Content uploaded by Martina Penke
Author content
All content in this area was uploaded by Martina Penke on Oct 19, 2018
Content may be subject to copyright.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
On the role of syllable components in German visual word identification
Sandra Beyermann and Martina Penke
8810 words
Sandra Beyermann and Martina Penke, Department of Rehabilitation and Special
Education, University of Cologne !
A visual lexical-decision experiment was conducted in order to find out whether syllable
components (onset, nucleus, and coda) are functional units in German single-word
reading at different stages of reading development. Three groups of readers (school
children in 3rd and 5th grade, and adults) were presented with trisyllabic pseudowords
created by exchanging a single-letter grapheme in different positions of the second
syllable, respectively in onset, nucleus, or coda position. Results reveal differences in
the task performances related to manipulations of the three different syllable
components, and suggest that syllable components constitute processing units in
German visual word recognition.
Keywords: syllable, subsyllabic units, letter status, visual word processing, German,
reading development
Correspondence concerning this article should be addressed to Martina Penke,
Department of Rehabilitation and Special Education, University of Cologne, Herbert-
Lewin-Straße 10, D-50931 Cologne. E-mail: martina.penke@uni-koeln.de
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
2
1. Introduction
In phonographic writing systems such as English, Dutch, French and German written
words represent spoken words by means of phonographic units. It is generally assumed
that when children learn to read in these languages they initially map letter units (e.g.,
<sh>, <ee>, <p>) to the corresponding sound units ([ʃ], [i], [p]), and then derive the sound
shape of printed words by blending the sound units ([ʃi:p]) to gain access to the word
meaning (e.g., Ehri 1992, Share 1995). With growing reading proficiency, this
phonological recoding procedure is more and more substituted with orthographic
processing mechanisms that allow for a fast and direct mapping of spelling patterns to
word meanings (e.g., Sprenger-Charolles, Siegel, Béchennec & Serniclaes 2003, Bijeljac-
Babic, Millogo, Farioli & Grainger 2004, Coltheart & Rastle 1994). Despite the
development of orthographic processing routines, numerous studies employing different
experimental paradigms have demonstrated that phonology continues to have an impact
even on skilled reading (e.g., Van Orden 1987, Ziegler & Jacobs 1995, Frost 1998, Nuerk,
Rey, Graf & Jacobs 2000, Penke & Schrader 2008). However, which phonological
constituents are functional units during visual word processing is a matter of ongoing
research (see e.g., Ziegler, Perry, Jacobs & Braun 2001, Morais 2003, Perry, Ziegler &
Zorzi 2010). In the past, among the studies investigating the role of phonology in single-
word reading the majority concentrated on the processing of small phonographic units
(graphemes) in monosyllabic words. The functional role of other phonological constituents
like syllables, foot structures, and stress patterns has only recently come into focus (e.g.,
Colombo 1991, Taft 1992, Fodor 1998, Kelly, Morris & Verrekia 1998, Rastle & Coltheart
2000, Stenneken, Conrad & Jacobs 2007, Yap & Balota 2009, Perry et al. 2010, Evertz &
Primus 2013, Beyermann & Penke 2014). Written language conveys information by
representing spoken language, and in phonographic writing systems such as English,
German and French written words represent the phonological form of spoken words. Basic
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
3
processing units of spoken language might hence well correspond to processing units of
written language. The syllable is a central processing unit in spoken language, and might
thus also influence the processing of written language. The aim of the present study
therefore is to explore whether syllable components (i.e., onset, nucleus, and coda)
constitute functional units in visual word processing.
A phonological syllable is composed of a nucleus (see example in Figure 1, here the
nucleus is filled with the vowel [ӕ]), an onset (here [k]), and a coda (here [t]). The nucleus
and coda form the syllable rime (here [ӕt]).
Figure 1. Structure of the syllable, subsyllabic units of the word <cat>
σ
!
onset
nucleus
Coda
k
ӕ
t
Languages differ with respect to restrictions on syllable structure which leads to
differences in the inventory of syllable structures that are present within words of that
language (for a detailed discussion see e.g., Blevins 1995, Kager 1999). German words for
example often contain rather complex syllables such as in Krampf [kχampf] ‘cramp’,
whereas Spanish words are often composed of CV or CVC syllables (Conrad & Jacobs
2004:3).
The syllable is considered a central unit in phonological theory (Kenstowicz 1994). It is
the basic domain of suprasegmental phenomena (e.g., stress, accent) as well as many
segmental phonological processes and regularities (e.g., phonotactic constraints,
assimilation, epenthesis), for a detailed discussion see e.g., Kenstowicz (1994:250ff). The
syllable is typically comprised of an opening and closing gesture (cf. Pompino-Marschall
2009:240) and is a basic articulatory unit in spoken word production. In auditory word
recognition, stressed syllables are prominent elements in the speech stream, and it is
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
4
assumed that the onsets of stressed syllables serve to single out and identify words from
continuous speech (cf. Grosjean & Gee 1987, Marslen-Wilson & Zwitserlood 1989,
Content, Kearns & Frauenfelder 2001). Developmental studies have shown that children at
the age of about 4 to 5 years already have a concept of the syllable and of words that
rhyme, whereas they are not yet explicitly aware of the phonemes in a word (Ziegler &
Goswami 2005:2ff; compare e.g., Liberman, Shankweiler, Fischer & Carter 1974,
Wimmer, Landerl & Schneider 1994). Thus, preliterate children perform better in
phonological awareness tasks that can be accomplished by operating with words, syllables,
onsets and rimes than in tasks requiring the manipulation of phonemes (Ziegler &
Goswami 2005:4f, and further references cited therein).
1.1 The syllable in written language processing
Besides this central role of the syllable in spoken language processing, syllables and
syllable components are also assumed to be functional units in reading and writing. The
assumption that syllables are relevant processing units during written word production is
for instance supported by studies on German skilled writing (Weingarten, Nottbusch &
Will 2004). Weingarten and colleagues reported that interkey intervals in typing two letters
that belong to different syllables are significantly longer than interkey intervals in typing
these letters within the same syllable. The authors suggest that letter strings forming one
syllable are typed in one planned motor action.
