[Spelling Progress Bulletin, Spring 1982 pp13-18]
[See dyslexia reports and extracts from books about dyslexia.]

SSS Conference 3: Teaching and Learning Spelling continued.

"The Spelling of Nasal Clusters by Dyslexic and Normal Children,"

by Dr. Margaret J. Snowling.*

*The National Hospitals College of Speech Sciences, London, England.

Abstract.

The spelling of nonwords containing nasal clusters either in final position (hent) or preceding syllabic 'l' (wemble) was explored. Subjects from 6 to 14 years old heard the word, repeated it, and then spelled it in a word completion task. A tendency to reduce clusters when writing but not when speaking was evident in less skilled spellers. The nasal consonants 'n' and 'm' were often omitted when occuring in clusters with other consonants.

Acknowledgements.

Thanks are due to Dr. Uta Frith and Dr. Dolores Perin for assistance and helpful discussions. Also to the dyslexic children who participated in the experiment and for the co-operation of the staff and pupils of Bromley High School, Kent, and Surrey Square School, Forest Hill, London.

Corpus.

In recent years it has become popular to carry out qualitative analysis of spelling errors in order to throw some light upon spelling difficulties (Peters 1975, Cromer 1980). It is customary to distinguish between a phonetic error which correctly preserves the sound sequence of a word (e.g. speshull for special, trafick for traffic) and a nonphonetic error in which the sound sequence is not preserved (e.g. deter for doubt, heyou for hay). This distinction is an important one which has been shown by many investigators to be of diagnostic significance. (Boder, 1971, Nelson & Warrington, 1974, Frith, 1979).

Phonetic errors are usually assumed to be less serious than nonphonetic errors because they are easily deciphered. A more liberal approach to spelling would regard these versions as acceptable. Furthermore, an individual who makes primarily phonetic errors shows evidence off the ability to segment the target words into appropriate speech units (phonemes) and of being able to translate these units into letters using phoneme-grapheme rules (Frith, 1980). In contrast, an individual whose errors are primarily nonphonetic may have difficulties at either or both of these initial stages. However, the available evidence suggests that these children have difficulty in organizing the speech sounds comprising words even before they begin to apply phoneme-grapheme translation rules to them. (Bradley & Bryant, 1978, Stringer & McKenzie, 1981). Thus, it is important to distinguish these two basic error types because they might point to the need for different sorts of remedial intervention. Whereas individuals who make primarily phonetic errors may require only a systematic introduction to conventional spelling patterns and spelling rules, individuals whose errors are primarily nonphonetic may require more specialized auditory skills training.

However, the ability to spell by ear is still a little understood process and before any definite remedial recommendations can be made upon the basis of spelling errors, it is necessary to know how well any individual of a given age and intelligence could be expected to spell 'by ear.' Even the first step in phonetic spelling - the ability to analyse the speech sounds in the target word, depends upon a variety of perceptual, linguistic and cognitive factors (Golinkoff, 1978). It would be unrealistic to expect these to be available in equal measure to young children just starting school and to school leavers. Moreover there is growing evidence that the individual's ability to segment words into phonemes is partly dependent upon their orthographic knowledge (Read, 1975, Marcel, 1980, Perin, 1982).

The importance of orthographic knowledge was first reported by Read (1971) following a study of the writings of pre-school children who had received no conventional instruction. Their spellings could be regarded as phonetically acceptable although they represented speech sounds in a far less conventional manner than that used by older children who were familiar with the orthography. For instance, they wrote 'trouble' as chrubl, 'train' as chran. Evidently, in analysing these words for spelling, they regarded the consonant cluster [tr] as similar to the acoustically-similar [tʃ]. As children learn to read, they presumably realize that words beginning with [tr] and [tʃ] are marked differently in the orthography and this influences their subsequent perceptual analysis of words containing these sounds. Indeed Read (1973) showed that whereas five and six year olds said that words beginning with [tr] sounded more like words beginning with ['t] than others beginning with [tʃ], seven and eight year olds: who were better readers, classified the words beginning with ['tr]' with those beginning with ['t]. Thus, it is undoubtedly important to take account of an individual's spelling knowledge when making a qualitative assessment of errors. Errors which at first glance may appear to be due to perceptual difficulties may in truth only arise because the individual is unfamiliar with the spelling convention being tested.

