Content uploaded by Heinz Wimmer
Author content
All content in this area was uploaded by Heinz Wimmer
Content may be subject to copyright.
Reading and Writing: An Interdisciplinary Journal 15: 653–682, 2002.
© 2002 Kluwer Academic Publishers. Printed in the Netherlands.
653
Phoneme awareness and pathways into literacy:
A comparison of German and American children
VIRGINIA MANN
1
& HEINZ WIMMER
2
1
Department of Cognitive Sciences, University of California, Irvine, California, USA;
2
Department of Developmental Psychology, Universität Salzburg, Germany
Abstract. Where American kindergartners are taught letters and letter sounds, German
kindergartners are not; where American first and second graders receive an eclectic blend
of whole language, whole word and phonics-based approaches, their German counterparts are
taught by an intensive synthetic phonics approach. As a probe to the consequences of these
pedagogical differences on the emergence of phoneme awareness, this study administered two
tests of phoneme awareness to kindergarten-, first- and second-grade children in Germany
and America, along with reading tests, the digit span test and a test of RAN color naming
ability. The American kindergarten children excelled on a phoneme identity judgement and a
phoneme deletion task that the German kindergartners found difficult. Their advantage held
equally whether the manipulated sound was a syllable onset or the initial part of a consonant
cluster. The first and second graders surpassed the kindergartners in both countries; however,
the German first and second graders equaled their American peers on both tasks and both types
of units. In addition, the German children were more accurate decoders of pseudowords by the
end of second grade, and the association between phoneme awareness and German decoding
ability was weaker. An increased emphasis on phonics and the greater transparency of the
German alphabet are discussed as possible factors in the decoding excellence of the German
second graders and its decreased association with phoneme awareness. The contrast between
the American and German kindergartners and the equivalence of the first and second graders
in the two countries are consistent with a view that phoneme awareness develops primarily as
a product of literacy exposure.
Key words: Cross-language comparison, Early literacy, Phoneme awareness
Introduction
The discovery of a strong relationship between children’s awareness of
phonemes and their progress in learning to read has been hailed as one
of the great successes of modern psychology (Bryant & Goswami 1987).
In study after study, English-speaking children who are successful begin-
ning readers consistently surpass the poorer readers of their classrooms in
the ability to count, compare, delete or otherwise manipulate phoneme-
sized segments within spoken words. Superior performance on such tests of
phoneme awareness has been a consistent characteristic of kindergartners and
even preschoolers who go on to become good readers in the early elementary
654 VIRGINIA MANN & HEINZ WIMMER
grades (for reviews, see, for example: Adams 1990; Goswami & Bryant
1990; Brady & Shankweiler 1991; Wagner & Torgesen 1987). Prompted by
this evidence that phoneme awareness is both an attribute and precursor of
reading success, we have explored one possible basis of individual differ-
ences in phoneme awareness. Where many studies have compared children
at different reading levels or different ages, we have compared children
from two different cultures, Germany and America. These two cultures
afford different paces with respect to the onset of alphabetic instruction, thus
comparison between them can offer a particularly opportune window into the
relation between early literacy exposure and the development of phoneme
awareness.
To expand upon this point, between kindergarten and the first two years
in school there exists a key difference between the onset of alphabetic
instruction and the pedagogical preferences of the German and Amer-
ican educational systems. In German-speaking countries (Germany, Austria,
Switzerland), the onset of literacy instruction is delayed until children are in
the first grade. The kindergarten year is spent on activities other than letters
and reading such that most children arrive in the first grade (after having
reached the age of six years) without any reading ability and with very limited
(if any) letter knowledge. Once in school, the pace quickens and children are
systematically and progressively led to self-reliant word recognition. Letters
are introduced not by name but by sound and initial reading is little else
than word decoding via sound blending. This synthetic phonics approach is
relatively easy to realize as the letter-sound relations are simple and the whole
writing system is quite consistent.
In contrast, the English-speaking world favors earlier, more gradual and
pedagogically varied instruction. The culture of home literacy and preschool
training is such that the majority of children are already acquainted with
letters and letter names by the time they enter kindergarten. Many kindergart-
ners can even recognize an appreciable number of written words. The kinder-
garten year typically offers many opportunities for learning letter names and
sounds, as well as exposure to ‘pre-reading’ activities that call attention to
the phonological segments of words. On entering the first grade (in many
places, the children need only be six by the first of December) the children
are exposed to an eclectic variety of teaching approaches. The pedagogical
methods range from meaning-based, whole-language approaches to sight-
word, basal approaches to phonics, with most teachers using a mixture. It
is tempting to link the pedagogical uniformity seen in German schools to
the transparency of the German writing system and the variability in Amer-
ican schools to the complex grapheme-phoneme relations embodied in the
relatively ‘deep’ English system.
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 655
The different pathways into literacy that are available to German and
American children can offer a particularly appropriate window into some
underlying assumptions about the emergence of phonemic awareness. If
phoneme awareness is triggered by some type of exposure to an alphabetic
writing system (e.g., Barron 1991, 1994; Morais, Cary, Alegria & Bertelson
1979; Read, Zhang, Nie & Ding 1986), then the different pathways into
literacy should lead to different pathways into phonemic awareness. In partic-
ular, we would expect the low or absent letter and word knowledge of the
German kindergarten children to be reflected in low or absent phonemic
awareness. Such was shown for Serbo-Croatian illiterates with varying levels
of letter knowledge (Lukatela, Carello, Shankweiler & Liberman 1995).
However, an alternative view is that the emergence of phoneme awareness is a
more natural concomitant of language development, reflecting developmental
changes in the organization of the lexicon. Several authors (e.g., Elbro 1996;
Fowler 1991; Walley 1993) have suggested that phoneme awareness arises as
a concomitant of a change from holistic to more explicit, segmental represen-
tations of words. From this view we would expect the phoneme awareness of
German children to be more or less equivalent to American children of the
same age. We may further speculate that equivalence despite widely different
levels of alphabetic instruction should be most evident for simple judgments
as opposed to extensive manipulations of phonemes. Exposure to alphabetic
instruction might also have less of an effect on tasks that involve more shallow
levels of phonological sensitivity, such as those focussed upon syllable onsets
as opposed to phonemes within a consonant cluster (see Stanovich, 1986, for
a discussion of levels of phonological sensitivity).
Before justifying and pursuing these predictions, we want to draw two
contrasts between the early reading skills of English- and German-speaking
children that substantiate our point about there being different pathways
to literacy. The first contrast involves letter knowledge, and the massive
advantage of American and British preschoolers over their German coun-
terparts. To illustrate, in one large sample of kindergarten children (five- and
six-year-olds) in Florida, mean letter name knowledge for upper case letters
was 21/26 (Wagner, Torgesen & Rashotte 1994), another Florida-based study
of four- and five-year-olds in a middle-income sample reports a mean of
20 (Lonigan, Burgess, Anthony & Barker 1998). Among one younger sam-
ple of four-year-olds in Britain, mean letter knowledge was 4.6 (Johnston,
Anderson & Holligan 1996). A recent study conducted in the UK (Muter,
Hulme, Snowling & Taylor 1998) reports letter knowledge of 4.3 at four
years, 12 at five years and 18.6 at six years. In contrast, to the data obtained
from English-speaking children, mean letter knowledge was between three
and six in a sample of German kindergarten children (mainly six-year-olds;
656 VIRGINIA MANN & HEINZ WIMMER
see Schneider & Näslund 1993). For a comparable level of letter knowl-
edge among English speaking children we must turn to the population of
four-year-olds in the UK (Johnston et al. 1996; Muter et al. 1998), or a
low SES population of four- and five-year-olds in America (Lonigan et al.
1998). These children recognized an average of four letters, possibly as a
consequence of being afforded fewer opportunities for formal and informal
instruction (Lonigan et al. 1998). In Germany, decreased letter knowledge
coincides with a delayed onset of formal education, but may also reflect a
decreased emphasis on literacy-related activities in the home and preschool.
For example, German children are not exposed to programs like ‘Sesame
Street’ and ABC books are not widely available (for a discussion of how
such cultural differences contribute to differing phonological awareness skills
among French vs. English children, see Bruck, Genesee & Caravolas 1997).
Whatever the source, formal schooling or home literacy environment, differ-
ences between the letter knowledge of American and German children appear
to be real and it is our intent to determine whether these differences provoke
differences in phoneme awareness.