The number of syllables was also shown to affect skilled single-word reading (e.g.,
Stenneken et al. 2007). In the reported visual lexical decision experiment the subjects’
response times on German mono-, di-, and trisyllabic words and pseudowords of the same
word length were compared. The response times on words increased significantly the more
syllables they contained. These findings suggest that printed words are pre-lexically
decomposed syllable by syllable, and this segmentation process is assumed to take longer
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
5
with an increasing number of syllables (cf. Álvarez, Carreiras & Perea 2004 and Carreiras,
Ferrand, Grainger & Perea 2005 for similar findings on Spanish and French). In a syllable
monitoring experiment with French beginning and skilled readers, Colé, Magnan &
Grainger (1999) found that CV or CVC printed syllables were detected significantly faster
when they were congruent with the first syllable of visually presented single words than
when they were not, compare SO in SOLEIL ‘sun’, SOL in SOLDAT ‘soldier’ (congruent)
vs. SOL in SOLEIL, SO in SOLDAT (incongruent). This effect of syllable congruency
was found in adult readers and in good beginning readers who had achieved high
performance results in a standard reading test. From these findings it might be concluded
that at least skilled readers tend to immediately associate the given letter strings to syllabic
units in visual word processing. Evidence for the role of the syllable in visual word
recognition also comes from the observation of syllable frequency effects in visual lexical
decision experiments reported for Spanish (e.g., Perea & Carreiras 1998), French (Mathey
& Zagar, 2002) and German (e.g., Conrad & Jacobs 2004). For these languages it was
shown that the recognition of polysyllabic printed words was inhibited when the frequency
of the first phonological syllable was high as compared to when it was low. This inhibitory
syllable frequency effect is attributed to the competition of candidate words during lexical
selection. Given that printed words are (serially) segmented into syllables, words with a
high-frequent first syllable are related to more lexical competition among activated word
units sharing the first syllable than words with a low-frequent first syllable.
Whereas these studies have provided evidence for the influence of syllables as a functional
unit in visual word recognition only few studies have yet focused on the relevance of
syllable constituents. The assumption that syllable components are relevant processing
units is reflected in some reading models (cf. e.g., Plaut, McClelland, Seidenberg &
Patterson 1996, Jacobs, Rey, Ziegler & Grainger 1998, Perry et al. 2010, Grainger &
Ziegler 2011) which assume that syllable components such as onsets, nuclei and codas
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
6
serve as organizational phonographic units for grouping graphemes during visual word
recognition. According to Grainger & Ziegler (2011:1) the optimal size of word
recognition units that allow for an accurate and fast recognition of printed words is shaped
by two principles: the “diagnosticity constraint” and the “chunking constraint”. The
“diagnosticity constraint” requires word recognition units to be maximally distinctive for
the identification of printed words, like prominent and characteristic letter combinations,
whereas the “chunking constraint” requires word recognition units to be functional units in
phonology and/or morphology, e.g., morphemes, syllables, or subsyllabic units.
Empirical evidence suggesting some role on subsyllabic units in visual word processing
comes from studies indicating that for early reading in English the subsyllabic units onset
and rime are of particular importance (e.g., Goswami 1991, Goswami & East 2000). For
example, Goswami (1991) found that young readers performed better in reading words
whose pronunciation could be generated by analogy to a previously presented training
word (e.g., <trim> and <trap> sharing two consonants in the onset vs. <limp> and
<camp> sharing two consonants as part of the rime, or <trim> and <trip> sharing the
onset and nucleus vs. <loft> and <soft> sharing the rime). The benefit from the presented
analogous word was larger when the shared unit was the onset or the rime of the word.
These findings lead to the assumption that onsets and rimes are important word recognition
units for young readers of English. However, there is empirical evidence in favor of the
assumption that the preferred size of word recognition units varies across languages (e.g.,
Ziegler et al. 2001, Yoon, Bolger, Kwon & Perfetti 2002).
Empirical findings from Ziegler et al. (2001), and Goswami, Ziegler, Dalton &
Schneider (2001, 2003) indicate that German readers rather rely on small phonographic
units during single-word reading. For example Goswami et al. (2003) compared the
performance of English and German beginning readers in reading two different types of
non-words: a) non-words that could be decoded by rime-analogies to existing words (e.g.,
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
7
<DAKE> like <CAKE>) and b) non-words that could only be decoded by grapheme-to-
phoneme translation (e.g., <DAIK>). The authors found that English children achieved
higher reading accuracy when only non-words of the same type were presented in one
homogenous block than when non-words of both types were presented in mixed blocks.
This blocking effect indicates that the English children had to switch between
phonographic processing units of large and small size (rime or larger unit vs. grapheme
units) in the mixed blocks condition which was related to higher cognitive load. In
contrast, the German children did not benefit from the presentation of homogenous blocks
of non-words which was interpreted to show that they rather relied on small phonographic
processing units, no matter whether rime-analogies with existing words could be made.
Thus, it is conceivable that for readers of German, small subsyllabic units are particularly
relevant processing units during single-word reading. This assumption is supported by
Ziegler et al. (2001) who conducted a naming study with German and English skilled adult
readers, and found stronger word length effects in the German than in the English readers
suggesting that the German readers rather relied on small phonographic units during visual
word processing. That phonographic units corresponding to syllable components (onset,
nucleus, coda) are processed in skilled German readers is suggested in a study by Nuerk et
al. (2000). Nuerk and colleagues found that the lexical decision performances of German
skilled readers were significantly affected by the subcomponent frequencies of the
contained syllable components: for monosyllabic low-frequency words the subjects’
responses were faster and more accurate when the test words had high subcomponent
frequencies as compared to low subcomponent frequencies. For monosyllabic non-words
the reversed pattern was observed: non-words with high subcomponent frequencies were
rejected more slowly than those with low subcomponent frequencies. A plausible
explanation for this observation is that printed words are segmented into syllable
components during visual word recognition, with high-frequent subcomponents being
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
8
processed faster due to frequent encounters in written language. Consequently, non-words
with high-frequent subcomponents might be harder to reject because they are more
wordlike than those with low-frequent subcomponents, whereas for words high
subcomponent frequencies seem to accelerate the lexical identification. Thus, the reported
findings indicate that printed words are segmented into subsyllabic units during reading.
1.2 Letter status as confounding factor in research on subsyllabic reading units
A factor that has to be considered in evaluating empirical findings on the influence of
subsyllabic units in visual word recognition relates to the observation that there are
differences between consonant and vowel letters in word processing. This so-called letter
status effect has been confirmed in a number of studies employing different experimental
techniques and paradigms (cf. e.g., Berent & Perfetti 1995, Lee, Rayner & Pollatsek 2001,
Cotelli, Abutalebi, Zorzi & Cappa 2003, Carreiras et al. 2009b, Carreiras, Vergara & Perea
2009; for an overview see e.g., Carreiras et al. 2009b). For example, in a recent ERP study
by Carreiras et al. (2009b) employing the delayed letter paradigm, the Spanish speaking
participants were slower in judging the lexical status of given words when the presentation
of two consonants was delayed (e.g., CHO_O_ATE) than when the presentation of two
vowel letters was delayed (e.g., CHOC_L_TE). Consonant letters usually are more
characteristic for the visual identity of a printed word.1 For one, there are usually more
different consonant letters than different vowel letters. Moreover, investigations of the
German phonographic system (e.g., Butt & Eisenberg 1990, Primus 2003, and Fuhrhop,
Buchmann & Berg 2011) demonstrate that letters with consonant status typically have
ascenders or descenders (e.g., <f, t, g, k>) that make them visually more prominent than
vowel letters. Typical vowel letters, in contrast, do not have ascenders or descenders, but
1 This parallels an empirical observation reported in the literature on speech processing which was captured
by the so called CV hypothesis (e.g., Toro, Nespor, Mehler & Bonatti 2008). According to this
hypothesis, there are differences with respect to the linguistic information that listeners extract from
consonants vs. vowels in continuous speech.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
9
appear only in central space (e.g., <a, o, u>). Since consonant letters are more distinctive
than vowel letters it is assumed that they play a major role during lexical search and are
given preference in processing before vowel letters (Berent & Perfetti 1995, compare
Grainger & Ziegler 2011).
