[Journal of the Simplified Spelling Society, J25, 1999/1 pp16-19]
[See other journal articles by Patrick Groff, and articles on Fonemes and Fonics.]

The Phoneme-to-Letter Route for Phonics Instruction

Patrick Groff.

Patrick Groff, professor of education emeritus at San Diego State University, has written extensively on his academic specialty, children's literacy development. This article, written in mid-1998, anticipates the current debate in the UK between advocates of an 'analytic' (ie, letter-to-sound) approach to initial literacy teaching and advocates of a 'synthetic' (ie, sound-to-letter) approach.


Whether phonics rules are best acquired by children by teaching them (1) to spell speech sounds, as versus (2) to sound out letters in words, remains an empirical question. So far, there do not appear to be experimental findings necessary to resolve this issue, Accordingly, a debate between reading instruction specialists continues in this regard, based exclusively on opinion and logic. One argument in favor of the second procedure, that it involves significantly fewer phoneme-letter correspondences for children to learn, is not substantiated, however. From the standpoint of simplified spelling, there appear to be inherent advantages in the speech sound-to-letter approach. Instructing children to spell words, as a means to develop their phonic skills, could concentrate teachers' attention on the need to simplify English orthography, and thus make them more hospitable to plans for implementing this reform. The overriding importance of regularizing speech sound-to-letter connections might receive more acceptance from educators if children's phonics knowledge ordinarily was developed as they learned to spell speech sounds.


In the past, the customary manner in which children were taught the relationships between speech sounds and how letters are used to represent them in writing was a letter-to-speech sound procedure. In this method, children first learn to recognize letters. Next, the traditional "phonics method introduces many of the sounds of letters and letter combinations so the child can put them together to make [ie, pronounce written] words" (Zintz, 1980, p189). Seldom offered to teachers of yesteryear, however, was advice as to how to be certain that the novice reader had become consciously aware of the separate speech sounds, a discovery that is necessary for him/her to successfully attach speech sounds to letters, ie, to sound them out.

It previously was noted in this respect, but only in passing, that the fledgling reader's "ability to discriminate among the various [speech] sounds...appears to be crucial" to his/her "success in reading" (Knight, 1983, p29). Teachers were encouraged by Smith and Robinson (1980, p70) to believe that the "conscious attention" of children "directed toward the [speech] sounds of language" was needed by them to learn phonics information. Nevertheless, these well-known reading teaching experts of their day provided only a way to test, not teach, this "auditory discrimination" ("pupils clap once whenever a word is said beginning with the [target speech] sound."

Defense of phoneme-to-letter teaching.

At the same point in time, however, a few reading instruction experts did maintain that a speech sound-to-letter instructional routine may be as effective (or more) for fostering beginning readers' phonics knowledge as the reverse practice. It thus was ventured that knowledge about associations between phonemes and letters formed in one direction, phoneme-to-letter, should be usable for decoding words, letter-to-phoneme (Barron, et al., 1980). If so, "only one set of print-phonology connections is logically necessary" for children learning to spell, as well as to read, it was held (Cossu & Marshall, 1985, p29).

Moats (1995, p43-44) also is confident that children progress faster in learning to read words if they are "taught to spell them by using speech sound-symbol correspondences."

Other recent commentary is even more direct about the superiority of a speech sound-to-spelling approach to children's acquisition of phonic skills. McCracken and McCracken (1996, p2) insist that "teaching children how to spell is the most direct way of teaching how print works." Wilde (1997, p75) agrees that "spelling, rather than reading, is the appropriate arena for children to focus on phonics relationships."

To this effect, McGuinness (1997, p97) asserts that the "true logic of the alphabetic code" demands that beginning readers first become consciously aware of the phonemes in English. Following this basic accomplishment, she (p102) goes on, students should learn "the most likely spelling for each phoneme," then "the next most likely spelling, and the next."

Three main assumptions.

The argument for teaching children to spell speech sounds, as the means by which to develop phonics skills, rests on three main assumptions. First, it is vouched that this practice develops novice readers' conscious knowledge of speech sounds (phonemic or phonological awareness) faster and more securely than otherwise is possible. This learning is an imperative goal, say several recent experimental studies. They found that prereaders' and beginning readers' phonemic awareness has a positive effect on their learning to accurately recognize written words.

