A2.

Phonology and Spelling

REBECCA TREIMAN 1

ABSTRACT

It has often been suggested that the English spelling system is so capricious that
children must rely largely on rote memory to learn how to spell in English.
However, Venezky's (1970) demonstration that the English spelling system is more
orderly than was thought before encouraged researchers to examine how children
go beyond rote memorization in learning to spell. In a series of publications
beginning in 1971, Read demonstrated that young children's spellings are often
systematic attempts to translate genuine phonological distinctions, even though
these spellings may be mistaken from the point of view of the conventional script.
For example, the fact that children sometimes spell the opening sound of truck as
ch reflects a real phonological difference between words whose spellings begin with
"tr" and those whose spellings begin with "t" followed by a vowel. More recent
research (e.g., Treiman, 1993) has amply confirmed that many of children's
beginning spellings are guided by their understanding of phonology. For example,
most Americans do not pronounce a separate vowel sound in words like girl, and
young American children often misplace the letter for the vowel sound in such
words or leave it out altogether, as in the errors GRIL and GRL. In British
English, in contrast, words such as "girl" and "better" are pronounced without an
"r" sound. Young British children are much more likely than their American
counterparts to leave out the r when writing these words. English has
morphological as well as phonological regularities, and some of these also
influence the spellings of quite young children. For example, when the "t" sound is
preceded and followed by a vowel in American speech, its pronunciation
approaches that of a "d." This is reflected in the writing of beginning American
spellers who are quite likely, for example, to write water as WODR. Such errors are
more likely to occur in one-morpheme words, like water, than in two-morpheme
words, like biting, where the first morpheme, in this case bite, ends in a "t" sound.
Children's willingness to maintain the t spelling in these latter words supports the

1 Preparation of this chapter was supported by grants from the National Science FoundatiOn (SBR-
9807736. BCS-0J30763). and the March of Dimes Birth Defects Foundation (12-FYOO-51). Thanks to
Lia Sotak for comments on a draft of the chapter.

31

T. Nunes, P Bryant (ed,.), Handbook oJ CllIldren's Lzteracy, 31-42

,[I 2004 Klul1'er Academzc Publzshers.
32 Rebecca Treiman

conclusion that from the start their spellings reflect their understanding of
linguistic patterns - phonological and morphological - and not just rote learning.

INTRODUCTION

The goal of this chapter is to review the research on the role of phonology in
children's spelling. I will focus on how children learn to spell in English,
particularly on the early stages of spelling development. The theme of the chapter is
that learning to spell is a linguistically guided process, not simply a process of rote
memorization. We must look closely at the structure of the language and at
children's knowledge of that structure if we wish to understand the spelling errors
that children make. One aspect of the structure of language - its phonological or
sound structure - plays an especially important role in children's early spelling. Its
role will be the focus of this chapter.

SPELLING AS ROTE VISUAL MEMORIZATION

Traditionally, it was thought that learning to spell in English is a process of rote

memorization. So capricious is the English writing system, so illogical are spellings
such as could and island, that learners have no choice but to laboriously memorize
words' spellings. On this view, learning to spell requires much diligence and
attentiveness but little linguistic skill or creativity. Children memorize words for
their weekly spelling tests in much the same way that they memorize other dull (to
them) facts, such as that Sacramento is the capital of California.
If learning the spelling of a word involves storing an arbitrary string of letters in
memory, then the same principles that govern other forms of serial learning should
apply to spelling. For example, the position of an item in the to-be-memorized string
should have the same impact on spelling as on other memorization tasks. Spellers
should show a serial position effect such that they perform relatively well on the
initial and final letters of a word and poorly on the middle letters. Indeed, a serial
position curve of this kind often appears in spelling (Jensen. 1962; Kooi. Schutz. &
Baker, 1965). Such findings could be taken to suggest that misspellings such as LRN
for learn, in which children omit letters from the middle of a word. are similar to
errors in which people omit items from the middles of other to-be-memorized
sequences. (Throughout this chapter, children's misspellings of words will be
indicated in upper-case letters.) Misspellings such as GRIL for girl. in which children
reverse two letters, are thought to be similar to reversal errors in other memorization
tasks, as when people interchange two digits of a memorized telephone number.
This view of spelling as rote visual memorization has implications for the
classification of children's spelling errors. On this view. misspellings should be
seen in relation to the letters of a word's conventional spelling. Errors should be
classified as omissions. reversals, or substitutions of particular letters in the to-be-
memorized form.
 Phonology and Spelling 33

