as AR).

vii. The shortend spelling TH was not a deletion made by writers, despite its obvious value for economy of paper as well as effort, nor was it popular with most readers, tho repetition within the same text appeared to increase tolerance. Since a significant number of respondents reported actual irritation at this ubiquitous change, which significantly affects the visual appearance of text, TH may not yet be advisable in texts written to acclimatise the general public.

Education.

It is clearly important that all children be given an understanding of fonics and TO's unknown under-lying spelling system, and that some way be found to enlighten literate adults as well, so that spelling improvement can be both understood and welcomed. (This understanding is one aim of the half-hour computer animated cartoon video Teach Yourself to Read, or find out where you got stuck, Yule, 1993.)

Subjects often did not seem aware of the functions of silent < e > as a modifier of a preceding long vowel, nor of double consonants as modifying short vowels, nor of CK as a special case of a doubld consonant, suggesting that the early learning of spelling may have been mere rote without understanding - or forgotten long since.

Related experiments that are needed:

i. Studies of writers' spelling mistakes compared with their own spelling preferences and tolerances in writing or reading as an exploration of useful orthografy for reading and writing.

ii. Similar experiments on reading in other modified spelling modes of English spelling reform must involve letter changes to repair unsatisfactory sound-symbol relationships not solvd by surplus-cut principls. Ives (1992) has suggested steps for this, including the encouragement of public use of the more fonemic spellings when dictionaries accept two or more alternativ spellings for words - as they do already for over 3,000 words (Emery 1973), and 2,000 of which are to be found in American college dictionaries (Deighton 1979). A further step in this direction can be testing and so encouraging public responses to fonemic letter changes such as Ives has suggested commencing with the consonants, as the most simpl and obvious to rationalise.

Footnote.

(1) Eight non-native speakers of English also participated, since the attitudes and opinions of international English-users are important, but altho they were all post-graduate University students, it was found that too much of the vocabulary was not known to many of them even in TO, and even by some who had been resident in Australia for ten to twenty years. It may help future researchers to exclude words likely to be unfamiliar to ESL subjects, which include MAGNATE, LETTERHEAD, MAMMAL, ACQUITTAL, SIEVE, INNOCUOUS, ZEALOUS, HORDE, UNPARALLELED and even COCOA.

References.

Baker, Robert. 1980. Orthographic awareness. In Uta Frith, (Ed.) Cognitive processes in spelling. London: Academic Press

Deighton, L.C. 1979. A comparative study of spelling. Pleasant Meltler, NY: Hardscrabble Press.

Emery, Donald. 1973. Variant spellings in modern American dictionaries. Revised edition. Urbana IL: National Council of Teachers of English

Frith Uta. 1979. Reading by eye and writing by ear. In P. Kolers, M. Wrolstad & H. Bouma (Eds). The processing of visible language. Vol 1. NY: Plenum Press

Frith, Uta. 1982. Cognitive processes in spelling and their relevance to spelling reform. Spelling Progress Bulletin, 22.1.6-9.

Henderson, Leslie, & Chard, Jackie. 1980. The reader's implicit knowlege of orthographic structure. In Uta Frith (Ed.) op cit.

Henderson, Leslie. 1985. Toward a psychology of morphemes. in Andrew Ellis (Ed.) Progress in the psychology of language. London: Erlbaum.

Ives, Kenneth. 1992. A Spelling Reform program for the 1990s for English Speaking adults. Journal of the Simplified Spelling Society, 4.1.11-13.

Smith, P.T. & Sterling, C.M. 1982. Factors affecting the perceived morphemic structure of written words. Journal of Verbal Learning and Verbal Behavior, 21: 704-721.

Smith, P.T. & Pattison, H.M. 1982. Models for letter cancellation performance and their implications for models of reading. Quarterly Journal of Experimental Psychology 34A 95-116.

Upward, Christopher. 1987b. Cut Speling - a linguistic universl? Journal of the Simplified Spelling Society.J5 17-24.

Upward, Christopher, Fletcher, Paul, Hutchins, Jean & Jolly, Christopher. 1992. Cut Spelling. A handbook for the simplification of written English by omision of redundant letters. Aston UK: Simplified Spelling Society.

Yule, V. 1982. Spelling as technology. Rewritten and retitled by John Bell as 'Shorter words mean faster reading' in New Scientist 96 1335.356-7. (That title is a mistaken and sweeping generalisation.)

Yule, V. 1986a. The design of spelling to match needs and abilities. Harvard Educational Review, 56: 278-297.

Yule, V. & Greentree, S. 1986b. Readers' adaptation to spelling change. Human Learning 5: 229-241.

Yule V. & MacKay, C.K. 1986. Practice effects for adults and poor readers in reading text in a modified spelling. (Unpublished manuscript).

Yule, V. 1991. Orthography and reading.- Spelling and Society. Doctoral dissertation. Monash University. See Dissertations Abstracts.

Yule, V. 1993. Teach yourself to read or find out where you got stuck Video. Melbourne: Literacy Innovations, Kew, Vic. Australia. This half-hour video is experimental and needs financial sponsorship for upgrading, but is still useful and entertaining. (Video with 2 manuals A$35 not including p&p.)


Back to the top.