These empirical findings consistently indicate that "children who discriminate and manipulate phonemes with ease learn to read and spell more successfully than peers who do not have these skills" (Goulandris, 1994, p410). If entry-level readers develop phonemic awareness before attempting to read written words, it is found, they thereafter will learn phonics information (how letters represent phonemes) better than otherwise is possible.

It therefore is indisputable that children's phonemic awareness "bears an important relationship to achievement in reading," and as a consequence, that "instruction [in phonemic awareness] improves reading skills" (Snider, 1995, p443). In this regard, Stanovich (1994, p284) believes that "a 7-minute phonological awareness test will predict ease of initial reading acquisition better than a 2-hour intelligence test!"

A second defense of the phoneme-to-letter route to children's phonics knowledge attainment rests on the deduction that it is more time-effective. In this respect, it is said that as students learn to spell words, during the same time they will gain the phonics knowledge required for sounding-out letters in words. As proof for this contention is cited the fact that students always can read words they are able to spell. While there is a single set of experimental findings to the contrary (Bryant & Bradley, 1980), it has been discredited as having a disabling flaw in its research design (Foorman, 1995). Learning to spell a word thus is deemed the most time-efficient manner in which to learn to read it.

Compounding this issue, however, is the incidence of children who can read normally, but still have serious problems in learning to spell (Thomson, 1990). Some children read well "but spell appallingly" (Treiman, 1993, p36). Thus, the "ability to read words does not assure the ability to spell them" (Dryer, et al., 1994, p309). "Being able to read a word is insufficient to enable misspelling to be detected reliably because reading does not establish a full orthographic description of a word," it is judged (Funnell, 1992, p. 98).

In any event, the reason that spelling-only retardation occurs in students appears to be a pedagogical one (Kirk, 1983). Retarded spellers/normal readers demonstrate they somehow acquire the ability to recognize familiar spelling patterns in words without having satisfactorily learned phonics generalizations. It is clear that the able reader/speller thus gains mastery of the phonics rules; the spelling-only retarded child relies on a partial number of phonics cues when reading words.

The experimental evidence indicates that reading instruction therefore must make certain that all children master all of the phonics rules. To be able to do so, students must give close attention to the sequence of speech sounds in spoken words. Speech sound-to-spelling instruction concentrates students' attention on the sequence of phonemes in words more intently than otherwise is possible. Hence, it will work better to prevent spelling-only deficiency among children learning to read than will spelling-to-phoneme instruction, it is submitted.

A third justification for phoneme-to-letter teaching is based on the close relationships found between students' learning to spell and to read. For example, relatively high coefficients of correlation (r's) have been calculated between these two variables. It is true these r's do not always denote a causal connection between the factors they examine. A sharp rise in the appearance of storks in a region, with an accompanying precise rate of increase in the area's birthrate, would calculate into a very high r between the two instances. But no causal link between them could be attributed, of course.

Nonetheless, when rationally related matters are of concern, such as different aspects of literacy attainment, r's between such factors do have legitimate interpretative values. Thus the relatively high average r of .83 found between children's spelling and reading achievement in the various grades (Ehri, 1997) suggests that students' acquisition of reading skills benefits significantly from their ability to spell words, and vice versa. In further support of this conclusion is the finding that an r obtained between children's phonics knowledge and reading is approximately the same as one computed between their phonics knowledge and spelling (Hammill & McNutt, 1981).

However, reading skills are not exactly the same as those employed in spelling, the r's between them suggest. In fact, about one-third of the variance in common between the two is not accounted for with an r of .80. Nevertheless, most literacy development specialists do not concur with Bosman's (1994, p122) judgment that "reading does not benefit from spelling."

To the contrary, reading and spelling are seen by them as "two sides of a coin" since a "logical symmetry" exits between the two processes (Perfetti, 1997, pp28-29). In this regard, experimental research is cited that suggests spelling ability is "a necessary component in a complete theory of early literacy acquisition" (Cataldo & Ellis, 1990, p101). Spelling therefore is inferred to be "an independent contributor to the emergence of reading," and accordingly has "the major role in promoting [children's] insight into the alphabetic nature of the written language" (p101). For instance, Barker, et al, (1992) found that by grade three (age 8) the spelling skills of children made a significant contribution to their performance on five types of reading tests.

Learning to spell doubtless helps make clear to beginning readers the relationships between written words and their pronunciations. Specifically, it forges a link between these children's phonological awareness and their letter-to-sound knowledge (Cataldo & Ellis, 1990). The service given in this regard contributes an explanation as to why children's phonemic awareness has such high utility for predicting how well beginning reading instruction with them will progress.