Read (1975) also reported the tendency for pre-school children to omit nasals from their spellings of nasal clusters, e.g. they spelled bent as bet, camping as capin. While this type of error could well be classified as 'nonphonetic', its occurence in the early stages of spelling development may well be acceptable. Read argued that, although the young children did not mark the nasals in their spellings, they were nonetheless aware of their presence. They could distinguish between minimal pairs such as camp-cap, bent-bet and could provide rhymes for words like 'trunk' without difficulty. Therefore it was suggested that while the children were aware of the nasal, they regarded it as a feature of the vowel. Thus, in spelling, as long as they had printed the vowel, they believed that they had represented the nasal characteristic. It would be only after sufficient exposure to the written word that children would realize that the nasal sounds [m] and [n] have always to be represented by graphemes in the standard Orthography.

Marcel (1980) observed the same tendency to reduce nasal clusters in the spelling of children who were poor spellers, in adult literacy students and in certain neurological patients. He argued from the nature of their errors that they were analysing speech in a manner similar to that used by young children. As they had not yet acquired spelling knowledge or, in the case of neurological patients, as they had lost this knowledge, they were making use of intuitions about the phonetic characteristics of words, uninfluenced by spelling conventions. The aim of the present study was to extend the findings of Marcel (1980) by investigating the ability of normal and dyslexic children of different ages to spell nasal clusters and to examine the contribution of spelling knowledge to this process. Furthermore the importance of perceptual factors was to be investigated. Both Read (1975) and Marcel (1980) reported that nasals were omitted more frequently from clusters in which they were paired with unvoiced consonants (e.g. tent, bank, bump) than from clusters in which they were paired with voiced consonants (e.g. tend, fence). They argued that this fits with the phonetic facts. In American English, nasals do not constitute true phonetic segments prior to unvoiced stops (Malécote, 1961). In standard English, a final unvoiced obstruent has the effect of shortening all preceding continuants. Thus, in most English accents, nasals are experienced as less perceptible in clusters where they preceded unvoiced stops and this may well be why they are omitted more frequently in these contexts. In the present study, by examining subjects' ability to spell all possible nasal clusters (nasals paired with voiced and unvoiced segments) the aim was to examine the role of perceptual factors in more detail.

Method.


The method chosen was a completion-spelling task. Following auditory presentation of a target word containing a nasal ending, the subject was required to repeat the word and then to add the appropriate ending to an incomplete version of the target word. All possible nasal endings were tested. In addition, stimuli endings in the sounds [m], [n] and [ŋ] were included to ensure that the graphemic responses for nasal sounds were available. Each of the nasal endings was tested in the context of a nonsense word to minimize the effect of knowledge of specific word spellings.

Two groups of subjects were tested and within each group, subjects of lower and higher spelling ability were included. There were three levels of task difficulty, the spelling of nasals (e.g. blem), final nasal clusters (e.g. clest), and medial nasal clusters followed by syllabic [l] (e.g. stemple). The number of spelling errors made by each subject under each level of difficulty was calculated.

Thus, there were two between subjects variables, Group (dyslexic and normal), Spelling Ability (High versus low), and one within subjects variable, Task Difficulty (nasal alone, nasal cluster, nasal cluster preceding [l]). The dependent variable was percentage spelling errors.

Subjects.


For the purpose of this study, an objective definition of dyslexia was adopted. Dyslexia is an impairment of the ability to read and to spell.

The children were all of at least average intelligence with reading and spelling ages which were significantly below the level to be expected given their age and intellectual ability. All had been referred for psychological assessment because of significant under-achievement at school. 23 dyslexic children were tested. They ranged in age from 7 years 8 months to 15 years. Reading ages (as measured by the Schonell Graded Word Reading Test) ranged between 7 years 2 months and 10 years 7 months and Schonell Spelling Ages ranged between 6 years 8 months and 9 years 7 months.

