"IIIWIIHHIIHI illlflll‘illililiil'l “W

H659
131293 00686 7661

i

I y *0 ' ‘ "'3. 1",? v
I ,f , . 3
J 2.: Q. .- 'v {n 9
~ 9_r‘u. .-"~-a§~'.'h

‘ Jf . ~ 5: ' " 4' lulu

‘ 9" V ,. ~ , - 4.1:.

I: 2, ‘. ) , _ :“.-f

I 1%” L I“ h ) d

V‘ f

This is to certify that the

dissertation entitled

CHINESE ADULT READERS: A PSYCHOLINGUISTIC
AND TRANSACTIONAL STUDY OF THE
READING PROCESS IN CHINESE,

WITH COMPARISON TO ENGLISH

presented by

Su-O Lin Tien

has been accepted towards fulfillment
of the requirements for

P11. D. degree in English

CM

Major professor

 

 

Date December 13, 1982

 

MS U is an Affirmative Action/Equal Opportunity Imn'mn‘on 0-12771

PLACE ll RETURN BOX to mwothb mum your "cord.
TO AVOID FINES Mum on or baton duo duo.

DATE DUE DATE—DUE DATE DUE

 

i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU loAnNflnndlvo MM Opportunity Institution

Copyright by
Su-O Lin TIEN
1983

C‘s/.32)

fl 0. Wax

CHINESE ADULT READERS: A PSYCHOLINGUISTIC
AND TRANSACTIONAL STUDY OF THE
READING PROCESS IN CHINESE,

WITH COMPARISON TO
ENGLISH

By

Su-O Lin Tien

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment
of the requirements

of the degree of

DOCTOR OF PHILOSOPHY

Department of English
1983

ABSTRACT
CHINESE ADULT READERS: A PSYCHOLINGUISTIC
AND TRANSACTIONAL STUDY OF THE
READING PROCESS IN CHINESE,

WITH COMPARISON TO
ENGLISH

By
Su-O Lin Tien

This study examined and compared the reading behavior
of readers of English and readers of Chinese using two
Western theories of reading, Kenneth Goodman's psycholin-
guistic model and Louise Rosenblatt's transactional theory
of the literary work, to determine if the processes in-
volved in reading an alphabetical and a non-alphabetical
writing system are the same or different. Two empirical
research projects were set up for this purpose.

Study 1, utilizing the framework of the Goodman/Burke

Reading Miscue Inventory (1972), investigated how readers

 

used their phonological, syntactic and semantic systems
while reading. Ten adult Chinese readers were selected
from students at Mishigan State University and asked to
read LU Syun's short story "Medicine" in Chinese. Their
taped readings were analyzed according to the RMI proce-
dures and then compared to the findings reported by Blair

(1977), Jensen (1972) and Devine (1980).

Study II focused on the experiential aspects of the

Su-O Lin Tien

reading process. It involved a quantitative analysis of
ten readers' response statements using the Purves-Beach
(1972) coding system, and a qualitative analysis of five
individual readers' protocols. Rosenblatt's transactional
theory served as a guiding principle for examining the
readers' performances. The same readers served as subjects
in both studies. In Study II, the readers were asked to
comment on what they read both during and after their read-
ing and the results were analyzed in comparison to relevant
studies done with English readers (e.g. Pollock, 1972).
Some conclusions reached in Study I and Study II are:
(1) There are no fundamental differences between readers
of English and Chinese as far as the basic reading process
is concerned.
(2) The Chinese differ from their English counterparts in
that their substitution miscues show less graphic and sound
similarity to the text items.
(3) The miscues of Chinese readers with high graphic simi-
larity tend to show high semantic acceptability scores.
(4) Chinese readers produced a set of miscues which are
high both in graphic similarity and semantic acceptability,
but very low in sound similarity.
(5) Perception is the dominant mode of response made by
'Chinese readers, followed by Interpretation, Self-involve-
ment and Evaluation.
(6) Qualitative analysis of individual readers' protocols

demonstrates that reading a literary text is an unique

Su-O Lin Tien

experience for each individual reader. It is also clear
that each of these unique personal reading experiences is
activated by the symbols of the one single text.

(7) The readers respond not only to the verbal signs
afforded by the text but also to their own evocation
activated by the verbal stimuli.

(8) The distinction between aesthetic and efferent reading
is usually a question of degree, rather than an absolute
difference.

(9) By combining the Goodman's psycholinguistic model of
reading and Rosenblatt's transactional theory of the
literary work, we find a productive way to reconstruct the
process of how a reader makes use of different levels of
information and combines such information in constituting

a literary work.

Dedicated to

Wei-hsin, John and Feng
with thanks and love.

ii

ACKNOWLEDGMENTS

I wish to express my deepest gratitude to a number of
people who have provided the essential help I needed to
initiate and to bring to completion this project.

Professor James Stalker, Chair of my doctoral commit-
tee, provided.much valuable counsel throughout my program
and contributed helpful suggestions and critical comments
in both the planning and writing stages of my dissertation.

Professor Alan Hollingsworth, Co-chair of my committee,
first made me aware of the professional contributions I
could make with my competencies in languages, linguistics
and literatures.

Professor Nancy Ainsworth aided me with her knowledge
of research procedures and reporting.

Professor Paul Munsell provided intellectual stimula-
tion and generous encouragement.

Professor John Yunck gave me helpful advice and
guidance in other areas of my doctoral program.

The ten Chinese students who participated in this
project shared with me their unique and valuable reading
experiences.

Ms. Reade Dornan, Dr. Jo Hollingsworth, Ms. Tamara
Sitterly and Ms. Syou-ming Wei helped me verify my data
analysis and codings.

Ms. Lois Hetzer tirelessly and patiently typed and

retyped this manuscript during many evenings and weekends.

iii

TABLE OF CONTENTS

Page
List of Tables ...................................... vi
List of Figures ..................................... vii
CHAPTER I: INTRODUCTION
Purpose of the Study ........................... l
A Brief Description of the Research Design ..... 9
A Brief Account of the Chinese Language ,,,,,,,, 11
CHAPTER II: THE PSYCHOLINGUISTIC VIEW OF READING
The Psycholinguistic View of Reading ........... 20
Miscue Analysis ................................ 33
Related Research ............................... 39
CHAPTER III: RESULTS OF STUDY I: THE INTERPLAY OF
READERS' LANGUAGE SYSTEMS
Design of the Studies and Analytic Instrument,, 46
Graphophonemic Cueing System ................... 49
Syntactic and Semantic Cueing Systems .......... 78
Grammatical Function ........................ 79
Syntactic Acceptability ..................... 84
Miscue Types and Levels of Structure ........ 91
Semantic Acceptability ...................... 113
Comprehending ............................... 122
Summary of the Findings ..................... 123
CHAPTER IV: THE TRANSACTIONAL THEORY OF READING
The Transactional Theory of Reading ............ 130
Content Analysis ............................... 144
Related Research ............................... 151
CHAPTER V: RESULTS OF THE STUDY II: FIVE
READERS' EXPERIENCES OF A LITERARY TEXT
Design of the Study and Analytic Instrument.... 158

iv

Presentation and Discussion of Results:

 

Quantitative Analysis of the Coding Data ...... 165
Presentation and Discussion of Results:
Responses of Individuals ...................... 177
Summary of the Findings ....................... 223
CHAPTER VI: CONCLUSIONS
The Relationship Between a Reader and a Text.. 228
Suggestions for Further Study ................. 238
BIBLIOGRAPHY ....................................... 240-
APPENDICES
APPENDIX A: Reading Miscue Inventory:
(Evaluation Form) ................ 249
APPENDIX B: An English Version of the
Reading Text: "Medicine" ........ 254

Table

1.

U'l-I-‘UJN

\DmVG

10.

11.
12.
13.
14.

15.

16.
17.
18.

19.

20.

LIST OF TABLES

Page
Graphic Similarity (Present Study) ........ 54
Sound Similarity (Present Study) .......... 55
Graphic Similarity (Blair Study) .......... 57
Sound Similarity (Blair Study) ............ 58
Graphic Similarity and Semantic
Acceptability (Present Study) ............. 67
Grammatical Function (Present Study) ...... 82
Grammatical Function (Blair Study) ........ 83
Syntactic Acceptability (Present Study)... 88
Syntactic Acceptability (Blair Study) ..... 89
Type of Miscue at the Bound Morpheme
Level ..................................... 104
Type of Miscue at the Word Level .......... 105
Type of Miscue at the Phrase Level ........ 106
Type of Miscue at the Clause Level ........ 107
Syntactic and Semantic Acceptability
in Z (Present Study) ...................... 118
Syntactic and Semantic Acceptability
in Z (Blair Study) ........................ 119
Comprehending Score (Present Study) ....... 124
Comprehending Score (Blair Study) ......... 125

Mean Scores of Responses
(Present Study) ........................... 166

Numbers and Percentages of Responses
Made After Each Division for Each
Category .................................. 172

Five Readers on Aesthetic/Efferent Rank... 219

vi

LIST OF FIGURES

Page
Figure 1. Graphic Similarity Mean Scores ............ 60
2. Sound Similarity Mean Scores .............. 61
3. A.Comparison of Miscue Type at
the WOrd Level ............................ 112
4. A Comparison of Syntactic and
Semantic Acceptability in Z ............... 120
5. A Comparison of Patterns of
Response of Pollock's American
Readers and the Chinese Readers
of the Present Study ...................... 170
6. The DevelOpment of Responses .............. 174

vii

CHAPTER I
INTRODUCTION

Purpose of the Study

In The Meeting of East and West, F. S. C. Northrop

 

argues persuasively that the East and the West could mutu-
ally benefit, perhaps even to the degree of helping each
other save civilization, by increased contact of an intel-
lectual kind. Each could learn something from the other
because the two are not saying the same thing. Each
expresses different yet complementary things, both of which
are necessary for an adequate understanding on a world

basis of man's personal and social fulfillment. As NorthrOp
put it, "the Orient, for the most part, has investigated
things in their aesthetic component; the Occident has inves-
tigated . . . things in their theoretical component" (1966,
p. 375). Northrop's implied suggestion that theoretical
concepts are not quite at home in the East is perhaps rele-
vant only to the last four or five hundred years of Chinese
history, the period of the Renaissance in the modern world.
There can be no doubt, however, that Northrop's desire to
bring about significant mutual understanding between East
and West is certainly praiseworthy, for the history of the
relationship is troubled by a number of deep and repeated

misconceptions despite the good intentions of many people.

Perhaps no single aspect of Chinese civilization has

been more misunderstood by the West than the Chinese
language in its written form. Some two hundred and sixty
years before Northrop published his book, the great
philOSOpher Gottfried Wilhelm Leibniz expressed similar
hopeful sentiments about the potential value of East
and West to each other:

If we are their equals in the industrial

arts, and ahead of them in contemplative

sciences, certainly they surpass us (though

it is almost shameful to admit this) in

practical philosophy, that is, in the precepts

of ethics and politics adapted to the present

life and use of mortals (Leibniz, 1957, p. 69).
Yet, along with such praise of China, Leibniz exhibited
some of the misunderstanding of the Chinese language and
writing system that had already become a tradition in the
Western World, going back at least to several of "the

Jesuits of Canton,‘ who believed that China got its writing
system, and much else, from the West (Derrida, 1976,

p. 330). Leibniz seemed to believe, or to want to believe,
that the Chinese writing system was a one-time invention by
one person and was, therefore, an artificial language (1949,
p. 287). He was apparently ready to believe that the Chi-
nese writing system grew out of ancient Chinese symbols
called trigrams that were themselves derived from ancient
Hebrew, possibly imparted during a supposed visit by Hebrew
Patriarchs to China in the remote past (1977, p. 23).
According to Father Joachim Bouvet, who spent the years

1688-1697 in China and who, on his return to Europe,

corresponded with Leibniz, the Chinese got their original

writing system by imitating Hebrew characters, but then at
some later time forgot that origin (Leibniz, 1977, p. 22).
Much has been written about the fascination in the
Western world in the seventeenth and eighteenth centuries
with the Chinese writing system, as philosophers sought for
the principles with which to construct an ideal universal
language. James Knowlson in Universal Language Schemes in

England and France (1975),Pau1 Cornelius in Language in

 

 

Seventeenth- and Early Eighteenth-Century Imaginary Voyages

(1965), and Jacques Derrida in Of Grammatology (1976) are

 

among the many who have written about the Western obsession

with the Chinese character script in the search for "the

I! H '1

real character, the universal characters, that was in-
tended to provide Western philosophers with the kind of
point-by-point, item-by-item system of naming and describing
in writing the contents of the universe that they supposed
the Chinese writing system offered the scholars of Asia.
"The concept of Chinese writing . . . functioned as a sort
of European hallucination," Derrida writes (1976, p. 80).
What Derrida calls "the Chinese prejudice" brought West-
erners to "hyperbolic admiration" of the Chinese script,
westerners as early as Francis Bacon and as recent as
Ezra Pound.

In Knowlson we read about Bishop Francis Godwin's
seventeenth century fantasy, The Man in the Moone or A

Discourse of a Voyage Thither by Domingo Gonsales, the

Speedy Messenger, in which the hero flies to the moon

helped by birds), learns that moon people speak a lan-

guage of "tunes,' and returns to the earth but lands in
China, where he is able to learn the language quickly,

since Chinese people also speak in tunes, each tune being
associated in some way with a Western alphabetical letter
(1975, p. 117).

The misunderstandings in the West of the Chinese lan-
guage, especially in its written form, are too numerous to
survey in even a preliminary way and are not the subject of
this dissertation. They are, however, evidence of the need
to clarify the differences and similarities that exist be-
tween the two writing systems, the Chinese character-script
system and the Western alphabetical system. Such an inves-
tigation inevitably involves the process of reading, for,
of course, writing systems came into being for the purpose
of being read. I

Undoubtedly the most famous modern example of the West-
ern tradition of misunderstanding the nature of the Chinese
writing system is that manifested by Ezra Pound and his
mentor, Ernest F. Fenollosa. Fenollosa admired Chinese
characters and character combinations as intensely as the
most devoted seventeenth-century enthusiast, but Fenollosa
turned the Western tradition of misunderstanding the Chinese
script upside down. The earlier enthusiasts had supposed
that the Chinese writing system was an essentially sci-
entific way of naming and describing each separate piece

of the universe, that each character directly represented

nature, and was thus a "real" character. Fenollosa took an
exactly opposite position; he believed that the Chinese lan-
guage as embodied in its writing system was essentially
creative and poetic, far more so than any Western alpha-
betical language could possibly be. In The Chinese Written

Character as a Medium for Poetry Fenollosa writes:

 

I have alleged all this because it enables me to

show clearly why I believe that the Chinese written
language has not only absorbed the poetic substance
of nature and built with it a second work of metaphor,
but has, through its very pictorial visibility, been
able to retain its original creative poetry with far
more vigor and vividness than any phonetic tongue
(1936, p. 24).

Western languages, because encased in a phonetic writing
system open to what Fenollosa believes to be the abuse and
erosion of speech, suffer by comparison with Chinese which,
because its written character remains the same despite
whatever pronunciation changes take place, retains its
original, primitive, metaphoric richness.

In this Chinese shows its advantage. Its etymology
is constantly visible. It retains the creative
impulse and process, visible and at work. After
thousands of years the lines of metaphoric advance
are still shown, and in many cases actually retained
in the meaning. Thus a word, instead of growing
gradually poorer and poorer as with us, becomes
richer and still more rich from age to age, almost
consciously luminous. Its uses in national phi-
losophy and history, in biography and in poetry,
throw about it a nimbus of meanings. These centre
about the graphic symbol . . . . Their ideographs
are like blood-stained battle-flags to an old
campaigner (1936, p. 25).

Chinese symbols are visual manifestations of an original
ancient metaphor, says Fenollosa, but he goes on to admit

that these ancient visual metaphors are nowadays forgotten

and no longer traceable. He rejects the very possibility
that in some cases perhaps they never existed.