Letter-to-phoneme teaching resistance.

Nonetheless, there remains active resistance to the contention that a speech-to-letter system for developing children's phonics skills is superior to its competitor. This is shown by the fact that teaching phonics to children by first having them learn to identify the letters, and then to attach speech sounds to them, clearly has the weight of conventional practice on its side. Popular texts over the years on phonics teaching (eg, Heilman, 1981) did not even contain the term spelling in their indexes. An up-to-date volume on this topic (Chall & Popp, 1996) devotes only a page and a half to the relationship of students' acquisition of writing/spelling and phonics skills.

Opponents of the phoneme-to-letter approach remain convinced that it has "been well-established" that "mastery of spelling is generally subsequent to [that of] reading" (Barry, 1994, p35). Henderson (1990, p88) concurs that "to study a word for spelling, it is necessary that the pupil [first] be able to read that word." A common contention over the years has been that children's identification of letters must be "thoroughly mastered" before phonics instruction for them is begun (Bloomfield & Barnhart, 1961, p36).

Bryant and Bradley (1980, p362) support the idea that "children start learning to read and spell in rather different ways." With this point in mind, it "would be quite wrong to regard the spelling process merely as the reverse of the reading progress," Nelson (1980, p491) warns teachers. Statements that "spelling and reading can't be simple inverses of each other" (Wilde, 1992, p27), ie, that "spelling cannot be simply the mirror image of reading" (Gough, et al, 1992, p42) are often expressed. Therefore, "spelling-to-sound rules just plain do not work in reverse," Adams (1990, p389) is convinced.

The major reason given for the belief that spelling and reading processes are in large part inimical is that letter-to-speech sound correspondences in English are more regular, predictable, consistent, dependable, reliable, etc, and therefore are easier for students to learn, than are phoneme-to-letter correspondences (Berninger, 1995; Bosman & Van Orden, 1997; Ehri, 1997; Henderson, 1990; Stanovich, 1993). It is relatively easy to locate such declarations of opinion on this matter. However, only rarely is any citation provided here as to comparative analyses of the two frequencies of correspondences that were consulted for making the declaration.

The Actual Number of Correspondences.

Therefore, it is appropriate to calculate precisely how accurate is the often expressed generalization that phoneme-to-letter correspondences are much more numerous than are letter-to-phoneme correspondences. A comprehensive source of evidence in that regard is the compilation of the frequency of phoneme-grapheme correspondences made by Hanna, et al. (1966).

These researchers compiled, for a corpus of the 17,000 most frequently used words, the total number of different ways speech sounds can be spelled with letters, digraphs, trigraphs, and quadrigraphs. For example, the speech sound of A as in sad is spelled 3 different ways. The speech sound of A as in late, is given 16 various spellings.

By cross-referencing these data, I calculated the total number of different speech sounds that can be attached to each of the letters, digraphs, etc. For example, the letter A in words can be sounded-out (decoded) with 6 different speech sounds. The letter E in words can be so decoded in 8 diverse ways, one of which is not to give it a speech sound.

My recalculations of the Hanna, et al. (1966) data did not include the correspondences between the schwa sound /ə/ and letters used to represent it, nor vice versa. It is not practical, ie, time-effective, to teach beginning readers these correspondences (Groff, 1983; Heilman, 1981).

My reanalysis of the Hanna, et al. (1966) data does not support the contention that the total number of letter-to-speech sound correspondences is much smaller than the number of phoneme-to-letter correspondences. In fact, I found that these numbers are approximately the same: 303 versus 290.

Discussion and conclusions.

The present review of the argument for use of the speech sound-to-letter technique for developing children's phonics knowledge, as versus that in favor of the letter-to-speech sound practice, reveals that neither side of this debate can cite findings of a teaching experiment, of at least a year's duration, as corroboration for their respective view on the issue. "I cannot come up with any reference pertaining to such a study," internationally known spelling researcher Anna Bosman (1998) writes to me.

A study by Thompson and Fletcher-Flinn (1993), which concludes that the speech sound-to-letter correspondence knowledge of 5-6-year-olds is not a source of their letter-to-speech sound knowledge, does not meet this research design criterion. Therefore, it remains an open, empirical question as to whether phonics knowledge of beginning readers is best developed by classroom teaching of one, as versus the other of these two procedures. Experimental studies of that question obviously are greatly in demand.