The second contrast that we wish to underscore concerns decoding
accuracy and the massive advantage of German first graders over their
American and British counterparts. A number of different studies of German-
speaking children at the end of first grade report decoding accuracy scores
of around 90% correct for grade appropriate words and pseudowords alike
(Schneider & Näslund 1993; Wimmer & Hummer 1990; Wimmer, Landerl,
Linortner & Hummer 1991; Wimmer, Mayringer & Landerl 1999). In
contrast, Wagner et al. (1994) report that at the end of first grade the children
in their Florida-based longitudinal sample could only read 7% of the 106
words on the Woodcock Word Identification Test and 7% of the 42 pseudo-
words on the Word Attack Test. At the end of third grade, their subjects’
reading accuracy scores were still only around 50% (Wagner, Torgesen,
Rashotte & Hecht 1997). For their sample of first-grade children in upper
state New York, Scanlon and Vellutino (1997) report mean scores of 35% of
the words on the Word Identification Test and 31% on the Word Attack Test,
still paling in comparison to the scores of the German children. Two studies
have the advantage of using more or less the same words and pseudowords
for assessment of reading ability in English and German. These have also
shown that German-speaking children have an enormous reading accuracy
advantage (particularly for pseudowords) over children in the UK (Frith,
Wimmer & Landerl 1998; Wimmer & Goswami 1994). In discussions of the
lower performance of the English children, these authors call attention to the
more transparent German orthography and to the possibility that the decoding
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 657
difficulties of the English-speaking children are in large part due to their less
extensive phonics-oriented instruction.
Thus there is ample evidence that, where American kindergartners know
more about letters than their German counterparts, and that by first or second
grade, German children are more successful decoders. Before we present our
design for investigating the possibility that these differences may associate
with differences in phoneme awareness, it is important that we discount
possible confounds. Perhaps the most obvious source to consider is the
phonological characteristics of English and German. Phonological influences
have been invoked by Cossu, Shankweiler, Liberman, Katz and Tola (1988)
to explain the somewhat higher syllable and phoneme counting scores of
Italian children compared to the American children originally studied by
Liberman, Shankweiler, Fischer and Carter (1974). However, Cossu et al.
found the largest cross-language differences in the first grade and, there-
fore suggested that the Italian advantage may be due to the more consistent
Italian orthography, above and beyond the simpler syllable and vowel system
of Italian. Language differences are also relevant to Caravolas and Bruck’s
(1993) comparison of the development of phoneme awareness in Czech-
and English-speaking children. They report that the Czech speakers achieved
better performance for onset-clusters than did the English speakers, which
could be related to the extremely high number of different onset clusters in
Czech. We note that the Czech advantage increased between kindergarten and
first grade, which Caravolas and Bruck interpreted as a likely consequence
of the consistent (i.e., shallow) Czech orthography. Language differences,
differences in orthographic transparency and different forms of instruction
are all evoked by Durgonoglu and Oeney (1999) to explain the superior
syllable and final-phoneme deletion of Turkish kindergarten and first-grade
children relative to American children of the same age. For our purposes,
English and German are a more appropriate comparison since they are far
more comparable than English and Italian, English and Czech or English and
Turkish. The phonological characteristics of English and German are quite
comparable due to their common Germanic roots. As a consequence, neither
the prevalence of open vs. closed syllables nor the number of consonant
clusters are confounds. English may have a more complex vowel system than
German, but we can minimize the impact of this difference by limiting our
tasks to the segmentation and judgment of consonants. German has a greater
use of compound words, but we will be using simple words. A further meth-
odological advantage of a German-English comparison is that the occurrence
of common root words allows us to present the same consonants in the same
positions (e.g., ball – ball, bread – brot). The German alphabet is clearly
more transparent than English and this could possibly have consequences
658 VIRGINIA MANN & HEINZ WIMMER
for the relationship between phoneme awareness and the ability to become
a proficient reader. However, from the perspective of the beginning reader,
the phonological structures of the spoken languages are comparable and any
differences in phoneme awareness would be presumed to arise from factors
other than the phonology of the spoken language.
Confounds aside, let us now return to the putative role of literacy in
the development of phoneme awareness. That alphabetic literacy is the
primary determinant of phoneme awareness was first proposed by Morais
et al. (1979), whose research on the phoneme awareness of alphabet-illiterate
adults suggested that the awareness of phonemes is a consequence of learning
to read an alphabetic orthography. This view is consistent with the spurt
in phoneme awareness that occurs when American children begin literacy
instruction (Liberman et al. 1974). It is supported by other research with
alphabet-illiterate adults (Read et al. 1986; Lukatela et al. 1995). It is
further supported by research on the effects of training (for a review see
Torgesen & Davis 1996). That direct training can facilitate phoneme aware-
ness accords with some of the French language research on the effects of
phonics vs. whole-word reading instruction on young children’s performance
on a phoneme reversal task (Alegria, Pignot & Morais 1982). However,
an emphasis on training is at odds with some of the English-language
research (Connelly, Johnston & Thompson 1999; Thompson & Johnston
2000; Tumner & Nesdale 1985). In these studies, while both pseudoword
reading and regularity effects in real words correlated with phoneme dele-
tion and improved with phonics training, phonics training did not have a
significant effect on phoneme deletion.
Barron’s (e.g., 1991, 1994) ‘Proto-literacy Hypothesis’ is a more recent
refinement of the literacy account. It assumes that the critical level of literacy
required for the inducement of phonemic awareness is rather minimal, being
letter knowledge as opposed to decoding ability. In this view, a child who,
for example, learns that a is /ae/ may become aware of /ae/ as an invariant
segment of words yet be unable to recover the sequence of phonemes that
a sequence of letters transcribes. Consistent with this hypothesis, previous
studies with English speaking children found very substantial correlations
(around 0.60) between letter knowledge and phoneme awareness measures in
pre-reading children (Bowey 1994; Johnston et al. 1996). A study of alphabet-
illiterate adults has also shown a direct link between knowledge of letters and
performance on phoneme awareness measures (Lukatela et al. 1995).
Contrasting with the accounts that emphasize the role of literacy are
those recent accounts that link phoneme awareness to changes in lexical
organization. These emphasize that the emergence of phoneme awareness
is an inherent aspect of language development, they propose that phoneme
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 659
awareness emerges when the need to store large numbers of vocabulary
items efficiently and adequately leads to some change in lexical represen-
tation. From word perception and word production studies, Fowler (1991),
for example, inferred that as children learn more and more similarly struc-
tured words, the lexicon changes from a syllable- or word-based format to
a format based on phonemes (e.g., the beginning of the word “bread” may
be re-formatted into two phonemes, when words like “bed” and “red” enter
the lexicon). Fowler further posited that the reorganization towards segmental
representations may well last into the early school years.
Fowler’s observations are echoed in Walley and Metsala’s ‘Lexical
Restructuring Model’ which emphasizes the impact of continued vocabu-
lary growth on spoken word recognition in childhood (e.g., Walley 1993;
Metsala & Walley 1998). This model raises the possibility that phonemes are
not preformed perceptual units whose conscious accessibility changes with
literacy exposure. Instead, it suggests that phonemes gradually develop over
childhood, as the growth of spoken vocabulary causes lexical representations
to become more segmental. With regard to phoneme awareness, Walley and
Metsala (1998, p. 102) state that “the phoneme emerges first at an implicit
level for the perceptual representation and processing of spoken words and
thus only later as a cognitive unit that can be consciously accessed and
manipulated.” They thus imply that lexical restructuring is necessary but not
sufficient for phoneme awareness. While they leave unresolved the factors
or experiences that allow awareness to develop once restructuring has taken
place, they do explicitly deny that reading experience will influence phonemic
awareness (p. 108).
A related but slightly different view of developmental phonological
organization owes to Elbro’s ‘Distinctness Hypothesis’ (Elbro 1996; Elbro
& Peterson 1998), which proposes that the lexicons of young children (and
dyslexic children in particular) may possess relatively indistinct phonological
representations that employ fewer phonetic features than those of mature
language users. In such indistinct representations, the boundaries between
phonemes can be blurred and this can lead to poor performance on phoneme
awareness tests.
Quite simply, a strong emphasis on literacy exposure predicts differences
between children in the two countries; a strong emphasis on lexical reorgan-
ization does not. An emphasis on the role of literacy predicts that different
levels of phonemic awareness should accompany differences in the onset and
pacing of early literacy and a decreased level of phoneme awareness should
accompany the reduced exposure to early literacy among German kinder-
garten children. At the same time, the intense exposure of German first and
second graders to synthetic phonics instruction should ameliorate any lag in
660 VIRGINIA MANN & HEINZ WIMMER
phoneme awareness by the early elementary grades. Indeed, to the extent that
first and second grade children in Germany are superior readers, they may
even excel with respect to first and second graders in America. An emphasis
on the role of lexical reorganization, however, predicts a more equivalent
profile of age-related changes in phoneme awareness performance across
the two countries. Presumably, normally developing German and American
children at the onset of kindergarten would be at a similar stage of vocabulary
development (Clark, 1995, provides data on the comparability of their lexical
creativity). They should be equally poised for reorganizing phonological
representations. There should be age-related changes as children’s lexicons
develop, but the profile of these changes should be more comparable for the
German and American children.
