Science in China Series C: Life Sciences
© 2009

SCIENCE IN CHINA PRESS

Springer

www.scichina.com
life.scichina.com
www.springerlink.com

Article

Reading comprehension without phonological
mediation: Further evidence from a Chinese
aphasic individual
HAN ZaiZhu & Bi YanChao†
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing 100875, China

An important issue in visual word comprehension literature is whether or not semantic access is mediated by phonological processing. In this paper, we present a Chinese individual, YGA, who provides
converging evidence to directly address this issue. YGA has sustained damage to the left posterior
superior and middle temporal lobe, and shows difficulty in orally name pictures and reading printed
words aloud. He makes phonological errors on these tasks and also semantic errors on picture naming,
indicating a deficit at accessing the phonological representations for output. However, he is intact at
understanding the meaning of visually presented words. Such a profile challenges the hypothesis that
semantic access in reading is phonologically mediated and provides further evidence for the universal
principle of direct semantic access in reading.
word reading, phonological mediation, Chinese, case study

1 Introduction
How readers access meaning from a visual word input is
a central issue in word processing research. One of the
debates has concerned whether or not this semantic access process is mediated by phonological access. The
phonologically-mediated hypothesis assumes that a visual stimulus only accesses its meaning by firstly activating a phonological form which subsequently accesses
the meaning[1−11]. An alternative hypothesis, the direct
access hypothesis, assumes that the semantic representation is directly activated by orthographic encodes[12−15].
The phonological mediation hypothesis was motivated and supported by a number of chronometric studies on normal subjects. The most important findings
were reported by Lukatela and Turvey[3−6] and Van Orden et al.[8−11]. For example, in a semantic category task
(e.g., Flower-is X a flower?), Van Orden et al.[8−9] found
that the participants had more difficulty rejecting words
and pseudo-words (e.g., “rows”) which are homophones
of a “yes” response (“rose”) than control words (“robs”).

In a semantic priming study, Lukatela & Turvey[5−6] observed that the target words (e.g., “frog”) benefited
equally from a semantic-related word (“toad”), a word
that is homophonic to the semantic-related word
(“towed”) and a pseudo-word that is homophonic to the
semantic-related word (“tode”). The semantic priming
effect produced by these homophone words was explained as followings: the visual input (e.g., “towed” or
“tode”) accesses its phonological representation (/toud/),
which is identical to the phonological representation of
the semantic neighbor (“toad”) of the target; this phonological representation serves as an input to activate the
corresponding semantic representations and therefore
activates both “towed” and “toad” to the same degree,
resulting in the priming of “frog”. In other words, the
results are readily explained by the phonological-meReceived September 2,2008; accepted December 19,2008
doi: 10.1007/s11427-009-0048-x
†
Corresponding author (email: ybi@bnu.edu.cn)
Supported by Grants PCSIRT (IRT0710), National Natural Science Foundation of
China (Grant Nos. 30770715, 30700224), Beijing Natural Science Foundation (Grant
No.7082051).

Sci China
Serphonological
C-Life Sci mediation:
| May. 2009
| vol.evidence
52 | no.from
5 | 492-9
Citation: HAN Z Z, BI Y C. Reading comprehension
without
Further
a Chinese aphasic individual. Sci China Ser C-Life Sci,
2009, 52(5): 492-499, doi: 10.1007/s11427-009-0048-x

diation hypothesis of semantic access in visual word
recognition.
Subsequent studies have challenged this hypothesis.
For example, using a broader semantic category in the
semantic judgment task (e.g., living thing?) than the
ones used in Van Orden et al.’s experiments[8−9], Jared
and Seidenberg[16] failed to find the homophone semantic priming effect. Jared & Seidenberg reasoned that the
homophone effect observed by Van Orden and colleagues was, at least in part, due to the subjects’ strategic
factors. In the experiments where narrow semantic
categories were used, the phonological representation of
their members might be pre-activated, and the phonological activation of homophonic words comes from the
semantics system (the narrow semantic category) but not
from the orthographic input, so successful reading comprehension is possible without phonological knowledge.
The most challenging evidence to the phonological-mediation semantic access hypothesis comes from
the studies of brain-damaged individuals. The logic is
straightforward. If phonological mediation is mandatory
in the semantic access of visual word comprehension, it
might be expected that impairment in retrieving the
phonological representation might result in poor performance concerning word comprehension. Contrary to
this prediction, some patients have demonstrated strong
reading comprehension ability in spite of phonological
output deficits [17−21]. For instance, EA[21] showed nearly
intact ability in visual and auditory word comprehension,
as revealed by strong performance on the tasks including
lexical decision, written synonym matching, the Pyramids and Palm Trees Test[22], and sentence-picture
matching (90%−99%, correct). By contrast, he was severely impaired and made predominantly semantic errors in tasks involving phonological retrievals, e.g., oral
picture naming, oral reading, and writing to dictation
tasks (0%−48%). He was also not able to read nonwords
or write to the dictation of nonwords (0%), indicating
severe disruption of the nonlexical grapheme-phoneme
conversion (GPC) procedure. EA’s performance showed
that written word recognition is possible without the
generation of phonology, supporting the direct access
hypothesis.
An important issue in understanding the lexical access processing in reading is whether the process is universal or language-specific. While the studies reviewed
above were all conducted in alphabetic languages (e.g.,
English), other languages differ with regard to vis-