Weingarten et al. (2004:11) claim that the segmentation of printed words into syllables
crucially hinges on the C/V status of the constituent letters: The vowel letter (sequence)
usually represents the syllable’s nucleus. And in German, the number of uninterrupted
vowel letter sequences determines the minimal number of syllables in a word (Neef
2004:223-225).
The impact letter status exerts on visual word recognition considerably complicates
research into the influence of subsyllabic units in reading. Processing differences related to
subsyllabic units might be due to the influence of the syllable components themselves or
due to the different letter status the different syllable components have. Consider for
example languages whose vocabulary is predominantly constituted by words with
structurally less complex CV syllables. Here, the syllable components onset and
rime/nucleus are confounded with letter status, since the onset is realized by a consonant,
whereas the rime or nucleus is realized by a vowel. In contrast, languages such as German
that exhibit a more complex syllable structure enable us to disentangle these two factors
since both syllable constituents onset and rime frequently contain consonants. Processing
differences between these two syllable constituents can thus not be attributed to a letter
status effect.
1.3 Aim of the study
The present study aims at providing further insight into the role of subsyllabic units during
single-word reading. We investigate the question whether onset, nucleus, and coda
constitute relevant processing units in German visual word recognition and whether
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
10
processing differences related to different subsyllabic units can be accounted for by a letter
status effect. We assume that if syllable components play a role in visual word processing,
an orthographic manipulation of single-letter graphemes in different syllable positions
(onset, nucleus, and coda) of German words should affect the subjects’ performances in
different ways. The rather complex syllable structure of German words allows to
disentangle observations that might be attributed to the letter status effect from those that
can be attributed to differences in the processing of consonant vs. vowel letters in onset,
nucleus, and coda position. Since both syllable positions are filled by consonant letters
processing differences related to onset vs. coda consonants cannot be attributed to a
different letter status but suggest an influence of subsyllabic elements on visual word
processing. Subjects from three different age groups, i.e. at different stages of reading
development participated in the experiment. The performances across the different groups
were compared in order to find out whether asymmetries in the processing of different
syllable components arise at different stages of reading proficiency.
2. Method
In order to investigate whether the subsyllabic units onset, nucleus, and coda constitute
processing units during German visual word recognition a visual lexical decision
experiment was conducted with German readers of different proficiency.
2.1 Subjects
Three different groups of German native speakers at different levels of reading proficiency
participated in the experiment. All subjects were monolingual speakers of German, and
participation was voluntary.
A group of 25 skilled adult readers (15 female, 10 male, mean age 27;7 years)
participated in the experiment. Furthermore, 25 school children in the 3rd grade (which
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
11
refers to the 3rd year-term of primary schooling in Germany) were recruited from primary
schools in Brandenburg and Berlin (11 female, 14 male, mean age 8;8 years). We tested a
second group of 27 children in the 5th grade from a primary school in Brandenburg (15
female, 12 male, mean age 10;10 years). Additional data from three adult readers, two 3rd
graders and two 5th graders were excluded from data evaluation due to the very slow mean
response latencies of these subjects (more than 1.5 SD above the mean response time of the
respective subject group).
The school children displayed no developmental language disorders according to their
class teachers. Additionally, the children passed a standardized reading test to ensure they
had normal word-reading skills (ELFE 1-6, Lenhard & Schneider 2006), except for five
3rd graders and four 5th graders who were tested separately and did not participate in the
standardized reading test. The experiment was run at the end of the school term.
2.2 Design and Stimuli
A set of 48 trisyllabic German words with stress on the second syllable provided the basis
for creating the stimulus material that consisted of 48 pseudowords. Each pseudoword was
created by exchanging one letter in the second syllable. The resulting phonological form
differed from existing words, i.e., the critical test items were not homophonous to real
German words. Out of the 48 pseudowords, 16 pseudowords resulted from a manipulation
in onset position of the second syllable (e.g., Schimpanse ‘chimpanzee’ > Schimhanse), 16
pseudowords displayed a manipulation in the nucleus position of the second syllable (e.g.,
Gymnastik ‘gymnastics’ > Gymnostik), and 16 pseudowords were created by exchanging
the coda letter of the second syllable (e.g., Kanister ‘can, canister’ > Kanigter). The
position of the exchanged letter in the trisyllabic word base was controlled for: it was
either the 5th letter (for 12 pseudowords in the coda condition, and 13 in onset and nucleus
condition) or the 6th letter (for 4 pseudowords in coda, and 3 in onset and nucleus
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
12
condition). None of the test items contained consonant clusters or diphthongs in the second
syllable, i.e., the exchanged single letter always corresponded to a single-letter grapheme
expressing one syllable component (onset, nucleus, coda). All word bases were familiar to
the tested children. 46 of the 48 trisyllabic base words are contained in Augst (1984, 1989),
an extensive list of school children’s active vocabulary. The young participants’ accuracy
scores and response times on the two test items not listed in Augst (1984, 1989) did not
differ from the remaining critical test items contained in this list.
Moreover, the following item factors were held constant across the three experimental
conditions: i) total word length: (F(2, 45) < 1, ns), ii) uniqueness point: (F(2, 45) <1, ns)2;
as well as iii) the CELEX3 lemma frequencies of the corresponding 16 trisyllabic base
words that were manipulated: (F(2, 45) < 1, ns). In addition, the visual similarity of the
exchanged letters was determined by a similarity index for low case letters (cf. Gibson
1969, Penke & Schrader 2008). Gibson developed a matrix that defines the graphemic
similarity of letters by means of binary features, for instance symmetry and discontinuity
of horizontal and vertical lines, or agreement in curves or straight lines. The minimal
similarity is indexed with 15 for dissimilar letters (e.g., <w> vs. <f>), and approaches 100
for very similar letters (e.g., <b> vs. <d>). The graphemic similarity of our critical
pseudowords and the corresponding base word was held constant across the critical
conditions (F(2, 45) < 1, ns). For an overview over the controlled item factors see Table 1.