It is a problem for the study of letter knowledge and phoneme awareness
that pre-reading children with high letter knowledge may come from stimu-
lating home and preschool environments and may possess a wide variety of
cognitive-linguistic competencies quite different from those of pre-reading
children with little or no letter knowledge. All of the above-mentioned studies
attempted to circumvent this problem by using control measures, most typi-
cally vocabulary, to assess cognitive-linguistic differences associated with
letter knowledge and phonemic awareness. However, these have met with
mixed results. Burhanpurkar and Barron (1997) and Johnston et al. (1996)
found that the associations between letter knowledge and phonemic aware-
ness were reduced but remained reliable when control measures of vocabulary
were partialled out. Bowey (1994) found that the massive phonemic aware-
ness difference between the high-letter- and the low-letter-knowledge group
was no longer reliable when vocabulary and verbal memory differences were
controlled. For the present cross-cultural comparison of phonemic awareness
in American and German kindergarten children there is little reason to expect
German kindergarten children to be either accelerated or delayed in cognitive
and language development more generally. Nevertheless, to confirm that the
German and American participants are equivalent in basic cognitive skills,
we included within our battery two commonly-used tests that readily lend
themselves to cross-language comparisons: the RAN test for colors (Denckla
& Rudel 1974) and the forward digit span (Wechsler 1967). As we were
aware of no expressive vocabulary test that was available in both English and
German, we opted against testing vocabulary directly, taking note of the fact
that others have shown that digit span to be well correlated with vocabulary
development in this age range (Gathercole, Hitch, Service & Martin 1997;
Gathercole, Service, Hitch, Adams & Martin 1999).
A fair assessment of phoneme awareness in pre-literate children requires
a task format that does not depend upon knowledge of letter sounds or other
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 661
abilities related to reading and writing. Ideally the task should also be similar
to those used in prior research. To these ends we employed two tasks: a
phonological identity judgement task and a phonological deletion task. Each
task focused exclusively upon consonants in utterance initial position. Half
of the items in each task focus upon a single consonant that is also the onset
of a syllable (e.g., ‘ball’ and ‘fire’) where half focus upon a consonant that
is part of a two-consonant cluster (e.g., ‘bread’ and ‘flat’). This was intended
to distinguish onset awareness from phoneme awareness, per se. The child
who performs less well on cluster items than on items that contain single
consonant onsets may be said to be more aware of onsets than of phonemes.
Awareness of the boundary between onsets and rimes is an aspect of phon-
ological awareness which, like phoneme- and syllable-awareness has been
linked to progress in learning to read (see Goswami & Bryant 1990; Treiman
& Zukowski 1996, but also see Muter et al. 1998). It is also regarded as a
lower level of awareness (see Stanovich 1986).
We chose to include a version of the deletion (elision) task because dele-
tion has been used quite widely in the literature. It was employed in studies
of alphabet illiterate adults (for example, Lukatela et al. 1995; Morias et al.
1979) as well as in many studies of beginning readers (see, for example, Mann
1986; Rosner & Simon 1971; Yopp 1988), including several we have cited
(for example, Durgunoglu & Oeney 1999; Elbro & Peterson 1998; Lonigan
et al. 1998; Muter et al. 1998). Unlike phoneme counting (e.g., Liberman et al.
1974), phoneme deletion requires relatively little training and memorization
on the part of the child. Many deletion tests are quite complicated in varying
both the size (syllable, onset, phoneme) and position (initial, medial, final)
of the unit being manipulated (e.g., Lonigan et al. 1998; Rosner & Simon
1971). We kept our task relatively straightforward by focussing only upon
the deletion of segments in initial position. In keeping with the design of our
deletion task, we examined deletion of single consonant onsets as well as
consonants at the beginning of two-consonant clusters. We further avoided
all use of letter names or sounds to minimize the demand on letter knowledge
that has been a confound in other research.
We chose a version of the phoneme identity judgement task, because it
constitutes a superficially simple, but nevertheless valid assessment of phon-
emic awareness as argued by Byrne (1998). In our version of that test, which
we adapted from Byrne and Fielding-Barnsley (1991), the child is given a
target word (e.g., “ball” or “bread”) and then asked to judge which of two
pseudowords (e.g., “baso” or “koso”) starts ‘in the same way’ as the target
word. For correct performance the child has to notice the categorical iden-
tity of the initial segment in the target and the probe word, despite different
articulatory and acoustic realizations of the phoneme. To us, the particularly
662 VIRGINIA MANN & HEINZ WIMMER
attractive feature of the phoneme identity judgement task is its simplicity.
It does not require manipulation of phonological structures as is the case
for the deletion task, it can be easily explained without reference to phon-
emes or sounds like ‘kuh’ or ‘buh’, and it makes relatively little demand on
either working memory or rote memorization. Consistent with this simpli-
city, Bowey (1994) found the phoneme identity task to be much easier for
preschool children than the more widely used oddity detection task. The ease
of the phoneme identity judgement task is importance for the present study as
it allows a fair assessment of phonemic awareness in German children, who
because of their limited letter knowledge may be at a disadvantage in tasks
that require explicit segmentation or manipulation of phonological structures.
To reiterate, we have employed both identity judgement and deletion as
measures of phoneme awareness in a cross-sectional comparison of German
and American children between kindergarten and second grade. The goal is
to probe the relationship between the development of literacy and phoneme
awareness. A primary question is whether earlier exposure to literacy will
lead to an advantage on the part of the American kindergartners. We will
also be asking whether any decoding advantage on the part of German
second graders will be accompanied by an advantage in phoneme awareness
and whether the expected relation between phoneme awareness and reading
ability obtains among first and second graders in each country.
Method
Subjects
The sample consisted of 100 German children (40 kindergartners, 20 first
graders and 20 second graders) and 60 American children (20 kindergartners,
20 first graders, 20 second graders) with approximately equal numbers of
boys and girls in each group at each age The American children attended a
public elementary school in a predominantly Caucasian middle-class, middle
to high SES neighborhood in southern California, and were volunteers who
participated with the written consent of their parents. On a questionnaire
exploring how reading was being taught in their classrooms, all teachers
indicated using a mixture of phonics and whole language methods. Their
questionnaires also indicated that training on the letters of the alphabet, their
names and the specific letter sounds are blended into story telling and singing
activities beginning in kindergarten. The German children were volunteers
who came from three different kindergartens and a public school in middle-
class middle to high SES neighborhoods around Munich and Salzburg. As
already noted, there is no reading instruction and letter teaching in kinder-
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 663
garten. In first grade and second grade, children are exposed to a systematic
synthetic phonics teaching approach that emphasizes letter sound blending as
a word recognition strategy. The larger size of the German population reflects
a desire to include at least some children with higher letter knowledge; an
initial sample of 20 had revealed none.
Tasks
Literacy measures
Letter knowledge. For letter knowledge assessment, children had to read a
page with 22 capital letters randomly sequenced and printed in 3 columns.
(The letters C, Q, X and Y were not used as these are very infrequent in
German.) A response of either the letter name or sound was considered
correct. Both accuracy and response time (in seconds) were scored. Since
recall of letter names or letter sounds may underestimate the letter knowledge
of German kindergarten children, the German children received an additional
letter recognition test in which they were shown the same 22 letters from the
letter reading task distributed over 5 pages with 4 or 5 letters on each page.
For each page children were presented two letter names and asked to point to
the corresponding letters. This resulted in 10 recognition trials. We also asked
children to write their names and identify the letters. If a child could not write
his or her name, the experimenter wrote the name and again asked the child
to name the letters.
Reading. All American children (i.e., those in kindergarten and grades one
and two) received the Word Attack and the Word Identification test of
the Woodcock reading mastery test (Woodcock 1998). To directly compare
decoding ability, we presented all the American children and the German first
and second graders with a decoding test employing real and nonsense words.