The normal subjects were selected from two schools in the London area to match the dyslexic children as closely as possible for reading and spelling achievement. They were selected by their teachers as being average for their age in reading and spelling. 19 children aged between 6 years 5 months and 9 years 6 months were tested. Schonell Reading Ages ranged between 6 years 3 months and 10 years 2 months, with Spelling Ages ranging between 6 years 8 months and 9 years 10 months.

Within each group of subjects (dyslexic and normal), subjects of high and low spelling ability were separated. Amongst dyslexics there were 15 children of lower and 8 children of higher spelling ability. Amongst the controls there were 12 children of lower and 7 children of higher spelling ability. The lower ability groups had Spelling Ages ranging from 6 years 8 months to 8 years. The dyslexics had a mean chronological age of 9 years 10 months. They were on average 18 months older than the normal controls whose mean chronological age was 8 years 4 months.

The higher ability groups had Spelling Ages from 8 years to 10 years. The dyslexics had a mean chronological age of 12 years 2 months, some three years older than their controls with a mean chronological age of 9 years 1 month.

There was no significant difference between the Reading Ages or the Spelling Ages of the dyslexic subjects and the normal controls. Normal children of Spelling Ages greater than 10 years were also tested but as they made no errors in the experimental task, they were not included in the analyses.

Stimuli.


A list of all possible nasal endings was compiled using Rockey's Phonetic Lexicon (1973). A phonetic classification of the exhaustive list is presented in Table 1. A list of nonsense words, each composed of a CV or CCV structure followed by a target nasal ending was devised. Each nasal ending was tested three times in the context of three different nonsense words (e.g. plankle, nunkle. minkle). Thus, there were 48 stimuli altogether. The experiment was carried out in three parts, each nasal spelling being tested once during each part. The testing order of the various nasal endings was randomized once but then presented in that order for all subjects. This allowed each subject to be given the same three page booklet with 16 incomplete spellings on each page. The order of presentation of the three parts of the experiment (pages of the booklet) was randomized across subjects.

Procedure.


The experimental procedure was explained to the children and sufficient practice was given to ensure that they were familiar with the task. First of all, the experimenter pronounced the target stimulus in a clear voice. The subject was then required to repeat the stimulus. Only very occasionally was the stimulus mispronounced and in these cases further repetition was elicited before proceeding. Having pronounced the stimulus satisfactorily, the subject completed the partially spelled version on the page in front of him.

Scoring.


Each subject's protocol was scored for phonetic accuracy. Provided that the subject's response was phonetically acceptable it was regarded as correct.

No account was taken of illegal spelling patterns (e.g. nj for [ndʒ]) and b/d reversal errors were ignored. Furthermore, no account of order of errors was made. So, if the subjects included both elements of a target cluster in their spelling, but the order was wrong (e.g. one subject added 'ten' to 'he-' for "hent"), their version was still marked correct. This type of error occurred only very occasionally.

Given the nature of the completion task, a maximum number of three errors per stimulus could occur. Errors could occur with respect to the nasal itself (e.g. blen for blem), with respect to the other consonant in a nasal cluster (e.g. hend for hent), or with respect to the syllabic [l] ending (e.g. stemper for stemple). In each case the target element could either be omitted or replaced. (The above gave examples of substitution errors - examples of corresponding omissions might be: twage for twange, hen for hent and stemp for stemple).

Examination of the total corpus of errors indicated that the nasal elements of the spellings provided the greatest source of difficulty (c.f. Marcel, 1980). Over all subjects 85.2% of errors were made with respect to the nasal segments. A much smaller proportion of errors, some 11.9%, occurred on the other consonants and a negligible 2.88% of errors were made on the syllabic [l] endings. Hence, a decision was made to concentrate primarily upon nasal errors in subsequent analyses. Furthermore, 87.4% of the nasal errors were found to be nasal omissions, i.e. reduction of nasal clusters. Nasal substitutions occurred in only 12.6% of cases. Nasal substitutions were made primarily by children with Spelling Ages of less than 9 years who tended to represent [ŋ] by 'n' instead of 'ng.' The phoneme [m] was never confused with [n]. Thus, since the main purpose of including nasal-alone spellings was to ensure that the graphemic responses 'm' and 'n' were available, it seemed justified to spend no further time in discussing the error category of nasal "substitutions."