As Hugh Kenner has shown in The Pound Era (1971,

 

p. 204), when Ezra Pound worked from Fendllosa's notebooks
in producing Cathay, he compounded Fenollosa's mistakes with
his own. Even with these mistakes, however, Kenner be-
lieves Pound's translations are better than Arthur Waley's,
though Waley knew Chinese and Pound did not (p. 195). But
it cannot be argued that Pound and Fenollosa were corrected
in their beliefs about the nature of the Chinese writing
system. Both rejected the possibility that the meaning of
some Chinese characters or combinations might be at least
in part phonetically, not visually, determined. Fenollosa
and Pound were still under that spell which the Chinese
writing system has cast over many Europeans since the seven-
teenth century--what Derrida calls "the Chinese prejudice."
The Chinese writing system does, indeed, offer traps
for the Westerner, not only for the unwary amateur but even
for the careful scholar, as John K. Fairbanks, the eminent
historian of China, has pointed out.
If asked, "Do you read Chinese?" one answers, "No,
only some kinds of Chinese." This is because the
Chinese writing system has a different vocabulary
of special terms for each special branch of learn-
ing or literature, yet each vocabulary may use the
same characters, which are therefore laden with
ambiguity, with the result that the special vocab-
ularies have to exist in the mind of the beholder.
They are not self-evident in the script . . . . Many
weaknesses stem from this difficulty of the writing
system. Reading widely in the styles of classical

Chinese is quite beyond the capacity of most of us
(1976, p. 222).

The difficulty of the writing system affects not only
Western scholars but native speakers as well.

As students of Chinese will tell you, the problem

in reading the language is not to recognize the

characters or even to find them in a dictionary

but to decide what shade of various possible

meanings they bring to the particular passage or

context in which they appear. A Chinese character

is an onion of many layers. The trick is to

decide which one you want to use--verbal,

nominative, or modifying; literal root meaning

or extended abstract meaning (1976. p. 241).

Given the misunderstanding on the part of the West about

the Chinese language in its written form, and given the real
difficulties of some written Chinese for many Chinese
speakers, we can legitimately wonder if the reading of the
Chinese script is a process different from that employed

in reading the alphabetical languages of the West.

This dissertation is designed to examine the reading
behavior of Chinese readers using two Western theories of
the reading process in an effort to answer the question of
whether the reading processes these two writing systems
involve are basically different or the same. The Western
theories under discussion are Kenneth Goodman's psycho-
linguistic theory of reading and Louise Rosenblatt's
transactional theory of the literary work. The present
dissertation appears to be the first application of these
two theories of reading to a non-alphabetical language.

In the past ten years, sixty major research projects in

miscue analysis have been conducted within the framework

of Goodman's psycholinguistic model. However, these studies

are centered on the oral behavior of subjects whose written

languages are alphabetical. Except for the present study,
apparently there has been no miscue research in non-
alphabetical writing systems, such as Chinese. Chinese is
a logographic writing system, the sound-spelling correspond-
ence of which is not so obvious as in an alphabetic orthog-
raphy. In this respect, to study how a Chinese reader
makes use of his language system to arrive at meaning may
shed some light on our understanding of the relationship
between sound and spelling in the English language, a
matter of controversy that has existed in the United States
for over sixty years. A

In her 1979 book Education and Popular Literacy in

 

Ching China, Evelyn Rawski shows that even in eighteenth

 

and nineteenth century China, a broad range of the male
populace, including both peasants and urban dwellers,
learned to read and write. Although the Chinese writing
system consists of hundreds of thousands of different
characters and character combinations, the Chinese have
maintained a high rate of literacy, according to Rawski.
If true, this is a fact of some importance in the matter of
reading, East and West, for it has been widely assumed in
the West that the Chinese writing system has kept the
people illiterate. It is both interesting and instructive
to apply the Goodmans' miscue analysis and Rosenblatt's
transactional theory of reading to a language which is as

utterly non-Western as Chinese. In accomplishing this

objective the procedure of this dissertation is two-fold.

In the first place, this study examines how Chinese readers
make use of their language system in the process of reading.
Secondly, it investigates how Chinese readers respond to
the reading of a short story, during and after their read-
ing of the text. On the basis of the evidence provided,
this dissertation asserts that THE READING PROCESS IN CHI-
NESE AND THE READING PROCESS IN ENGLISH ARE ESSENTIALLY THE
SAME PROCESS DESPITE THE DIFFERENCES IN THE WRITING SYSTEMS.

A Brief Description of the Research Design
Study I
The first study, Study 1, is primarily designed to
provide some preliminary answers to the question of how
Chinese readers make use of their underlying language
systems in the process of reading. Study I duplicates the
research design originally developed by Yetta Goodman and

Carolyn Burke and presented in their 1972 book, Reading Mis-

 

cue Inventory Manual: Procedures for Diagnosis and Evalu-

 

agign (RMI). The Goodman and Burke RMI procedures were
based on the psycholinguistic model of reading originated
by Kenneth Goodman and other psycholinguists who are inter-
ested in the reading process. The basic assumption of this
reading model is that reading is systematic and rule-
governed behavior which involves the interplay of all the
parts of the whole language system. According to the Good-
mans and other researchers, by looking at the overt oral re-

sponses of the readers in question, the researcher should

10

be able to discover patterns which describe and explain the
product and the process of reading as well. There is much
research in reading miscues done with English speakers which
supports the Goodmans' argument. The present study is the
first attempt to apply this psycholinguistic model of
reading to the readers of Chinese to test whether there is

a universal reading process as proposed by K. Goodman
1973c). Ten mature adult readers of Chinese were used as

the subjects of this study.

Study 11

Study II makes use of Rosenblatt's transactional the-
ory of the literary work to answer the following questions:

(1) What kinds of responses do Chinese adult readers
make to a short story?

(2) Is there any evidence which indicates that reading
in Chinese involves simultaneously the reader and
the text? If there is, how, where and when does
this evidence appear?

(3) Does the unique nature of the Chinese writing
system.have any special and obserable effects on
the readers' reading behavior? .

The ten readers who did the oral reading for the RMI in
Study I were the same readers for this Study II. These ten
readers were asked to report any feeling, or thought, that
came to their minds both during and after their reading.
After they had completed the reading of the whole story,
they were asked to comment on any additional feelings or
thoughts about the story as a whole. The responses made by

the readers were recorded. All of the data thus collected

were then subjected to a content analysis, a system

ll

originally developed by Purves and Rippere as presented in

Elements of Writing about a Literary Work: A Study of

 

 

Response to Literature (1968), and later revised by Purves
and Beach (1972). The findings obtained from the content
analysis were further compared to some relevant studies

done in English (e.g., Pollock's 1972 study). In order to
study the Chinese readers' reading behavior in greater depth,
the content analysis of the data was complemented by some

closely-scrutinized case studies.

A Brief Account of the Chinese Language

Before going into a detailed discussion of the data
collected for this research, a brief description of how
Chinese characters are constructed is necessary.

Since Westerners first encountered the Chinese writing
system, they have given it many names. It has often been
called a pictographic writing system. According to people
who hold this particular view, the basic units of the script
are pictographs, which originated as pictures of objects
and still embody those pictures in their present state.
This judgment of the nature of the Chinese writing system
is true only as far as its origin, and perhaps also its
early stages of development, are concerned. Very few
characters can still be seen as pictures today. Another
common belief about Chinese is that it is an ideographic

writing system in which the visual symbols (ideographs)

12

are directly related to "ideas" or abstract concepts of
objects, acts or qualities. The claim that Chinese is
ideographic writing is also valid only to a certain extent,
because the number of Chinese characters that are purely
of this type is also very small.

Chinese characters are constructed in several different
ways. They can be classified into six categories according
to the traditional division of the Chinese writing system
now almost two thousand years old. Only five classes will
be introduced here since the class called T? 31 ijn
ju has never been clear in its nature and there are very
few established examples of characters that can be clearly
assigned to this class. The other five classes are as
follows:

1. Pictographs: Pictographs are characters that are

 

derived directly from pictures of objects. Their pictorial
aspects are not so clear in the present forms, but the
straight-forward pictures of objects can still be seen intflua
primitive shapes, as in the following examples:

Primitive forms Present forms

SUN

EYE

O
MOUNTAIN * LLJ
qfl

l3

HORSE % %
cow .I’ if;

2. Ideogrgphs: Ideographs are characters meant to

 

stand for various abstract concepts, such as acts or events,
as in.)$»shang 'up,"1: sya 'down,'—/-yi 'one,';:.er
'two,'jk_da 'big' (having the connotation of a man looking
big or trying to look big).

3. Compound Ideographs: Compound ideographs are
characters which rely for their meaning on the combination
of the meanings of their parts. One famous example is
Eva ming 'bring,‘ consisting of E3 'sun' and )3 ywe 'moon.‘
Another example isfi wii 'military,‘ consisting of‘fl; jr/
'to stop' and :2, gé 'arms,‘ meaning "a war to end all wars”
j¥1 dung 'east' is said to consist of Ej'sun' rising behind
ch mb 'tree.'

These three types of characters are often thought to
be representative of Chinese writing, but in fact, they
represent types that are used in the formation of only a
small number of the total Chinese character inventory.

This important fact was not realized in the Western
world for almost three hundred years. As a consequence

there has persisted a profound misunderstanding in the

West about the real nature of the Chinese writing system.

14

4. Phonetic loans: Phonetic loans occurred when a

 

character was borrowed from one word for a word which is
pronounced exactly the same way but does not have a written
form. The meanings of the two words are nevertheless, quite
different. For example, the form fig. 151, originally a
pictographic character meaning '(a kind of) grain' began

to be used as a homophonous word for 'come'. This class of
characters clearly demonstrates the error of the notion

that Chinese script never really got past its pictographic
state.

5. Phonetic Compound: In a phonetic compound, a group

 

of characters is phonetic in conception. About nine-tenths
of modern Chinese characters come under this category. In
this class, each character consists of two parts. One part
is a phonetic which suggests the pronunciation, while the
other part, a signific (or radical) will give, in a very
general way, some part of the meaning of the character.

For example,%§7 fahg 'to visit' consists of a sigpific
which means speech, indicating that it has something to do
with speech, and a phonetic 25 fang 'square' serving as a
guide to the pronunciation. With alteration of the
signific, the character can becomejfi fang 'district' or
fangfito spin' , or’% fang 'house' (upper part is the
signific, lower part the phonetic), or28t_fang 'to put'
(left sounding, right signifying). In this example, the
principle of phoneticization seems to work, because the

constant phonetic elementZZi(f§ng) does really give a clue

15

to the pronunciation of the various characters which use it,
if we are willing to disregard the tone. However, tone in
Chinese is phonemic. Changes in tones may result in changes
in meaning. In this regard, as a clue to sound, the
phonetic, at its best, can only suggest the basic sound
itself. There is no indication in the phonetic concerning
which tone to choose.

But the example cited above can be considered a good
one. It is good in the sense that the phoneticja fang
at least gives some indication of how to pronounce the sound
segments of these words. There are cases in which the
phonetic can not even give a reliable clue to the sound
components. For instance, the phonetic/fii.jwéi, in the
characteréfi, 'an awl' is to be pronounced as jw'e'i, but
in the wordfiffii 'only' or 'however' it is pronounced as wéi,
ing]? 'who' as shéi, in'fi/to 'proceed' as jyin, in ifi
'to pile' as dwEi, and in %E_ 'a Chinese family name'
as tswéi, etc. One phonetic may have more than one
possible pronunciation. A reader does not have any access
to the pronunciation unless he or she already knows how
the word should sound. On the basis of this inherent
confusion and the fact that this phonetic-compound class of
characters accounts for about 80-90Z of modern Chinese
characters in the research it might be expected that there
would be some graphic and sound miscues of the Chinese

readers which would come under this category. In fact, a

16

considerable number of miscues of this type were found in
the data analyzed in the present study. A more detailed
discussion of how to deal with this type of miscue is
presented in Chapter III.

The preceding passages present a short description of
how Chinese characters are put together. What now follows
is a brief account of some specific features of the Chinese
language which may have some effects on the Chinese readers'
reading behavior relevant to this study.

Chinese is not an inflectional language. Chinese does
not have number distinctions in its nouns. Since the forms
of nouns are the same for both singular and plural meanings,
the concept of agreement between noun and verb does not ap-
ply. In addition to this, the Chinese verb does not conju-
gate, and has neither tenses nor moods. Instead, Chinese
uses word-order and markers to make structural distinctions
that other languages make by means of inflection. Further-
more, the clear distinctions between verbs, nouns and adjec-
tives which exist in English do not exist in Chinese. In
short, a word in Chinese may serve as a verb, noun or adjec-
tive depending on its immediate context. Sincezthese specific
characteristics of the Chinese language may facilitate the
task of reading, especially in oral reading, it should not be
surprising that the readers of Chinese analyzedixithis study
did not produce many grammatically unacceptable structures

when miscues occurred. The deviations from the text the

17

readers made could change the meaning of the original
sentence or affect the semantic acceptability of the text
words, but as far as the grammaticality was concerned, it
was seldom affected by the misread structure.

The written-language style of modern Chinese seems to
be another feature which was reflected in the patterns of
miscues of the Chinese readers in their oral reading.
Chinese is a morpheme-syllable language in which each
syllable,known.as a character, corresponds closely to a
morpheme. Since the overwhelming majority of Chinese
morphemes are mono-.syllabic, each character is almost always
a syllable. This does not mean that the words of Chinese
are all mono-syllabic. In modern spoken Chinese,even though
the majority of morphemes are still mono-syllabic, words
often contain more than one syllable.

The classical writing style, calledji;ég wényén
'literary 1anguage,‘ represented the norm of writing in
China until the beginning of the twentieth century. It
was the impact of Western civilization that made some
Chinese scholars realize the importance of mass education
and social progress. The main obstacle to social progress
was, according to the revolutionists of the May 4th
Movement in 1919, "the gulf between common speech and
literary usage." They rejected the classical §i_€§ wéhyéh
as a proper means of written communication. As a result,

- —,>; . .
the classical 5L3 we/nyan was abandoned and in its place the

18

standard spoken Chinese was used as a new written language

style. However, in spite of the fact that the classical

\

5k

\ J' / I .
language, many of the phrases ofié wenyan and some of Its

QM!

/ / . .
wenyan was rejected as the norm for written

features are still being used today in writing. Classical
Chinese has been valued by the Chinese people for thousands
of years, and they have been unwilling to turn away from
its vast literary heritage. Most of the educated people
felt that the written language was verbose and redundant
when its style conformed to the style of the spokenlanguage.
Consequently, the spoken language never really became the
norm for written style in China. The accepted new written
style of modern Chinese is a mixture of the new and the old,
a mixture of the spoken language and wenyan. This mixture
has developed not only as a standard for written.language
but also for daily oral language as well. As P. Kratochvil
observed:

The classical wenyan was rejected as a means of

written communication as a whole, but some of its

features have still been utilized in writing, and

through writing, they have also penetrated into

the oral standard (1968, p. 136).
In the study of the Chinese readers' miscues in their oral
reading, examples emerge which demonstrate the fusion of the
classical writing style and standard spoken Chinese as
described by Kratochvil. The general features oftim2Chinese

language presented here will prove usefulixithe presentation

and discussion of the actual data in the subsequent

 

19

chapters, particularly in Chapter III, the analysis of the
miscue data. In Chapters II and IV, two theories of
reading which are utilized as the theoretical foundations
for the two studies in this dissertation along with some
related research are presented. Chapter V presents and
analyzes the Chinese readers' responses to the reading of
a short story. A description of the sample selection and
data analysis as well as a comparison of research done
with readers who are native English speakers are included.
Finally, in Chapter VI a discussion of the relationship
between the text and the reader based on the findings

of the present study along with conclusions and recom-

mendations for future research are proposed.

CHAPTER II
THE PSYCHOLINGUISTIC VIEW OF READING

Two theories of reading are used as the theoretical
basis for investigating the Chinese readers' reading
behavior in this dissertation. The first theory is the
psycholinguistic view of reading developed by Kenneth
Goodman and many other psycholinguists whose research has
led them to the conclusions that reading is an active and
meaning-oriented process. The second theory of reading,
which will be presented in Chapter IV, is the transactional
theory of reading developed by Louise Rosenblatt as a result
of her actual experiences as a reader and as a teacher. In
spite of differences in orientation, these two theories of
reading share some very important insights into the process
of reading, a process which is examined in this dissertation
by studying the reading performance of the Chinese readers.