ual-sound and visual-meaning mappings. For instance,
in Chinese, a typical logographic language, written
characters have virtually no grapheme-phoneme-correspondence on a segmental level, yet often contain components which provide cues for the semantic properties.
Therefore, it is of theoretical importance whether the
phonological mediation plays a less crucial role in semantic access in written Chinese. Nevertheless, it has
been repetitively observed in chronometric studies that
there is early and automatic phonological activation in
written Chinese word recognition tasks, and that the activated phonological information plays a critical mediating role in accessing meaning[23−28]. For instance, using a backward masking paradigm with Manderin
speakers, Tan and Perfetti[27] found that a homophone of
a target’s synonym facilitated target identification if the
prime had few homophones but not when the prime had
many homophones. The direct synonym of the target
produced a greater priming effect than the homophone
of the synonym in short SOA conditions (across 129 −
243 ms) and only the synonym priming effect persisted
under the SOA 500 ms condition. These results, while
compelling, were not replicated by subsequent studies[29−32]. In a recent study, Chen et al.[33] did not also
observe the homophone-mediated semantic priming effect in the recognition task of Japanese Kanji, a logographic script.
Paralleling the aphasic study of alphabetic languages,
Zhou et al.[34] reported a Chinese-speaking individual
who speaks to the target issue, LY, who had relatively
preserved visual word comprehension ability yet showed
poor performance on oral production tasks where she
mostly made phonological errors. The authors therefore
argued that the profile of LY provided counter-evidence
for the phonological-mediation semantic access hypothesis in Chinese. However, the data of that case was
rather inclusive. First, LY’s visual comprehension was
indeed impaired, making errors with relatively easy
word comprehension tasks (e.g., association decision for
visually presented word pairs, 32/50; auditory wordpicture matching, 14/17). She also presented deficits in
the oral word repetition task (43/52). In this case, the
pattern of LY’s impairments seemed jumbled, and she
might have been subject to multiple impairments.
Therefore LY’s data might be insufficient to challenge a
phonologically mediated semantic access hypothesis.
In this paper, we present a Chinese case, YGA, who
was studied in greater detail and indeed constitutes evi-

HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

493

dence which directly addresses this issue. YGA showed
a clear deficit in phonological retrieval tasks such as oral
picture naming and oral word reading. Nevertheless, he
was perfect at visual word comprehension and oral repetition tasks, suggesting the existence of a direct access
route of semantic representation from visual input.

2 Case background
YGA is a 58-year-old, right-handed, Mandarin-speaking
man with a middle school education. He had been the
manager of an auto service company before he retired
three years ago. Ten days before be admitted to a hospital, he was suddenly not able to read the newspaper, and
had occasional sharp pain on the top of left brain hemisphere. He had no dizziness or nausea. His family
members reported that five days later he had some
paraphrasias in speech, without symptoms of headache,
nausea, limb paralysis, dysarthria or ataxia. A brain
magnetic resonance imaging (MRI) scan on the day of
hospitalization revealed that YGA had an infarct in the
left posterior superior and middle temporal lobe (Figure
1). He was diagnosed with anomic aphasia according to
the Chinese adaptation[35] of the Western Aphasia Battery[36].