Hence, the number of nasal omissions made by each subject at each of the three levels of difficulty was calculated. (Nasal alone, nasal cluster, and nasal cluster +[l]). These scores were divided by the total number of errors possible at each level of difficulty (9, 12, and 18 errors respectively) and the results were expressed as percentage errors.

Results.


The data describing each subject's performance under each level of difficulty of the experimental task was transformed using a (log + 1) transformation. These data were then subjected to an ANOVA with two between and one within subjects variable (see Table 2).

The results of the ANOVA indicated that there was a significant difference between the various levels of difficulty of the task F (2,76)=48.1, p<0.001. For all subjects significantly fewer nasals were omitted from nasal-alone endings than from nasal-cluster items. Furthermore, significantly fewer nasals were omitted from final nasal clusters than from medial nasal clusters followed by syllabic [l] (F 0.5: 2,80=83-008, p<0.001).

There was a significant effect of Spelling Ability, F (1,38) =5.925, p<0.05 which confirmed that subjects of higher spelling ability made fewer nasal omission errors than subjects of lower spelling ability. However, the Groups effect did not reach significance, which indicated that dyslexic subjects omitted no more nasals during the experiment than their Spelling Age controls. None of the interactions were significant.

Qualitative Analysis of Errors.


Having confirmed that the Task Difficulty Effect was significant, it was of interest to establish the hierachy of difficulty of the various nasal cluster spellings within each level. (Nasal alone, Nasal + voiced stop, Nasal + unvoiced stop, Nasal + fricative, Nasal + affricate, Nasal + voiced stop + [l], Nasal + unvoiced stop +[l]. Thus, the percentage of error rate for each subject on each ending-type was tabulated (see Table 3 for subject means). A Friedman two-way analysis of variance by ranks indicated there was a significant difference in error rate across the various nasal cluster endings (χ2= 13.26, df =6, p< 0.025). However, the variability of the data was such that none of the more detailed comparisons (e.g. nasal + voiced vs nasal + unvoiced stop) reached significance. Therefore it was only possible to discuss the apparent trends tentatively.

Thus, amongst nasal cluster endings, the endings, nasal + affricate caused most difficulty (nch, nge). The remaining endings, nasal + voiced stop (nd), nasal + unvoiced stop (nt, nk, mp), and nasal + fricative (ns), appeared to be of equivalent difficulty. Amongst nasal cluster and [l] endings, those in which the nasal was followed by an unvoiced stop (-nkle, -ntle, -mple) appeared to cause more difficulty than those in which it was followed by a voiced stop (-ngle, -ndle, -mble).

Discussion.


The results of the experiment confirmed that there was a significant improvement in the ability to consistently represent nasals in spelling with an increase in spelling ability. The absence of a dyslexic-normal group difference suggested that this improvement was due primarily to increased spelling knowledge and was not dependent upon chronological age. Moreover the improvement could not be attributed to perceptual development because all of the subjects were able to repeat the stimuli accurately in all conditions.

The Task Difficulty Effect is of interest because it points to an important factor determining spelling accuracy, namely the number of phonetic segments in the target word. For all subjects, accuracy was greatest when only one phonetic segment had to be identified and transcribed as in the nasal alone endings. Performance in this condition also confirmed that graphemic responses for [m] and [n] were available for all subjects. Accuracy was less when two separate phonetic segments had to be identified and transcribed as in the final nasal clusters and least when it was necessary to deal with more than two segments as in the nasal cluster preceding syllabic [l] endings.

The observed order of difficulty cannot be attributed to a deterioration of phoneme-grapheme translation over time. This explanation would predict that the majority of errors on the nasal cluster + syllabic [l] spellings (e.g. stemple) would occur on the syllabic endings. However, the data shows that relatively few errors were made in these positions and the majority of errors were reductions of the medial nasal clusters. The Task Difficulty Effect also rules out the possibility that the accuracy with which a nasal is represented is entirely dependent upon its perceptual salience. If this were so, then nasals should be omitted as frequently from the final nasal clusters as from the same nasal clusters in medial position preceding syllabic [l]. The data show clearly that this is not the case, for many more reductions of medial clusters were made than of the same clusters in final position. A further argument against the perceptual salience explanation is that, although there was a tendency for nasals to be omitted more than when they were less "perceptible", i.e. prior to unvoiced stops (tantle) than when they were more "perceptible", i.e. prior to voiced stops (dundle), this tendency did not reach significance (F (1,36) = 3.76). Thus, while perceptual factors undoubtedly have a part to play in determining spelling proficiency, their role may be less important than previously suggested (Read, 1975; Marcel, 1980).