An example for a test item in each condition including the corresponding controlled factors
is given in Table 2 below.
2 The uniqueness point (UP) of a pseudoword is the letter position in the word at which the lexical status as
non-word can be determined unambiguously; for example, the UP of the pseudoword <Krawotte> is located
at the 5th letter (o), due to the fact that no lexical entry can be found starting with the letter sequence
<Krawo>, but there are words starting with <Kraw>, e.g., <Krawall ‘riot, rumpus’, Krawatte ‘tie, cravat’>.
3 Baayen, R. Harald, Piepenbrock, Richard & Hedderik van Rijn (1993): The CELEX lexical database. [CD-
ROM]. Philadelphia, University of Pennsylvania. Linguistic Data Consortium.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
13
Table 1. Controlled item factors across experimental conditions
Controlled factor
Manipulation of
Onset
Nucleus
Coda
Word length
(n of letters)
Mean = 8.75
(SD = 0.78,
range 8-10)
Mean = 8.69
(SD = 0.87,
range 8-10)
Mean = 8.44
(SD = 1.03,
range 7-10)
Uniqueness point
(letter position)
Mean = 5.25
(SD=.45,
range 5-6)
Mean = 5.44
(SD=.81,
range 5-7)
Mean = 5.38
(SD=.50,
range 5-6)
Lexical frequency of
base word
Mean = 67.56
(SD = 94.67,
range 0-296)
Mean = 62.56
(SD = 44.36,
range 9-152)
Mean = 67.88
(SD = 72.67,
range 1-255)
Visual similarity of
exchanged letter
Mean = 76.72
(SD = 12.2,
range 65-100)
Mean = 76.56
(SD = 11.79,
range 60-85)
Mean = 76.56
(SD = 10.68,
range 68-100)
Table 2. Example for test items and controlled factors
Controlled factor
Manipulation of
Onset
Nucleus
Coda
Blongine
(from Blondine
‘blonde’)
Krawotte
(from Krawatte
‘tie, cravat’)
Kanigter
(from Kanister
‘can, canister’)
Word length
8
8
8
Uniqueness point
5
5
5
Lexical frequency of
base word
18
32
13
Visual similarity of
exchanged letter
d-g: 85
a-o: 85
s-g: 75
Position of
exchanged letter
5
5
5
The 48 critical pseudowords were embedded in a total set of 144 different tri- and
disyllabic items presented to each participant. Half of the presented items were words and
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
14
the other half were pseudowords. All pseudowords obeyed the German phonotactic and
graphotactic rules.
2.3 Procedure
The experiment was run on a computer. Subjects were tested in small groups (1-4 test
persons) in a quiet room. Each participant was comfortably seated in front of a computer
screen. The participants were instructed to decide as fast and as accurately as possible
whether or not a given word was a real German word by pressing a green/red labeled
keyboard button for “yes/no” respectively. To avoid confounding effects by left- or right-
handedness, the green button for ”yes” responses was placed on the right hand side for
right-handers and on the left hand side for left-handers such that it was always pressed by
the subject’s dominant hand.
The critical pseudowords were presented in randomized order in the center of the
computer screen. Item presentation and response registration was controlled by the
software DMDX (Forster & Forster 2003). Each single item was presented until key press
or timeout. The maximal duration of item presentation and inter stimulus interval (ISI) was
varied according to the assumed differences in the cognitive capacities of the different
subject groups: For 3rd and 5th graders, items were maximally presented for 5 seconds with
an ISI of 1.8 second. For students, items were maximally presented for 4 seconds with an
ISI of 1 second. Reaction-times were measured from the onset of the stimulus presentation.
To familiarize the subjects with the task, each subject was presented with 10 practice trials
consisting of words and non-words. The total duration of the experimental session varied
between 10 to 20 minutes. No feedback was provided during the experiment.!
The number of correct “no” and incorrect “yes” responses was counted for each subject
in the critical onset, nucleus and coda condition. Reaction-times obtained for correct “no”
responses in the three critical experimental conditions (onset, nucleus, coda manipulation)
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
15
were compared. On the basis of these data the error rates, and mean reaction-times for
correct responses were computed for each subject and subject group.
3. Results
3.1 Global comparisons
A global comparison of the mean response times and total error rates in each subject group
reflects the progress in reading acquisition from readers in 3rd grade (mean RT = 2507 ms,
17.4% total error rate), over more advanced readers in 5th grade (mean RT = 1742 ms, 11.1
% total error rate) to skilled adult readers (mean RT = 836 ms, 1.4% total error rate). A 3x3
analysis of variance (ANOVA) with the position of the manipulated letter as within-subject
factor (onset, nucleus, coda) and reading proficiency as between-subject factor (3rd, 5th
graders, and adult readers) revealed a significant main effect of position in the response
accuracy by subject (F1(2, 148) = 14.933, p < .001, ŋp2 = .168), but no such effect in the
response times (F1(2, 148) < 1, ns), as well as a significant main effect of reading
proficiency on accuracy rates by subject (F1(2, 74) = 25.045, p < .001, ŋp2 = .404), and on
response times (F1(2, 74) = 94.222, p < .001, ŋp2 = .718). In addition, a significant
interaction of position*reading proficiency was detected in the accuracy scores by subject
(F1(4, 148) = 3.323, p < .05, ŋp2 = .082), but no such effect was found in response times
(F1(4, 148) < 1, ns). The global comparison of the performance accuracy in the three
critical conditions indicates that a manipulation of the three different syllable components
(onset, nucleus, and coda) is associated with different accuracy scores in the subjects’
lexical decisions, and that subjects in the different age groups are affected differently.
Since the global comparison yielded that response latencies were not significantly affected
by the position of the manipulated letter, the next section only reports results of within-
group comparisons for the accuracy rates achieved by the participants.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
16
3.2 Within-group comparisons
An ANOVA (repeated measures) detected significant differences between the critical
conditions (onset, nucleus, or coda letter manipulated) in the performance accuracy in 3rd
graders (F1(2, 48) = 11.432, p <.001, ŋp2 = .323). The accuracy rates of 3rd graders are
depicted in Figure 2.