SPELLING AS A PHONOLOGICALLY GUIDED PROCESS

The work of Richard Venezky (Venezky, 1970) set the stage for a change in thinking
about spelling and spelling acquisition. Venezky's analyses of the spelling-sound
relationships in some 20,000 English words showed that the writing system of this
language is more predictable than often believed. To be sure, the English writing
system is not characterized by the simple one-to-one relationships between
graphemes (letters or letter groups) and phonemes (units of sound) that prevail in
certain other alphabetic writing systems, such as Finnish. In English, a single
grapheme often corresponds to more than one phoneme, as when c stands for "k" in
cake but "s" in city. (Units of sound will be indicated in quotation marks throughout
the chapter.) Also, a single phoneme often has more than one potential spelling, as
when the "long a" sound is written as ay in bay and mayor but as ai in bait and aid.
What is often overlooked is that these variations are typically not random. As
Venezky pointed out, one can often predict which pattern will occur based on the
position of the letter or phoneme in the word and the word's morphological structure
(i.e., the smaller meaningful units or morphemes that it contains). For example, the
ay spelling of "long a" usually occurs at the end of a morpheme, as in bay, or before
a vowel, as in mayor. The ai spelling generally appears in other positions, as in hait
and aid. As another example, a medial "short e" sound in a monosyllabic word is
generally spelled as e, as in bed. However, the "short e" sound of health is spelled as
ea rather than as e, corresponding to the fact that the word is related to heal. Not all
English spellings are predictable, of course. The s in island and the ai in plaid seem to
be genuine exceptions. Still, the English writing system is more principled than often
believed (see Kessler & Treiman, 2001, 2003).
Venezky's work focused primarily on spelling-to-sound translation rather than on
sound-to-spelling translation. The work thus has more direct implications for
theories of word pronunciation and its development than for theories of spelling.
However, Venezky's discussion of the linguistic patterns in the English writing
system encouraged researchers to examine how spellers learn and use the patterns.
The work set the stage for viewing spelling errors not only in relation to the
conventional spelling of a word but also in relation to the word's linguistic structure.
Charles Read (1971, 1975a, 1975b, 1986) was one of the first investigators to
study the development of spelling in English as a linguistic process rather than a
process of rote visual memorization. Read's conclusions were based, in part, on a
detailed study of young children who began to spell before they had received much
formal instruction in spelling or reading. Preschoolers such as these may compose
messages such as WAN YOU GAD I CHANS SEND IS OL I LADR. AD DOW
GT ANE CHRIBLS, messages that contain many unusual spellings and that may
be difficult for adults to read. This one says: When you get a chance send us all a
letter. And don't get any troubles (Read, 1975b). Read examined the commonalities
among different children's spellings and the ways in which the children represented
various linguistic features. To bolster his conclusions, he also carried out
experimental studies with young children. This combination of naturalistic and
34 Rebecca Treiman