In order to provide a parsimonious explanation of these results, it is necessary to consider the various stages in the process of "spelling-by-ear." There are at least two possible ways in which this process could proceed. The target word could first be segmented into phonemes. Following this, each of the segments could be encoded and held in short term store for transcription in a left to right sequence. However, introspection suggests that it is more likely that the transcription process begins as soon as the phoneme has been segmented. In this case, the content of the "working store" is the unanalysed target word. A detailed examination of the time course of phonetic spelling could possibly shed light upon these two alternatives. For present purposes, the important consideration is that the processing demands presented by the nasal cluster + syllabic [l] spellings are greater than those presented by the final nasal clusters. More phonemes have to be segmented and also more 'bits' of information have to be held in short term store. As already argued, medial clusters cannot of themselves be more difficult to segment than similar final clusters but they may be more difficult to analyse when an additional segment, (e.g. [l] has to be stored simultaneously. In such cases, less attention can be devoted to their analysis and subsequent transcription. Hence, the nasal clusters are analysed in a superficial phonetic manner guided by perceptual factors instead of drawing upon knowledge of conventional spelling patterns. As spelling knowledge becomes more automatic, these spelling patterns are utilized more easily.

If this theory were to be accepted, it could also explain a discrepancy between the present results and those of Marcel (1980). Marcel reported a strong tendency for nasals to be omitted from final clusters in which they preceded unvoiced stops. This effect was absent from the present results. A possible explanation lies in the difference between the two experimental tasks. Marcel required his subjects to spell complete nonsense words while in the present study, subjects had only to add the target spelling patterns to incomplete versions. Thus, Marcel's subjects had to deal with a greater number of phonetic segments than the subjects in the present study. It is interesting that a similar tendency to that reported by Marcel, i.e. the reduction of more nasal clusters preceding unvoiced than voiced stops arose in the present study when the clusters were followed by syllabic [l]. In these cases, processing demands were higher just as they were in Marcel's free-spelling situation.

If the theory is a plausible one, then nasal reductions should occur in other instances when information processing demands are high. For instance, nasals should frequently be omitted if other phonemes with which they occur are difficult to analyse or transcribe. There is at least preliminary evidence that this is true in that nasals were frequently omitted when they were paired with affricates, ([ntʃ], [ndʒ]). Affricates pose difficulty for several reasons. First, they are more complex sounds and each appears to be composed of features common to more than one other phoneme. Thus, /dʒ/ starts like the phoneme /d/ but is released with affrication common to /ʒ/. Similarly, /tʃ/ starts like the phoneme /t/ but is released with the friction associated with /ʃ/ Secondly, the affricates /dʒ/ and /tʃ/ are not only similar to other phonemes but they also sound very similar to one another particularly in unfamiliar contexts (e.g. nonsense words). Thus, they are often confused by children (Ingram, 1976). In the present experiment, a significant number of substitution errors occurred in the nasal + affricate spellings (9%). These did not occur at any significant rate for any of the other final nasal clusters. The most common substitutions were [tʃ] for [dʒ] or vice versa, confirming their confusability. However, the substitution of [d] for [dʒ] and [t] for [tʃ] was noticed, which lends credence to the previous argument.

Finally, the affricates [dʒ] and [tʃ] also cause more problems for the speller than the phonemes like stop consonants because their orthographic representation is less straight forward. Whilst the phoneme /d/ can only be represented as d or dd, the phoneme /dʃ/ can be represented as j, ge, gi, gy. Similarly, while the phoneme /t/ can be represented by a single grapheme t or by tt, the phoneme /tʃ/ is represented by a consonant digraph ch. Children usually learn digraphs later than single letters (Perfetti & Hogaboam, 1975) and may confuse them with other digraphs e.g. sh for some time. Therefore there are several reasons why the nasal + affricate endings place heavy demands upon processing capacity. The difficulty posed by their analysis and transcription means that most of the speller's attentional resources are directed towards the affricate. This causes only superficial processing of the nasal and consequently it is frequently omitted.