In his influential and controversial book, The Structure

 

of Scientific Revolutions (1962), Thomas Kuhn challenged the
traditional "accumulation mode" of scientific knowledge.
Kuhn rejected the notion that progress in science is the
result of the accumulative growth of factual discoveries,
improvement in methods and theoretical generalizations.
According to Kuhn, science advances not in a linear but in

an alternating fashion. Kuhn's concepts of the growth of

20

 

21

scientific knowledge might be true for advanced, mature
science, especially for the physical sciences, but in the
field of social science, this would not always be the case.
In the area of reading, for example, what makes the
psycholinguistic model of reading important is not an
outcome of a "sudden flash of insight" but of the continuous,
accumulative contributions of numerous psychologists,
linguists and psycholinguists. The invaluable research in
eye movements conducted about one hundred years ago
certainly contributed to our understanding of the reading
process today. In other words, the existence of the
psycholinguistic view of reading today is the result of the
efforts and contributions of a large number of theorists and
researchers interested in a model that takes both mind and
language into account. The phrase "psycholinguistic view of
reading" defines an area larger than the research of any one
person. That having been said, it must also be made clear
that the version of psycholinguistic reading theory to which
this dissertation is most indebted is that originated by
Kenneth Goodman and developed by him and his associates

through two decades of substantial research.

H

According to Kenneth Goodman, reading is a

phycholinguistic process, one in which the reader functions
as a user of language" (1973, p. 1). In describing this

model of reading he states:

 

 

 

 

22

reading is not a precise process of letter

or word recognition but is in fact a process in

which prediction, selection, and sampling of cues,

and subsequent testing by syntactic and semantic

screens occurs (1973, p. 2).
Several important concepts which are central to Goodman's
miscue analysis and to the psycholinguistic view of reading
are revealed in this statement. Especially crucial are the
three key notions, namely, sampling, prediction and selec-
tion, which are the main focus of the present discussion.

One of the key operations which forms the nucleus of
the Goodman's reading model and the psycholinguistic view
of reading as well, is that reading can be accomplished by
sampling the ”visual information" or "graphic signals,”
without going through the stages of letter or even word
identification. The viewpoint that visual information is
sampled rather than exhaustedly analyzed is extremely
important, because it is one of the fundamental concepts
which distinguishes this model of reading from the structural
or phonic methods. The psycholinguists who claim that
reading does not involve a precise identification of
letters and words are able to draw some evidence from
empirical research to support their arguments. Studies
in eye movements are probably the earliest proof that readers
cannot possibly process all the printed material in a
continuous, sequential way. Classical experiments on eye

movements were conducted by Raymond Cattell in 1885 and 1886.

It was found that within an exposure of lOmsec only three

23

to four unconnected letters could be read, whereas in this
same time two unconnected words could be read and as many
as four connected short words. Furthermore, it was also
found that it took a longer time to name letters than to
name words. Catell concluded from these results that
words were not read letter by letter, but as a whole
(reported in Huey, 1908; Gibson and Levin, 1975). These
results were further confirmed by Erdman and Dodge in 1898.
They observed that readers did not read in a sequential,
step—by-step, linear order. Rather, the eye movements in
reading were quick jumps (saccades) from one fixation to the
next. In addition, the reader's eyes did not always jump
forward. Very often the jumps ahead were interruped by
equally sudden movements back to the lines of the text
that had been scanned (see Huey, 1908; Gibson and Levin,
1975). There are also a number of other sources<xfevidence
supporting the idea that the identification of words need'
not depend on prior letter identification. For example, there
is evidence that word identification is too fast for letter-
by-letter analysis (Pierce and Karlin, 1957; Neisser and
Beller, 1965). Also relevant are the demonstrations made
by Frank Smith (1969) that words can be identified in
conditions under which none of their component letters are
individually discriminable.

Reading units at least as large as words are obvious.

But does a reader have to process every single word?

 

24

According to the psycholinguistic theory of reading, the
answer is negative. Research on the rate of reading
indicates that the average reading rate of college students
is about 300 words per minute. If a reader proceeded by a
serial scan, he would not be able to read as fast as 300
words in a minute (Kolers, 1970). Research on the rate of
reading also shows that there is a correlation between
speed and comprehension (Huey, 1908). The slow readers
who labored for exact understanding of every word and
phrase were usually comprehending less than the faster
readers who skipped words and phrases here and there but
were able to read through a passage rapidly. Such empirical
studies seem to support the claim that reading is not a
linear visual process.

Another strong support for the view that reading is not
a linear process comes from recent linguistic theory. A
major linguist who has had definite influence on the
development of the psycholinguistic view of reading is
Noam Chomsky (1957). His distinction of the surface and
deep structures together with his claim that there is no
simple correspondence between the physical aspects of a
sentence and its meaning provide some theoretical basis
for the psycholinguistic reading model. The statement
made by Frank Smith in the following quotation serves as a
good example. "You cannot simply take a sequence of

sounds or of printed words and immediately derive a meaning

 

 

25

from them, because there are aspects of meaning that are not
present in the spoken or written words" (1973, p. 29).

That is to say, to know how words are joined together to
form a sentence it is necessary to understand the full
meaning of a sentence. ”Grammar is the key to comprehending
language” (Smith, 1971, p. 38). The Chomskian notions of
surface structure and deep structure and of syntax as a
bridge between these two levels of language form the
theoretical foundations for the position that the meaning

of a sentence is not the sum total of the meanings of the
individual words that comprise it. And by implication, if
reading is for meaning, then it cannot be just a simple
process of visual information processing.

The second key notion which characterizes the
psycholinguistic view of reading is the idea that reading
involves a process of prediction. According to Kenneth
Goodman, reading is a "psycholinguistic guessing game" in
which the reader

as a user of language has both syntactic and seman-

tic input to relate to graphic cues and interact with

them. He uses graphic cues, perhaps supported by re-
lated phonological cues, to help predict rammatical
sequences; he uses graphic and grammatica cues to
trigger the search of his memory for related

meaning; and he uses all, in turn, to predict

subsequent input (1972, p. 154).

There are several important assumptions in this notion of
prediction. First, the thesis that a reader is able to

predict language structure is built upon the assumption

that a reader brings into his reading his knowledge of the

 

 

26

oral language. K. Goodman (1970) described the relation
between oral language and reading: ”When a child undertakes
to learn to read at the age of five or six, he is already

a skilled user of language." From his past experience in
acquiring the oral language system a child has already
learned that language is systematic and rule-governed. A
child does not put words together randomly even as early as
the age of two. Instead, he constructs his first two-word
sentences according to a rule, his first grammatical rule.
This first grammatical rule is not found in adult language,
but a child persists in using it, testing it and modifying
it until finally he gets to the adult grammar. This
hypothesis-testing is the fundamental mechanism of a child's
oral language acquisition. The child who reads carries

this hypothesis-testing process together with his knowledge
of the oral language to the reading activity. "His ability
to recreate meaning depends on this ability to associate
those experiences and the concepts he has formed through
them with the language" (Goodman, 1970),

Central to the notion of prediction in reading is the
assumption that language is systematic and rule-governed.
Strongly influenced by Chomskian Transformational Grammar,
the psycholinguists took the view that the grammarcflfEnglish
consists of a set of rules for producing all the grammatical
sentences and only the grammatical sentences. According to
this view, grammar consists of a set of rules for distinguish-

ing the grammatical sentences from the ungrammatical

 

 

 

27

sentences in the language. These rules, known to both the
writer and the reader, are vital sources of the information
that is necessary for effective communication.

English, like all natural languages, is systematic.
There are certain kinds of regularities observed. For
example, a verb tends to follow its subject and to precede
its object; a preposition tends to be followed by its
substantive. And if one reads the phrase no more, he
expects to come upon a Eggp; if one reads however at the
beginning of a sentence, it is a ygp, a gpill, or a

nevertheless that he expects. A noun in a certain

 

position requires a verb in a certain mood and number; in
another position it demands a relative pronoun. Adjectives
call for nouns, and so forth. English words are thoroughly
organized according to a set of general grammatical rules
and logical usages. Furthermore, hundreds of phrases and
sentence structures have occurred so often in speech that
they have a place in the mind and only a slight glimpse is
needed to start the recognition. This kind of syntactic
regularity in language allows readers to predict the
sentence structures that they are reading. In Frank Smith's
judgment syntax has a special function in this regard:
Analysis of meaning at the deep structure level
leads to the analysis of the surface visual structure.
Syntax is a tool that the fluent reader uses to
predict what the surface representation should be,

and he needs only a minimum of visual cues to provide
a confirmation of that prediction. . . (1971, p. 221).

 

 

28

Miller and Isard (1963) have also shown that the
intelligibility of a sentence depends on whether it follows
syntactic and semantic rules. Sentences which break
semantic and syntactic constraints, such as "Tox ims
glerf onmo nerbs” are more difficult to process than
"Thetxncims have glerfed the onmo nerbs," which in turn
is more difficult than ”The bad boys have chased the poor
cats." The Miller and Isard experiment was conducted to
demonstrate that readers do possess a knowledge of grammar
and make use of this knowledge when it is necessary.

According to the psycholinguistic view of reading,
syntactic prediction is a common practice among proficient
readers. When a reader comes to a printed page, he comes
with certain syntactic expectations. If the expectations
fail to be correct, for example if what follows a noun
phrase is not a verb or a verb phrase, he regresses or stops
and reads the line again.

A cognitive process depends to a very great extent
on the human capacity to anticipate elements in the sequence
of experience. Because of the connections and relationships
among incoming elements and of their tendency to cluster
together in predictable ways, it is expected that certain
kinds of things will follow from certain others. For
instance, if a person hears someone speak his name, he
usually looks in the speaker's direction. And he would be

shocked, or at least surprised, if his expectation failed

 
 

,1
\
i
-'\
\\\‘ ‘
\\ ‘
\
-. \
1‘ ‘ ‘ ‘
\\:' ‘\
‘.\
‘ ‘ ‘\ '\‘ It
\ f\ \
k \
. § ‘ _ _
\ ‘ .X ‘ \‘ “ \ Q
\:,.\,
‘ ‘ V‘ \‘ \~
\‘ ‘1, \
\::‘.I‘
\'\{ \\
‘ \ \ \
‘ \ \\‘ \ ‘ ‘\‘\‘
\
\
\‘rs .‘
K \\\ ‘\\t ““:
Q
‘ \
\_ ‘
‘t‘ O \ ‘.
‘ ‘ ‘ 94‘ 33(‘\:\1¢“\:
\S ‘
K‘ ¢ ’
N “01.1: tonsh‘lp
\
\‘\\
\ \
‘ ‘0 cluster
\
3‘;
\\\ers For
\
\
\
\

 

 

29

to be correct. The same kind of capacity is also found in
reading.

In addition to the grammatical knowledge, a reader
also brings with him his knowledge of the world into
reading. A reader not only appears to seek out, either
consciously or unconsciously, the underlying patterns that
govern language as he reads, but he also forms his
prediction of meaning based on his past experience and
conceptual backgrounds. Huey (1908) pointed out in his

The Psychology and Pedagogy of Readipg that "in the second

 

reading nearly one-third of the total time was saved."

What this finding demonstrates is that some prior knowledge
about the subject matter of the written material can be very
helpful both in increasing the understanding of the text and
the reading speed. In reading for comprehension, according
to the psycholinguistic view of reading, there must be a
"point of contact" between what is expected and what is
already known. As Smith (1975) has noted: "Comprehension
means relating new experience to the already known."
Anything that cannot be associated with a personal "theory
of the world” is nonsense, regardless of how meaningful it
might appear to anyone else. Ulric Neisser and other
researchers in cognitive psychology have offered versions
different from, but compatible with, Frank Smith's account

of the theory of the world.

30

One psycholinguist who has greatly emphasized the
importance of the reader's prior knowledge about the
material he is reading is Paul Kolers. In his ”Intro-

duction" to Huey's The Psychology and Pedagogy of Reading

 

Kolers states "what the reader 'finds' in the printed pages
reflects his knowledge, his assumptions and many other
cognitive events that carry him beyond and give significance
to the text itself" (Huey, 1968, p. xxv). Kolers seems to
suggest that what the reader brings to the written material
is as importanteun probably even more important than, the
text itself. It might explain why when a reader is
confronted with totally unfamiliar material, his compre-
hension spirals down. It is not because he does not
recognize the words, but because he is not familiar with
the material. There is no "point of contact" that he can
establish between what he expects to see and what he knows.
K. Goodman :is expressing the same idea when he says "all
readers are illiterate in some senses,since no one can read
everything written in his native language." (1973c, p. 25).

In sum, in order to make the strategy of prediction
possible in reading, one needs to have two kinds of infor-
mation. The reader must have the knowledge of his language,
and the reader must know something about the world in

general.

31

The third key operation in the psycholinguistic view
of the reading process is selection. If reading is not
simply a visual activity, then how does a reader handle
all of the visual display in front of his eyes? Frank Smith
proposed an answer to this question by saying

it does not matter how much information enters

the visual system, a reader can get only four

or five items through the processing bottleneck

into short-term memory. So, he must choose the

four or five items that will best meet his

information needs (1971, p. 94).

According to Frank Smith, it istfimabrain that makes the
decision about what is seen. And what the brain sees
depends on what is known, what is expected to be seen or
what is anticipated (Smith, 1975, pp. 26-27). The brain
is selective. Usually when too much visual information is
exposed to the eyes at once, whatever is beyond the
capacity of the brain to handle is ignored. The brain
selects and processes only that information which is
thought to be crucial and necessary. How precisely a
reader selects his items is unknown because there is no
reliable indicator of how the brain is working while
reading. Nevertheless, the fact that the more familiar
the subject matter of the text is, the faster and easier
a reader can read seems to suggest that the reader's

prior knowledge about the material may play an important

role in the reading process.

32

Yetta Goodman and Carolyn Burke in their 1980 book
Reading Strategies: Focus on Comprehension mention some
factors which might affect readers' choices about which
chunks of information are to be retained. Two of the most
important ones are "the purpose of reading" and'Hfluarelation-
ship of what I am reading to my view of the world" (1980,

p. 8). The selections that a reader will make are always
related to the purpose he sets for himself. If the
information is not important or is irrelevant, according

to Goodman and Burke, the reader will either ignore it or
forget it. As for those chunks of information which are
selected, they will be further integrated into the reader's
belief system or his storehouse of knowledge.

We have discussed the three basic principles which
form the nucleus of the psycholinguistic view of reading
separately. However, it is important to remember that when
a reader is in the act of reading these three key operations
are working together as a unified single process. Indeed,
the psycholinguistic view of reading can be described in
the following way: reading is an active interaction between
the reader and the text (created by the author) as the read-
er strives to get at meaning. The reader, any reader, comes
to a piece of written material with certain kinds of
expectations. The reader, guided by his knowledge of the
language, looks at the text, forms his prediction, and

samples it to see whether the meaning he predicts is the

33

same as that presented by the text. It is when his expec-
tation is frustrated that he readjusts his reading. He
regresses to correct his expectation or resorts to sounding
out letters, words or using any visual information that is
available to him. This process of reading is most vividly
summarized by Kenneth Goodman in the following passage:
the efficient language user takes the most
direct route and touches the fewest bases necessary
to get his goal. He accomplishes this by sampling,
relying on the redundancy of language and his
knowledge of linguistic constraints. He predicts
structures, tests them against the semantic context
which he builds up from the situation and the
ongoipg discourse and then confirms or disconfirms
as he processes further language (1973c, p. 23).
It is on the basis of these premises that the Goodman
Taxonomy and the Reading Miscue Inventory (the RMI),
shorter and simpler version of the taxonomy, were developed
as a means to examine and to provide some concrete evidence

about how a reader processess the printed material before

his eyes.

Miscue Analysis

Miscue Analysis was originally developed by the Good-
mans as a tool to investigate the nature of the reading pro-
cess by analyzing the reading errors produced by readers in
their oral reading. As Kenneth Goodman said, miscue analy-
sis has provided "windows" to look into the nature of the
reading process. The central notion here is that miscues, or

deviations from the text, are not simply random errors on

34

part of the reader, but rather parts of the total reading
strategy. By looking at the overt responses of the readers
in question, patterns which describe and explain the product
as well as the process are discovered. By discovering the
patterns of miscues that characterize a given reader, valu-
able insights into how this reader reads are gained.