All stimuli were written in lowercase 36-point font black ink. Accuracy was
recorded by hand, and reading time was recorded (in seconds). The ‘Real
Word List’ contained common words in each language subdivided into nine
Content and nine Function words (see Appendix A). Each group of nine
words appeared on a single sheet of paper, and reading was timed separ-
ately with the content words given first, followed by the function words. The
Nonsense word list consisted of nine monosyllables and nine bisyllables that
were permissible in either language (see Appendix A). These simple pseudo-
words are phonologically plausible in either of the two languages and contain
neither digraphs nor clusters. Each list was presented on a single sheet of
paper and was timed separately. Both real and nonsense lists presented six
practice words before the actual test items.
664 VIRGINIA MANN & HEINZ WIMMER
For the German kindergarten children, who were expected to have very
limited reading ability, two simplified reading tasks were used to see if any of
the children had connected some letters or letter sounds with word pronun-
ciations without possessing any productive reading ability. As an extremely
simple recoding test, children were presented with two-letter sequences, with
either MA and AM or with SO and OS (depending on the child’s letter
knowledge). AM and SO are function words, MA and OS are pseudowords.
In addition, an environmental print recognition task was designed for the
German children based on 32 sight words (Stop, Taxi, etc.) and logos (Mickey
Mouse, Twix, Coca-Cola, Ford, etc.) from their environment. A response was
counted as correct if the child produced the word represented by the letters.
Control measures
To assure the comparability of the two populations we examined short-term
memory and rapid automatized naming of colors (RAN, Denckla & Rudel
1974) because these two tests readily lend themselves to the comparison of
English- and German-speaking children. Short-term memory was measured
using the WISC-R Forward Digit Span (Wechsler 1967), while RAN was
measured by asking children to name five rows of five colored dots (with each
row presenting a different order of colors). The five colors were the same for
the two language groups and each term consisted of a single closed syllable:
blue – blau, green – grün, red – rot, brown – braun, black – schwarz. The
children were practiced on three rows of these colors. Timing began when
the test sheet was displayed and ended when the child had named the last dot
on the page.
Phoneme awareness measures
Phoneme identity judgement. The materials appear in Appendix B. As
outlined in the Introduction, each item in this test required the child to
choose which of two pseudoword alternatives begins in the same way as the
target word, for example, “Which word begins in the same way as /flute/?
/fidu/ or /lidu/?” To make sure that the children understood this test ques-
tion, we presented a pre-test training using the same question format for
lines of colored dots. After this training, eight practice trials – four with the
target word “cow” (German: “Kuh”) for the onset condition, four with the
target word “cross” (German: “Kreuz”) were given to introduce the phoneme
identity judgment task. Each target word was presented together with a corre-
sponding picture to reduce memory load. On the first practice trial for each
word (“cow” and “cross”), the experimenter demonstrated and explained the
correct choice. On the following practice trials, the child chose and the exper-
imenter reinforced or corrected. The test items were then presented without
feedback starting with the target words for the onset identity judgements –
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 665
“ball” and “fish” (German: “Ball” and “Fisch”) – each with 8 judgements.
The test trials for the phoneme identity judgements with the target words
“bread” and “flute” (German: “Brot” and “Flöte”) followed, each with 8
judgements. The assumption was that children might profit from having the
same phonemes that were first presented as single consonant onset (i.e., “f”
in “fish”) also be presented as the initial segment of a two-consonant cluster
onset (i.e., “f” in “flute”). The two target words that started with the same
sound in the onset and the phoneme identity condition also involved similar
pseudoword pairs. The full list of items appears in the Appendix.
Phoneme deletion. The materials for the Phoneme Deletion Task also appear
in Appendix B and consisted of 12 words each to be repeated without the
first ‘sound’. (E.g., “Fire. Fire without the first sound becomes
?”) As
a probe to the deletion of onsets vs. phonemes, six of the items started with
a single consonant onset, and six started with a two-consonant cluster. As a
training device, we again presented a visual analogy of the task using colored
dots: four cardboard dots of different colors were grouped in a horizontal
line, and children were asked what colors were left once the first dot was
removed from the array. This training was followed by eight practice trials
with feedback that included four single consonant items and four items with
two-consonant clusters. For the 12 test words that followed, no feedback was
given. All test words appear in the Appendix.
Procedure
All testing was done between March and July of the academic year. Parti-
cipants were tested individually in a quiet room in their schools. The RAN
test was administered first, followed by the phoneme identity judgement task,
the phoneme deletion test, the letter knowledge test, the reading test and the
Digit Span. All instructions for all tests were carefully matched between the
two languages.
Results
Table 1 lists the means and standard deviations for raw scores on each task,
separately for the American and German children at each age. For ease in
comparing the two groups, we begin with a discussion of the control measures
before proceeding to the measures of phonological awareness and reading
ability.
666 VIRGINIA MANN & HEINZ WIMMER
Table 1. Means, standard deviations of the English- vs. German-speaking children
on reading and general cognitive measures
English German
Grade level: K 1 2 K 1 2
Age (months): 74 88 98 75 89 101
Reading
Letter identification
Percent correct 94.1 98.0 97.5 28.8 100.0 100.0
SD: 4.7 2.7 2.7
Time (seconds) 18.7 15.1 13.3 16.5 13.4
SD: 5.0 3.0 2.5 4.5 3.1
Real word decoding
Percent correct 51.1 94.2 99.7 95.8 99.4
SD: 44.2 12.9 1.2 6.7 1.7
Time (seconds) 12.7 8.0 5.5 10.8 5.3
SD: 12.2 5.8 1.7 7.3 1.5
Woodcock word ID 22.8 45.0 58.6 na na
SD: 21.2 14.5 10.0
Nonsense word decoding
Percent correct 49.7 80.1 85.0 90.8 92.8
SD: 39.4 25.5 14.8 6.1 3.6
Time (seconds) 27.6 17.2 15.8 13.6 13.4
SD: 15.2 9.5 10.6 7.6 3.0
Woodcock word attack 9.3 17.9 25.1 na na
SD: 10.4 8.6 7.0
General cognitive
Digit span 4.4 6.3 5.6 5.4 6.4 7.6
SD: 0.9 2.2 1.3 1.5 1.4 1.6
RAN (seconds) 28.4 21.6 21.8 35.5 25.4 22.3
SD: 17.4 5.7 4.6 11.8 4.5 3.3
Control measures
The means for the control measures suggest that the German children have
an advantage on the digit span test (F
1,134
= 5.74, p < 0.018), which inter-
acts with grade (F
2,134
= 4.87). Age differences between the German and
American children do not underlie this result; covaried for age, there was still
a significant effect of country, (F
1,33
= 6.44, p < 0.012), and a country by
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 667
Table 2. Means, standard deviations of the English- vs. German-speaking
children on measures of phoneme awareness
English German
Grade level: K 1 2 K 1 2
Phoneme Awareness
Phoneme judgment 91.6 98.7 98.4 68.5 95.9 99.5
SD: 13.2 2.7 1.9 15.7 7.6 1.5
Single consonant 92.2 99.7 99.7 71.1 96.9 99.4
SD: 14.9 1.4 1.4 17.6 8.7 1.9
Consonant cluster 90.0 97.8 96.9 65.9 95.0 99.7
SD: 13.4 4.7 3.8 18.1 7.2 1.4
Phoneme deletion 48.3 74.2 82.1 17.5 68.7 82.5
SD: 35.8 21.1 14.6 25.8 29.0 21.1
Single consonant 59.2 84.2 85.5 17.1 70.8 90.8
SD: 43.3 23.2 13.5 26.8 38.6 14.8
Consonant cluster 37.5 64.2 78.3 18.8 66.7 74.2
SD: 39.7 32.1 18.8 26.5 34.2 38.4
grade interaction, (F
2,133
= 5.13, p < 0.007). Looking at Table 2, the Amer-
ican kindergarteners and second graders have lower average scores than their
German counterparts while the first graders are equivalent. The variance is
also higher for the American second graders suggesting that there could have
been a sampling problem. On the rapid color-naming task, children made
virtually no errors hence we have concentrated on the speed of response.
There was a slight advantage of the American children (F
1,134
= 4.04, p <
0.046), which persists when age differences are controlled, (F
1,134
= 4.98,
p < 0.027). This did not interact with age.