Such investigations are needed to answer pressing questions such as: (1) Is the d ifficulty children have in mastering the handwriting of letters, which is needed for spelling, so great that it nullifies an advantage found in spelling words as a means of acquiring phonics rules? and (2) To what relative degree does the speech sound-to-spelling process overcome the disruptive effects of spelling irregularities on children's acquisition of reading skills? If this process was found to be more effective in this respect than is a letter-to-speech sound procedure, its reputation among educators doubtless would be enhanced.

In the interim, a simplified spelling point of view about this matter suggests that teachers should not be hesitant about giving priority to a phoneme-to-letter instructional approach to developing their pupils' phonics skills. That is to say, simplified spelling constructs are based on the principle that the number of phonemes in a language is of primary consideration in determining its optimum spelling system. In that respect, simplified spelling holds the phonology of English constant, while experimentally augmenting, to the least degree possible, the number of written symbols needed to represent it. Hence, from the simplified spelling vantage, the most advantageous spelling system for anyone learning to be literate is one in which a static number of phonemes, and their protean written representations are equal in number.

In this regard, the likelihood that a phoneme-to-letter approach will put more emphasis on a development of children's phonemic awareness, ie, attentiveness to speech sounds, than will its competitor, should impress advocates of simplified spelling. The consistent empirical verification of the important influence that children's phonemic awareness has on their acquisition of traditional spelling skills, and the close relationships discovered between children's reading and spelling ability, lends substance to a simplified spelling precept that phoneme identification may be the first step in the spelling reform process.

Then, as noted, in traditional spelling there is about an equal number of correspondences in both phoneme-to-letter and letter-to-phoneme directions. This fact should help reduce apprehension by teachers that their emphasis of the former will unnecessarily complicate students' task of attaining knowledge about phonics rules. Thus, if the spelling reform movement were to endorse speech sound-to-letter instruction, this decision probably would not risk creating greater difficulty for children learning to read than they historically have experienced. As an added bonus, this ratification of a speech sound-to-letter approach may help reduce the number of spelling deficient-normal reading students.


Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: MIT.

Barker, T. A., Torgesen, J. K., & Wagner R. K. (1992). The role of orthographic processing skills on five different reading tasks. Reading Research Quarterly, 27, 335-345.

Barron, J., Treiman, R., Wilf, J. F., & Kellman, P. (1980). Spelling strategies. In U. Frith (Ed.), Cognitive processes in spelling (pp159-193). New York, NY: Academic.

Barry, C. (1994). Spelling routes (or roots or rutes). In G. D. A. Brown & N. C. Ellis (Eds.) Handbook of spelling (pp27-49). New York, John Wiley.

Berninger, V. W. (1995). Introduction to the varieties of orthographic knowledge: Relationships to phonology, reading, and writing. In V. W. Berninger (Ed.), Varieties of orthoqraphic knowledqe, Volume II (pp1-22) Boston MA: Kluwer Academic.

Bloomfield, L. & Barnhart, C. L. (1961). Let's read: A linguistic approach. Detroit, MI: Wayne State University Press.

Bosman, A. M. T. (1994). Reading and spelling in children and adults: Evidence for a single‑route model. Amsterdam, Netherlands: University of Amsterdam.

Bosman, A. M. T. (1998). Letter to the author.

Bosman, A. M. T. & Van Orden, G. C. (1997). Why spelling is more difficult than reading. In C. A. Perfetti, L. Rieben, & M. Fayol (Eds.), Learning to spell (pp173-194). Hillsdale, NJ: Erlbaum.

Bryant, P. E. & Bradley, L. (1980). Why children sometimes write words which they do not read. In U. Frith (Ed.), Cognitive processes in spelling (pp355-370). New York, NY: Academic.

Cataldo, S. & Ellis, N. (1990). Learning to spell, learning to read. In P. D. Pumphrey & C. D. Elliott (Eds.), Children's differences in reading, spelling and writing (pp101-125). New York, NY: Falmer.

Chall, J. S. & Popp, H. M. (1996). Teachinq and assessing phonics. Cambridge, MA: Educators Publishing Service.

Cossu, G. & Marshall, J. C. (1985). Disassociation between reading and written spelling in two Italian children: Dyslexia without dysgraphics? Neuropsychologia, 23, 697-700.

Dryer, L. G., Luke, S. D., & Milican, E. K. (1995). Children's acquisition and retention of word spellings. In V. W. Berninger (Ed.), The varieties of orthographic knowledge, Volume II (pp291-320). Boston, MA: Kluwer Academic.