Miscue analysis was first reported in a study by
Kenneth Goodman in 1965. Goodman studied the oral reading
of 100 subjects in grades one, two and three in Detroit.
The results revealed that young readers at times regress
in their reading primarily to get at meaning. Goodman
found that children were able to recognize words read in
context that they could not read in isolation. The study
concluded that children are increasingly efficient in
the use of the syntactic and semantic cue systems.

A result of this initial study in miscue analysis

was the first version of the Goodman Taxonomy of Cues and

 

Miscues in Reading (Kenneth Goodman, 1969). Yetta Goodman

 

and Carolyn Burke (1972) designed the Reading Miscue Inven-

tory based on the Goodman Taxonomy. A shorter form of RMI

 

has also been designed by Y. Goodman, Burke and Lindberg
(1974). It is this short version of RMI that is utilized
as an instrument for analyzing the Chinese readers' oral
miscues in the present dissertation (Appendix A).

The techniques and procedures of miscue analysis

research are based on the theoretical assumptions embodied

35

in the psycholinguistic view of reading discussed in the
previous section of this chapter. One of the most fundamen-
tal principles underlying miscue analysis research is the
notion that reading is a systematic and rule-governed
behavior which involves the interplay of all the parts of
the whole language system. This interplay of the whole
language system is described by Kenneth Goodman as follows:
Three kinds of information are available to the reader.
One kind, the graphic information, reaches the reader
visually. The other two, syntactic and semantic infor-
mation, are supplied by the reader as he begins to pro-
cess the visual input. Since the reader's goal is
meaning, he uses as much or as little of each of these

kinds of information as is necessary to get to the

meaning from the structure. In turn his sense of
syntactic structure and meaning make it possible

to predict the graphic input, so he is highly

selective, sampling the print to confirm his

prediction. In reading, what the reader thinks

he sees is partly what he sees, but largely what he

expects to see. As readers become more efficient,

they use less and less graphic input (1973. p. 9).

In order to examine the interplay of the three
language systems one studies the oral miscues of readers.
In such an analysis both the full Taxonomy and the RMI
provide a series of questions to ask about each miscue.
Each miscue has to be examined through all these questions
so that the researcher can determine to what extent and
with how much success a reader can use each of the language
systems available.

The shorter version of the RMI adopted for this study
consists of six categories. The first and second categories

concern the questions of how much a miscue "looks like" and

"sounds like" the expected response. When judging

36

the graphic and sound similarity, the reader's observed
response and the expected response are identified as be-
ginning, middle, and end. If two of the three parts are
similar, a "high" degree of similarity is assigned. If one
of the three parts is similar, a "some" degree of similar-
ity is rated. The miscue will be coded as "none" if none
of their three parts is similar. The third category is
grammatical function. When evaluating this category, the
researcher compares the grammatical function of the miscue
with that of the text word. The miscue and the text

word are rated ”same” if the grammatical functions of

both words are the same; they are rated "questionable" if
the miscue is "so disrupted" that it is impossible to de-
termine whether the grammatical function of the miscue is
the same or different from the grammatical function of the
text word; the miscue and the text word are rated "differ-

ent if the grammatical functions of the two words are
different. The fourth category concerns syntactic accept-
ability. In coding this category, the researcher has to
decide whether the sentence containing the miscue is syn-
tactically acceptable within the context of the story. If
the sentence produced by the reader is syntactically ac-

ceptable, it is marked as "yes,' and the sentence is marked
as "no" if it is not syntactically acceptable in the story.
Semantic acceptability constitutes the fifth category of
this short version of the RMI. In this category, the re-

searcher wants to know if the sentence involving the miscue

37

is semantically acceptable in the story. As in evaluating
syntactic acceptability, the researcher has two choices.
The sentence is marked "yes" if it is semantically accept-

I!

able, and is marked "no if it is not.
The sixth category is comprehending. Evaluation of
this category indicates the degree to which the reader
changes the intended meaning of the author as he reads.
The central question a researcher seeks to answer in
coding this category is "Is there a change in meaning
involved in the sentence?" The researcher is left with
three Options. When the sentence as read by the reader
causes no change in meaning, it is marked as "no." If
there is a ”change, inconsistency or loss of minor inci-
dents, characters or sequences in the story" the sentence
is coded as "minimal." If the sentence read by the
reader involves a major change in terms of meaning or
causes a loss of major incidents, characters or sequences
in the story, the sentence is marked ”yes."
The following is an example of how this coding
system works in evaluating a reader's oral miscues.
TEXT: He hesitantly produced a five—dollar bill.
READER: He hesitantly produced a give-dollar bill.
If the reader produced an observed reSponse "give" for the

expected response of "five," the first three categories

would be coded as follows:

38

 

Reader Text Graphic Sound Gramm. Func.

 

High Same None High Same None Same Quest.Dif.

 

give five X X X

 

 

 

 

 

 

The substitution miscue "give' for the expected response of
"five” changes the text sentence. The grammatical function
of "give” and "five" are obviously different, since "five”

in the original sentence functions as an adjective while the
substituted word "give" can only function as a verb. Conse-
quently, the resulting sentence is neither syntactically“nor
semantically acceptable and constitutes major loss in mean-

ing. Categories four, five, and six involving the sentence in

which substitution miscue occurred would be coded as follows:

 

 

 

Syntactic Semantic Meaning
Accepta- Accepta- Change
bility bility
Yes No Yes No No Mini- Yes
mal
X X X

 

 

 

 

 

 

 

 

 

When the six inventory questions have been answered for all
the miscues which will be coded, percentages are computed

for each vertical column for statistical data. Statistical
counting yields data both for the readers as a group as well

as for the individual reader.

 

39

Related Research

Since 1967 a number of studies have been conducted
within the framework of the Goodmans' miscue analysis.
Despite the fact that these studies focused on the reading
of a wide variety of materials in a wide variety of situa-
tions, the results generally support the Goodmans' hypothe-
sis that reading involves an active reconstruction of a
conceptual representation of the situation described in the
text. When reading takes place, a reader makes use of the
graphic, syntactic and semantic systems of the language
interrelatedly and simultaneously in his search for meaning.

In this section, two groups of research projects were
reviewed. The first group dealt with the readers who are
native speakers of English. The second group are studies
which investigated the reading behavior of the readers
whose first language was not English. Since one of the
primary concerns of this dissertation was to explore the
nature of the reading process in general, the researcher
selected those studies which investigated the reading
process from different perspectives.

Allen (1970) analyzed the oral reading of a group of
readers in grades two, four and six. His study confirmed
the hypothesis that reading involves the interplay of the
graphic, syntactic and semantic information systems. Ac-
cording to Allen, "One of the broadly stated conclusions of

this study is that readers make use of all the available cue

40

systems in the reading process" (Allen, 1970, p. 125). He
found that all of the children in his study seemed to exhibit
a strong awareness of syntax in their reading. Some of the
readers seemed to pay more heed to syntax than to meaning.
All of the three grade levels made substitution miscues

that were totally syntactically acceptable 70 percent of the
time and 90 percent of the time syntactically acceptable to
some degree. Readers at all levels made use of the

strategy of correction, although there appeared to be no
consistent relationship between the number of miscues and
the degree of correction. Finally, Allen pointed out that
phrase level substitutions constituted the largest number

of substitutions at all three levels, which might suggest
that a larger unit of syntax and meaning than the word

level was processed in the activity of reading.

Carlson (1971) examined the oral reading patterns
generated by some average fourth graders reading a basal
reader, science material and social studies material. His
conclusions indicated that readers were able to use grapho-
phonic and syntactic cues as they read all the selections.
However,it appeared that his subjects made better cues in
the reading of basal stories. The ratio of non-correction
was also much higher in the content area material than for
the basal reader.

Yetta Goodman (1971) directed :1 longitudinal study

of four elementary school readers over a period of one year

41

and nine months. All these four readers were black and
represented a relatively wide socio-economic range in
background. The study found that dialect offered no
hindrance for the subjects in learning to read. Her study
also revealed that there were some developmental changes
which occurred as readers grew from beginning reading
toward more proficient reading. In discussing the major
differences between average and low readers Yetta Goodman
concluded that although all readers used the same basic
reading strategies, the average readers seemed to have better
control over when to emphasize one strategy over the others
while using all the language cue systems at the same time.

The effect of the context on comprehension was examined
by Menosky (1972). Her study consisted of eighteen subjects
reading various portions of text. The major finding of
Menosky's study was that miscues change quality as readers
progressed through a given piece of reading material. The
length of a text was crucial, especially for poor readers.
The task of reading became easier as context was developed.
Menosky's study had direct implications for the construction
of reading materials and reading comprehension tests.

In 1972, Jensen completed the first major RMI study
at Michigan State University. In this study, she compared
the oral reading errors of some proficient readers with that
of readers who used reading strategies less effectively. The
major findings were: The proficient readers retained

meaning better than the less proficient readers did; and

42

proficient readers made fewer substitutions but a higher
percentage of omissions, insertions and reversals. Her
explanation for this phenomenon was that the more proficient
readers had "freer use of optional transformations."

Bean (1976) conducted a study to examine the oral
reading behavior of Hawaiian Island Dialect speakers in
grades four, five and six. He concluded that dialect
miscues were mainly phonological in nature and did not
interfere with meanings. Below average readers across
grades made greater use of graphic and sound cues than the
average readers. As a group, sixth graders were more
adept at applying semantic and syntactic information in
reading than the fourth and fifth graders. And the fourth
grade children relied more heavily on graphic-phonic cues
than the two other groups.

A study conducted by Blair (1977) involved a group
of young mature readers and a group of older mature readers.
The assumption that reading is a psycholinguistic process
is further substantiated by the results of this study. All
the subjects in this research used the three cueing systems.
Nevertheless, according to Blair, syntactic information
seemed to be the most important cueing information for both
groups of his readers.

Mabel G. Smith (1978) researched the oral reading
behavior of normal and educable mentally retarded readers.

Her conclusions revealed that the mentally retarded readers

43

exhibited a tendency to make a greater percentage of
meaning—change miscues. Consequently, they produced a
great number of structures which resulted in partial or
complete loss of comprehension. However, it was also
clear that the mentally retarded. readers were able to
utilize the reading strategies suggested by the Goodmans,
though not as efficiently as the normal readers did.

The research studies reviewed above are a few examples
of studies in miscue analysis done within the framework of
a psycholinguistic model. Those studies are centered on
the oral reading behavior of those subjects whose written
and oral language is English. There are also a number of
studies which deal with the readers whose native languages
are other than English.

The miscue study by Romatowski (1973) was the first
to include subjects whose first language was not English.
She compared the reading of three bilingual fifth grade
subjects in both Polish and English. Romatowski reported
that for all of the subjects the influence of the Polish
phonemic system was apparent, however, "In no instance did
the phonemic differences disrupt the syntactic or semantic
organization of the passage" (1973, p. 153). She also
found that the influence of Polish syntax on the retellings
and reading of her subjects resulted in omissions of de-
terminers in their reading and omissions of determiners and

word order changes in their retelling. Like the subjects

44

in Allen's study, the bilingual subjects relied heavily on
the syntactic information. The predictions made by these
bilingual readers in their reading of the English story
were syntactically acceptable for about 65 percent of the
miscues which is somewhat less than the 70 percent level
of syntactically acceptable miscues reported in the Allen
study.

Devine (1980) studied the miscues of fourteen
Spanish-speaking adults reading English. Although
Devine's primary interest is to find whether there are
any "developmental patterns in native English and non-native
reading acquisition" her analysis of data yielded some
strong evidence to confirm the hypothesis proposed by the
Goodmans. "A comparison of the two groups of readers with
respect to their use of the cueing systems in language
show that while there are a few differences in their
behavior, for the most part, their oral reading is remark-
ably similar" (1980, p. 139).

Over the past decade, several dozen research projects
have been carried out by examining the oral miscues of
readers whose native language is either English or another
Western language. The following are examples of the signifi-
cant observations about the reading process made in these
studies:

1. There is but one reading process common to all

of the readers regardless of differences in

language background or varying levels of
proficiency among individual readers.

45

2. Dialect does not result in a disruption of
meaning.

3. All readers make use in varying degrees of the
three cueing systems available to them. Pro-
ficient readers have a better control over the
syntactic and semantic cue systems than the less
proficient readers do.

4. Readers regress primarily to arrive at meaning.

5. The frequency of miscues itself is not a signifi-
cant measure of reading proficiency.

These findings involve readers whose native languages fall
within the alphabetical writing systems of the Western
world. The need for such a study involving readers of a

non-Western, non-alphabetical writing system is obvious.

CHAPTER III

RESULTS OF STUDY 1: THE INTERPLAY OF
READERS' LANGUAGE SYSTEMS

This chapter is divided into two parts. The first
part of this chapter describestfimageneral design of the
Blair study cited in Chapter II and the current study. The
Blair study is the closest one to the present study in
terms of scope and nature; therefore, it was chosen as a
basis for making a comparison between the reading behaviors
of the English readers and the Chinese. The second part
of this chapter discusses the findings of these two studies.
The focus of this section is a comparison of these two
studies made by looking at the oral reading miscues
produced by these two groups of readers. A comparison is
made on the basis of graphic and sound similarity,
grammatical function, syntactic and semantic similarity
and meaning change. In addition to the Blair study,
reference is also made to Devine's and Jensen's study,

both of which were cited in Chapter II.

Design of the Studies and Analytic Instrument

The Blair Study

 

The Blair study was designed to analyze and compare
the oral strategies of twelve young mature readers and

eleven older mature readers selected from reading classes

46

47

at Glendale Community College in Glendale, Arizona. All
of these subjects were English speakers. The readers were
chosen on the basis of ranking on percentile scores of

the Nelson-Denny Standardized Reading Test (1960). Those
subjects all had a reading score comparable to a reading
grade level of 13.

The text selected for the reading was a short story,
”Poison" by Raol Dahl. A pilot study was carried out with
four students who were not in the study to ensure the
material would pose some difficulty without leading to
extreme frustration. The reading of the text usually took
30-40 minutes. The subjects in this study were instructed
to read the material by following the procedures described
in the Goodman and Burke RMI (1972). The first 25 sub-
stitution miscues were coded and analyzed according to the

short version of the RMI as presented in Chapter II.

The Present Study

 

The current investigation was designed primarily to
provide some preliminary answers to the question of how
Chinese readers make use of their underlying language
systems in the process of reading. Two research questions
were posed to guide the investigation.

(1) Are there any significant differences in
the oral reading strategies of the Chinese
readers and English readers on any of the
following categories: (a) Graphic and
Sound similarity, (b) Grammatical function,
(c) Syntactic acceptability, (d) Semantic
acceptability and (e) Meaning change?

48

(2) What is the relationship between meaning
and Chinese print? Can the readers get
at the meaning of a character even if
they assign an "unconventional" sound
to the character?

The present study of this dissertation duplicated the
research design originally developed by Y. Goodman and
Burke in 1972. In the present study, ten adult Chinese
readers were selected from the Chinese students who are
currently living in the East Lansing area. These ten
subjects were asked to read a short story in Chinese
characters (Appendix B). The readers were chosen on the
basis of one criterion: each of the readers came from
Taiwan and had finished a bachelor's degree in one of the
universities in Taiwan. This factor insured that the
readers shared a similar educational background and had
achieved a high level of reading proficiency.

These ten readers were asked to read a selection which
they had never read previously and with which they-were not
at all familiar. The text selected for the present study
was a short story written by Lh Syhn, whose works have been
banned in Taiwan for political reasons. In order to
control the differences in structure and conceptual load of
the material, all of the readers were required to read the
same text. An informal pilot study was conducted in order
to ensure that the text was challenging enough so that it

would elicit sufficient miscues. The actual reading time

used by each reader was about 40 minutes.

49

In general, the present study followed the standard
procedures outlined in the RMI except for the retelling
part. Instead of instructing the readers that they were
required to retell the story in their Own words when they
finished reading the text, the current investigator asked
the readers to make comments on what they had read both
during and after their reading.