Fig. 2 Mean Accuracy Rates (in %) of the 3rd Graders in Critical Conditions
70
80
90
100
Onset Nucleus Coda
Mean4Accuracy4(%)
3rd4graders
The performance accuracy of the tested 3rd graders was lowest when the manipulated letter
was in nucleus position (mean accuracy rate = 76%, SD = 16,7), higher when it was in
coda position (83.5%, SD = 15.9), and highest for words with a manipulated onset letter
(88.3%, SD = 8.9). In other words, they mistakenly accepted pseudowords as words more
often when the exchanged letter was the nucleus letter (e.g., Gymnostik > Gymnastik
‘gymnastics’) than when it was the coda letter (e.g., Kanigter > Kanister, ‘can, canister), or
onset letter of the second syllable (e.g., Schimhanse > Schimpanse ‘chimpanzee’). The
differences in the performance accuracy reached significance in all comparisons across the
experimental conditions (onset vs. nucleus, nucleus vs. coda, and coda vs. onset letter
manipulation, see Figure 2).
The separate ANOVA for 5th graders shows that the factor position of the manipulated
letter significantly affected the accuracy rates (F1(2, 52) = 3.49, p < .05, ŋp2 = .118). The
p < .001
p < .05
p < .05
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
17
comparisons across conditions revealed significant differences in the performance
accuracy between onset vs. nucleus, and nucleus vs. coda but no difference in the rejection
of pseudowords with a manipulation in onset vs. coda position. The accuracy rates for each
condition achieved by the 5th graders are depicted in Figure 3.
Figure 3: Mean number of correct responses in critical conditions given by 5th graders
70
80
90
100
Onset Nucleus Coda
Mean4Accuracy4(%)
5th4graders
Like in the 3rd graders, the accuracy scores for correctly rejecting the presented
pseudowords were significantly higher when the manipulated letter was in the orthographic
onset (mean accuracy rate = 91.9%, SD = 7.7) or coda position (91%, SD = 8) than when it
was in the nucleus position (85.9%, SD = 12.7) of the second syllable. In contrast to the 3rd
graders, 5th graders displayed no significant difference in the performance accuracy with
respect to the onset vs. coda position of the manipulated letter.
The adult participant achieved an overall high accuracy rate close to ceiling (see Figure
4 below). However, in the accuracy data of skilled adult readers, a tendency towards
significant differences between accuracy rates in the three conditions was detected (F1(2,
48) = 2.761, p = .073), with a significant difference between the onset vs. coda condition.
Thus, the task performances of the adults show the same pattern as those of the school
children from our experiment. Figure 4 illustrates the mean accuracy rates achieved by the
skilled readers.
Figure 4. Mean Accuracy Rates of the Adults (in %) in the Critical Conditions
p < .05
p < .05
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
18
70
80
90
100
Onset Nucleus Coda
Mean4Accuracy4(%)
adults
The figure illustrates that like in the primary school children who participated in our
experiment, the number of correct “no” responses on pseudowords given by the adult
readers were highest when the manipulated letter was located in the onset of the second
syllable (mean accuracy rate = 99.8%, SD = 1.3). In the nucleus condition (98%, SD = 4.3)
and coda condition (98%, SD = 3.9), the performance accuracy was slightly lower
confirming to the behavioral pattern observed in school children. Note also, that the
response time data of the adult participants point to the same pattern with a mean RT= 851
ms, SD = 189 in the onset condition vs. mean RT = 812 ms, SD = 135 in the nucleus, and
mean RT = 846 ms, SD = 181 in the coda condition.
4. Discussion
The aim of our study was to investigate whether subsyllabic units (onset, nucleus, and
coda) are functional units in German visual word processing. Our data show that
manipulations of different subsyllabic units were associated with differences in the
accuracy with which these manipulated words were judged by German subjects at different
stages of reading acquisition. In order to provide a full account of our observations, we
need to carefully consider two different potential sources causing these detected
differences: The syllable position (onset, nucleus, or coda) as well as the letter status
(consonant vs. vowel) of the corresponding orthographic manipulation might have had an
impact on the subjects’ lexical decision performances.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
19
A first result of our study is that subjects achieved lower accuracy scores for
pseudowords that were created by exchanging a single-letter grapheme in the nucleus (e.g.,
Gymnostik > Gymnastik) as opposed to the onset (e.g., Schimhanse > Schimpanse), and
the coda position (e.g., Kanigter > Kanister). This indicates that the subjects noticed the
orthographic manipulation in the syllable nucleus less often, and then mistakenly accepted
the presented letter strings as words. Both children groups were significantly better in
rejecting pseudowords in the coda and onset condition than in the nucleus condition. The
adults showed more subtle differences between the critical experimental conditions, which
due to a ceiling effect, can only be interpreted with considerable caution. They displayed
no differences in the performances on pseudowords with an exchanged letter in the syllable
nucleus vs. coda, but like the young readers they achieved lower accuracy scores on
pseudowords in the nucleus vs. onset condition.
There are two conceivable explanations for these observations. On the one hand, the
detected differences in the lexical decision performances on words with orthographic
manipulations in the syllable onset vs. nucleus, and nucleus vs. coda might be attributed to
an orthographic effect of letter status. According to Grainger & Ziegler (2011) one of the
main principles influencing the shape of a word recognition unit is the “diagnosticity
constraint” that requires word recognition units to be maximally distinctive with respect to
word identity. Recall that vowel letters are considered to contribute less distinctive
orthographic information on a given word than consonant letters. This might be due to the
smaller number of vowel letters as compared to consonant letters that typically have more
salient visual features and are more prominent than vowel letters. According to the
“diagnosticity constraint” vowel letters are hence less important during word identification
as compared to consonant letters. As a result, the subjects might have paid more attention
to the consonant letters in the presented words, leading to more correct rejections of
pseudowords in the onset and coda condition than in the nucleus condition.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
20
On the other hand, an alternative account for our findings is based on the assumption
that basic units of orthographic processing are shaped by phonology, an assumption that is
captured by the “chunking constraint” (Grainger & Ziegler 2011) requiring word
recognition units to be functional in phonology and/or morphology. In our study, there are
two observations that indicate an influence of subsyllabic phonographic units over and
above the potential letter status effect, and suggest that German printed words are
decomposed into small phonographic units during word recognition. For one, if the letter
status were the source of the observed differences in the lexical decision performances in
our experiment, one would have expected clear-cut differences in the onset vs. nucleus and
in the nucleus vs. coda condition in the subjects’ performances. However, we found no
such difference in the university students who performed equally well on pseudowords
with orthographic manipulations of the vowel in the nucleus position and the consonant in
the coda condition. Thus, for the group of adult readers, the observed differences in the
accuracy rates across conditions do not suggest an influence of letter status. This leaves the
possibility that there is a developmental change between the children groups and the adult
readers, in the sense that a letter status effect initially operative in young readers is no
longer influential in visual processing routines of skilled adult readers. A second
observation, however, argues against this potential explanation and indicates that a letter
status effect is not crucial in explaining the behavioral pattern observed in young readers:
The 3rd graders (and the adults) demonstrated a poorer performance accuracy in rejecting
pseudowords with manipulations in the coda (e.g., Kanigter > Kanister) than in the onset
position (e.g., Schimhanse > Schimpanse), although in both cases a consonant letter was
exchanged. This pattern of performance cannot be explained in terms of a letter status
effect. Also, this observation does not support the view of a developmental change in
visual processing strategies between young and experienced readers since the young
readers as well as experienced adult readers displayed behavioral patterns that cannot be
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
21
accounted for by a letter status effect. Given these two observations, we interpret the
differences in the accuracy scores to be due to the manipulation of different syllable
components that contribute differently to the word recognition process. Our findings thus
support the assumption that letter strings grouped into subsyllabic units constitute
phonographic processing units during single-word reading for young as well as skilled
German readers.