experimental research proved fruitful, with the naturalistic data pointing to certain
phenomena that were then examined in greater detail in the experiments.
Read concluded that learning to spell is a linguistic process, more akin to
learning to talk than to memorizing arbitrary sequences such as telephone numbers.
When children acquire their spoken language, they learn about the patterns in the
system and apply their generalizations to new instances, sometimes with surprising
results. For example, a preschooler might say "I holded the baby" rather than "}
held the baby." "Holded" is an error when judged against the conventional system,
but it reveals an appreciation of how the English past tense is typically formed.
Similarly, the preschool orthographer cited above wrote CHRIBLS instead of
troubles. This is an error when judged against the conventional writing system, but it
reveals the child's belief that the first part of troubles sounds like the first part of
chubby. The sounds are indeed similar: The articulation of "t" becomes close to that
of "ch" when "t" precedes "r." A similar phenomenon occurs for "d," which is
pronounced similarly to "j" when it comes before "r."
The child who wrote CHRIBLS for troubles is not unique. Other precocious
spellers studied by Read produced similar renditions of "t" before "r." Moreover,
children sometimes misspelled "d" before "r" as) or g, as in JRAGIN for dragon.
Such misspellings reflect the way in which children classify the sounds of their
language, one aspect of their linguistic knowledge. Supporting this interpretation,
experimental work reported by Read (197 Sa) showed that some young children do
not consider the first sound of truck to be the same as the first sounds of turkey
and tie, as adults do and as the conventional English writing system assumes.
Instead, these children consider the first sound of truck to match the first sounds
of church and chicken. The children's invented spellings testify to their own
understanding of English phonology.
Are only gifted preschoolers able to invent spellings that reflect their conceptions
of words' phonological forms? To determine whether more typical children do the
same, I examined the writings produced by middle-class first graders who attended a
state-supported school in the Midwestern United States (Treiman, \993). These first
graders were not precocious or advanced. What distinguished them from many
other children was that their teacher was a strong believer in the whole-language
approach to reading and writing instruction. Advocates of this view (e.g., Goodman,
1986) believe that children will pick up what they need to know about the relations
between spellings and sounds from the reading and writing that they do. It is felt
that children should not work with isolated words or isolated sounds but that they
should read meaningful and interesting texts. Independent writing is thought to be
important, especially writing that grows out of the children's own experiences. In
line with the whole-language philosophy, the teacher in the first-grade classroom
that I studied expected her students to write daily. She encouraged them to spell
words on their own, and she did not tell the children how to spell words even if they
asked. Invented spelling is a teaching tool in other classrooms as well, although this
teacher's refusal to provide the correct spelling of a word even when a child requests
it is rather unusual. The school district did mandate that the children memorize a
 Phonology and Spelling 35

list of six spelling words each week, beginning in the second half of the first-grade
year. The teacher complied with this requirement.
The first graders in my 1993 study made many of the same kinds of spelling
errors discovered by Read (1971, 1975a, 1975b, 1986). For example, the first graders
produced spellings such as CHRAP for trap and JEAD for drowned. Thus, it is not
just precocious children who begin to write at an early age who show a sensitivity to
the phonological structure of language in their spelling. My findings further revealed
that the spelling errors discovered by Read were just the tip of the iceberg. A
number of other intriguing errors that were motivated by phonology emerged in the
first graders' classroom spellings. My co-workers and I have carried out a number of
experiments to examine these phenomena in more detail (Bernstein & Treiman,
2001; Cassar & Treiman, 1997; Reece & Treiman, 2001; Treiman, 1985a; Treiman,
1985b; Treiman, 1991; Treiman, 1994; Treiman, Berch, Tincoff, & Weatherston,
1993a; Treiman, Berch, & Weatherston, 1993b; Treiman, Broderick, Tincoff, &
Rodriguez, 1998; Treiman & Cassar, 1996; Treiman, Cassar, & Zukowski, 1994;
Treiman, Goswami, Tincoff, & Leevers, 1997; Treiman & Tincoff, 1997; Treiman,
Zukowski, & Richmond-Welty, 1995). These studies, like my naturalistic study of
first graders, were carried out with children from the Midwestern United States.
Together, the naturalistic and experimental work provides a picture of how early
spelling is guided by phonology and other types of linguistic knowledge, what kinds
of errors occur among beginning spellers, and why children make these errors.
Consider the words girl and her. The first graders in my naturalistic study often
omitted the vowels when they spelled such words, producing errors such as GRL
and HR (Treiman, 1993). Studies in which children were asked to spell dictated real
and made-up words revealed the same phenomenon among other groups of
kindergarten and first-grade children from the United States (Reece & Treiman,
2001; Treiman et ai., 1993a; Treiman et ai., 1997). The results of these studies show
that children are much more likely to omit the vowels of words such as girl and her
than the vowels of words such as kept and him. Such differences emerge even when
the two types of words are comparable in length (e.g., girl and kept are both four
letters), consonant-vowel spelling pattern (e.g., her and him both contain a
consonant letter followed by a vowel and another consonant), and frequency of
occurrence in the English language. It is difficult to explain the difference in vowel
omission rates if we view spelling as purely a matter of rote memorization of letter
sequences. On that view, one might expect a fair number of vowel omissions in both
girl and kept because the vowel letter is the third letter of a four-letter sequence in
both cases. However, it would be difficult to explain why vowel omissions are more
common with girl than kept. The difference in the rate of vowel omissions becomes
easier to understand if we view spelling as guided by phonology. In American
English, the phonological form of girl contains a syllabic "r" (i.e., an "r" that takes
the place of the vowel at the centre of the syllable); it does not contain a separate
vowel phoneme preceding an "r." Because the spoken form of the word does not
include a separate vowel, it makes sense that children would often fail to include a
vowel letter in their spellings. Children who have begun to grasp the alphabetic
36 Rebecca Treiman