Literacy
Kindergarten
The mean percentages that appear in Table 1 confirm the expected advantage
of the American over the German kindergarten children with regard to letter
knowledge, (F
1,59
= 16,42, p < 0.000). There was hardly any overlap between
the two groups: All 20 American kindergarten children read more than 18 of
the 22 letters correctly, whereas only 3 of the 40 German child did so. Half
of the German kindergarten children read fewer than four letters, with six
children unable to read any letters at all. The German kindergarten children
also did poorly on the simplified letter recognition test. The mean recognition
668 VIRGINIA MANN & HEINZ WIMMER
score was 69% (guessing rate: 50%). The majority of the children (i.e., 55%)
made fewer than 4 correct responses on the 10 letter recognition trials and
can be considered to have been guessing. Interestingly, 85% of the German
kindergarten children could write their first name correctly, but most often
could identify only a few of the letters in their names. The mean for correctly
recognized letters in the children’s names was 2.5, the mean number of letters
in each name was 6.3.
From the mean scores for Word Identification and Word Attack in Table 1
it is evident that a substantial number of the American kindergarten children
had acquired some level of reading ability. Thirteen of the 20 American
kindergartners (i.e., 65%) could read at least one item correctly on the Word
Attack test. In contrast, only 15 of the 40 German kindergartners (i.e., 37%)
could read one of the two two-letter sequences (MA and AM or SO and OS).
On the ‘environmental print’ test the German kindergarten children provided
the correct label to only 13% of the logos and sight words.
First and second graders
On the letter identification task, we found a significant main effect of country
on accuracy for first and second graders combined (Mann–Whitney Uz =
4.774, p < 0.000, owing to the ceiling performance of the German children),
and a main effect of grade on letter identification speed (F
1,76
= 11.2, p <
0.001). The German children achieved perfect scores, while the American
children made a few errors (<3%); the main effect of grade comes from
the trend for second graders to be faster than first graders in general. The
tests of word and pseudoword decoding ability were analyzed separately with
2-factor ANOVAS that considered country and age. These indicated effects
of age and effects of country. As would be expected, the second graders
were significantly more accurate than the first graders on real words (F
1,76
=
7.75, p < 0.007); but they showed no advantage in accuracy for pseudo-
word reading (p > 0.1) The reading speed of second graders was faster for
both words (F
1,76
= 13.8, p < 0.004) and pseudowords (F
1,75
=3.2,p <
0.077). While the German children showed no significant advantage in either
accuracy or speed of reading words (p > 0.1), perhaps owing to a ceiling level
of performance on this test (but the test did at least show effects of age), they
did show a strong advantage for both accuracy (F
1,76
= 18.3, p < 0.000),
and time of pseudoword reading (F
1,76
=8.9,p < 0.004). In light of the
greater variance for the American children we also computed nonparametric
analysis, which confirmed the German advantage for pseudoword accuracy
(Mann–Whitney Uz = 5.788, p < 0.000) and reading time (Mann–Whitney
Uz = 4.608, p < 0.000).
On the control measures of digit span and automatized color naming, it
will be remembered that the German-speaking children excelled on the digit
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 669
span and the American children on the RAN. We thus turned to an analysis
of covariance to test the possibility that digit span or RAN differences might
underlie these differences in reading ability. Both Digit Span (effect size =
0.08; F
1,74
=6.3,p < 0.02) and RAN time (effect size = 0.07; F
1,74
=5.4,p <
0.03) related to reading accuracy in the German-speaking children but not
the American ones (p > 0.1). This pattern of correlation, together with the
German advantage on digit span and disadvantage on RAN makes it unlikely
that their reading advantage was due to cognitive ability. Consistent with
this line of reasoning, the German advantage in both pseudoword accuracy
(F
1,73
= 10.5, p < 0.002) and speed (F
1,73
=6.4,p < 0.014) was confirmed
in an analysis of covariance which control for differences in age, RAN time
and digit span.
In summary, the children from the two cultures conformed to expectation.
The American kindergarten children showed consistently higher letter knowl-
edge and a substantially greater number of them had acquired a beginning
level of reading ability. In contrast, the majority of the German kindergarten
children exhibited very low letter knowledge and few could read. By the first
and second grades, the German-speaking children possessed letter identifi-
cation skills superior to those of their American counterparts (though both
groups performed at high levels). As first and second graders, the German
children also tended to be superior decoders of pseudowords.
Phoneme awareness
On the phonological awareness tasks, older children performed better, in
general and American children averaged higher scores, in general. However,
once the kindergartners’ data were removed from the ANOVA, the perfor-
mance of the first and second grade children was virtually the same and
any effect of grade was limited to the deletion task (F
1,176
=4.8,p <
0.031). The kindergartners, then, are the primary source of any differences
within and between populations. From the means of the combined judgement
scores, it is obvious that the American children massively outperformed their
German peers, and our analysis will focus upon this result. As the single
consonant onset items were easier than the consonant-cluster items for judge-
ment (F
1,159
= 11.1, p < 0.001) and deletion (F
1,159
= 12.6, p < 0.001), we
have considered type of item as a repeated factor in the analyses that follow,
except where noted.
Kindergarten
ANOVAS considered country and the single vs. cluster items in each test.
On the phoneme judgement task, it is clear from Table 2 that the Amer-
ican children strongly outperformed their German peers, averaging 92%
670 VIRGINIA MANN & HEINZ WIMMER
correct compared to 69% for the German children (F
1,58
= 29.5, p < 0.000).
Moreover, the advantage of the American children held equally for the single
phoneme and consonant cluster items and there was no interaction (p > 0.1).
A general advantage for single phoneme onsets falls just short of significance
(F
1,58
=3.6,p < 0.06) . It is also apparent in Table 2 that, while scores on
the deletion task were lower, in general, the American kindergartners had a
strong advantage, achieving 48% correct as compared to 18% for the German
children. For the deletion task, there was a main effect of country (F
1,58
=
17.1, p < 0.000, one-way ANOVA), a significant advantage for the items
with single phoneme onsets (F
1,58
=5.7,p < 0.021), and a slight interaction
between country and type of item (F
1,58
=4.2,p < 0.046). Post hoc t-tests
indicated that the scores of the American children were significantly greater
for both single (t
26.3
= 3.82, p < 0.001), and cluster items (t
45.2
= 4.85, p <
0.000).
First and second graders
Among children in this age range there were no significant differences in
either judgement (F
1,79
= 0.24, p > 0.9) or deletion performance (F
1,79
=
1.8, p > 0.08). Children performed at roughly the same level independent
of grade level or country of education. To better understand the relationship
between reading and phonological awareness we turn to linear models of
reading ability.
Separate step-wise multiple regressions of the German and American
children’s reading performance (the sum of z-transformed scores on the word
and pseudoword reading tests) were conducted using phonological awareness
(the sum of z-transformed scores on the phoneme judgment and phoneme
deletion tests), Digit Span, and RAN z-scores as predictors of variance in
reading speed and accuracy. Data were combined across the two grades for
the sake of simplicity. Table 3 lists the weights Beta scores of each variable
and the amount of variance accounted for in each country. Most notice-
able was the difference in percent of decoding variance accounted for in
the two countries. The English-speaking children’s decoding accuracy was
quite robustly accounted for by the combined variables (r
2
English
= 0.70, p <
0.01), yet of the three variables, only phonological awareness significantly
accounted for reading (Beta = 0.79, p < 0.01). In contrast, the reading
accuracy of the German children was not as strongly accounted for by the
three variables (r
2
German
= 0.27, p < 0.01) and no single variable made a signifi-
cant contribution. In regard to reading speed, there was once again a greater
degree of variance accounted for in the case of the American children’s
performance (r
2
English
= 0.49, p < 0.01) as compared to the German children
(r
2
German
= 0.29, p < 0.01). For the American children, phonological awareness
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 671
Table 3. Prediction of reading ability by phonological awareness, Digit Span and RAN
Accuracy Beta r
2
English: 0.70
∗∗
F(3,56) = 44.3
German: 0.27
∗∗
F(3,36) = 4.5
Phonological awareness
English: 0.79
∗∗
t(59) = 8.0
German: 0.29 t(39) = 2.0
Digit span
English: 0.05 t(59) = 0.7
German: 0.29 t(39) = 2.0
RAN time
English: –0.04 t(59) = –0.5
German: –0.24 t(39) = –1.7
Speed Beta r
2
English: 0.49
∗∗
F(3,48) = 15.4
German: 0.29
∗∗
F(3,36) = 5.0
Phonological awareness
English: –0.64
∗∗
t(51) = –5.6
German: –0.14 t(39) = –1.0
Digit span
English: 0.00 t(51) = –0.01
German: –0.24 t(39) = –1.7
RAN time
English: –0.15 t(59) = 1.4
German: 0.42
∗∗
t(39) = 3.0
∗
p < 0.05;
∗∗
p < 0.01.
was the only independently significant variable (Beta = –0.64, p < 0.01).
for the German children, RAN speed was the only independently significant
variable (Beta = 0.42, p < 0.01).