An additional observation well in line with the suggestion that the presented letter strings
were grouped in subsyllabic units is that the readers were more sensitive to orthographic
manipulations in the syllable initial position than to those in the other syllable positions
(nucleus, coda). All subject groups rejected pseudowords with exchanged letters in the
onset position more accurately than those with exchanged letters in the nucleus or coda
position (compare Figure 2-4). This finding suggests that the syllable onset plays a more
important role than the nucleus and coda (i.e. the rime) during visual word recognition.
This is consistent with the assumption that the onsets of stressed syllables are highly
prominent elements in the speech stream that serve to single out and identify words from
continuous speech (cf. Marslen-Wilson & Zwitserlood 1989, Content et al. 2001). Note
that an alternative explanation for this finding that is closely related to the letter-status
effect can be ruled out: Thus it could be argued that a consonant letter is more prominent
when it is located between two vowel letters (vCv) which are visually less salient. As this
intervocalic consonant position often coincides with the onset of a syllable (e.g., <hotel>
'hotel'), one might assume this coincidence to account for the subjects’ higher sensitivity to
letter exchanges in the onset position in our data. However, the exchanged consonant
letters in 15 out of 16 stimulus words from the onset condition in our experiment were
preceded by a consonant and followed by a vowel letter (CCv, e.g., Schimhanse >
Schimpanse ‘chimpanzee’). Therefore, an intervocalic environment could not account for
the detected performance differences in our experiment (for a detailed discussion of the
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
22
influence of the relative letter position, see e.g., Brand, Giroux, Puijalon & Rey 2007).
Compare this also to our stimulus words in the coda condition in which the exchanged
consonant letter was always preceded by a vowel and followed by a consonant letter (vCC,
e.g., Kanigter > Kanister ‘can, canister’). Thus, in both the experimental conditions onset
and coda, the exchanged consonant letters were located in a similar orthographic
environment.
The findings of our study provide empirical support for the relevance of sublexical
processing routines as proposed in a number of reading models. Some of these models
integrate a mechanism of subsyllabic decomposition that segments a given letter string into
syllable components (e.g., Plaut et al. 1996, Jacobs et al. 1998, Perry et al. 2010). Thus, for
example in the CDP++ reading model of Perry et al. (2010) a sublexical network is
assumed, assigning the graphemes of encountered printed words to a graphosyllabic
template which is mapped onto syllabic groupings of phonemes. In their model for
multisyllabic word reading, both the syllabic grouping of graphemes and phonemes is
based on an onset-nucleus-coda template. Our observation of processing differences
related to orthographic manipulations in different syllable positions supports the
assumption of a subsyllabic decomposition process during visual word processing, and
suggests that a subsyllabic processing routine should be implemented in models on visual
word recognition.
Our study results are well in line with previous findings on German visual word processing
reported by Nuerk et al. (2000), Ziegler et al. (2001), and Goswami et al. (2001, 2003).
The cross-language studies on English and German by Ziegler et al. (2001), and Goswami
et al. (2001, 2003) confirm the assumption that young as well as skilled German readers
rely on smaller phonographic units than English readers during visual word processing (cf.
e.g., Wimmer et al. 1994, Ziegler et al. 2001). Whereas German readers were more
affected by word length, English readers were shown to be more sensitive to rime spellings
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
23
(e.g., by neighborhood effects), indicating that the preferred phonographic processing units
for decoding printed words for English readers are larger than for German readers. The
non-word reading accuracy of English children was influenced by the presentation mode of
the stimuli list (in homogenous or mixed blocks of items that could or could not be
decoded by analogy to the rime or larger word units), whereas German children showed no
such blocking effect. A plausible interpretation of these results is that the English readers
demonstrated switching costs between phonographic processing units of large and small
size, whereas the German readers relied on small phonographic processing units
throughout. The study by Nuerk et al. (2000) provides further support for this assumption
by reporting positional frequency effects for orthographic syllable components in German
skilled reading. However, these naming and lexical decision studies investigated the
processing of subsyllabic phonographic units in German without considering the different
subsyllabic units (onset, nucleus, and coda) individually. Thus, the results from Nuerk et
al. (2000) demonstrate a processing effect for subsyllabic units: The subcomponent
frequency effect which is based on subsyllabic units in the phonological representation of a
word and calculated by means of positional letter (cluster) frequencies has an impact on
German visual word recognition over and above purely orthographic frequency measures
(such as bigram or trigram frequencies). The authors showed that skilled German readers
were affected by this influential phonographic frequency measure by comparing lexical-
decision performances on words with high vs. low subcomponent frequencies. However,
Nuerk and colleagues did not consider processing effects separately for individual syllable
components (onset, nucleus, coda). To our knowledge, the present study is the first to
investigate the functional role of different syllable components by comparing the impact of
an orthographic manipulation separately for each of these syllable components: onset,
nucleus, and coda.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
24
In sum, the issue whether subsyllabic units such as onset, nucleus and coda are
functional in visual word processing is confounded with the influence of the different letter
status these subsyllabic components have. The present study illustrates an attempt to
disentangle the influence of subsyllabic components on visual word processing from the
effects of letter status. We observed that manipulations of different subsyllabic units were
associated with differences in the accuracy with which these manipulations were detected
by German subjects at different stages of reading acquisition. The comparatively lower
accuracy rates for pseudowords with a manipulation in the nucleus position might be
attributed to purely orthographic factors (the letter status) because typical vowel letters are
considered less prominent and thus less important for the identification of printed words.
However, additional findings indicate an influence of subsyllabic units over and above the
potential letter status effect in visual word processing. The observation that the highest
accuracy rates were achieved for words with a manipulation in onset position indicates that
the syllable onset plays a central role during word identification, as proposed e.g., by
Content et al. (2001). Moreover, the different performances on words with manipulations
in onset vs. coda position indicate that syllable components play a role in German visual
word recognition. The present study provides empirical evidence for the relevance of
subsyllabic units during German single-word reading and thus further supports the
assumption that word recognition units are shaped by phonology.