principle expect it to be implemented in a simple fashion such that each phoneme in

a spoken word is translated with a single letter. Girl - with its initial "gil, medial
syllabic "r," and final "1" - is most naturally spelled as GRL. In contrast, kept has a
true vowel phoneme in its pronunciation. Children therefore usually include a vowel
letter when spelling this word. As these examples show, we can begin to understand
why children make the spelling mistakes they do if we consider the phonological
structures of words.
In girl and her, the syllabic "r" is stressed. Syllabic "r" also occurs in unstressed
form in American English. For example, doctor and tiger end with an unstressed
syllabic "r." Children from the United States often fail to include a vowel letter in
such cases. For example, they may produce errors such as DOCTR for doctor and
TIGR for tiger, omitting the vowel that should appear in the second syllable of the
word. We may compare children's omissions of the second vowel in words such as
doctor and tiger, on the one hand, and words such as basket and salad, on the other.
The two types of words are similar in frequency, length, and consonant-vowel
spelling pattern. However, doctor and tiger do not contain a separate vowel in the
second syllable, whereas basket and salad do. Correspondingly, children are much
more likely to include a vowel letter when spelling the second syllables of basket and
salad than the second syllables of doctor and tiger. This result points, again, to an
effect of phonology on spelling. It further implies that labelling an error as an
omission of a letter from the word's conventional spelling may not shed much light
on the basis for the error. One must consider the word's phonological form - in this
case, whether it contains a syllabic "r" - as well as its standard spelling.
As children gain experience with the English writing system, they observe that
the conventional printed forms of her, girl, and learn contain vowel letters. These
are letters that the children did not anticipate based on their phonological
representations of the words. In some cases, children are also explicitly taught
that all words should be spelled with at least one vowel. As a result of such
teaching (if provided) and their exposure to conventional print, children begin to
include vowel letters in their spellings of words like her, girl, and learn. Children
sometimes make an interesting kind of spelling error during this learning process:
They use a vowel letter, but they put it in the wrong place. For example, children
may spell girl as GRIL or teacher as TETRE (Reece & Treiman, 2001; Treiman,
1993; Treiman et aI., 1993a; Treiman et aI., 1997). Looked at superficially, these
errors involve the reversal of two letters from the conventional spelling of the
word - i and r in the case of girl, and e and r in the case of teacher. However,
errors such as GRIL for girl and TETRE for teacher are more common than
errors such as KPET for kept and BASKTE for basket. If the misordering of
letters from the memorized spelling of the word were all that was involved, one
would not expect to find such a difference. The different error patterns for the
two types of words reflect differences in the words' phonological structures.
Because there is no separate vowel phoneme preceding the syllabic "r" in the
spoken form of girl, the word's phonological form provides no guidance on where
any vowel letter should go. Given this lack of phonological underpinning,
 Phonology and Spelling 37