Discussion
Before discussing the implications of the present research for theories about
the emergence of phoneme awareness and its relation to literacy and phono-
logical reorganization, we offer a brief summary of the critical findings and
how they relate to data from comparable studies.
First, we found the expected differences in the pacing of literacy educa-
tion in the two countries. On the one hand, the American kindergarten
children showed an earlier progression into literacy. Their letter knowledge
672 VIRGINIA MANN & HEINZ WIMMER
was close to perfect and some of them showed substantial reading ability
as well. In contrast, the large majority of German kindergartners knew few
letters, and could neither decode the simplest possible letter sequence nor
name a substantial number of environmental print patterns. At the same
time, the German first and second graders gave evidence of being more
successful decoders. By the end of first grade, the German children exhibited
more perfect letter knowledge than their American counterparts and superior
pseudoword decoding as well.
Second, we have found that an advance in phoneme awareness accom-
panied the earlier literacy of the America kindergarten children. This is
indicated by their high performance on both the phoneme identity judgement
task and the phoneme deletion task. On the judgement task, it is particularly
impressive to note the contrast between the near perfect performance of the
American kindergarten children and the near chance-level performance of the
German kindergarten children. The advantage of the American kindergarten
children held for two different tasks, it held whether the focus was upon
syllable onsets or upon phonemes, per se. Thus the advantage holds for two
levels of awareness and for two different tasks.
Third, despite the strong advantage of the American kindergartners, our
testing of older children indicates that, by that time they are completing the
first grade, the German children are able to perform phoneme awareness
tests as well as the American children. Every first and second grader scored
near ceiling on the judgement task whether or not they came from a culture
that had stressed literacy exposure before the first grade. On the deletion
task, performance is far from ceiling yet the German and American children
are comparable. There is room for differences in phonological awareness
to emerge (e.g., second graders surpass first graders), but the country-based
differences are not there.
Fourth, although they do not appear to be more aware of phonemes, there
is some indication that the German children are superior decoders of pseudo-
words by the end of first grade. We do not replicate previous findings about
a German advantage for either the speed or accuracy of reading of words
(e.g., Schneider & Naslund 1993; Wimmer & Hummer 1990; Wimmer et
al. 1991, 1999), perhaps because the words in our real word test were too
easy. We only show that German first and second graders have an advantage
for the reading of pseudowords, both accuracy and speed are superior. We
might attribute this advantage for pseudoword decoding to the intensity of
the German phonics curriculum, given the advantages that Alegria et al.
(1982) observed when they compared phonics and basal approaches. But this
would seem at odds with Thompson and Johnston’s (2000) recent finding
that phonics instruction did not improve performance on a phoneme dele-
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 673
tion task. These authors suggest that an advantage for pseudoword reading
in the face of equivalent phoneme awareness could reflect a superior ability
to use lexical knowledge as a source of phonological recording. This view is
compatible with our results if we attribute the greater lexical knowledge to
the increased regularity of the German orthography. For example, the more
transparent German orthography may help to explain our observation that,
although the first and second grade German children were equivalent to the
American children in performance on phoneme awareness tests, phoneme
awareness was more strongly associated with decoding in the case of our
American population (as it was in Naslund, Schneider & Van Den Broek
1997). A more regular orthography German may somehow ease the stress on
individual differences phoneme awareness (for a similar view, see Frith et al.
1998).
Our primary result concerns phoneme awareness and the strong advantage
of the American kindergarten children who also had greater knowledge
of letters and greater ability to read. It is interesting to note how closely
our results regarding the phoneme awareness tests and letter naming scores
resemble those of other researchers. We may first compare our kindergarten
results to those of Muter et al. (1998) who observed British preschoolers
over a three-year period and used a deletion task very similar to the one we
employed. At age 6.3 years, the letter naming (72% accuracy for the entire
set of letters) and initial phoneme deletion scores (52% accuracy for single
phoneme deletion) of the Muter et al population are virtually identical to
the data of our American kindergartners. Like the American kindergartners,
the British children tested at this time had completed one year of literacy
instruction. At age 5.3 years, the scores of the British children (42% accuracy
for letters, 25% accuracy for phoneme deletion) are roughly equivalent to
those of our German kindergartners. Like those kindergartners, the five-year-
olds were tested before formal literacy instruction began. Elbro and Peterson
(1998), who had used the same type of deletion task to examine Danish
kindergartners at the beginning of formal instruction, report similar values
of 45% correct for letter naming and 16% for phoneme deletion.
Our kindergarten and first-grade results may also be compared to Duron-
golu and Oeney (1999), who compared the syllable and phoneme deletion
skills of American and Turkish children. They report letter accuracy scores
(capital letters) of 61% for Turkish kindergartners as compared to 96%
for American kindergartners and 99–100% accuracy for all first graders.
Phoneme deletion scores for the Turkish children were 43% in kindergarten
and 93% in first grade, where scores for American children were 43% in
kindergarten and 80% in first grade. As in America, Turkish children are
674 VIRGINIA MANN & HEINZ WIMMER
exposed to literacy instruction beginning in kindergarten, as in America,
phoneme deletion scores are superior.
We may compare our first and second grade results to Naslund et al.
(1997), who tested German and American children in the first and second
grades. Like them, we see equivalence between children in the two countries
on the phoneme deletion task. Their American participants had scored 74%
correct as first graders and 86% as second graders; their German children
had scored 63% as first graders and 89% as second graders. These levels
of performance are virtually identical to the levels of performance that we
observed. In their study as well, the German children were faster pseudoword
decoders and they also report that, while phonemic tasks were related to
reading performance in both countries, the relationship was stronger for the
American children.
Turning to a principal concern, which is the relationship between phoneme
awareness and the onset of literacy exposure, we note that poor phoneme
awareness performance among German kindergarten children has been
observed before, in other samples (e.g., Schneider, Roth & Ennemoser 2000;
Wimmer et al. 1991). However, in these other studies more traditional
phoneme awareness tasks were used, which require explicit segmentation
or manipulation of phonological forms. We reasoned that these tasks may
depend on letter knowledge and reading and writing competencies and that
the present phoneme identity judgement task could assess phoneme aware-
ness in a way that is less dependent on early literacy. Nevertheless the German
kindergarten children achieved lower scores on both the phoneme identity
judgement task and the more traditional judgement task. The task demands
on the judgement task were simple and the children were carefully intro-
duced to the task. The German kindergarten children had no difficulty with
the nonverbal pre-training items and appeared to understand the test question
which required only that they match words on the basis of their beginnings
(e.g., “Which word begins in the same way as /fish/? /fidu/ or /lidu/?”) The
saliency of the segments was kept high by having the judgement involve the
beginnings of the to-be-matched words and for half of the items (e.g., in /fish/)
the judgement involved the full syllable onset. We made efforts to employ
a perceptual matching task that should have been particularly easy, yet the
German children nonetheless erred.
The widespread failure of the German kindergarten children to achieve
above chance performance on the phoneme identity judgement task has two
implications. One is that the task apparently demands more than simple
perceptual matching. The most plausible interpretation is that appreciating the
identity of the segments at the beginning of /fish/ and /fidu/ or of /flute/ and
/fidu/ requires that the child possess a conception of phonological structure
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 675
and that the low-literacy German kindergarten children had not yet acquired
such a conception at the level of their American counterparts.
From this first implication follows the second one, namely that the poor
performance of the German kindergarten children on the phoneme identity
judgement task speaks against a strong version of the phonological reor-
ganization hypothesis. As explicated in the Introduction, this hypothesis
assumes that phoneme awareness arises due to a change from syllable-based
to phoneme-based word representations, a change that is triggered by the
demands of increasing vocabulary size. Since the present German kinder-
garten children were about 6 years old, the major spurt in vocabulary growth
must be well under way and with it the change towards phoneme-based phon-
ological representations. Their digit span scores, which we expect to correlate
with vocabulary scores (Gathercole et al. 1997, 1999), are actually superior to
those of the American kindergartners. If vocabulary growth leads to phoneme
awareness then the present German kindergarten children should have exper-
ienced no more difficulty with the phoneme identity judgement tasks than
their American counterparts. Obviously, this was not the case. Neither on
phonemes that comprise the onset of a syllable nor on phonemes at the onset
of a cluster did the German kindergartners equal the performance of their
American counterparts. Neither our manipulation of task difficulty nor of
level of phonological structure altered the difference between kindergartners
in the two countries.