Acknowledgements
This research was supported by grant G.0292.09N from FWO (Fonds Wetenschappelijk
Onderzoek, Vlaanderen) to Martina Penke. The data was collected thanks to support from
Grundschule am Koppenplatz (Berlin), Grundschule Klosterfelde (Brandenburg), and
Potsdam University (Brandenburg). We also thank Eva Neuhaus (University of Cologne)
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD IDENTIFICATION
25
for helpful comments. A special thanks is dedicated to Uwe Manthey, Antje Jessa, Sandra
Rex, Elli Kappler and Solveig Teich for their help in data collection.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD RECOGNITION
26
References!
Álvarez, Carlos J., Carreiras, Manuel & Perea, Manuel (2004). Are syllables phonological units in visual word
recognition? Language & Cognitive Processes 19: 427-452.
Augst, Gerhard (1984). Kinderwort: Der aktive Kinderwortschatz kurz vor der Einschulung. Frankfurt a.M.:
Lang.
Augst, Gerhard (1989). Schriftwortschatz. Untersuchungen und Wortlisten zum Orthographischen Lexikon bei
Schülern und Erwachsenen. Vol. 10. Frankfurt a. M.: Lang.
Baayen, R. Harald, Piepenbrock, Richard & Hedderik van Rijn (1993). The CELEX lexical database. [CD-
ROM]. Philadelphia, University of Pennsylvania. Linguistic Data Consortium.
Berent, Iris & Charles A. Perfetti (1995). A rose is a REEZ: The two cycles model of phonology assembly in
reading English. Psychological Review 102: 146-184.
Beyermann, Sandra & Martina Penke (2014). Word stress in German single-word reading. Reading Psychology
35:6, 577-600.!
Bijeljac-Babic, Ranka, Millogo, Victor, Farioli, Fernand & Jonathan Grainger (2004). A developmental
investigation of word length effects in reading using a new on-line word identification paradigm. Reading
and Writing: An Interdisciplinary Journal 17: 411-431.
Blevins, Juliette (1995). The syllable in phonological theory. In J.A. Goldsmith (ed.), The handbook of
phonological theory, 206-244. Oxford: Blackwell.
Brand, Muriele, Giroux, Ibrahima, Puijalon, Carole & Arnaud Rey (2007): Syllable onsets are perceptual reading
units. Memory & Cognition 35(5): 966-973.
Butt, Matthias & Peter Eisenberg (1990). Schreibsilbe und Sprechsilbe. In Christian Stetter (ed.), Zu einer
Theorie der Orthographie. Interdisziplinäre Aspekte gegenwärtiger Schrift- und Orthographieforschung, 33-
64. Tübingen: Niemeyer.!
Carreiras, Manuel, Ferrand, Ludovic, Grainger, Jonathan & Manuel Perea (2005). Sequential effects of
phonological priming in visual word recognition. Psychological Science 16(8): 585-589.
Carreiras, Manuel, Vergara, Marta & Manuel Perea (2009a). ERP correlates of transposed-letter priming effects:
The role of vowels versus consonants. Psychophysiology 46: 34-42.
Carreiras, Manuel, Gillon-Dowens, Margaret, Vergara, Marta & Manuel Perea (2009b). Are vowels and
consonants processed differently? ERP evidence with a delayed letter paradigm. Journal of Cognitive
Neuroscience 21(2): 275-288.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD RECOGNITION
27
Colé, Pascale, Magnan, Annie & Grainger, Jonathan (1999). Syllable-sized units in visual word recognition:
Evidence from skilled and beginning readers of French. Applied Psycholinguistics 20: 507–532.
Colombo, Lucia (1991). The role of lexical stress in word recognition and pronunciation. Psychological
Research 53: 71-79.
Coltheart, Max & Kathleen Rastle (1994). Serial processing in reading aloud: evidence for dual-route models of
reading. In Journal of Experimental Psychology: Human Perception and Performance 20(6): 1197-1211.!
Conrad, Markus & Jacobs, Arthur M. (2004). Replicating syllable frequency effects in Spanish in German: One
more challenge to computational models of visual word recognition. Language & Cognitive Processes 19:
369-390.
Content, Alain, Kearns, Ruth K., & Frauenfelder, Uli H. (2001). Boundaries versus onsets in syllabic
segmentation. Journal of Memory & Language 45: 177-199.
Cotelli, Maria, Abutalebi, Jubin, Zorzi, Marco & Stefano F. Cappa (2003). Vowels in the buffer: A case study of
acquired dysgraphia with selective vowel substitutions. Cognitive Neuropsychology 20(2): 99-114.
Ehri, Linnea C. (1992). Reconceptualizing the development of sight word reading and its relationship to
recoding. In Gough, Philip B., Ehri, Linnea E. & Rebecca Treiman (eds.), Reading acquisition, 105–143.
Hillsdale, NJ: Erlbaum.
Evertz, Martin & Beatrice Primus (2013). The graphematic foot in English and German. Writing Systems
Research, [DOI:10.1080/17586801.2013.765356].
Fodor, Janet D. (1998). Learning to parse? Journal of Psycholinguistic Research 27(2): 285–319.
Forster, Kenneth I. & Forster, Jonathan C. (2003). DMDX. A Windows display program with millisecond
accuracy. Behavior Research Methods, Instruments and Computers 35: 116-124.
Frost, Ram (1998). Toward a strong phonological theory of visual word recognition: True issues and false trails.
Psychological Bulletin 123: 71–99.
Fuhrhop, Nanna, Buchmann, Franziska & Kristian Berg (2011). The length hierarchy and the graphematic
syllable. Evidence from German and English. Written Language & Literacy 14(2): 275-292.
Gibson, Eleanor J. (1969). Principles of perceptual learning and development. New York: Meredith Corporation.
Goswami, Usha (1991). Learning about spelling sequences in reading: The role of onsets and rimes. Child
Development 62: 1110–1123.
Goswami, Usha & East, Martin (2000). Rhyme and analogy in beginning reading: Conceptual and
methodological issues. Applied Psycholinguistics 21: 63–93.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD RECOGNITION
28
Goswami, Usha, Ziegler, Johannes C., Dalton, Luise, & Schneider, Wolfgang (2001). Pseudohomophone effects
and phonological recoding procedures in reading development in English and German. Journal of Memory
and Language 45: 648–664.
Goswami, Usha, Ziegler, Johannes C., Dalton, Luise, & Schneider, Wolfgang (2003). Nonword reading across
orthographies: How flexible is the choice of reading units? Applied Psycholinguistics 24(2): 235-247.
Grainger, Jonathan & Johannes C. Ziegler (2011). A dual-route approach to orthographic processing. Frontiers
in Psychology 2(54): 1-13.
Grosjean, François & James P. Gee (1987). Prosodic structure and spoken word recognition. Cognition 25: 135-
155.