children sometimes put the vowel letter in the wrong place. Kept, in contrast, has
a vowel phoneme after the "k" and before the final cluster in its spoken form.
Children who treat spelling as a map of phonological structure thus know that
they should place the vowel letter between the letters corresponding to the initial
and final consonants. We gain a deeper understanding of children's spelling errors
by acknowledging the importance of phonology in the spelling process than by
classifying errors into such superficial categories as omissions and reversals.
Further evidence of the role of phonology in spelling comes from comparisons
between children who speak different dialects of the same language. If phonology
contributes importantly to spelling, then phonological differences among dialects
may show themselves in spelling. To determine whether this is so, my colleagues and
I (Treiman et aI., 1997) compared the spellings of children who spoke two dialects of
English - American English, as spoken by children from the Midwestern United
States, and Southern British English. As discussed above, young children from the
Midwestern United States are more likely to misspell girl as GRL than as GIL or
GUL. Young children from Southern Britain, we found, show the opposite pattern.
They often include a vowel letter but omit the r, as in GUL. Differences are also
found on words like doctor. For American children, omissions in the second syllable
tend to involve the vowel. For British children, it tends to be the r that is omitted. If
spelling were only a matter of rote visual memorization, then we would expect to see
similar errors by children from the United States and children from Britain. After
all, the two groups of children encounter virtually the same conventional spellings
(with a few differences such as color/colour and recognize/recognise). The finding
that American and British children make different types of spelling errors on certain
words supports the idea that spelling is guided by phonology. The key phonological
difference in this case is that Southern British English is a non-rho tic dialect.
Speakers of this dialect do not include an "r" when saying a word such as girl.
When they pronounce doctor in isolation (and it appears to be the way in which
words are pronounced in isolation that is critical for spelling), they also omit the
"r." It is thus not surprising that young British children often produce spelling
errors such as GUL and DOCTE. Most varieties of American English, including the
Midwestern dialects that my colleagues and I have studied, are rho tic. This means
that an "r" is present in words such as girl and doctor. American children typically
include an r when spelling such words.
Intriguingly, dialect-related differences in spelling are not confined to words that
are pronounced differently in American English and British English. Differences are
found, too, on certain words that are pronounced alike in the two dialects. Consider
the word pizza. In both American and British English, this word ends with a short
unstressed vowel, called a schwa, when it is pronounced in isolation. However, the
British children in our study (Treiman et aI., 1997) frequently spelled such words
with a final r, as in PISER. The American children rarely made such errors. What is
the reason for this difference? For British children, the spoken form of pizza ends
with the same unstressed schwa as the spoken forms of doctor, tiger, after, and so
on. British children learn that schwa is often spelled not with a single-letter
38 Rebecca Treiman

grapheme but with a digraph, or two-letter spelling. The digraph consists of a vowel
letter, most often e, followed by r. British children generalize this pattern to the final
vowel of words like pizza, producing errors such as PISER. These errors show that
children pick up the relationships between phonology and spelling and generalize
these patterns to new instances. They sometimes create spellings based on words'
phonological forms rather than reproducing spellings from memory.
Another example of a dialect-related generalization error is BARTH for bath.
My colleagues and I (Treiman et aI., 1997) observed errors of this kind among the
British children that we studied. Other examples are PARS and PRS for pass. In
Southern British speech, bath has the same "ah" vowel sound as card. Children
sometimes use the ar spelling (or just the r) that they have seen in card when
attempting to spell words like bath. That is, they generalize the ar spelling that
they have associated with the "ah" vowel of card to the "ah" vowel of hath. Note
that bath does not contain an "r" sound in its pronunciation in non-rhotic
dialects. Errors like BARTH for bath reflect a generalization process rather than
an attempt to represent an "r" phoneme that is heard in the word itself.
In our cross-dialect study (Treiman et aI., 1997), dialect-related generalization
errors in spelling did not necessarily become less common as spelling ability
increased. For example, British children with spelling ages of about 8 and 9 years
old according to a standardized test actually made more errors like PISER for
pizza than British children with spelling ages of about 6 and 7. Misspellings such
as BARTH for bath were also found among the more advanced spellers. A certain
amount of experience with the conventional writing system is necessary to learn
that the schwa vowel may be spelled with the digraph er and that "ah" may be
spelled as ar. These patterns are fairly complex in that a single phoneme is
represented with a two-letter grapheme rather than a single letter. Once children
have learned these digraph spellings, they sometimes extend them to cases in
which they do not apply, much as a child learning to talk sometimes extends the
regular past tense marker to irregular verbs.
With enough experience, children learn the conventional spellings of common
words such as pizza and bath. However, Treiman and Barry (2000) found that
even adults sometimes produce dialect-related errors when attempting to spell less
common words. Consider the following misspellings from British university
students: SCUBER for scuba, KARKI for khaki, and CARSKET for casket.
These misspellings are similar to the PISER and BARTH errors that are found
among British children. American college students rarely make these particular
types of spelling errors, although they sometimes misspell the words in other
ways. These results suggest that phonology continues to be involved in spelling
even into adulthood. This conclusion fits with other work that has found
phonological effects in adult spellers (e.g., Kreiner, 1992; Kreiner & Gough,
1990); it does not support the view that skilled spellers generally bypass
phonology (e.g., Burt & Fury, 2000).
In English, it is not enough to learn which grapheme or graphemes may be used
to symbolize each phoneme. When a phoneme may be represented with more than
 Phonology and Spelling 39