The present findings, at the group level, are most consistent with the
literacy hypothesis on phoneme awareness: American kindergartners show
an advantage but at the end of first grade, both the German and the American
children showed more or less perfect phoneme judgement performance. In
general, the kindergarten results correspond with the proto-literacy hypoth-
esis of Barron (1991), in showing that the critical level of literacy for the
inducement of phonemic awareness may be rather low. The American kinder-
garten children could read only half of the words and pseudowords yet
still achieved a significantly higher level of phoneme awareness. However,
we would not claim strong support for a view that letter knowledge and
not reading ability is the critical aspect of literacy for the initial stages of
phoneme awareness. For our American population, we actually noted that
both phoneme judgment and phoneme deletion were slightly more correlated
with reading ability on the Woodcock tests (pearson correlations between
0.48 and 0.80) than with letter identification (correlations between 0.42 and
0.43), analogous to Bowey (1994). Nor would we claim that either knowledge
of the alphabet or decoding skill is the sole basis of individual differences
in phoneme awareness. For example, one German child without any letter
knowledge or reading ability, exhibited perfect performance on the phoneme
676 VIRGINIA MANN & HEINZ WIMMER
judgement task. Such variation has been noted by Mann (1986) who found
some Japanese children surprisingly adept at phoneme manipulation tasks
despite a lack of alphabetic literacy or a Kana-based strategy.
The general point following from the support for the literacy hypothesis is
that, for most children, a conception of phonemes as perceptual-categorical
segments of speech arises from something above and beyond the experiences
that support primary language development. It is most often a consequence
of some extra stimulation. As stressed by the literacy hypothesis, this extra
stimulation typically comes from learning to read an alphabetic writing
system, which makes a conception of phonemes mandatory. Evidence to this
effect can be seen in Mann’s (1986, Experiment 3) finding that American
first graders surpass their Japanese peers in performance on both phoneme
counting and phoneme deletion tasks. Of course, the extra stimulation could
also consist of phoneme awareness training without letters, as is provided in
some forms of language play, for example. Our point is that, for the majority
of children, phoneme awareness must be triggered by something above and
beyond the experiences that are sufficient to support primary language devel-
opment, otherwise we are at loss to explain the poor performance of the
German kindergartners.
In the Introduction we pointed out that one major strength of Fowler’s
(1991) phonological reorganization hypothesis and Elbro’s (1996) ‘Distinct-
ness Hypothesis’ is the natural way they integrate findings about the phoneme
awareness deficits of dyslexic children with findings about dyslexic children’s
more general language problems. A conclusion that a metacognitive concep-
tion of phonemes arises from some stimulation above and beyond the
requirements of spoken language development misses this integrative feature
and is in need of explaining the dyslexia findings. Here it becomes important
to distinguish between what triggers a metalinguistic awareness of phon-
emes and what is responsible for differences in the ease with which such
an awareness is induced. Our data indicate that the initial push towards a
conception of phonemes comes from outside of spoken language develop-
ment (typically from alphabetic literacy). Yet we would follow the spirit of
Elbro, Fowler, Walley and their colleagues in recognizing that that the child’s
response to such a push is nevertheless dependent on the state of phonological
representations. The lexical and phonological restructurings that they speak
of may be essential for making phoneme-sized segments available for the
inducement of phoneme awareness. If the initial contact with literacy (e.g.,
learning some letter sounds) meets a system of advanced, phoneme-based
representations, then phoneme awareness may quickly follow. In contrast, no
phoneme awareness may be induced when letter-sound information, word
play, etc. hits a system of largely syllable-based or otherwise impoverished
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 677
representations. Other theoretical accounts have been offered to explain how
the language system responds to alphabetic literacy (for example, Ehri &
Wilce 1980). It is beyond the scope of our work to evaluate their success.
We merely suggest that while the inducement of phoneme awareness requires
more than the experience of spoken language development, the maturational
status of phonological representations may also be critical. Our empirical
contribution has been to show that aspects of literacy experience do indeed
influence phonemic awareness in normally developing children whose ages
suggest that restructuring has taken place.
Acknowledgements
Partial financial support was provided by the Austrian Science Foundation
(Grant No. P12481-SOZ). We would like to thank Maria Singson, Gina
Rappeleye, Sheela Shah, Petra Patzke and Sabine Hölzl for help in data
collection and/or analysis. We thank the elementary schools who allowed us
to work with their pupils and we also thank our anonymous reviewers for
their very constructive comments.
Appendix A
Word decoding test
Content words
English: German:
one kind (child)
high brot (bread)
home oma (grandma)
school maus (mouse)
boy auto (car)
bird mutter (mother)
play esel (donkey)
yellow wasser (water)
birthday torte (tart)
Function words
English: German:
my der
them mit
about bis
678 VIRGINIA MANN & HEINZ WIMMER
all und
your was
when du
how sind
out es
many wer
Nonsense words (bilingual)
Monosyllabic Disyllabic
ol ebol
mu abi
tis rela
ak awit
su edu
kep soti
ip orus
ko uki
dus mipu
Appendix B
Phoneme identity judgement task
Single consonant onset items
Target: ball (German: Ball) Target: fish (German: Fisch)
Items: pesa-besa Items: fime-pime
bopi-wopi seimo-feimo
dilu-bilu foka-schoka
bako-mako fidu-lidu
bafo-safo fetu-wetu
liwa-biwa gimo-fimo
schomu-bomu nalo-falo
bemo-remo dosa-fosa
Phoneme cluster onset items
Target: bread (German:Brot) Target: flute (German: Flöte)
Items: pase-base Items: pemi-femi
wipo-bipo fomei-somei
buli-duli schako-fako
moka-boka ludi-fudi
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 679
sofa-bofa wute-fute
bawi-lawi fomi-gomi
bumo-schumo fola-nola
rome-bome faso-daso
Phoneme deletion task
Single phoneme onsets: Phoneme cluster onsets:
Items: dolphin/dolphin Items: dragon/drache
fire/feuer friend/freund
foot/fuss flat/flach
garden/garten glacier/gletscher
goose/gans green/grun
thief/dieb three/drei
References
Adams, M.J. (1990). Beginning to read: Thinking and learning about print. Cambridge: MIT
Press.
Alegria, J. Pignot, E. & Morais, J. (1982). Phonetic analysis of speech and memory codes in
beginning readers, Memory & Cognition 10: 451–456.
Barron, R.W. (1991). Proto-literacy, literacy, and the acquisition of phonological awareness,
Learning and Individual Differences 3: 243–255.
Barron, R.W. (1994). The sound-to-spelling connection: Orthographic activation in auditory
word recognition and its implications for the acquisition of phonological awareness and
literacy skills. In V.W. Berninger (ed.), The varieties of orthographic knowledge. I:
Theoretical and developmental issues (pp. 219–242). Dordrecht, The Netherlands: Kluwer
Academic Publishers.
Brady, S. & Shankweiler, D. (1991). Phonological Processes in Literacy. Hillsdale, New
Jersey: Erlbaum.
Bowey, J.A. (1994). Phonological sensitivity in novice readers and nonreaders, Journal of
Experimental Child Psychology 58: 134–159.
Bruck, M., Genesee, F. & Caravolas (1997). A cross-linguistic study of early literacy acquisi-
tion. In B. Blachman (ed.), Dyslexia and its treatment (pp. 45–62). Mahwah, New Jersey:
Erlbaum.
Bryant, P. & Goswami, U. (1987). Beyond grapheme-phoneme correspondence, Cahiers de
Psychologie Cognitive/Current Psychology of Cognition 7: 439–443.
Burhanpurkar, A. & Barron, R.W. (1997). Origins of phonological awareness in pre-readers:
Roles of language, memory and proto-literacy. Paper presented at the Society for Research
in Child Development, Washington, DC.
Byrne, B. (1998). The foundation of literacy: The child’s acquisition of the alphabetic
principle. Hove, UK: Psychology Press.
Byrne, B. & Fielding-Barnsley, R. (1991). Evaluation of a program to teach phonemic
awareness to young children, Journal of Educational Psychology 83: 451–455.
680 VIRGINIA MANN & HEINZ WIMMER
Caravolas, M. & Bruck, M. (1993). The effect of oral and written language input on
children’s phonological awareness: a cross-linguistic study, Journal of Experimental Child
Psychology 55: 1–30.
Connelly, V., Johnston, R.S. & Thompson (1999). The influence of instructional approaches on
reading procedures. In G.B. Thompson & T. Nicholson (eds.), Learning to read: Beyond
phonics and whole language (pp. 103–123). New York: Teachers College Press.
Clark, E.V. (1995). The young word maker: A case study of innovation in the child’s lexicon.