Jacobs, Arthur M., Rey, Arnaud, Ziegler, Johannes C. & Jonathan Grainger (1998). MROM-P: an interactive
activation, multiple read- out model of orthographic and phonological processes in visual word recognition.
In Grainger, Jonathan & Arthur M. Jacobs (eds.), Localist Connectionist Approaches to Human Cognition,
Mahwah, NJ: Lawrence Erlbaum Associates, pp. 147–188.
Kager, René (1999). Optimality theory. Cambridge: Cambridge University Press.
Kelly, Michael H., Morris, Joanna & Laura Verrekia (1998). Orthographic cues to lexical stress: Effects on
naming and lexical decision. In Memory & Cognition 26(4): 822-832.!
Kenstowicz, Michael (1994). Phonology in Generative Grammar. Oxford: Blackwell.
Lee, Hye-Won, Rayner, Keith & Alexander Pollatsek (2001). The relative contribution of consonants and vowels
to word identification during reading. Journal of Memory & Language 44: 189-205.
Lenhard, Wolfgang & Wolfgang Schneider (2006). ELFE 1-6. Ein Leseverständnistest für Erst- bis
Sechstklässler. Göttingen: Hogrefe.
Liberman, Isabelle Y., Shankweiler, Donald, Fischer, F. William & Bonnie Carter (1974). Explicit syllable and
phoneme segmentation in the young child. Journal of Experimental Child Psychology 18: 201–212.
Marslen-Wilson, William & Pinie Zwitserlood (1989). Accessing spoken words: The importance of word onsets.
Journal of Experimental Psychology: Human Perception and Performance 15(3): 576-585.
Mathey, Stéphanie & Daniel Zagar (2002). Similarity in visual word recognition: The effect of syllabic
neighborhood in French. Current Psychology Letters: Behaviour, Brain & Cognition 8: 107-121.
Morais, José (2003). Levels of phonological representation in skilled reading and in learning to read. Reading
and Writing: An Interdisciplinary Journal 16(1-2): 123-151.
Neef, Martin (2004). The relation of vowel letters to phonological syllables in English and German. Written
Language and Literacy 7(2): 205-234.
Nuerk, Hans-Christoph, Rey, Arnaud, Graf, Ralf & Arthur M. Jacobs (2000). Phonographic sublexical units in
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD RECOGNITION
29
visual word recognition. Current Psychology Letters 2: 25-36.
Nottbusch, Guido, Grimm, Angela, Weingarten, Rüdiger & Udo Will (2005). Syllabic structures in typing:
Evidence from deaf writers. Reading and Writing 18: 497-526.
Penke, Martina & Kathrin Schrader (2008). The role of phonology in written word recognition: Reading
acquisition vs. skilled reading. Written Language & Literacy 11(2): 167–190.
Perea, Manuel & Manuel Carreiras (1998). Effects of syllable frequency and syllable neighborhood frequency in
visual word recognition. Journal of Experimental Psychology: Human Perception and Performance 24(1):
134-144.
Perry, Conrad, Ziegler, Johannes C. & Zorzi, Marco (2010). Beyond single syllables: Large scale modeling of
reading aloud with the Connectionist Dual Process (CDP++) model. Cognitive Psychology 61(2): 106-151. !
Plaut, David C., McClelland, James L., Seidenberg, Mark S. & Karalyn Patterson (1996). Understanding normal
and impaired word reading: Computational principles in quasi-regular domains. Psychological Review
103(1): 56-115.
Pompino-Marschall, Bernd (2009). Einführung in die Phonetik. Berlin: De Gruyter.
Primus, Beatrice (2003). Zum Silbenbegriff in der Schrift-, Laut-, und Gebärdensprache – Versuch einer
mediumübergreifenden Fundierung. Zeitschrift für Sprachwissenschaft 22(1): 3-55.
Rastle, Kathleen & Max Coltheart (2000). Lexical and nonlexical print-to-sound translation of disyllabic words
and nonwords. Journal of Memory and Language 42: 342-362.!
Share, David L. (1995). Phonological recoding and self-teaching: Sine qua non of reading acquisition. Cognition
55: 151–218.
Sprenger-Charolles, Liliane, Siegel, Linda S., Béchennec, Danielle & Willy Serniclaes (2003). Development of
phonological and orthographic processing in reading aloud, in silent reading, and in spelling: a four-year
longitudinal study. Journal of Experimental Child Psychology 84(3): 194-217.
Stenneken, Prisca, Conrad, Markus & Arthur M. Jacobs (2007). Processing of syllables in production and
recognition tasks. Journal of Psycholinguistic Research 36: 65-78.
Taft, Marcus (1992). The body of the BOSS: Subsyllabic units in the lexical processing of polysyllabic words.
Journal of Experimental Psychology: Human Perception & Performance 18: 1004-1014.
Toro, Juan M., Nespor, Marina, Mehler, Jacques & Luca L. Bonatti (2008). Finding words and rules in a speech
stream. Functional differences between vowels and consonants. Association for Psychological Science 19(2):
137-144.
Van Orden, Guy C. (1987). A ROWS is a ROSE: spelling, sound, and reading. Memory and Cognition 15: 181–
198.
ON THE ROLE OF SYLLABLE COMPONENTS IN GERMAN VISUAL WORD RECOGNITION
30
Weingarten, R., Nottbusch, G., & Will, U. (2004). Morphemes, syllables, and graphemes in written word
production. Trends in linguistics studies and monographs, 157, 529-572.
Wimmer, Heinz, Landerl, Karin & Wolfgang Schneider (1994). The role of rhyme awareness in learning to read
a regular orthography. British Journal of Developmental Psychology 12: 469–484.
Yap, Melvin J. & David A. Balota (2009). Visual word recognition of multisyllabic words. Journal of Memory
end Language 60: 502-529.
Yoon, Hye-Kyung, Bolger, Donald J., Kwon, Oh-Seek & Charles Perfetti (2002). Subsyllabic units in reading: A
difference between Korean and English. In Verhoeven, Ludo, Elbro, Carsten & Pieter Reitsma (eds.),
Precursors of Functional Literacy, 139-164. Amsterdam: John Benjamins.
Ziegler, Johannes C. & Usha Goswami (2005). Reading Acquisition, Developmental Dyslexia, and Skilled
Reading Across Languages: A Psycholinguistic Grain Size Theory. Psychological Bulletin 131 (1): 3–29.
Ziegler, Johannes C. & Arthur M. Jacobs (1995). Phonological information provides early sources of constraint
in the processing of letter strings. Journal of Memory and Language 34(5): 567-593.
Ziegler, Johannes C., Perry, Conrad, Jacobs, Arthur M., & Mario Braun (2001). Identical words are read
differently in different languages. Psychological Science 12(5): 379-384.