Ehri, L. C. (1997). Learning to read and learning to spell are one the same, almost. In C. A. Perfitti, L. Rieben, & M. Fayol (Eds.), Learning to spell (pp. 237-269). Hillsdale, NJ: Erlbaum.

Foorman, B. R. (1995). Research on "the great debate": Code-oriented versus whole language approaches to reading instruction. School Psychology Review, 24, 376-392.

Funnell, E. (1992). On recognizing misspelled words. In C. M. Sterling & M. Robson (Eds.), Psychology, spelling and education (pp87-99). Philadelphia, PA: Multilingual Matters.

Gough, P. B., Juel, C., & Griffith, P. L. (1992). Reading, spelling, and the orthographic cipher. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp35-48). Hillsdale, NJ: Erlbaum.

Goulandris, N. K. (1994). Teaching spelling: Bridging theory and practice. In G. D. A. Brown & N. C. Ellis (Eds.), Handbook of spelling (pp407-423). New York, NY: John Wiley.

Groff, P. (1983). A test of the utility of phonics rules. Reading Psychology, 4, 217-225.

Hammill, D. D. & McNutt, G. (1981). Correlates of reading. Austin, TX: PRO-ED.

Hanna, P. R., Hanna, J. S., Hodges, R. E., & Rudorf, E. H. (1966). Phoneme‑grapheme correspondences as cues to spelling development. Washington, DC: U.S. Office of Education.

Heilman, A. W. (1981; 1990). Phonics in proper perspective. Columbus, OH: Charles E. Merrill; New York, NY: Macmillan.

Henderson, E. H. (1990). Teaching spelling. Boston, MA: Houghton Mifflin.

Kirk, U. (1983). Introduction: Toward an understanding of the neuropsychology of language, reading, and spelling. In U. Kirk (Ed.), Neuropsychology of lanquaqe, readinq, and spelling (pp3-31). New York, NY: Academic.

Knight, L. N. (1983). Readiness for beginning reading. In J. E. Alexander (Ed.), Teachinq readinq. Boston, MA: Little, Brown.

Macmillan, B. (1997). Why schoolchildren can't read. London, UK: The Institute for Economic Affairs.

McCracken, M. J. & McCracken, R. A. (1996). Spelling through phonics. Winnipeg, Canada: Peguis.

McGuinness, D. (1997). Why our children can't read. New York, NY:

Moats, L. C. (1995). Spelling: Development, disabilities, and instruction. Baltimore, MD: York.

Nelson, H. E. (1980). Analysis of spelling errors in normal and dyslexic children. In U. Frith (Ed.), Cognitive processes in spelling (pp475-493). New York, NY: Academic.

Perfetti, C. A. (1997). The psycholinguistics of spelling and reading. In C. A. Perfetti, L. Rieben, & M Fayol (Eds.), Learning to spell (pp21-38). Hillsdale, NJ: Erlbaum.

Smith, N. B. & Robinson, H. A. (1980). Reading instruction for today's children. Englewood Cliffs, NJ: Prentice-Hall.

Snider, V. E. (1995). A primer on phonemic awareness: What it is, why it's important, and how to teach it. School Psychology Review, 24, 443-455.

Stanovich, K. E. (1993). The language code: Issues in word recognition. In S. R. Yussen & M. C. Smith (Eds.)§ Reading across the life span (pp111-135). New York, NY: Springer-Verlag.

- (1994). Romance and reality. Reading Teacher, 47, 280-291.

Thompson, G. B. & Fletcher‑Flinn, C. M. (1993). A theory of knowledge sources and procedures for reading acquisition. In B. Thompson, W. E. Tunmer, & T. Nicholson (Eds.), Reading acquisition processes (pp20-73). Philadelphia, PA: Multilingual Matters.

Thomson, M. (1990). Evaluating teaching programmes for children with specific learning difficulties. In P. D. Pumfrey & C. D. Elliott (Eds.), Children's difficulties in reading, spelling, and writing (pp155-172). New York, NY: Falmer.

Treiman, R. (1993). Beginning to spell. New York, NY: Oxford.

Wilde, S. (1992). You kan red this! Spelling and punctuation for whole language classrooms, K-6. Portsmouth, NH: Heinemann.

- (1997). What's a schwa sound anyway? Portsmouth, NH: Heinemann.

Zintz, M. V. (1980). The reading process: The teacher and the learner. Dubuque, IA: Wm. C. Brown.

Back to top.