Figure 1 MRI scan of YGA

YGA’s spontaneous speech was fragmented with
word finding difficulties. He produced the following to
describe the Cookie theft picture from the Boston Diagnostic Aphasia Examination[37]: “At home, children, his
mother in, in kitchen, prepare, just that, wipe stuff, son
and daughter go onto window, cabinet to take food. At
home, at, there is big cabinet, small closet, that is about
it. (Where is it?). /chui2fang21/ (a nonword phonologi1

cally similar to kitchen, /chu2fang2/), room to make
food stuff, wash clothes. (在家里头, 孩子他妈, 在, 在
那个厨房, 准备, 就是那个, 擦东西, 儿子和闺女就
去上窗, 柜子上拿吃的, 在屋里头, 在, 有大柜柜, 小
柜橱, 差不多了. (问: 这发生在哪里?) chui2 房, 做吃
的房间, 洗衣服的房间). We started testing him about
one month post onset of his stroke using the
BNU-CNLAB language screening battery (see Bi, Han,
Shu, & Caramazza[38]). He had no errors on the
Bucco-facial apraxia task (15/15 correct), and only one
error in the oral repetition (39/40), had a mildly reduced
auditory digital span (forward: 5, backward: 3). He
showed intact auditory comprehension abilities (auditory
lexical discrimination: 40/40; auditory lexical decision:
19/20; auditory word/picture matching: 50/50), and was
also spared concerning visual comprehension tasks
(visual word/picture matching: 50/50; visual sentence
completion where he needed to choose one of four
printed words to complete a visually presented sentence:
20/20). He had difficulty with oral picture naming
(35/68) and oral word reading (47/57). He was poor at
written picture naming (3/10) and writing to dictation
(2/10), although he could perform direct-copying well
(10/10). His main error types in oral picture naming
were semantic substitutions (16/33, e.g., 拖 鞋
(/tuo1xie2/, slipper) → 脚 (/jiao3/, foot); 扫 (/sao3/,
sweep) → 擦 (/ca1/, wipe) and phonological errors
(13/33, e.g., /shu4/ (树, tree) → /lu4/; /shui3bei1/ (水杯,
cup) → /shui3pei1/). Other errors included unrelated
lexical errors (1/33, 跳绳 (/tiao4sheng2/, rope skipping)
→ 相片 (/xiang4pian1/, photo) and no responses (3/33).
In word reading-aloud, all of his erroneous responses
were phonologically related to the targets (8/10, e.g., 祝
(/zhu4/, wish) → /chu4/; 签 (/qian1/, signature) →
/jian1/) except for the “don’t know” response trials
(2/10).
YGA’s good performance on comprehension tasks in
the screening test showed that his semantic knowledge
was relatively spared. The prevalence of semantic and
phonological errors in oral picture naming and word
reading suggests that his deficit was due to the phonological (output) lexicon or the access to it. Critically,
despite this impairment of accessing phonological lexical representations, his understanding of visual words

The number denotes one of four tones in Chinese.

494

HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

did not seem to be affected. However, one may argue
that the picture-word matching task used in probing the
comprehension abilities should have been relatively
easier, as the subject could complete the task by either
successfully identifying the target or rejecting the foil,
and therefore the tasks were not sensitive enough to detect mild semantic impairment.
Hence, in experimental study, we investigated the
reading comprehension and phonological access abilities
of YGA using a much larger set of stimuli and more
sensitive comprehension tasks.

3 Experimental study
To closely examine YGA’s reading comprehension and
phonological production ability we presented the following five tasks using the same stimuli set[39]: visual
word/picture verification, auditory word/picture verification, oral picture naming, oral repetition, and oral
word reading. In three experimental sections we studied
these relevant functions by analyzing YGA’s overall
performances and the contributing variables. Note that a
further set of stimuli was constructed for the reading
task due to specific nature of the reading process (see
3.3).
3.1 Experiment a: visual and auditory word comprehension
The two verification tasks were constructed using 162
pictures from the Chinese adaptation of the Snodgrass
and Vanderwart set[40]. Each of the critical items (e.g.,
tiger) was presented in three separate blocks, once along
with the target word (e.g., 虎 (/hu3/, tiger)), once with
a semantically related word (豹 (/bao4/, leopard)) and
once with a phonologically and/or orthographically related word (鼓 (/gu3/, drum)). The inclusion of these
related foils ensures the sensitivity of the task in detecting mild semantic, phonological, or orthographic impairments in word comprehension. The three words associating with each picture were assigned into three
blocks using the Latin-square design. In the visual
word/picture verification the word was printed below the
picture on the same sheet and in the auditory version the
word was spoken by the experimenter. The subject
needed to say “yes” or “no” whether the word matched
the picture. An item was scored as correct only if he was
correct in all three blocks, i.e., correctly accepting the
target and rejecting the two foils. This task is found to be
more sensitive to word-picture matching because here