Hence, it is proposed that children reduce nasal clusters whenever they have to deal with novel materials which place heavy demands upon their processing capacity. The simplifications which they make are systematic and rule governed and can be likened to 'phonological rules.' In the present experiment, the children were found to omit nasals from their spelling when a large number of phonetic segments had to be transcribed. However, it is interesting that a similar phenomenon is noticed at an earlier stage in child speech.

In the same way that pre-school children can distinguish minimal pairs such as 'cap' and 'camp' but spell 'camp' as cap (Read, 1975), young children have been shown to perceive certain phonemic distinctions but not to produce them (Smith, 1973, Dodd, 1975). For instance a child may recognize that 'pay' and 'play' are different but may produce [pei] for both. Another analogy between spelling and speech can be drawn in that in the present study, nasal clusters would be spelled correctly in certain contexts but not in others. In child speech development, it is often the case that a given sound may be produced in one context but not in another (Ingram, 1976). For instance, a child may produce a correct rendering of the cluster [st] in star [sta]. However, the same child may resort to a less mature pronunciation when learning a new word, e.g. stable [steibl].

Thus, the explanation proposed may be a general one. Basically, each individual is assumed to possess a system of phonological rules. The status of an individual's phonological system is dependent upon several factors, including age, language experience and, as argued in the present paper, orthographic knowledge. When an individual has to process familiar verbal materials, either for speech or for spelling, automatic motor programmes are available so there is no reason to call upon the phonological system. However, when unfamiliar materials have to be processed, phonological rules are brought to bear. These are basically just simplification devices which allow the individual to handle processing demands which exceed their processing capacity. Examples would be [nt] → [t], [mp] → [p], and so on. While this hypothesis could offer an attractive explanation, it must await further evidence and remain highly speculative at the present time.

Conclusions.


The present study has shown that the ability to consistently represent nasals in the spelling of nasal clusters is primarily dependent upon spelling knowledge.

The perceptual salience of the nasal has a part to play but its role is minor in comparision to that played by the overall phonological complexity of the target spelling. The study highlights the importance of taking Spelling Age into account when assessing the quality of spelling errors. For instance, had comparison been made between the spelling of the dyslexic subjects and normal spellers of the same chronological age, a preponderance of nasal omissions would have been observed in the dyslexics' spelling. These spellings would have been classified as 'nonphonetic' under some schemes or as arising because of perceptual difficulties under others. In turn, this classification may have led to the prescription of auditory skills training for the dyslexic children and such training may, in many cases, have been misdirected. First, there is evidence that children could already 'perceive' the nasal segments and secondly, perceptual salience has itself been shown to be of minor importance. The present study suggests that a more appropriate course of action might be initially to familiarize the children with the nasal spelling patterns visually (by analytic reading) or kinaesthetically (by copying or tracing). It may only then be reasonable to expect them to be able to organize spoken versions of words containing nasal clusters into the form required for accurate spelling.

Hence, the purpose of the present paper is not to suggest that a distinction between phonetic and nonphonetic spelling categories should be abandoned. Rather, it is meant to suggest that much more information is required about the development of the ability to spell-by-ear before this potentially fruitful approach to remediation can be pursued.

References.

Boder, E. (1971), "Developmental dyslexia: Prevailing diagnostic concepts and a new diagnostic approach." In Myklebust, H.R. (Ed.), Progress in Learning Disabilities, 2, N. Y.: Grune & Stratton.

Bradley, L. & Bryant, P.E. (1978), "Difficulties in auditory organization as a possible cause of reading backwardness." Nature, 271, 746-747.

Cromer, R.F. (1980), "Spontaneous spelling by language-disordered children." In Frith, U. (Ed.), Cognitive Processes in Spelling. London: Academic Press, 1980.

Dodd, B. (1975), "Children's understanding of their own phonological forms." Quarterly Journal of Experimental Psychology, 27.