In E. Wanner and L.R. Gleitman (eds.), Language acquisition, the state of the art (pp. 390–
425). Cambridge: Cambridge University Press.
Cossu, G., Shankweiler, D., Liberman, I.Y., Katz, L. & Tola, G. (1988). Awareness of phon-
ological segments and reading ability in Italian children, Applied Psycholinguistics 9:
1–16.
Denckla, M.B. & Rudel, R. (1974). Rapid “automatized” naming of pictured objects, colors,
letters and numbers by normal children, Cortex 10: 186–202.
Durgunoglu, A.Y. & Oeney, B. (1999). A cross-linguistic comparison of phonological
awareness and word recognition, Reading & Writing 11: 281–299.
Ehri, L.C. & Wilce, L.S. (1980). The influence of orthography on reader’s conceptualization
of the phonemic structure of word, Applied Psycholinguistics 1: 371–385.
Elbro, C. (1996). Early linguistic abilities and reading development: A review and a
hypothesis, Reading & Writing 8: 453–485.
Elbro, C. & Peterson, D.K. (1998). Predicting dyslexia from kindergarten: The importance of
distinctness of phonological representations of lexical item, Reading Research Quarterly
33: 36–60.
Fowler, A.E. (1991). How early phonological development might set the stage for phoneme
awareness. In S. Brady & D. Shankweiler (eds.), Phonological processes in literacy: A
tribute to Isabelle Y. Liberman (pp. 97–117). Hillsdale, New Jersey: Erlbaum.
Frith, U., Wimmer, H. & Landerl, K. (1998). Differences in phonological recoding in German-
and English-speaking children, Scientific Studies of Reading 2: 31–54.
Gathercole, S.E., Hitch, G.J., Service, E. & Martin, A.J. (1997). Phonological short-term
memory and new word learning in children, Developmental Psychology 33: 966–979.
Gathercole, S.E., Service, E., Hitch, G.J., Adams, A. & Martin, A.J. (1999). Phonological
short-term memory and vocabulary development: Further evidence on the nature of the
relationship, Applied Cognitive Psychology 13: 65–77.
Goswami, U. & Bryant, P. (1990). Phonological skills and learning to read. London: Erlbaum.
Johnston, R.S., Anderson, M. & Holligan, C. (1996). Knowledge of the alphabet and explicit
awareness of phonemes in pre-readers: The nature of the relationship, Reading and Writing
8: 217–234.
Liberman, I.Y., Shankweiler, S., Fischer, F.W. & Carter, B. (1974). Explicit syllable and
phoneme segmentation in the young child, Journal of Experimental Child Psychology 18:
201–212.
Lonigan, C.J., Burgess, S.R., Anthony, J.J. & Barker, T.A. (1998). Development of phono-
logical sensitivity in 2- to 5-year old children, Journal of Educational Psychology 90:
294–311.
Lukatela, K., Carello, C., Shankweiler, D. & Liberman, I.Y. (1995). Phonological awareness
in illiterates: Observations from Serbo-Croatian, Applied Psycholinguistics 16: 463–487.
Mann, V.A. (1986). Phonological awareness: The role of reading experience, Cognition 24:
65–92. [Also published under the same title in: P. Bertelson (ed.) (1987) The onset of
literacy. Cambridge: MIT Press.]
PHONEME AWARENESS AND PATHWAYS INTO LITERACY 681
Metsala, J.L. & Walley, A.C. (1998). Spoken vocabulary growth and the segmental restruc-
turing of lexical representations: Precursors to phonemic awareness and early reading
ability. In J.L. Metsala and L.C. Ehri (eds.), Word recognition in beginning literacy
(pp. 89–120). Mahwah, New Jersey: Erlbaum.
Morais, J., Cary, L., Alegria, J. & Bertelson, P. (1979). Does awareness of speech as a sequence
of phones arise spontaneously?, Cognition 7: 323–331.
Muter, V., Hulme, C., Snowling, M. & Taylor, S. (1998). Segmentation, not rhyming, predicts
early progress in learning to read, Journal of Experimental Child Psychology 71: 3–27.
Naslund, J.C., Schneider, W. & Van Den Broek, P. (1997). Beginning reading in Germany
and the US: A comparison of phonological segmentation, decoding, lexical access and
comprehension. In C.K. Leong & R.M. Joshi (eds.), Cross-language studies of learning to
read and spell (pp. 103–120). Dordrecht, The Netherlands: Kluwer Academic Publishers.
Read, C., Zhang, Y., Nie, H. & Ding, B. (1986). The ability to manipulate speech sounds
depends on knowing alphabetic writing, Cognition 24: 31–44.
Rosner, J. & Simon, D.P. (1971). The auditory analysis test: An initial report, Journal of
Learning Disabilities 4: 384–392.
Schneider, W. & Näslund, J.C. (1993). The impact of early metalinguistic competencies and
memory capacity on reading and spelling in elementary school: Results of the Munich
Longitudinal Study on the Genesis of Individual Competencies (LOGIC), European
Journal of Psychology of Education 8: 273–287.
Schneider, W., Roth, E. & Ennemoser, M. (2000). Training phonological skills and letter
knowledge in children-at risk for dyslexia: A comparison of three kindergarten interven-
tion programs, Journal of Educational Psychology 92: 284–295.
Scanlon, D.M. & Vellutino, F.R. (1997). A comparison of the instructional backgrounds and
cognitive profiles of poor, average, and good readers who were initially identified as at
risk for reading failure, Scientific Studies of Reading 1: 191–215.
Stanovich, K.E. (1986). Matthew effects in reading: Some consequences of individual
differences in the acquisition of literacy, Reading Research Quarterly 21: 360–406.
Thompson, G.B. & Johnston, R.S. (2000). Are nonword and other phonological deficits
indicative of a failed reading process? Reading & Writing 12: 63–97.
Torgesen, J.K. & Davis, C. (1996). Individual difference variables that predict response to
training in phonological awareness, Journal of Experimental Child Psychology 63: 1–21.
Treiman, R. & Zukowski, A. (1996). Children’s sensitivity to syllables, onsets, rimes, and
phonemes, Journal of Experimental Child Psychology 61: 193–215.
Tunmer, W.E. & Nesdale, A.R. (1985). Phonetic segmentation skill and beginning reading,
Journal of Educational Psychology 77: 417–427.
Wagner, R.K. & Torgesen, J.K. (1987). The nature of phonological processing and its causal
role in the acquisition of reading skills, Psychological Bulletin 101: 192–212.
Wagner, R.K., Torgesen, J.K. & Rashotte, C. (1994). Development of reading-related phono-
logical processing abilities: New evidence of bidirectional causality from a latent variable
longitudinal study, Developmental Psychology 30: 73–87.
Wagner, R.K., Torgesen, J.K., Rashotte, C.A. & Hecht, S.A. (1997). Changing relations
between phonological processing abilities and word-level reading as children develop
from beginning to skilled readers: A 5-year longitudinal study, Developmental Psychology
33: 468–479.
Walley, A.C. (1993). The role of vocabulary development in children’s spoken word recogni-
tion and segmentation ability, Developmental Review 13: 286–350.
Weschler, D. (1967). Preschool and primary scale of intelligence. Cleveland, Ohio: The
Psychological Corporation.
682 VIRGINIA MANN & HEINZ WIMMER
Wimmer, H. & Goswami, U. (1994). The influence of orthographic consistency on reading
development: Word recognition in English and German children, Cognition 51: 91–103.
Wimmer, H. & Hummer, P. (1990). How German speaking first graders read and spell: Doubts
on the importance of the logographic stage, Applied Psycholinguistics 11: 349–368.
Wimmer, H., Landerl, K., Linortner, R. & Hummer, P. (1991). The relationship of phonemic
awareness to reading acquisition. More consequence than precondition but still important.
Cognition 40: 219–249.
Wimmer, H., Mayringer, H. & Landerl, K. (1999). The double-deficit hypothesis and learning
to read a more regular orthography. Unpublished manuscript, University of Salzburg.
Woodcock, R.W. (1998). Woodcock reading mastery tests C revised. Circle Pines, Minnesota:
American Guidance Service.
Yopp, H.K. (1988). The validity and reliability of phonemic awareness tests, Reading Research
Quarterly 23: 159–177.
Address for correspondence: Dr. Virginia Mann, Department of Cognitive Sciences, Univer-
sity of California, 3151 Social Science Plaza, Irvine, CA 92697-5100, USA
Phone: +1-949-824-6680; Fax: +1-949-824-2307; E-mail: vmann@uci.edu