one grapheme - as most English phonemes can - children must also learn when
each spelling is appropriate. As Venezky (1970) pointed out, one can often predict
which of several possible graphemes will be found in a particular word based on
various linguistic factors. One example, mentioned earlier, is that the "long a" sound
is generally spelled as ay at the end of a morpheme or before a vowel, as in bay and
mayor. In other positions, the ai spelling is more likely (e.g., aid, bait). The
alternation between ay and ai is one instance of a general pattern that applies to
digraphs. Digraphs ending with y and w typically occur at the ends of morphemes
and before vowels, whereas those ending with i and u are more often found in other
positions. This pattern does have some exceptions. For example, growl has ow
before a consonant in the middle of a morpheme, where u would normally be
expected. Still, the pattern holds true in many cases. Another example of a context-
conditioned spelling alternation involves "k." The ck spelling of "k" occurs in the
middles and at the ends of words but not at the beginnings, where c or k is more
common. This pattern has no exceptions, in that no English words begin with ck.
Evidence from Treiman (1993) suggests that children begin to learn about the
effects of positional context on spelling as early as first grade. For example, the
first graders in that study were more likely to produce spellings such as SEILF
and PLEW, which follow the alternation pattern described above, than spellings
such as AI and EWT, which do not follow this pattern (by having a digraph
ending in i at the end of a word or a digraph ending in w before a consonant).
Also, the first graders did not often produce spellings such as CKES, with ck at
the beginning of a word.
Further evidence of children's sensitivity to the effects of position on spelling
comes from experimental work. In one study, Treiman (1993) asked kindergartners,
first graders, second graders, and college students to look at pairs of nonwords and
decide which item in each pair looked more like a real word should look. Sample
pairs are nuck-ckun and moil-moyl, where the first member of each pair follows one
of the patterns described above and the second member of the pair does not. Even
the kindergartners, who were tested near the end of the school year and who were
reading at a beginning first grade level, were slightly but significantly above the level
expected by random guessing (56% correct, where 50% is expected by chance). By
second grade, correct responses occurred 83% of the time, and by adulthood, 95%.
These results, together with those of Cassar and Treiman (1997), suggest that even
young children have some inkling that spelling patterns vary in a systematic way
with word position. This knowledge expands and deepens as children encounter a
variety of words and make generalizations from them.
It is not just a phoneme's position in a word that can affect its spelling. Even
when position is held constant, the identity of the surrounding units may
influence a phoneme's spelling. Kessler and Treiman (2001) documented this
phenomenon. As one example, the "short e" sound is sometimes spelled as ea
before "d," as in bread and head. It is rarely so spelled before "p". Adults appear
to use the surrounding consonants as a cue to the spelling of a vowel. For
example, having learned that the "e" sound of friend is spelled as ie, they are
40 Rebecca Treiman

more willing to generalize this spelling to "chend," in which the "nd" after the
vowel is maintained, than to "cheth," in which different consonants surround the
vowel (Treiman & Zukowski, 1988; see also Treiman, Kessler, & Bick, 2002). It is
not clear when and how children begin to use phonemic context as a cue to
spelling. The results of Goswami (1988) suggest that units consisting of a vowel
and final consonant play a special role from early on in the development of
spelling. In that study, children who were given a "clue word" such as beak were
able to use the clue to help them spell related words; the benefit was significantly
greater for words that shared the vowel and final consonant of the clue word
(e.g., peak) than for words that shared the initial consonant and vowel (e.g.,
bean). However, the findings of Nation and Hulme (1996) suggest that young
children show no particular priority for vowel + final consonant units. In their
research, children did not make significantly more analogies when a vowel +
final consonant unit was shared than when some other unit was shared. Bernstein
and Treiman (2001) found similar results. Further research is needed to study
how children use context in choosing among alternative spellings for phonemes
and when the ability to use different types of context emerges.
The research reviewed so far shows that spelling is guided to a large extent by
phonology. One important aspect of learning to spell is learning how the
phonological structure of language is symbolized in print. When children's
conceptions of phonological structure do not match those embodied in the
conventional writing system, or when the links between phonology and spelling
are complex, spelling errors may occur. Importantly, the errors are usually not
random or haphazard. The misspellings reflect children's belief that the visual
forms of words map the phonological structure of the language.