the subjects cannot make correct decisions based solely
on the knowledge of a foil[41]. Given that visual word
comprehension is the focus here, we specifically compared YGA’s visual word verification performance to the
corresponding control data. The available control data
(data unpublished) were based on the performance of 15
college students at Beijing Normal University (Mean
age: 23).
The first complete responses of controls and YGA’s
were scored. YGA’s correct percentages in all tasks are
given in Figure 2. His performance on the two
word-picture verification tasks had few errors. In the
visual verification task, there was even a non-significant
trend that YGA (96%) performed better than the controls
(mean: 94%, range: 86%−98%) (P = 0.31). The statistical program developed by Crawford & Garthwaite[42]
was used here for this comparison. Therefore, YGA’s
semantic knowledge, at least for the concrete objects
that we assessed, was preserved and the semantic access
from visual input was spared.
3.2 Experiment b: oral production I (oral picture
naming and oral word repetition)
Both of the two oral production tasks used the stimuli of
all 232 items from the Chinese adaptation of the
Snodgrass and Vanderwart set[40]. YGA showed evident
impairment in oral word picture naming (Figure 2).
Given that he made no errors in oral repetition of the
same items, his difficulties were not due to a peripheral
oral production deficit. As shown in Table 1, in oral picture naming, he made semantic errors and phonological
errors.
We carried out a multiple logistic regression analysis
to elucidate the variables that influenced YGA’s naming
performances. The accuracy of the responses was treated
as a dependent variable (1 for correct and 0 for incorrect). The independent variables included word length
(number of syllables), name consistency, image consistency, familiarity, visual complexity[39], and log frequency of words[43]. Using the “enter” method, we found
that word frequency was the only significant predictor of
his oral picture naming accuracy (P = 0.02). This result
further confirms our deduction that YGA was impaired
at the stage of retrieving phonological (output) lexical
representation for oral production. This is because, it is
generally agreed that familiarity is a variable affecting
semantic/conceptual representations and word frequency
is a variable playing a role at the stage of lexical access[44−48].

HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

495

YGA’s semantic errors in oral picture naming, despite
having spared semantic knowledge, may be attributed to
the deficit in accessing phonological output lexicon. As
firstly articulated in Caramazza and Hillis[49], semantic
errors in production may either arise from the semantic
level or from the lexical retrieval stage. This is because
the semantic representation of a target during oral production activates a cohort of semantically related lexical
nodes in phonological output lexicon. In a normal situation, the target phonological representation is most
highly activated, reaching the selection threshold and is
produced correctly. If the target lexical representation is
not available due to brain damage, however, a semantic
related word is then most likely to be selected, leading to
the occurrence of semantic errors.
3.3 Experiment c: oral production II (reading words
aloud)
Material Set 1: The same set of items used in Experiment b was presented to YGA and he was asked to read
the words aloud. YGA had impairment of this task (Figure 2), although his performance was better than oral
picture naming of the same items (χ2(1) = 51, P <
0.00001). In oral reading, most of his errors were phonological (Table 1). We also conducted a similar multiple logistic regression analysis to understand the nature
of his reading difficulty. The dependent variable was the
reading accuracy of a particular item. The independent
variables included word length, familiarity[39], and log
lexical frequency[43]. No variables were observed to significantly predict his reading accuracy (P >0.25).
The lack of lexical frequency effect in YGA’s reading
might be due to the existence of a reading pathway that
bypasses the phonological lexicon. It is widely accepted
that at least two routes underlie the reading aloud process – the semantic route and the nonlexical grapheme-phoneme conversion route[50,51]. In Chinese the
nonlexical conversion process might be achieved by
means of the phonetic radicals. A majority of Chinese
characters are composed of a semantic radical that provides cues of meaning and a phonetic radical that gives
cues of sound. In the case of so-called regular characters,
where the phonetic radical is pronounced the same to the
whole character, successful reading aloud of the whole
characters may be deduced from the phonetic radical
without the involvement of the target phonological lexical representation. Because the stimuli set we used contained a mixture of regular and irregular words, the effect of the lexical frequency of entire words might be
496

shadowed.
Material set 2: Given that the current issue is whether
or not YGA is impaired with accessing the phonological
(output) lexicon and at least partly underlies the reading-aloud difficulty, we designed a further set of stimuli
that manipulated both the lexical frequency and the
phonetic regularity of characters. A 2 (word frequency:
high, low) × 2 (phonetic regularity: regular, irregular)
design was used and forty items were selected for each
cell. We observed that the reading of regular characters
was not affected by frequency (34/40 vs. 30/40, χ2(1)
< 1), whereas concerning irregular characters, high frequency characters were read significantly more accurately than low frequency ones (35/40 vs. 23/40, χ2(1) <
9.03, P <0.003). Thus, the impairment in the phonological lexicon indeed played a role in YGA’s reading impairment.