By following the RMI procedures, the first 25 miscues
Of each reader were coded for analysis, which yielded a
total of 250 miscues for the ten readers, an amount
sufficient to find general patterns concerning how Chinese
readers make use of their language systems. It seemed to
be more apprOpriate to study the first 25 miscues, which
included various types of miscues, rather than simply to
concentrate on one type Of miscue, as did Blair. The data
were then subjected to statistical computation which yield-
ed percentages for each individual reader as well as for
all readers as a group in the categories of graphic-
phonemic similarity, grammatical function, syntactic-
semantic acceptability and meaning change.

The sections that follow include a presentation and
discussion Of the results Obtained from the analysis of

miscues from the readers of the Blair and the present study.

Graphophonemic Cueing System
In the RMI, the category of graphic similarity is

designed to investigate a reader's effectiveness in

50

processing the visual information available to him. This
category asks the question, how much does the miscue look
like the expected response? When judging the graphic
similarity of the Observed response and the expected
response, each single word or non-word substitution is
broken into three parts-~beginning, middle and end, and
further compared to the text word. Three different ratings
could be assigned depending upon the amount of similarity
among each of them. This system of coding needs some
modification when applied to Chinese to accommodate the
specific features of the Chinese writing system. It is
very difficult,and usually impossible, to break each
Chinese character into three equal portions. This diffi-
cultyixspartly due to the fact that Chinese characters
are constructed by six different principles. Characters
composed by each of these different principles may require
different methods to decompose them. In this regard, it
is extremely hard to set up a single or fixed criterion
to apply to all of the characters. It is true that the
majority of the Chinese characters belong to the category
of phonetic-compounds in which most of the characters
consist of at least two identifiable components. However,
even with this group of phonetic-compound words, we find
the criterion suggested in the RMI is still not directly
applicable. The number of similar parts shared by two

characters is not always a reliable measure to decide

51

how a miscue should resemble the original word in the

text. The following examples should illustrate this point.

Observed Expected Common - Degree Of
miscue response component similarity

(1) 2i f2}: High

 

 

49.—
m it

(2) fi‘ 4; Partial

 

 

 

 

 

 

(”(5% 5 if, E, High
(4) 533‘; ‘72)? > Partial
(5)23‘ 32: None Partial

 

Among these five examples, four pairs are phonetic-compound

characters. Each character consists Of two parts,aisionific

 

__Q_____

and a phonetic. Each character in each pair shares one
identical component with its counterpart (e.g.,£i:_in the
charactersfii andfii; i in the characters fifl and 31% ) .
According to the RMI, all of these four pairs should be
coded as having ”high" graphic similarity since each of
them has the same number of similar parts. However, if we
examine each pair closely, it is clear that they do not
have the same degree Of similarity. The characterii looks
more similar tofii thanjf’?‘ tow—v. Obviously, the

common elementzi1in.the charactersjfit'andzyiiis dominant
in the sense it is larger in size, while in the case Off]?

andjg , the salient component is not the identical
g.

52

component 4" which they both share, but if" in the character
3?, , and E in the characterig. For this reason, the
graphic similarity betweenfili andfiiis high, whilefj’: and
jgi-is only partial. Miscues (3) and (4) are coded by
following the same principle. The miscue (5) is a different
matter. In the example of.z;., andggé, none Of the compo-
nents Ofthese two characters can be considered similar;
nevertheless, taken as a whole, they resemble each other to
a certain degree. As a result, they are coded as having
"partial" graphic similarity.

In coding the graphic similarity in the present study,
several different criteria have been considered depending
on the types of characters involved. Not only was there a
comparison of the components of the characters but also
the visual configuration of the words was taken into
conSideration. In order to eliminate subjective and
arbitrary decisions as much as possible, a reader other than
the researcher herself was asked to code the miscues. The
percentage of agreement was about 9OZ. In the cases where
the two readers disagreed, a third reader was used. The
evaluation of the miscues by the third reader served as the
last resort for coding the problematic miscues. Using more
than one reader to determine the graphic similarity can
make up for the disadvantage of not being able to set up a
more systematic scheme.

In coding the sound similarity, no such complication

as the one with the graphic similarity was found in Chinese

writing. The coding scheme designed by Y. Goodman and
Burke was directly applicable to the Chinese language.
Therefore, no modification was made for the current study.

Table I summarizes the data on graphic similarity in
the present study. The data indicate that 18.2Z of all the
substitution miscues of the Chinese readers contained
"high" graphic similarity and 19Z contained "partial”
similarity while 62.8Z of miscues showed no graphic
resemblance to the original text words.

Table 2 presents the degree of sound similarity
employed by Chinese adult readers. As the figures show,
the majority of the miscues (76.9Z) produced by the
Chinese readers have no similarity in sound sequences to
the expected words in the text. And only about 24Z of
the observed responses contained either high or some
similarity to the original words presented in the text.

This preliminary analysis of the graphic and Sound
similarity of the Chinese readers reveals a number of
general features of the reading of Chinese readers. First,
the data clearly demonstrate that reading requires some
references to the print. Readers made use of the grapho-
phonemic cues which are available to them. Readers produced
miscues that both 1995 like and sgppd like the text words.
On the other hand, the data also indicate that although the
graphophonemic cues are the first information available to
the readers, reading does not operate on this level alone.

Of 106 substitution miscues analyzed in the present study,

54

TABLE 1.

 

GRAPHIC SIMILARITY
CHINESE ADULT READERS

 

 

SUBJECT NoHIGH Z NO?OMB Z N011]ONE Z
1. 2 14 4 29 8 57
2. 2 15 o o 11 85
3. 2 15 3 23 8 62
4. 3 3o 4 4o 3 3o
5. 2 l8 2 l8 7 64
6. 3 33 o o 6 67
7. o o 2 22 '7 78
8, o o 2 29 5 . 71
9. o o o o 6 100
10. 8 57 4 29 2 14
TOTAL 22 21 63
MEAN 2.2 18.2 2.1 19 6.3 62.8

 

55

TABLE 2.

 

SOUND SIMILARITY
CHINESE ADULT READERS

 

 

SUBJECT N011IGH Z No?OME Z N011]ONE Z
1. 1 7 l 7 12 86
2. 0 0 0 0 13 100
3. 0 0 4 31 9 69
4. 1 10 2 20 7 70
5. 0 0 1 9 19 91
6. 1 ll 2 22 . 6 67
7. 1 ll 1 22 6 67
8. l 14 2 29 4 57
9. 0 0 l 17 5 83
10. 0 O 3 21 ll 79

TOTAL 5 18 83

MEAN 0.5 5.3 1.8 17.8 8.3 76.9

 

56

about five-eighths bear no graphic resemblance to the

text and three-fourths bear no resemblance to the text in
sound. The finding seems to suggest that readers also

used sources other than just graphic and sound cues in their
word recognition. Thirdly, the fact that mean scores of the
high and some graphic similarity categories (18.2Z and

l9Z) are higher than that of the same sound similarity
categories (5.3Z and 17.8Z) suggests that these readers

made greater use of visual information than the sound
property of the language in their oral reading.

Visual and sound properties of a language constitute
an indispensable element in the reading process. This is
Obvious in the fact that in all taxonomic and RMI studies
reviewed, readers have shown some dependency on the graphic
and sound information. It is important to keep in mind that
this generalization is true only as far as the fundamental
reading process is concerned. This statement should not be
taken to mean that all readers are using the grapho-phonemic
cues in the same way.

Tables 3 and 4 are taken from Blair's study and
summarize the data on graphic and sound similarity that he
found in his research. Blair reported that 80.7Z of all
miscues produced by American mature readers contain high or
some graphic similarity while 74.5Z of their miscues contain
high or some sound similarity. When these figures are
compared with 37.2Z of all miscues for the Chinese readers

that contain high or some graphic similarity and 23.1Z of

57

TABLE 3.

 

GRAPHIC SIMILARITY

AMERICAN MATURE READERS

 

 

SUBJECT No}?IGH Z NosOME Z N01?ONE Z
l. 6 24 12 48 7 28
2. 12 48 6 24 7 28
3. 13 52 8 32 4 l6
4. 14 56 6 24 5 20
5. 12 48 9 36 4 28
6. 12 48 6 24 7 28
7. 8 32 10 40 7 28
8. 15 60 9 36 l. 4
9. 15 60 6 24 4 16
10. ll 44 10 4O 4 16
ll 10 40 12 48 3 12
TOTAL 128 94 53
MEAN 11.6 46.5 8.5 34.2 4.8 19.3

 

58

TABLE 4.

 

SOUND SIMILARITY
AMERICAN MATURE READERS

 

 

SUBJECT NO§IGH % NoéoME Z NoNONE Z

1. 6 24 8 32 11 44

2. 6 24 14 56 5 20

3. 6 24 15 60 4 16

4. 11 44 8 32 6 24

5. 13 52 1o 40 2 8

6. 5 20 11 44 9 36

7. 6 24 12 48 7 28

8. 11 44 9 36 5 20

9. 13 52 4 16 8‘ 32

1o. 12 48 9 36 4 16

11. 7 28 9 36 9 36
TOTAL 96 109 70

MEAN 8.7 34.9 9.9 39.6 6.3 25.

 

59

all the miscues for the group that contain high or some
sound similarity, the differences between these two
groups of readers immediately become obvious. As shown
in Figures 1 and 2 the Chinese readers recorded the highest
percentage in the category of "None” for both graphic and
sound similarity while the English-speaking readers
seemed to show a very different pattern. The Chinese readers,
in general, seemed to show less dependency on the visual
display than did their English counterparts.

One possible explanation for this high occurrence of
"no similarity" in graphic and sound categories found in the
Chinese readers is that those ten subjects are all mature
and proficient readers. According to the psycholinguistic
view of reading, fluent readers need only minimum visual
cues to get at meaning (e g., K. Goodman, 1973c; F. Smith,
1971). As a result, their reliance on the graphophonemic
information tends to be low. In this regard, it is possible
to attribute the high percentage of no graphic and no
sound similarity in the scores of the Chinese readers to
their high reading proficiency. However, it is important
to note that the readers in the Blair study can also be
considered as fluent readers. Blair points out that
all of his 11 adult readers have a comparable reading
grade level of 13. If high reading proficiency is the only
factor which makes the Chinese readers in the present study
produce miscues with no resemblance to the original text

words both in physical forms and in sound sequence one

60

FIGURE 1.

GRAPHIC SIMILARITY MEAN SCORES

 

HIGH SOME NONE

 

100
90
80
7O
6O

 

50
40
3O
20

 

 

 

 

lO

 

 

 

 

 

 

 

 

 

 

American readers

:>
ll

Chinese readers

61

FIGURE 2.

SOUND SIMILARITY MEAN SCORES

 

HIGH SOME NONE

 

100
90
80

 

7O
6O
50
4O
30
20
10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

American readers

11>
II

Chinese readers

0
ll

would expect to find a similar pattern of performance from
the readers in Blair's study.

Furthermore, the figures reported in Blair's study
seem to correspond closely to the results of other miscue
studies. Devine (1980), for example, reports that 58Z
of the miscues of her high group readers (Spanish-speaking
adults reading English) show high graphic similarity and
52Z of the miscues have high phonemic similarity to the
text words. In her 1972 study, Jensen reveals that more
than half of the total miscues produced by her highly
proficient readers (native speakers of English reading
English) have high or at least moderate similarity to the
expected words of the text. From the findings of these
three studies, it is possible to see a common practice
which characterizes how the readers of English make use
of the visual and sound information in their oral reading.
The common practice tells us that the miscues produced
tend to share some physical properties with the text words.

On the basis of the findings reported in the three
studies cited above, it becomes obvious that the different
patterns produced by the readers of the Blair and the
present study in their use of visual information cannot
be fully explained simply from the point of view of
reading proficiency. Differences regarding the two
language systems involved may offer a better and more

convincing explanation. The high degree of graphic and

63

sound similarity found in the Blair, Devine and Jensen
studies may reflect the relationship between the phonemes
of the English language and the written alphabetic symbols
used to represent them. Although the Written language of
English is not a direct representation of the sounds of the
spoken language, nevertheless, the correspondence between
the oral and written forms of language is still much more
obvious than what one finds in Chinese. In English, graph-
ically similar words also tend to be similar in sound, as
in though and through, better and letter. No such general-
ization can be made for the Chinese language. In Chinese,
characters of similar form can be also similar in sound,
such as 16 £838 and {(5 fa_/r_1g, a% 1181 andg‘m 1351. However,
in most situations characters which look similar are often
distinctive in sound (e.g.;figkpwé andj%E-b§i;égzti§g and
fiifigg). As pointed out in Chapter I, Chinese characters
are constructed in six different ways. Pictographs, for
example, are derived directly from pictures of objects.
Ideographs are characters meant to represent abstract
concepts or ideas. Compound-Ideographs are characters
which rely for their meaning on the combination of the
meanings of their parts. The combination of E3 'sun' and

H 'moon, ' for example, form the word EH 'bright.‘ It is
very clear that these three types of characters are semantic
in conception. When these words were constructed, they
were created to represent meaning rather than sounds. The

sound value of a particular character was determined on the

64

basis of which object or concept the character was to
represent. Of these three types of characters there is no
internal consistency as far as sound system is concerned.
A reader does not have any access to the pronunciation of
a particular word unless he or she already know the
meaning of the word and how it should sound.

Phonetic-Compound is a group of characters which
contain some sound elements in their structures. In this
class, each character consists of two parts. One part
is a phonetic which suggests the pronunciation, while the
other part, a signific, will give some part of the meaning
of the character. However, even within this group, the
correspondence between the oral and written forms of
language is not always direct and obvious. In the first
place, a phonetic may have more than one possible
pronunciation (see discussion on pagesllb15). Secondly,
most of those sound indicators are independent characters
themselves. Unless a reader is already familiar with these
characters, it is not likely that he will be able to figure
out their pronunciation simply by looking at the printed
forms.

In sum, the relation between written and spoken forms
in English is much more direct and obvious, and we find
that readers of English are greatly affected by the "visual
convenience" offered by their written language and thus

do produce more miscues which both look like and sound like

65

the expected responses in the text.

In comparing the studies conducted independently by
Blair, Devine and Jensen with the present study, it seems
that some significant differences exist between the
readers of Chinese and English in terms of the way they
use the graphophonemic cues available to them, which
suggests that forms of written language may have some effect
on how readers process visual information. Readers of
English obtain cues from the visual properties of words to
a much greater extent than Chinese readers do. It is
obvious from the evidence that the majority of miscues
produced by the readers of English contain high or at
least some degree of graphic similarity to the text material.
Deviations from the text made by the Chinese readers, on
the other hand, tend to show no visual or acoustical
relation to the text words.

The emphasis thus far has been on the discussion of
the physical properties which might be shared by an observed
response and the expected response. But is there any
semantic relationship between these two? When a reader
produces a miscue which looks very much like the original
word in the text is he also able to preserve the meaning
of the individual text word? In the investigation of the
reader's efforts to preserve the meaning of the text words
and its relation to graphophonemic similarity, Devine
(1980) reports that there exists a very interesting

semantic relationship between the miscue and the text word.

66

Devine points out for readers at all proficiency levels
"high similarity" graphic miscues tend to have lower than
average semantic acceptability (1980, p. 109). Devine

uses miscues which result in non-words to demonstrate her
observation. "In a taxonomy analysis, non-words are always
coded as semantically unacceptable since the researcher can-
not determine with any certainty, what, if any, semantic
relationship exists between the miscue and the text

word. When non-word miscues are separated from the rest

of the miscue data and coded for graphic and phonemic
proximity, the data indicate that these miscues have a
higher than average degree of similarity to the text
material [@hicé] shows that non-words, which have no
acceptability semantically, more closely resemble and

sound like the text words than miscues in general" (1980,
p. 110). A similar observation is made by Jensen (1972,
pp. 25-6).

When comparing Devine's analysis with that of the
present study with respect to the meaning relationship
between miscues and text words, one finds similarity as
well as difference among these two groups of readers.

The data obtained from readers of Chinese support Devine's
finding that "no similarity" graphic miscues tend to have a
high percentage of semantic acceptability. As has been
mentioned earlier in this section, the Chinese adult

readers have produced a greater number of miscues with no

67

graphic similarity to the text words. Of a total of 106
substitution miscues, 63 of them are coded under this
category. And about 71Z of these 63 "no similarity"
graphic miscues are semantically acceptable in the
context of the story. Only l6Z result in total loss of
meaning.