BEYOND PHONOLOGY

Learning to spell in English involves learning about the way in which the
phonological form of the language is symbolized in print. However, this is not all
that it involves. English spelling also reflects other aspects of linguistic structure,
including the morphological structure of the language. Children must learn about
this aspect of the system as well. As one example of how this takes place, consider
the American English pronunciation of medial "t." This sound is typically
pronounced with a quick tap of the tongue against the upper part of the mouth,
called aflap, when it occurs before an unstressed vowel. Flaps are voiced (i.e., the
vocal cords vibrate during their articulation), and in this way are similar to "d."
American children often spell a word like "water" as WODR, representing the
flapped "t" with d (Read, 1975a; Treiman, 1993; Treiman et aI., 1994). This error
makes sense given the pronunciation of the word. With "biting," children can
avoid making an error on the flap if they consider the word's morphological
structure. "Biting" is composed of the stem "bite" (which ends with a "t" that is
not flapped when the stem is pronounced by itself) and the inflectional ending
 Phonology and Spelling 41

"ing." If children maintain the spelling of the stem when spelling the inflected
word, they can avoid errors like BIDING for "biting." Studies have shown that
children begin to be able to do this from an early age (Treiman et aI., 1994; see also
Treiman & Cassar, 1996). Further discussion of the role of morphology in spelling
may be found in the chapter by Bryant and Nunes (this volume).

CONCLUSIONS

We have seen, in this chapter, that learning to spell in English is to a large extent a
linguistic process. Memorization plays some role, for example in learning about the
s of island, but there is much more to spelling than rote memory. From an early
age, children treat spellings as maps of linguistic structure. They create spellings
that reflect their knowledge of linguistic form. This chapter has focused on one
particular aspect of linguistic form - phonology. Children's conceptions of
phonology do not always match those that are assumed by the conventional
English writing system. For example, children may classify the first sound of
"troubles" as a member of the "ch" category rather than as a member of the "t"
category. In addition, children take time to master the more complex mappings
between phonology and spelling, as when a single phoneme is represented with two
letters. By considering these and other factors, we can begin to understand the logic
behind children's misspellings. Even when children's spellings are incorrect, as they
often are, the spellings may reveal a good deal of linguistic knowledge.

REFERENCES
Bernstein, S., & Treiman, R. (2001). Learning a novel grapheme: Effects of positIOnal and phonemIc
context on children's spelling. Journal of Experimental Child Psychology, 79, 56-77.
Burt, J. S., & Fury, M. B. (2000). Spelling in adults: The role of readmg skills and experience. Reading
and Writing' An Interdisciplinary Journal, 13, 1-30.
Cassar, M., & Trelman, R. (1997). The beginnings of orthographic knowledge: ChIldren's knowledge of
double letters in words. Journal of EducatIOnal Psychology, 89, 631-644.
Goodman, K. S. (1986). What's whole in whole language. Portsmouth, NH: Heinemann.
Goswami, U. (1988). Children's use of analogy in learning to spell. British Journal of Developmental
Psychology. 6, 21-33.
Jensen, A. R. (1962). Spellmg errors and the serial-position effect. Journal of Educational Psychology, 53,
105-109.
Kessler, B., & Treiman, R. (2001). Relationships between sounds and letters in English monosyllables.
Journal of Memory and Language, 44, 592-617.
Kessler, B., & Treiman, R. (2003). Is English spelling chaotic? Misconceptions concerning its Irregulanty.
Reading Psychology, 24, 291-313.
Kooi, B. Y., Schutz, R. E., & Baker, R. L. (1965). Spelling errors and the serial-position effect. Journal oj
Educational Psychology, 56, 334-336.
Kreiner, D. S. (1992). Reaction time measures of spelling: Testing a two-strategy model of skilled
spelling. Journal of Experimental Psychology. Learning, Memory, and Cognition, 18,765-776.
Kreiner, D. S., & Gough, P. B. (1990). Two ideas about spelling: Rules and word-specific memory.
Journal of Memory and Language, 29, 103-118.
Nation, K., & Hulme, C. (1996). The automatic activatIOn of sound-letter knowledge: An alternative
42 Rebecca Treiman