Figure 2 YGA’s correct percentages across tasks in the Experiment
Table 1 The error distributions in YGA’s picture naming and word
reading performances (number of errors in parentheses)
Error type

Picture naming

Word reading

Semantic error

52% (62)

4% (2)

Phonological similar error

32% (38)

93% (42)

Visual similar error

2% (2)

0% (0)

Neologism

2% (2)

2% (1)

Unrelated errors

1% (1)

0% (0)

No response

12% (14)

0% (0)

Total

100% (119)

100% (45)

4 Discussion
We present a clear profile showing dissociation between
visual word comprehension and visual word reading/oral
production in a Chinese case. YGA performed normally
in a variety of visual and auditory word comprehension

HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

tasks, including word/picture matching, sentence/picture
matching, sentence completion, and word/picture verification. Critically, in the visual verification task which
has proven to be sensitive to the breakdown of the semantic system[41], YGA’s performance was even higher
than the control subjects who were of a higher education
level and younger age than him. However, he had difficulty reading aloud the same visual words which he had
correctly comprehended. He also made semantic errors
and phonological errors in oral picture naming for the
same items. Furthermore, he presented a word frequency
effect in these oral production tasks including picture
naming and word reading aloud. Together these results
demonstrated that YGA’s deficit was localized to the
access of the phonological output representation and/or
the phonological representation itself.
YGA’s error-free visual word comprehension ability
despite his phonological output lexicon (or access to it)
deficit challenges the phonological mediation semantic
access hypothesis for reading. This hypothesis assumes
that successful reading comprehension is mediated by
phonological access and hence predicts that semantic
access in reading will be compromised due to phonological processing deficit. The proponents of this theory
might argue that YGA achieves an error-free performance task via partially preserved phonological representation, which was impaired in oral reading aloud and
picture naming, yet preserved enough to sustain accurate
semantic access. However, in our word comprehension
test, foils were semantically or orthographically/phonologically related to the target. It is unlikely that the
partial phonological knowledge would allow YGA to
correctly refuse the foils in this task but lead to errors in
production tasks.
YGA’s performance is readily explained by the direct
access hypothesis, which assumes that there is a direct
connection from the orthographic input lexicon to the
semantic system. If we assume that the brain-damage in
YGA disrupted access to the phonological output lexicon and spared the direct pathway from the orthographic
input lexicon to semantic representation, YGA’s behavioral patterns may be accounted for (e.g., Coltheart et al.,
1980[12], 1997[13]; see also some connectionist theories

for a similar notion[52−54]).
In the following sections we discuss the implications
of this conclusion in a more general framework of lexical access in reading firstly by addressing the following
two aspects of YGA’s performances. First, why did YGA
show superior performance on oral reading over oral
picture naming? If the phonological representation was
impaired, why did not it affect oral reading and oral picture naming to the same degree? This discrepancy in the
two tasks may be attributed to the classical dual route
model of reading aloud[13], which assumes a nonlexical
grapheme-phoneme conversion route to generate phonology from visual input and a semantic route that goes
through the semantic representation of words. The output of these two routes may also integrate with each
other to optimize the production accuracy and to prevent
semantic errors in reading aloud[50,51,55,56]. In oral picture
naming, by contrast, the only processing pathway is the
semantic route – activation from the semantic system
activates corresponding the phonological representation
and there is no additional compensation mechanisms
available.
The other issue is why YGA was intact in auditory
word comprehension and oral word repetition while he
had difficulty in accessing the phonological output lexicon information, as shown by impaired performance in
oral picture naming and reading aloud. The current consensus of the lexical system includes two functionally
independent phonological lexicons: the phonological
input lexicon and the phonological output lexicon,
which are respectively involved in auditory comprehension and oral production[57]. A great number of cases
have been reported of having selective impairment to
one of these two lexicons with the other spared[17-21, 49,
58−61]
. Oral repetition is accomplished by either retrieving the phonological output representation of the to-berepeated words, or through a nonlexical route by converting phonemic input directly to phonemic output,
bypassing the phonological output lexicon. Indeed, it is
not uncommon for brain-damaged patients (e.g., patients
with anomic aphasia) to be able to repeat words (and
nonwords) normally in spite of several deficit concern-