However, when one looks at the semantic relationship
between miscues with high graphic similarity and the text
words, a somewhat different picture from the one reported
by Devine emerges. According to Devine, miscues with
high graphophonemic proximity, such as non-words, reveal
a high degree of semantic unacceptability. No such
generalization can be drawn on the basis of the data of
the present study.

Table 5 illustrates the distributions of the 106
substitution miscues analyzed by graphic similarity and
their relationship to semantic acceptability. The data
indicate that Chinese adult readers produce a relatively
small number, about 21Z, of miscues closely resembling the

text words.

TABLE 5

 

GRAPHIC SIMILARITY AND SEMANTIC ACCEPTABILITY
CHINESE ADULT READERS

 

 

l
GRAPHIC SIMILARITY SEMANTIC ACCEPTABILITY
YES N0
N0. Z No. Z No. Z
HIGH 22 21 19 86 3 l4
SOME 21 20 15 71 6 29
NONE 63 59 45 71 18 29

 

 

 

 

68

Among the miscues with high graphic similarity, 86Z are
semantically acceptable in the story as opposed to 42Z
in Devine's study. Quite contrary to the finding in the
Devine study, the data suggest that for Chinese readers
reliance on graphic cues, as in those miscues with high
graphic similarity, does not necessarily result in
disruption in meaning.

This disparity between these two groups of readers
is probably attributable to the differences in their
written language systems. To explain the reason for the
high correlation between high graphic similarity and
semantic acceptability, we need to consider the
question of how the Chinese characters are constructed.
The majority of miscues with high graphic similarity in
the present study are characters of phonetic+compounds.
As was mentioned in Chapter 1, characters of the
phonetic-compound group consist of two parts; one part
(the radical or signific) indicates the meaning of the
word in a very general way, and the other part, the
phonetic, gives some clues to the pronunciation of the
character. For example, the symbol; is a radical which
means 'hand; of the hand or having something to do with
the hand.’ Characters comprised of this radical 3’
in general have something to do with the hand. In
modern Chinese, more than three hundred characters have

this hand radical as part of their components.

69

Characters likej‘é pai 'to strike with the hand, ' or 393'
chyan 'to hold; to grasp,‘ or *5 j‘f 'to point with the
finger' are a few examples of this hand radical category.
For a proficient Chinese reader, sometimes a glimpse of
the radical? will give a sufficient clue as to what kind
of words would fit the meaning in the context. In the
present study, there are many miscues which clearly
demonstrate that readers were making their prediction
with the help of the radicals of the text words. For
example, more than one reader substituted the observed
response 32:73 chya 'to pinch; to cut with fingernails'

for the expected response Enye 'to press with fingers.‘
In this example, both the deviation $6,; and the text
wordii contain the hand radical? . In addition,

there is also a reader who produced 5'] ti 'to, separate
bones from meat' instead of the expected response§?j

ts 'to pierce; to stab.‘ One reader miscuedi’v'] shéng
'left over; residue' for the original wordgiq bw6 'to
peel; to make bare.‘ The four characters g1], f'JIjE']
£1} consist of the same radical L] dau, meaning 'knife or
having something to do with a knife.‘ It may become
obvious from the few examples given above that Chinese
characters comprisedwxfan,identical radical may share not
only some physical features among them, but also some
meaning elements. This is probably one of the main reasons
that miscues of the Chinese readers with high graphic

similarity also tend to be highly semantically acceptable.

70

In dealing with the questions of graphic and sound
similarity and their relationship to meaning, the research-
er has to face an interesting problem. Chinese is not a
phonemic writing system, yet there are some phonetic
elements built in some of the Chinese characters,
especially in the group of phonetic-compounds. Almost
every literate Chinese has the experience of being told
by someone who knows how to read Chinese that most of the
Chinese characters consist of an indicator of meaning and
an indicator of sound. It has become a common practice
among beginning Chinese readers as well as proficient ones,
to try to find out the sound indicator of the word and
pronounce it accordingly, when they encounter an unfamiliar
Chinese character. In the present investigation, seven of
the ten readers made at least one miscue in their first
twenty-five coded miscues which can be recognized as
readers' effort to make use of the phonetic element given
by the word. The following are some instances of char-
acters on which more than one reader miscued. The
charactenfifiii§n_'ripped seam; a crack' is a phonetic-com-
pound. It is composed of two parts. The right side of
the character is the radicalé , which originally means
'a line; a connection link.' The left side of the
character is the symbol §E_ digg which is supposed to
serve as a sound indicator to the word. Three readers

‘

in the present study did takeE ding as the clue to the

7l

pronunciation ofé‘fiand produced the observed response
digg while the correct pronunciation is 1&3. In addition
to the charactersgijfl, the symbolfiding is also found
in two other modern Chinese characterSZfiELdigg 'an anchor
(for a ship),' andégfiidigg, 'ingot of gold or silver; a
spindle.‘ For both/EE'and/fi, the phonetic L? ding does
give a reliable clue to their pronunciation, but why this
symbol does not work for the character fig, 183 needs
further study. It is possibly because of historical
reasons, and it is also possibly because of dialectal
influence. However, one thing is certain: when the
readers miscued & ding for the expected response $95!.
1&2, they were not making random errors. They assigned
the sound value to the word on the basis of their
intuitive knowledge of the Chinese phonological rules.
Another instance of a multi-reader miscue is the
characterjgé'. The correct pronunciation of this character
is Eé 'to hold with hand,‘ but several readers read it
as nfi by relying on the phonetic elementfifiLgé, which
constitutes the upper portion of the character??? .
There are at least seven other examples in modern
Chinese which also utilize the symbolfflfiLas part of their
components. These seven characters are'géigé,é§i_gfi,j%;
1.193% 111,451 {11% mi, and’fié‘rfii. Although these seven
words differ in tones, they share with each other the

same basic sound elements g2 suggested by the sound

72

indicator1§1.gfi. Again, it is not clear how and why
the characterzgé is pronounced as Eé instead of gfi. It
might have been at one time that the phonetic part of the
character matched the sound of the character perfectly,
but sound changed over time. Sound—change does not have
the same effect on a writing system like that of English
as it does on the Chinese writing system, because sound
changes can be more readily accommodated by the alphabetical
system than by the character system without necessitating
a major change in the shape of the word. In Chinese,
the sound might change, but the symbol would stay the
same.

Thus, from the examples given above, the reliability
of using the sound indicator of a character as a guide
to its pronunciation seems dubious. On the other hand,
mispronuncing a certain word in Chinese is not necessarily
a strong indication that a person does not know the meaning
of the word. In fact, mispronunciation of certain words
is not uncommon among Chinese. Most Chinese seem to
demonstrate unusual tolerance toward this "unscholarly"
performance. When a person makes an error in his
pronunciation, it might be the case that the person is
able to arrive at the meaning of the word, but just does
not pick out the right pronunciation. Or it is also
possible that the reader already knows the word and

had assigned a constant sound value to the word ever since

73

the first time he learned it by following the phonological
rules of Chinese. Every time he sees the word, he will
read it the same way, even though he does not pronounce
it in the conventional way. In other words, it had
become some kind of "idiolect." In analyzing the data
for this present study, the researcher found that there
were instances of correction. Often, when miscues did not
make sense in the context, readers went back and corrected
it for meaning. Interestingly enough, none of these seven
readers made any attempt to correct their pronunciation
when they produced miscues of this type. One possible
explanation for this phenomenon is that the readers have
sufficient confidence in their pronunciation to accept
what they read when it makes sense to them.

The Chinese writing system has often been praised
for its unifying power, which enables Chinese with
different dialect backgrounds to communicate with each
other. A literate man from.Canton province,for example,
should be able to comprehend a written text by associating
the characters with his dialect as much as a literate man
who speaks Mandarin. In fact, it is still a common
practice among members of the older generation for a
Taiwanese to read written Chinese in the Taiwanese dialect,
for a Cantonese in Cantonese, and for people from Shanghai
in Shanghai accent. Differences in pronunciation do not
seem to interfere with the understanding of the text.

Research in the influence of English dialects on reading

74

also demonstrates dialect does not result in disruption
of meaning (e.g., Y. Goodman, 1971; Bean, 1976). Further-
more, according to the RMI, pronunciation deviations are
acceptable as long as they can be explained within the
sound system of the speaker's dialect.

The situation is somewhat more complicated in the
present study, because it is not simply a question of
dialectal differences. All of the readers in this
study speak Mandarin. Their graphic and sound miscues
cannot be attributed to the differences in regional
dialects. However, in a sense, their graphophonemic
miscues are one kind of dialect--an "idiolect"--since
readers construct them by utilizing some kind of
phonological rules of the Chinese language. In this
regard, to decide whether a certain graphic and sound
miscue should be coded as a non-word as a semantically
acceptable word turnscnn:to be the major analytical
problem.

Two criteria are set up to deal with this problem.
First, when a miscue of a phonetic-compound occurs, all
the possible pronunciations which this particular
phonetic might carry are listed. Then, the reader's
miscue is compared with those sounds. If the miscue
falls within this sound range, then the miscue has
satisfied the first condition to be regarded as the

reader's "idiolect" which would be coded both syntactically

75

and semantically acceptable if it also meets the second
condition.

The second condition that a miscue of the phonetic-
compound character has to meet to be qualified as a
reader's idiolect is that the meaning the reader assigned
to it has to be similar to or at least close to the
original word in the text. To determine whether a
reader is able to retain the meaning of a word when
be assigned a wrong pronunciation to it, Y. Goodman and
Burke have suggested a very useful guideline: "Any
mispronunciations or name changes which the reader has
instituted should be retained by the teacher" in collecting
data (1972, p. 25). The method employed in this study
to check whether a reader is still retaining the meaning
of a character regardless of the mispronunciation is
based on this principle. The researcher examined the
reader's concepts of a word by reproducing the sound
which was just produced by the reader and asked the
reader to give some explanation of the sound.

In coding English readers' miscues, graphic and
sound similarity can be judged independently of, or
prior to, syntactic and semantic acceptability. To
code the miscues of phonetic-compound characters in
Chinese, the four categories must be considered together
before one can make a decision on phono-graphemic

similarity. This coding difference shows that Chinese

76

readers are encouraged by their written symbols to rely
more on syntactic and semantic systems than on the
grapho-phonic system.

It is important to note that the coding device just
described might have some impact on the general patterns
of Chinese readers' reading behavior as compared to that
of readers of English. Some of the miscues which would
be normally coded as non-words in the RMI system are
treated as both syntactically and semantically acceptable,
if they meet the two criteria. The influence of this
modification in methodology will further show up in the
analysis of syntactic and semantic acceptability.

In this study, miscues coded as a reader's idiolect
take up about 15Z of the total single word substitutions.
Since an idiolect has to share the same phonetic with the
text word, it should not be too surprising to find that
about 7 Z of this type of miscue look highly similar
to the text material. However, when judging sound.similarity
of idiolect miscues one finds that only 14Z have high
and 24Z have partial similarity to the expected responses.
These figures clearly demonstrate how unreliable and
problematic it is to use a phonetic as a guide to the
sound of a word. As a matter of fact, the Chinese have
already been aware of this problem for many decades,
and several attempts have been made to reform the writing
system. Unfortunately, there has been a great deal of

confusion as to the goals of reform. Should they make the

77

existing system easier to learn? Or should they create

a new writing system? The use of Pinyinj/nysaia type of
Romanization) in China now is one of the products of this
reform attempt.

In spite of this drawback of the Chinese writing
system, the data obtained from the readers of Chinese do
provide the researcher an excellent avenue to look into
the nature of the relationship between meaning and sound.
That "meaning is independent of sound" is hardly a new
concept among psycholinguists and researchers who are
interested in the reading process, but it seems that
the finding of this present study is significant in the
sense that it is able to provide concrete examples to
support the argument. The Chinese readers seem to be
able to extract meaning from the printed forms even if
they fail to assign "conventional" pronunciation to
the words. Chinese readers produce a set of miscues
which score high both in graphic similarity and semantic
acceptability, but very low in sound similarity. This
factor appears to suggest that Chinese orthography can be
mapped onto meaning at the level of words rather than of
phonemes. Consequently, it is not necessary for a Chinese
reader to translate the written symbols into spoken

language in order to get at meaning in reading.

78

Syntactic and Semantic Cueing Systems

By the age of five and six when a child is about to
read, he has already learned almost all of the fundamental
syntactic structures of the spoken language. This syntactic
knowledge possessed by the child consists of sentence
patterns, function words, inflections and transformational
rules. The knowledge of his oral language systems a
person brings into his reading, according to the
psycholinguistic view of reading, is crucial to the
process of reading, because the written language is
basically the same language as that of oral language. It
consists of a set of rules for producing all the grammatical
sentences and gply the grammatical sentences of the language.
It also consists of a set of rules for distinguishing the
grammatical sentences from the ungrammatical ones in the
language. Those rules, known to both the writer and
the reader, are needed for effective communication. Just
as the reader brings his knowledge of language into
reading, so the writer brings his into writing. "Reading
is a meaningful interaction between the language of the
reader and the language of the author" (Y. Goodman and Burke,
1972, p. 5). But reading is not an exact process. An
efficient reader does not make any attempt to use all of
the visual information available to him. He samples and
selects from the available information on the basis of

what he knows about the language and the world. Since

79

reading is a process of prediction and sampling, there
will at times be miscalculations. The prediction a reader
makes does not always match the ideas or language patterns
which the author presented in the text. These mismatches
may lead to miscues. According to the Goodmans and many
other researchers, if we know how to evaluate the devia-
tions made by a reader from the text, we should be able to
understand the strength and weakness of that particular
reader. In this way we can also learn more about the
nature of the reading process in general.

In addition to the categories of graphic and sound
similarity which concentrate on a reader's use of visual
information, the Goodman Taxonomy and the RMI provide a
series of questions through which a researcher is able
to study how a reader's knowledge of grammar interacts
meaningfully with that of the author in the text. The
three major categories chosen for detailed discussion
in the present study are (l) grammatical function, (2)

syntactic acceptability and (3) semantic acceptability.

Grammatical Function

To evaluate a reader's use of the syntactic cueing
system, a researcher usually begins with an examination
of the grammatical function of the miscues that have
resulted from the deviations from the text. To determine

the grammatical function of a miscue, the researcher

80

compares the observed response with the expected response.
If the miscue has the same grammatical function as that of

the text word, it is coded ”same" as in the following

example' “a 4:225, 3 3%??? 21? 232 621% n] 29675

)3,
a.*é12fé1.$731133.

In this sentence, the reader substitutes the observed

\

\
response e. tsu 'to draw together' for the expected

N \
responseig’: bye 'to limp.‘ The substitution word/Ix";E

 

despite the fact that it does not make sense in the
context of the whole sentence, fills the same grammatical
role as the text word

Amiscue can be judged as "different" if the reader's
production is grammatically different from the text

word. For instance, in the sentence 2K F737 {1: tb$3j

£1m@fi7',§é 945.45. 2 31,15. $144124 Zc‘hé‘ifiifiz 6.7.2

45 .
9. 1 752227 7.29 2272 6 1 "8: [3 2%”; mamas-78372 296:8.

the word;¥£i,b§i_'to cover; to shroud; because of' was

used in place oflflt_pw§"broken; ruined.‘ Although these
two words are highly similar graphically, nevertheless,
they have very distinctive grammatical functions. The
original wordiégz.pw§_'broken; ruined' functions as an
adjective, but the observed responsejpt. pg; is hardly ever
used as an adjective. These two words differ not only in
the grammatical roles they play, but also in meaning. The

result is a nonsense sentence.

81

The third subcategory, in which the grammatical func-
tion of a miscue is regarded as ”questionable,” is provided
for those instances when it becomes impossible for the
researcher to determine the grammatiCal functions of the
miscues confidently. In the present study, one reader pro-
duced an observed response ilé, instead of the expected
response'fig: pyé 'to limp.’ It is not clear to the
researcher exactly what word the reader intended to
produce when the miscue occurred because there are more
than a dozen Chinese characters with the pronunciation of
iyé. Furthermore, the miscue iyé can not be coded as the
reader's "idiolect" for it fails to meet the two
conditions described in the previous section.