Frith, U. (1979), "Reading by eye and writing by ear." In Kolers, P.A., Wrolstad, M. & Bouma, H. (Eds.), Processing of Visible Language, 1, N.Y.: Plenum Press.

Frith, U. (1980), "Unexpected spelling problems." In Frith, U. (Ed.) Cognitive Processes in Spelling, London: Academic Press.

Golinkoff, R.M. (1978), "Phonemic awareness skills and reading achievement," In Murray, F.B. & Pikulski, J.J. (Eds.) The acquisition of reading: cognitive, linguistic and perceptual prerequisites. Baltimore: Univ. Park Press.

Ingram, D. (1976), Phonological Disability in Children. London: Edward Arnold.

Malécote, A. (1960), "Vowel nasality as a distinctive feature in American English." Language, 36, 222-229.

Marcel, T. (1980), "Phonological awareness and phonological representation: Investigation of a specific spelling problem." In Frith, U. (Ed.) Cognitive Processes in Spelling. London: Academic Press.

Nelson, H.E. & Warrington, E.K. (1974), "Developmental spelling retardation and its relation to other cognitive abilities." British Journal of Psychology, 65, 265-274.

Perfetti, C.A. & Hogaboam, T. (1975), "Relationship between single word decoding and reading comprehension skill." Journal of Educational Psychology, 67, 461-469.

Perin. D. (1982), Phonemic segmentation and spelling. In British journal of Psychology (In prr ss)

Peters, M.L. (1975), Diagnostic and Remedial Spelling Manual. London: MacMillan.

Read, C. (1971), "Preschool children's knowledge of English Phonology," Harvard Educational Review, 41, 1-34.

Read, C. (1973), "Children's judgements of phonetic similarities in relation to English spelling. "Language Learning, 23, 17-38.

Read, C. (1975), "Lessons to be learned from the preschool orthographer." In Lennenberg, E. (Ed.). Foundations of Language Development - a multidisciplinary approach, Vol 2. UNESCO, Paris.

Rockey, D. (1973), Phonetic Lexicon. London: Heyden & Son, Ltd.

Smith, N.V. (1973), The Acquisition of Phonology: a Case Study. London: Cambridge University Press.

Stringer, J.A. & McKenzie, B.E. (1981), "Phonological analysis of monosyllabic words by backward spellers." Unpublished manuscript, La Trobe University, Australia.


TABLE 1.

PHONETIC CLASSIFICATION OF NASAL ENDINGS.

CLASSIFICATIONCONVENTIONAL
SPELLING PATTERN
  
Final nasal [n], [m], [ŋ]
Nasal + unvoiced stop [nt], [nk], [mp]
Nasal + voiced stop [nd]
Nasal + fricative [ns]
Nasal + affricate [ntʃ], [ndʒ]
Nasal + unvoiced stop + [l*] [ntl], [nkl], [mpl]
Nasal + voiced stop + [l**] [ndl], [ngl], [mbl]
-n, -m, -ng
-nt, -nk, -mp
-nd
-ns -nse -nce
-nch, -nge
-ntle, -nkle, -mple
-ndle, -ngle, -mble
 
* This character should have a stop below it.
** This character should have a comma below it.


TABLE 2.

Log percentage error rate of dyslexic and normal subjects of high and low spelling ability under three levels of task difficulty.
 Nasal alone
(blem)
Nasal cluster
(lound)
Nasal cluster
+ /l/, (wemble)
Low Spelling Ability
(S.A. < 8.0)1
 
Dyslexics0.6271.1811.539
Controls0.2940.8981.264
 
High Spelling Ability
S.A. 8.0 -10.0)
Dyslexics0.0.7021.039
Controls0.4640.7500.891

TABLE 3.

Spelling of nasal endings by normal and dyslexic children.
Qualitative assessment: Mean percentage of nasal reductions errors.

Type of EndingDyslexicsControls
Nasal alone7.512.0
Nasal + voiced stop or fricative13.516.5
Nasal + unvoiced stop13.012.0
Nasal + affricate28.020.0
Nasal + voiced stop + [L*]29.518.0
Nasal + unvoiced stop + [L**]37.325.5
* This character should have a stop below it.
** This character should have a comma below it.


Back to the top.