2

Another caveat is that the majority of the experimental stimuli in the present study are concrete nouns and one may argue that YGA’s deficits are confined
to concrete nouns, but not to other words including abstract nouns or verbs. We examined this possibility by asking YGA to read aloud a set of 99 monosyllabic words including 1/3 concrete nouns (e.g., 碗, bowl), 1/3 abstract nouns (e.g., 祸, disaster) and 1/3 abstract verbs (e.g., 借, borrow), which were
［50］
). YGA’s reading
matched on various linguistic variables such as word frequency, number of strokes, and phonetic regularity (see details in Bi, et al.
2
accuracy on these tasks were 23/33, 28/33 and 25/33. There was no significant difference among the three types of words (χ (2) = 2.15, P =0.34), suggesting that YGA’s reading aloud was not affected by grammatical class.
HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

497

ing phonological lexical processing[62].
One caveat2 to consider here is that we have granted
that the phonological mediation semantic access hypothesis assumes that the phonological mediation happens by going through the phonological output lexicon
because there is no phonological input available in the
visual word stimulus, and it is the same phonological
output lexicon which underlies the oral production
processes such as picture naming and reading aloud.
Nevertheless, as discussed earlier, it is a generally accepted view that the phonological input process and the
phonological output process are functionally autonomous and are therefore independently damaged[17−21, 49,
58−61]
. It has been posited that there is a direct connection
from the orthographic input lexicon to the phonological
input lexicon, and the phonological input lexicon may
serve as a mediating component for semantic access[19].
1

Such a theory would indeed accommodate our data because YGA had good performance on auditory word
comprehension and lexical tasks showing that his phonological input lexicon was preserved and might mediate successful semantic access. However, the premise of
the theory – the direct connection between the orthographic input lexicon and phonological input lexicon –
is not intrinsic to the lexical representation framework
and would require support from independent motivation
and evidence.
In conclusion, the case presented here parallels those
cases reported in alphabetic languages. They together
support a universal principal that successful reading
comprehension does not rely on phonological mediation,
but rather is achieved through a direct connection between the orthographic input lexicon and semantic representations[21].

Frost R. Toward a strong phonological theory of visual word recognition: True issues and false trails. Psychol Bull, 1998, 123: 71—99

2

Gough P B. One second of reading. In: Kavangh , J P, Mattingly I G.

3

Lukatela G, Turvey M T. Phonological access of the lexicon: Evi-

Routledge & Kegan Paul, 1980. 197—226
13

reliant upon phonological representation. Cogn Neuropsychol. 1997,

(Eds.). Language by Eye by Ear. Cambridge: MA: MIT Press, 1972

4

14: 167—175
14

data and a preliminary model. J Verbal Learn Verbal Behav, 1975, 14:

Psychol: Learn, 1991, 17: 951—966

17—29

Lukatela G, Turvey M T. Similar attentional, frequency, and associa-

15

1971

Lukatela G, Turvey M T. Visual lexical access is initially phono-

16

logical: I. Evidence from associative priming by words, homophones,
Lukatela G, Turvey M T. Visual lexical access is initially phono-

358—394
17

logical: 2. Evidence from phonological priming by homophones and

8

128—174
18

processing: A case study in cognitive neuropsychology. Cognition,

1971, 10: 645—657

1983, 15: 111—144

Van Orden G C. A rows is a rose: spelling, sound, and reading. Mem

19

pairment. Cogn Neuropsychol, 1997, 14: 3—33
20

word recognition. New York: Lawrence Earlbaum, 1991. 77—103

Further neuropsychological observations. In Ellis A W. (Ed.), Normality and pathology in cognitive function. London: Academic Press,

Learn, 1988, 14: 371—386

1982
21

Van Orden G C, Pennington B F, Stone G O. Word identification in
Rev, 1990, 97: 488—522

498

Neuropsychol, 1997, 14: 105—129
22

Coltheart M. Reading, phonological recoding and deep dyslexia. In
Coltheart M, Patterson K E, Marshall J. (eds), Deep Dyslexia. London:

Shelton J R, Weinrich M. Further evidence of a dissociation between
phonological and orthographic output lexicons: A case study. Cogn

reading and the promise of subsymbolic psycholinguistics. Psychol
12

Patterson K E. The relation between reading and phonological coding:

Van Orden G C, Johnston J C, Hale B L. Word identification in
reading proceeds from spelling to sound to meaning. J Exp Psychol:

11

Hanley J R, McDonnell V. Are reading and spelling phonologically
mediated? Evidence from a patient with a speech production im-

Van Orden G C. Phonologic mediation in fundamental to reading. In
Besner D, Humphreys G W. (Eds.). Basic processes in reading: Visual

10

Ellis A W, Miller D, Sin G. Wernicke’s aphasia and normal language

recoding in visual word recognition. J Verbal Learn Verbal Behav,

Cogniti, 1987, 15: 181—198
9

Caramazza A, Berndt R S, Basili A G. The selective impairment of
phonological processing: A case study. Brain and Language, 1983, 18:

pseudohomophones. J Exp Psychol: Gen, 1994 , 123: 331—353
Rubenstein H, Lewis S S, Rubenstein M A. Evidence for phonemic

Jared D, Seidenberg M S. Does word identification proceed from
spelling to sound to meaning? J Exp Psychol Gen, 1991, 120:

and pseudohomophones. J Exp Psychol: Gen. 1994, 123: 107—128

7

Smith F. Understanding reading: A psycholinguistic analysis of
reading and learning to read. NewYork: Holt, Rinehart & Winston,

1993, 19: 166—178

6

Johnson N F. On the function of letters in word identification: Some

dence from associative priming with pseudohomophones, J Exp

tive effects for pseudohomophones and words. J Exp Psychol: Human,
5

Coltheart M, Coltheart V. Reading comprehension is not exclusively

Howard D, Patterson K E. The Pyramids and Palm Test. Themes
Valley Test Company, Bury St Edmunds, 1992

23

Perfetti C A, Liu Y, Tan L H. The lexical constituency model: Some

HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

implications of research on Chinese for general theories of reading.

43

Psychol Rev, 2005, 112: 43—59
24

nese). Beijing: Beijing Language Institute Press Christopher Hogan,

Perfetti C A, Tan L H. The time-course of graphic, phonological, and
semantic activation in Chinese character identification. J Exp Psychol:

1986. (in Chinese)
44

Learn, 1998, 24: 101—118
25

921—950
45

Spinks J A, Liu Y, Perfetti C A, et al. Reading Chinese characters for
B1—11

form. J Exp Psychol: Learn, 1994, 20: 824—843
46

Tan L H, Perfetti C A. Visual Chinese character recognition: does
phonological information mediate access to meaning? J Mem Lang,

47

Wingfield A. Perceptual and response hierarchies in object identify-

48

Wingfield A. Effects of frequency on identification and naming of

49

Caramazza A, Hillis A E. Where do semantic errors come from?

cation. Acta Psychologica, 1967, 26: 216—226

Tan L H, Perfetti C A. Phonological codes as early sources of constraint in Chinese word identification: A review of current discoveries
and theoretical accounts. Reading Writing, 1998, 10: 165—200

29

objects. Am J Psychol, 1968, 18: 226—234

Chen H C, Shu H. Lexical activation during recognition of Chinese
characters: Evidence against early phonological activation. Psychono

Cortex, 1990, 26: 95—122
50

Bull Rev, 2001, 8: 511—518
30

ropsychologia, 2007, 45: 2660—2673
51

Psychol, 2000, 42: 65—86

Brain lang, 2005, 95: 235—236
52

the 18th Annual Conference of the Cognitive Science Society NJ:

logical processes. Psychol Rev, 2004, 111(3): 662—720

32

Zhou X, Marslen-Wilson W. Phonology, orthography, and semantic

33

Chen H C, Yamauchi T, Tamaoka K, et al. Homophonic and semantic

53

activation in reading Chinese. J Mem Lang, 1999, 39: 579—606
priming of Japanese kanji words: A time course study. Psychon Bull

New York: Academic Press, 1995
54

1996, 22: 48—62
55

Gao S, Wang Y, Shi S, et al. Aphasia. Beijing: the Union Press of
1993

deficit. Brain Lang, 1991, 40: 106—144
56

semantic and sublexical phonological information in accessing lexical

Corp, 1982

representations for spoken output. Cogn Neuropsychol, 1995, 12:

Goodglass H, Kaplan E. The assessment of aphasia and related dis-

38

Bi Y, Han Z, Shu H, et al. A. Nouns, verbs, objects, actions, and the

39

Shu H, Cheng Y, Zhang H. Name agreement, familiarity, image

187—227
57

16: 5—27
58

agreement, and visual complexity for 235 pictures. Acta Psychol

nemic level. Brain, 1982, 105: 271—300
59

Caramazza A, Hillis A E. Lexical organization of nouns and verbs in

60

Franklin S, Turner J, Lambon R M A, et al. A distinctive case of word

Norms of name agreement, image agreement, familiarity, and visual
complexity. J Exp Psychol: Hum Learn, 1980, 6: 174—215

the brain. Nature, 1991, 349: 788—790

Breese E L, Hillis A E. Auditory comprehension: Is multiple choice
really good enough? Brain Lang, 2004, 89: 3—8

meaning deafness? Cogn Neuropsychol, 1996, 13: 1139—1162
61

Crawford J R, Garthwaite P H. Comparison of a single case to a control or normative sample in neuropsychology: Development of a
Bayesian approach. Cogn Neuropsychol, 2007, 24: 343—372

Auerbach S H, Allard T, Naeser M, et al. Pure word deafness:
Analysis of a case with bilateral lesions and a defect at the prepho-

Sinica, 1989, 21: 389—396 (in Chinese)
Snodgrass J G, Vanderwart M. A standardized set of 260 pictures:

Shelton J R, Caramazza A. Deficits in lexical and semantic processing:
Implications for models of normal language. Psychon Bull Rev, 1999,

animate/inanimate effect. Cogn Neuropsychol, 2007, 24: 485—504

42

Hillis A E, Caramazza A. Converging evidence for the interaction of

Kertesz A. Western aphasia battery. San Antonio, TX: Psychological

orders. Philadelphia: Lea & Febiger, 1972

41

Hillis A E, Caramazza A. Mechanisms for accessing lexical representations for output: Evidence from a category-specific semantic

Beijing Medical University - the Beijing Union Medical College,

40

Seidenberg M S, Petersen A, MacDonald M C, et al. Pseudohomophone effects and models of word recognition. J Exp Psychol: Learn,

Zhou X, Bai X, Shu H, et al. Non-semantic anomia: A case study of a
Chinese patient. Psychol Sci, 1999, 4: 289—292 (in Chinese)

37

Seidenberg M S. Visual word recognition: An overview, In Eimas P,
Miller J L. (Eds.). Handbook of perception and cognition:Language.

Rev, 2007, 14: 64—69

36

Harm M W, Seidenberg M S. Computing the meanings of words in
reading: Cooperative division of labor between visual and phono-

Erlbaum, 1996. 714—719

35

Han Z Z, Bi Y C, Shu H, et al., The interaction between semantic and
sublexical routes in reading: Converging evidence from Chinese.

Zhou X, Marslen-Wilson W. Direct visual access is the only way to
access the Chinese mental lexicon. In Cottrell G. (Ed.). Proceedings of

34

Bi Y, Han Z, Weekes B, et al., The interaction between semantic and
the non-semantic systems in reading: Evidence from Chinese. Neu-

Wu J T, Chen H C. Evaluating semantic priming and homophonic
priming in recognition and naming of Chinese characters. Chinese J

31

Levelt W J M, Roelofs A, Meyer A S. A theory of lexical access in
speech production. Behav Brain Sci, 1999, 22: 1—75

1997, 37: 41—57
28

Jescheniak J D, Levelt W J M. Word frequency effects in speech
production: Retrieval of syntactic information and of phonological

meaning: The role of phonological information. Cognition, 2000, 76:
27

Howard D, Gatehouse C. Distinguishing semantic and lexical word
retrieval deficits in people with aphasia. Aphasiology, 2006, 20:

Perfetti C A, Zhang S. Phonological processes in reading Chinese
characters. J Exp Psychol: Learn, 1991, 17: 633—643

26

Xiandai Hanyu Pinlu Cidian (Frequency dictionary of modern Chi-

Gazzaniga M S, Glass A A, Sarno M T, et al. Pure word deafness and
hemispheric dynamics: A case history. Cortex, 1973, 9: 136—143

62

Bi Y, Han Z, Shu H. Compound frequency effect in word production:
Evidence from anomia. Brain Lang, 130: 55—56

HAN ZaiZhu et al. Sci China Ser C-Life Sci | May. 2009 | vol. 52 | no. 5 | 492-499

499