Table 6 summarizes the grammatical function data
from the Chinese readers. The data indicate that 89.1Z
of all substitution miscues of the Chinese readers
retain the same grammatical function as the original
text words. Only 5.7Z of the miscues result in a change
of the grammatical function.

Table 7 presents the finding reported from the
Blair study on the category of grammatical function. The
figures show that 87.6Z of the total miscues produced
by Blair's mature readers play exactly the same grammatical
role as the expected response, a finding very similar to
that of the present study. The high percentage of same

grammatical function of the miscues produced by these two

82

TABLE 6.

 

GRAMMATICAL FUNCTION
CHINESE ADULT READERS

 

 

SUBJECT NoéAME % NOOUES.% NODIFF.
1. 12 86 1 7 1
2. 12 92 0 o 1
3. 13 100 o o 0
4. 9 90 O 0 1 10
5. 10 91 1 9 o
6. 8 89 1 11 o
7. 7 78 1 11 1 ll
8. 6 86 0 O 1 14
9. 6 100 0 o 0
10. ll 79 2 14 1
TOTAL 94 6 6
MEAN 9.4 89.1 0.6 5.2 0.6

 

83

TABLE 7.

 

GRAMMATICAL FUNCTION
AMERICAN MATURE READERS

 

 

SUBJECT NOSAME 76 NOQUES . o/o NODIFF . cya
1. 21 84 1 4 3 12
2. 20 8O 3 12 2 8
3. 19 76 0 0 6 24
4. 22 88 1 4 2 8
5. 24 96 1 4 O O
6. 22 88 O 0 3 12
7. 21 84 1 4 3 12
8. 25 100 o O O -O
9. 21 84 o 0 4 16
1O. 23 92 O O 2 8
11. 23 92 O o 2 8
TOTAL 241 7 27
MEAN 21.9 87.6 .6 2.6 2.5 9.8

 

84

groups of readers of very different language backgrounds
suggests, that readers, in general, are very sensitive

to the syntactic features of their languages, regardless
of what kinds of writing systems they are exposed to. It
can also be used as a strong evidence to support the
claim made by the Goodmans and other psycholinguists that
readers bring their syntactic knowledge with them into
reading.

The analysis of the grammatical function provides a
researcher with an important tool with which to evaluate a
reader's strategies in dealing with grammatical structures
of his language. Nevertheless, this analysis, like the
analysis of grapho-phonemic similarity, allows the research-
er to study miscues only on a word-for—word basis. It
should not be too difficult to see the insufficiency and
imcompleteness of studying miscues in isolation. Therefore,
in addition to the grammatical function, the Goodman
taxonomy and the RMI also contain the category of Syntactic
Acceptability. When the two categories are considered
together, a researcher can get an idea of how well a
reader is handling syntactic cues available to him in the

context of sentences and in the context of whole discourse.

Syntactic Acceptability

A reader may produce miscues which preserve the

grammatical function of the text words, but when they are

85

considered along with other elements within the context of
sentences, they become syntactically unacceptable, as in
the following example.

TEXT: 4

(may *9; 277? 59.4. P, fi’fll’é’x J—o

Ta de1 jing shenz, syan dzéi 3jr4 dzai5 yi ge6 bau7 shangS:

(His spirit (interest)2 now only4 is on5 a6 parcel7 (on)8:)

l 3

READER 232652777, 991227; 7% “4'39? 2:

Ta de lljirqgshenz, syan dzéi 3jr4 you5 yi ge6 bau7 shangs.

(H18 spirit (interest)2 now only4 has a6 parcel7 on8.)

l 3 5
In this example, the reader substituted a yo/u 'have; has'
for the expected response2%E2d25i 'to be at or on.‘ Both

the observed response and expected response are verbs. How-
ever, the miscue is grammatically acceptable only with the
portion that comes before it, because the noun phrase *‘7

{Q (a); yi ge bau shang' is a complex place word consisting
of the noun phrasev/gfflg'y'i ge bau and a locator stem .1"—
shang. In Chinese, place words or place word phrases share
the general properties of nominals. They are distinguished
from other nominals by the fact that they occur as objects
of a set of verbs to indicate the general circumstances or
location of the action expressed in the sentence predicate.
The verbZELdzai of the original sentence is one of this kind
of verbs which require a place word or a phrase as the object,

while the miscueJéi ygh is not subjected to this restriction.

As a result, when the miscuefifi ygh is considered along with

86

the portion which follows:it,namely, the complex place word
"453@U11'y1 ge bau shang' the whole sentence becomes
syntactically unacceptable.

Unlike the category of grammatical function which
evaluates word level substitution miscues, the category of
syntactic acceptability involves an analysis of miscues
in the context of sentence and whole discourse. According
to the short version of the RMI which is adopted for the
present study, the syntactic acceptability allows for two
choices. First, the sentence involved with a miscue can
be coded "Yes" if the sentence as finally produced by the
reader is syntactically acceptable in the story. The
alternative is to code "No" as being syntactically unaccept-
able if the sentence as finally produced by the reader is not
syntactically acceptable in the story. To demonstrate how
this coding scheme works, the researcher would once again

use 4a QB if? 3?, 3E, 273':— F“ g— , {if} a": as an example.

Ta de jing shen, syan daii j; dzai yi ge ban shang.

If a reader reads
4172897117? 273’ aimfififizfilévi—
"Ta de jing shen, syan dzdi jr y__ yi ge bau shhng"
the sentence will be coded "No" as being syntactically unac-
ceptable, even if both the miscue and the text word share the

same grammatical function. If a reader reads

may 3% 27* ii) A fi? a ,173 Z J; the sentence

"Ta de jing shen, syan dzai dou dzai yi ge bau shang.‘

will be coded "Yes" as being syntactically acceptable,

87

because the substitution of7g'P _d_§_u 'all' for the text
word,F{ if 'only' does not change the syntactic structure
of the original sentence.

Table 8 summarizes the data on the syntactic accepta-
bility of the miscues for the readers of Chinese of the
present study. The figures indicate that 87.6Z of the
sentences involved with miscues are syntactically accept-
able in the story. These figures do not include all
instances in which the readers appear to use syntactic
information in their predictions which are acceptable only
with the prior portion of the sentence. If we combine the
totally acceptable predictions with those which are accept-
able with the prior portion, a somewhat different picture
is seen. With the combined scores, about 96Z of the ob-
served responses preserve at least partially the grammati-
cal structures of the text, which indicates the readers'
awareness of sentence structures and their ability to
make use of syntactic knowledge in making predictions.

Table 9 presents the finding reported in the Blair
study in terms of syntactic acceptability. According to
Blair, 93.5Z of miscues analyzed are syntactically accept-
able in the story. This figure is somewhat higher than
the figures reported in Devine's study (66Z) and Jensen's
(82.12Z). The high percentage of syntactic acceptability
found in the Blair study may be due to the type of miscues

analyzed. As has been pointed out previously, Blair focuses

88

TABLE 8.

 

SYNTACTIC ACCEPTABILITY
CHINESE ADULT READERS

 

 

1. 19 76 6 24
2. 22 88 3 12
3. 22 88 3 12
4. 21 84 4 l6
5. 21 84 4 l6
6. 22 88 3 12
7. 23 92 2 8
8. 24 96 l 4
9. 25 100 O O
10. 20 80 5 20
TOTAL 219 31

MEAN 21.9 87.6 3.1 12.4

 

89

TABLE 9

 

SYNTACTIC ACCEPTABILITY
AMERICAN MATURE READERS

 

 

SUBJECT NOTES % N0 N0 %
1. 24 96 1 4
2. 24 96 1 4
3. 24 96 1 4
4. 23 92 2 8
5. 22 88 3 12
6. 24 96 1 4
7. 21 84 4 16
8. 24 96 1 4
9. 23 92 2 8
10. 24 96 1 4
11. . 24 96 1 4
TOTAL 257 18
MEAN 23.3 93.5 1.6 6.5

 

90

only on the substitution miscues, while Devine, Jensen and
the present study analyze all kinds of miscues identified
in the readers' oral reading.

In spite of the minor variations in percentage reported
from these four studies, a broad generalization based on
these data may still be drawn. All readers, no matter what
language they are dealing with, seem not to have any
difficulty in responding to the syntactic cues available to
them, which confirms the hypothesis set up by the Goodmans
and other psycholinguists that syntax is one of the most
crucial resources used by readers in reading.

The statistical analysis of the categories of
grammatical function and syntactic acceptability yields
very important information regarding to what extent a
reader is using the grammatical features of his language.
However, in order to get a better view of how a reader is
making use of syntactic information, it becomes necessary
to look more closely at the miscues themselves and their
corresponding text words. As Rudine Sims points out "The
syntactic constituents of our language are interrelated in
such a way that a change within one constituent may also
cause change in another. A miscue which involves one
word, for example, may at the same time alter the phrase
structure of the clause in which it occurs" (Sims, 1976,
p. 46). One common method used by researchers working

under the framework of miscue analysis to study the complex

91

interaction of a reader's syntactic system is to investigate
types of miscues and the levels of language structures on
which a miscue is operating. To examine miscues in terms

of their relationship to the structural levels of language
enables a researcher not only to treat a miscue locally,

but at the same timetxwlook into the complex interplay of

a reader's grammatical system while he reads.

Miscue Types and Levels of Structure

Substitution, insertion, omission and reversal are the
four types of miscues identified and analyzed in a taxonomy
analysis. When a miscue occurs, it is first classified
into one of these four categories and then the researcher
decides at which language level the miscue occurs. A
miscue may cause changes at any, or sometimes all, of the
following levels: submorpheme, bound morpheme, word or
free morpheme, phrase or clause. The use of this five-
category system allows a researcher to capture the inter-
relationship between categories as demonstrated in the
following example.

Text: They drove until they found a good place to camp.

 

Reader: They drove until they found a place to camp.
(Jensen, 1972). In this example, the reader omitted the
word "good." According to the Taxonomy, this miscue is an
omission on the word level, a substitution on phrase level

(a good place is substituted by a place) and finally an

 

92

omission on clause level where the deep structural
adjectival clause is lost.

The interrelation between categories of structural
levels demonstrated in this example is by no means an
uncommon phenomenon in readers' oral reading. In fact,
according to the various miscue analysis research projects
reviewed above in the present study, most miscues involve
more than one level of language structure. Actual statisti-
cal figures reported from some of these studies will be
given in the following section together with the presenta-
tion of the data obtained from the oral reading of the
Chinese readers.

A few notes concerning the methodology for examining
miscue types and the levels of operation of the present
study are needed before the presentation and discussion of
the actual data. In the standard form of a taxonomy
analysis, each miscue is judged on a five-structure-level
basis. But in the present study, each miscue will be
examined on a four-language-level scale, namely, bound
morpheme, word, phrase and clause. The reason for
excluding the submorphemic level from analysis is quite
obvious. Since Chinese is not an alphabetical writing
system, its morphemes and words are not composed by the
combinations of alphabetical letters. But the justifica-
tion for using the other four levels of this system to

study a language which is totally non-western is not so

93

obvious. A brief account of the internal structures of
Chinese morphology may be helpful at this point.

Chinese is commonly described as a mono-syllabic
language due to its lack of inflectedness and the physical
independence of each individual written symbol. On the
printed page, each Chinese character is equally spaced off
and there is no functional marker to serve as a word
boundry. In this regard, each written symbol has every
right to become an independent word. However, that
Chinese is a mono-syllabic language is true only as far
as the classical Chinese is concerned. As Chao has said
"The so-called 'mono-syllabic myth' is in fact one of the
truest myths in Chinese mythology" (Chao, 1968, p. 139).
The words of modern Chinese are by no means all mono-
syllables. WOrds have become largely dissyllabic or
polysyllabic, and many formerly free mono-syllabic
morphemes now occur only as bound morphemes in compounds.

In modern Chinese, not even all of its morphemes are
mono-syllabic. There are instances of morphemes which are
represented by two or more syllables, none of these
morphemes having any meaning or function without the other.
One such example isfifi fifi putau 'grapes.'%§ €§ putéh
is an example of a polysyllabic morpheme. There are also
a very small number of morphemes which are represented by
parts of syllables. Cases of morphemes less than a

syllable, however, are very rare. The only impOrtant ones

94

are the non-syllabic "diminutive suffic" gfiLI-r as in
7397565111: 'knife' and the plural suffix 4?? -men as in 4‘53 1??
tamen 'they.‘ Generally speaking, the modern Chinese
morphemes can be classified into three major types:
mono-syllabic morphemes which correspond with single
syllables, polysyllabic morphemes which are represented by
the arrangement of two or more syllables and subsyllabic
morphemes which are represented by parts of syllables.
Chinese, like its English counterpart, contains morphemes
which can exist by themselves as independent units.
Chinese also contains morphemes which function very much
like ”bound” morphemes in the western linguistic sense.
In short, in spite of the drastic difference in symbols used
to form morphemes, the internal morphological constructions
of English and Chinese are quite similar. 2

To consider the status of "word” in Chinese is a
much more difficult theoretical problem. In the first
place, the Chinese term wordlgldé has never been clearly
defined. In the traditional use, all of the mono-syllables
are calledH§;.dz. Thus, what we call morphemes today are
referred to asgdi. Furthermore, the term ’37- de is also
used to refer to bisyllabic morphemes such as% 3% putau
.grapes , 33—11% ddufu 'beancurd, 'ffl iflméigwei 'rose. ' This
lack of an established tradition in Chinese as to what
constitutes a word is central to the confusion. The other

source of the problem comes from the fact that the Chinese

95

writing system does not have phonological or morphemic
markers to indicate the word boundaries. Therefore, it is
very difficult to establish a firm borderline between words
and other segments. As Kratochvil has stated, at the
present stage of the Chinese language "it is next to im-
possible to set up MSC (Modern Standard Chinese) words, or
rather to find words in MSC by using the gauge of the set
of features popularly associated with words in European
languages" (Kratochvil, 1968, p. 94). Nevertheless,
despite the difficulties, both linguists and Sino-logues
believe that it is both necessary and possible to establish
a linguistic unit which can serve as an intermediate link
between the morpheme and the phrase or sentence. Instead
of defining words in terms of precisely conceived features
of morphological structure, of syntactic behavior, and

most important of all, of meaning, words are defined from

a more functional point of view. Chinese words are
established on the basis of actual structural features of
modern Chinese utterances.

The definition adopted for this present study is
proposed by Kratochvil. He defines a modern Chinese word
as "the smallest unit which may function as an immediate
constituent of MSC segmental sentences" (1968, p. 94).

To treat the Chinese word as a unit which may function as
an immediate constituent of segmental sentence makes it

possible to take into account the different morphological

and syntactic behaviors of a word in different contexts.

96

Furthermore, to define the Chinese word in this way allows
us to determine the size or numbers of morphemes which
constitute a word unit on the basis of the actual struc-
tural features of the utterances rather than on any pre—
determined and fixed criteria. In other words, what is a
word in one context is not necessarily a word in another.
context. The following example will demonstrate this
point. In the sentence ‘7( $ F1?) ’1 th ché kai la. 'The
train has left' the segmentW%<.hw6 'fire' is a "sometimes
bound" morpheme which combines with another morpheme a?

ch'é 'vehicle' to form the polysyllabic wordi‘? hwo/che

'train.‘ However, the same morpheme'7<,hw5 when it occurs
in the context of ‘)\ XE 3 Iwo' syf le. 'Fire extin-
guished,‘ has become a single morpheme free word.

This dependency of word status on the environment in
which the given segment occurs is not a feature specific
to the modern Chinese language. In fact, in the Goodman
Taxonomy, the terms word and free morpheme are also defined
from a functional perspective: "Free morphemes are oral
meaning bearing units within the language which can
function independently or in combination with other free
or bound morphemes. Words are graphic representations of
free morphemes, and free and bound morpheme combinations"
(K. Goodman and Burke, 1973). The definition of word
given in the Goodman Taxonomy is comparable to the one es-

tablished for the Chinese word. Both of these definitions

97

try to establish an operational unit rather than to
characterize the word by some specified and fixed features.

At this point we may try to pin down the differences
and similarities between English and Chinese syllable to
word relations. The differences start with the lack of
inflectedness of the Chinese language. The various
English ways of enlargement of the root by affixed deriv-
atives are seldom found in Chinese. In Chinese, on the
other hand, there are a large number of morphemes which
possess dual functions. They are "sometimes free" and
”sometimes boundfl'depending on the linguistic environments
they are in, a phenomenon found less often in English
morphological constructions.