InterpretatIOn of analogy and priming effects In early spellIng development. JourIlal oj Expemnental
Child Psychology, 63, 416-435.
Read, C. (1971). Pre-school children's knowledge of English phonology. Harvard EducatIOnal Rel'lelt', 41,
1-34.
Read, C. (1975a). Children's categori::atlOn of speech sounds in English (NCTE Research Report No. 17)
Urbana, IL: NatIOnal Council of Teachers of EnglIsh.
Read, C. (1975b). Lessons to be learned from the preschool orthographer. In E. H Lenneberg & E.
Lenneberg (Eds.), Foundations of language development (Vol. 2, pp. 329-346). New York: AcademiC Press.
Read, C. (1986). Children's creative spelling. London: Routledge & Kegan Paul.
Reece, C., & Treiman, R. (2001). ChIldren's spellIng of syllabiC /r/ and ofletter-name vowels. BroadenIng
the study ofspe1lIng development. Applied PsycholingUlstics, 22,139-165.
Treiman, R. (\985a). Phonemic awareness and spelling: Children's Judgments do not always agree with
adults'. Journal of Experimental Child Psychology, 39, 182-201.
Treiman, R. (\985b). Spelling of stop consonants after /s/ by chIldren and adults. Applied
Psycholinguistics, 6, 261-282.
Trelman, R. (\991). Children's spelling errors on syllable-imtIal consonant clusters. Journal of
Educational Psychology, 83, 346-360.
Treiman, R. (1993). Begmning to spell A study offirst-grade children. New York: Oxford UnIversity Press.
Treiman, R. (\ 994). Use of consonant letter names in beginmng spellIng. Developmental Psychology, 3(),
567-580.
Treiman, R., & Barry, C. (2000). Dialect and authography: Some differences between American and BntIsh
spellers. Journal of Expel'lmental Psychology: Learning, Memory, and Cogl1ltlOn, 26, 1423-1430.
Treiman, R., Berch, D., TIncoff, R., & Weatherston, S. (\ 993a) Phonology and spellIng: The case of
syllabic consonants. Journal of Experimental Child Psychology, 56, 267-290
Treiman, R, Berch, D., & Weatherston, S. (1993b). Children's use of phoneme-grapheme correspon-
dences in spelling: Roles of position and stress. Journal of EducatIOnal Psychology, 85,1-12.
Treiman, R., Broderick, V., Tincoff, R., & Rodriguez, K. (\998) ChIldren's phonological awareness
Confusions between phonemes that differ only In voicing. Joumal of Expenmental Cluld Psychology,
68, 3-21.
Treiman, R., & Cassar, M. (\996). Effects of morphology on children's spellIng of final consonant
clusters. Journal of Experimental Child Psychology, 63, 141-170
Treiman, R., Cassar, M., & Zukowski, A. (1994). What types of lInguistIc informatIon do children use In
spelling? The case of flaps. Child Development, 65, 1318-\337.
Treiman, R., Goswaml, U., TIncoff, R., & Leevers, H. (1997). Effects of dialect on AmerIcan and BrItish
chIldren's spellIng. Chtld Development, 68, 21\-227.
Treiman, R., Kessler, B., & Blck, S. (2002). Context senSitivity In the spellIng of EnglIsh vowels Joumal
of Memory and Language, 47, 448-468.
Treiman, R., & Tincoff. R. (1997). The fragility of the alphabetIc prInCiple. Children's knowledge of
letter names can cause them to spell syllabically rather than alphabetIcally. Jotlmal of Experimental
Child Psychology, 64, 425-451.
Treiman, R., & Zukowski, A. (1988). Umts in readIng and spelling. Journal of Memory and Language, 27,
466-477.
Treiman, R., Zukowski, A., & Richmond-Welty, E. D. (1995). What happened to the "n" of SInk?
Children's spellings of final consonant clusters. Cognition, 55, 1-38.
Venezky, R. L. (\970). The structure of English orthography. The Hague: Mouton

SUGGESTED FURTHER READINGS

Brown, G. D. A. and N C. Ellis (1994). Handbook oj spelllllg Theory. process and Intervention.
Chichester, England: Wiley.
Treiman, R. (1993). Beginning to spell A study of first-grade chzldren. New York' Oxford Umvcrslty
Press.