Early in its history, English was a ”synthetic"
language and had more bound syllables, but it has become
more "analytic" as its free forms increased. Chinese has
been and still is ”analytic," but as the number of its
bound syllables and compounds seems currently to be growing,
so the language must in time become more "synthetic." In
other words, these two languages are developing in opposite
directions in such a way as to become more alike. As
English becomes more "analytic" and less "synthetic" and
Chinese more "synthetic" and less "analytic" the distance
between these two languages shortens, and, as a result,
they become more similar in the aspects of morphological

constructions.

98

The following are illustrations of types of miscues
generated from the Chinese readers' oral reading and the

structural levels at which these operations take place.

1. Morphemic Level

In a taxonomy analysis, any change of the miscue at
the morphological level requires an analysis of a two-part
coding. The first part deals with the process involved,
namely, substitution, insertion, omission and reversal.
The second part determines the types of morphemes involved:
inflectional,derivational,contractional or non-inflectional.
These subcategories used to characterize the types of
bound morphemes that occur in the miscues do not apply to
the Chinese morphological structures due to the lack of
inflectedness of the Chinese language and_the way that
Chinese characters are written down. Rather, miscue types
at this level will be assigned on the basis of a morpheme's
relation with another morpheme with which it co-occurs.
Chinese morphemes can be roughly classified into two
categories: free morphemes and bound morphemes. However,
it is important to note that the term "free" has to be
understood in its potential sense because Chinese "free"
morphemes are not always free. They are sometimes free and
sometimes bound, depending on the role they play in
constructing morphological units. By the same token, the

term "bound" cannot be taken in its absolute sense either.

99

The ”bound" morphemes in Chinese consist of two sub-
categories. One type of "bound" morphemes consists of
those ”sometimes bound” and "sometimes free" morphemes.
The second type of bound morphemes consists of those
bound in the truest sense of the term. They are always
bound. All such morphemes constitute a morphological
subtype of morphemes which are generally called affixes,
word-formation affixes or grammatical affixes (such as
affixes 4?? -men, $1, -r andmarkers 3 le,-% jé, etc.).
To describe the ways in which Chinese can be repre-
sented morphologically, it is obvious that just to distin-
guish between free and bound morphemes is not enough. It
seems to be necessary to observe how individual morphemes
behave in relation to the other morphemes with which they
represent a word. To understand which of the two kinds of
bound morphemes is used to construct a given word becomes
the key to describing the complex morphological composition
of a Chinese word. In determining the types of bound
morphemes involved in a miscue, two categories are used
in the present study. A miscue can be coded either "some-
times bound" or "always bound." The following examples

illustrate the operation of this coding scheme.

515
(1%- fi 47% 49.2. *7}? “F —2 fifififififlfiz

L56 shwin hdu ta péng jing sya chyh, tsai d1 dide iju dah:
Substitution of sometimes bound morpheme for another some-

times bound

100

2 _. " J; ‘
“$in 37228: $47,146”? ’fi,’i&1
Yi ge shr yi shah bah chywan bang, jEi lyghg sya , yi ding

757 (if—’1 ’37. a 3..
gou ta shou w 1e.

Omission of a sometimes bound morpheme

(”73 ’EWJ Ear J:- 6% 455 63‘? -... >9.

Bah yi hwe, shén shang ije de you dyan léhg.

Insertion of an always bound morpheme

«awax 478 465; m .2212 3.

J1 ge ren tsung ta myan chyanp gwo chyu 1e.

Reversal of sometimes bound morpheme

2. Word or Free Morpheme

Like the morphemic level miscues, miscues involving a
word/free morpheme are examined first for the physical
qualities of the miscues--substitution, insertion, omission,
reversal--and then are analyzed for the kind of morphemic

involvement. Here are some examples:

Text= 93% fling ”£18 5— phdz $9132? 656 j‘r
Reader‘ '3 Y‘l’ [35’ dyan % a‘fiL déi ba/n
When the reader reads EL yi for;§€§_j1ng, he substitutes a
single morpheme word for another single morpheme word. In

. \ . .
the second example, the reader's miscue,fiB dyan, 1s a Single

lOl

morpheme word, while the text wordétfiémh dz‘ has two
morphemes. In the third example, the reader has substituted
a multiple-morpheme wordgijhfidai.bah for another multiple
morpheme wordjfifing‘bgn jr. I

As pointed out earlier in this section, in spite of
the difference in symbols which are used to form words in
Chinese and in English, there are some common features
shared by these two languages in the morphological
constructions. Like English, Chinese words can be roughly
grouped into four major types: single morpheme words,
multiple morpheme words, complex words and compound words.
In this regard, the system originally developed for
analyzing the structures of English words can be directly

applied to the study of Chinese language.

3. Phrase Level

The analysis of miscues at the phrase level is pro-
posed under the assumption that miscues may involve larger
language units than morpheme and word. It may be a noun
phrase, a verb phrase, or an adverbial phrase. When a
miscue occurs at the phrase level, it may take the form
of substitution, insertion, omission or reversal. For
instance, when a reader readsflga’wL ,7; yau dai 'belt' for 2&1
Eg'flfi' KT; sywan se yau dai 'black belt' he has substituted

one phrase structure for another.

102

4. Clause Level

The influence of transformational grammar on the
Goodmans' miscue analysis is most clearly seen in the
analysis of miscues at the clause level. As stated in Thg
Goodman Taxonomy of Reading Miscues ”The surface structure
of a sentence can be composed of varying combinations of
independent, dependent and embedded clauses. At the deep
structure level, a clause is considered to be composed of
a noun phrase and a verb phrase. At the surface level,
a clause might retain both its noun and verb phrases or
might be represented by any one or several of its
constituents" (K. Goodman and Burke, 1973). By following
this logic, an adjective embedded within a noun phrase
represents a deep structure clause. An omission or inser-
tion at the word level may cause some changes at the clause
level. For example, when the reader deletes the adjective‘ii
a sywan 322 'black' from the adjective phraseggfig'é, sywan
se ya'u dai and results in miscuefl$yau dai he has altered
the clause structure of the sentence. This miscue causes
an omission of a deep structure clause. Miscues can also
affect the dependency relationship between clauses or

sentences. For example, if the reader reads

«if! 111%] “Z @135] 1354' ,4‘E% 36 WO/men ch; fan de shf hOu,

I. \ O O I
ta la1 1e. 'He came whilevnawere eating, instead of

47:. 7:4 99 727 2726, 2218479 22 872%.

- I. l v - ~
Ta la1 de shr hOu, women chr fan ne. 'When he came we

103

were eating.‘ He alters the clause dependency within
the sentence. The first clause, formerly dependent,
becomes independent in the new sentence.

Tables 10, ll, 12 and 13 summarize the data for
miscue types and levels of structure at which these miscues
take place. As the data show, the readers of Chinese
make use of almost the same types of operations in dealing
with the syntactic information available to them as their
English counterparts do. They substitute, delete, insert
or reverse the order of the items. The data also indicate
that miscues produced by the readers of Chinese involve
various levels of language structures as described in
the Goodman Taxonomy. In the present study, we find
instances of miscues that involve more than one Structural
level. About llZ of the total miscues involve the phrase
level, and about 13Z the clause level. Most of the miscues
of the current investigation occur at the word level (88.8Z)
and the next most frequent miscues are the bound morphemes.
The observation that most miscues take place at the word
level confirms the findings reported in both Devine (88.4Z)
and Jensen (89Z) studies.

But the high degree of involvement at the bound
morphemic level of the Chinese readers' miscues is something
of particular interest. Jensen reported that 94 percent
of her H group readers' miscues do not involve the bound
morphemes. A somewhat lower percentage but similar pattern

is observed in Devine's readers. Her high group subjects

104

TABLE 10.

TYPE OF MISCUE AT THE BOUND MORPHEME LEVEL

 

 

SUBJECT $3888 61133.“ 2:16? 83le $8888“
No. 11 3 1 0 10
1. Z 44 12 4 O 40
No. 12 0 3 1 9
2. Z 48 0 12 4 36
No. 8 5 O 2 10
3' Z 32 20 0 8 10
No. 7 2 4 2 10
4. Z 28 8 16 8 40
No. 8 6 3 2 6
5. Z 32 24 12 .8 24
No. 6 3 5 1 10
6. Z 24 12 20 4 40
No. 7 4 2 5 7
7. Z 28 16 g 20 28
No. 7 6 1 2 9
8. Z 28 24 4 8 36
No. 6 6 4 3 6
9. Z 24 24 16 12 24
No. 6 5 2 0 12
10.
Z 24 20 g 0 48
TOTAL 78 4O 25 18 89

MEAN 31.2 16 10 7.2 35.6

105

TABLE 11.

TYPE OF MISCUE AT THE WORD LEVEL

 

 

888- 838- 88- $888-

No. 18 4 1 0 2
1' Z 72 16 4 o 8
No. 18 1 3 o 3
2' Z 72 4 12 o 12
No. 19 o 2 1 3
3' Z 76 o 8 4 12
No. 17 o 4 1 3
4' Z 68 o 16 4 12
No. 20 1 0 O 4
5' Z 80 4 o o 16
No. 19 o 2 o 4
6' Z 76 o 8 o 16
No. 19 1 o 3 2
7' Z 76 4 o 12 8
No. 20 4 1 O 0
8' Z 80 16 4 o o
No. 22 1 1 1 o
9' Z 88 4 4 4 o
No. 15 o 2 1 7

10.
Z 60 o 8 4 28
TOTAL 187 12 16 7 28

Mean 74.8 4.8 6.4 2.8 11.2

 

106

TABLE 12.

TYPE OF MISCUE AT PHRASE LEVEL

 

 

8898- 838- EEEER' 888-
No 5 o o o 20
1' Z 20 o o o 80
No. 4 0 O 0 21
2° Z 16 o o o 84
No. 2 o o o 23
3' Z 8 o o 0 92
No. 2 O 1 O 22
4' Z 8 o 4 o 88
No. 1 o 1 1 22
5° Z 4 o 4 4 88
No. 2 o o ‘0 23
6' Z 8 o o o 92
N0. 1 0 o o 24
7° Z 4 o o o 96
No. 2 0 0 O 23
8' Z 8 o o o 92
No. 4 o o o 21
9' Z 16 o o o 84
No. 2 0 O 0 23
10.
Z 8 o o o 92
TOTAL 25 o 2 1 222
MEAN 10 o 0.8 0.4 88.8

 

107

TABLE 13 .

TYPE OF MISCUE AT CLAUSE LEVEL

 

 

$88- 88- 88- 81%888'
NO. 0 4 1 0 20
1. ‘73 0 16 4 0 80
NO. 0 o 3 O 22
2' Z O O 12 O 88
No. 0 0 1 0 24
3. Z O 0 4 0 96
No. 0 0 1 O 24
4. Z O O 4 o 96
NO. 0 2 2 0 21
5' 7° 0 8 8 O 84
NO- 0 1 3 O 21
6' Z 0 4 12 o 84
No. 0 2 O O 23
7. Z 0 8 0 o 92
NO. 0 4 1 O 20
8' Z o 16 4 o 30
No. 0 2 2 O 21
9. Z O 8 8 o 84
NO. 0 O 4 O 21
10.
Z 0 O 16 O 84
TOTAL 0 15 18 O 217
MEAN 0 6 7.2 O 86.8

 

108

have a mean of 76.1Z of the total miscues which are free
of bound morphemes involvement. However, in the present
study, we find that about 64Z of the miscues take place at
the bound morphemic level. 1

The result of the bound morpheme analysis may provide
some interesting information about the morphological
constructions of the Chinese language. The miscues
produced by Chinese readers at the morphological level
can be used as good examples to demonstrate that Chinese
is no longer a mono-syllabic language. Of the 250 miscues
analyzed in the present study, 161 involve some sort of
changes in their polysyllabic structure. There are 31.2Z
of them related to the substitution of one bound morpheme
for another. The most common type of morphemic substitution
is a "sometimes" bound for another "sometimes" bound. As
mentioned earlier in this section, "sometimes" bound
morphemes are morphemes which can exist independently in
some other context as free morphemes. The observation that
these "sometimes" bound morphemes occur as bound morphemes
rather than as free morphemes seems to provide some evidence
that Chinese is becoming more "synthetic."

The data obtained from the analysis of the miscues
involving the operations of insertion and omission also
support this observation. In fact, among the syntactic
deviations produced by the readers of Chinese the research-

er discovers that the examples in which omission and

109

insertion are involved are the most revealing as far as
the current development of the Chinese morphology is
concerned. 0n the basis of the data generated from
readers' miscues at the morphological level, the instances
of omission and insertion are two sides of one phenomenon.

The following examples demonstrate this point.

Text: (1) $1.12 (2) iii- syEm (1in

iing
Reader: (1)3, (637.: yT jing (2) fig. syén
In the first example, the character 6“}: jing meaning
'already' is a mono-syllabic word while its counterpart EL
gfifi'yf'jing is a bisyllabic word which is made up by the
"sometimes" bound morphemes EL y! andgfi jfng both of which
contain the same semantic element 'already.‘ The words ED
ytgrzfijing, and afigyf ji'ng convey the same meaning, but
only differ in the sense that the former two are more the
classical literary style, whereas the latter is more
colloquial. In the present study, several readers
miscue on this character. They insert the "sometimes"
bound morphemeflyf and produce the bisyllabic wordaag‘g:
y! jing.

In example 2, the operation of omission is involved.
The Chinese wordjjtfi§yén dzai 'now' is a polysyllabic word
which consists<xftwo morphemes. It has a mono-syllabic
counterpart £31,8y5n 'now.‘ In this example, readers
alter the author's choice, which is more colloquial and

redundant, by omitting the bound morphemefigidzéi and

110

constructing. more compag%fistructure
“343971861 2) .“4133'54-35-
Ta de jing shen, syan jr dzai yi ge baushang.

Miscues as the result of omission and insertion make
up 26Z of the total miscues. The fourth and last kind of
miscue type identified in the present study is reversals,
which constitutes a little more than 7Z of the total
miscues of the current study. Despite the small number
in percentage, this type of miscue needs to be noted in
this context because such miscues represent one specific
group of modern Chinese morphological composition not
found in an alphabetical language. In modern Chinese
there is a group of words in which the two morphemes within
the given word have a similar meaning and the meaning of
the whole word is again similar to the meaning of both
the constituent morphemes. For instance, the wordjfij‘fib
chyén myén 'the front; ahead' is composed of the morpheme
1%] chyan 'front' and the morpheme 6E7 myan 'face; front.‘
Since the two components of the word are similar in meaning,
the reversal of the order of the morphemes does not change
the meaning of the original text words. Some other
examples of this type of miscue found in the present study

are:
\ . L’ s , \
Text: £78; dung ye’lu; fig 3%? ché ij6; fig 4i bei hou
I . I it \ \,
Reader:,f% % yau dung ”13;: 2% Jywe ch5;¢f?{ fl hou bei
Miscues involved in reversal differ from miscues of insertion

and omission in one aspect, namely, the two morphemes such

111

asii] 57 chyén mya‘in have to stay together as one word.

The positions of the two morphemes are reversible, but
the deletion of either one of the two constituents as in
the case of omission will result in an unacceptable word.

Aside from the major differences that are noted at
the bound morphemic level, there is no significant pattern
of differences emerging from our data, as can be seen in
Figure 3. Figure 3 summarizes the data of miscue types
for Devine, Jensen and the present study at word level.
For all these three studies, substitution miscues constitute
the majority of the miscues produced by the readers. Both
Devine's subjects and those of the present study show
about 75Z substitution miscues at this level, and Jensen's
show 47.8Z. The biggest difference between groups is in
the insertion category with Jensen 21.1Z, Devine 3Z and
the present study 4.8Z. Jensen attributes the larger
percentage of this type of her readers' miscues to their
application of optional transformations. Based on the
present data, no generalization can be made to account for
the difference in the performance of the readers' in this
category.

Miscues involving reversals are the least common among
these three groups of readers, Both Devine and Jensen have
very few reversals in their data. In the present study
there are eighteen reversals at the bound morpheme level,

seven at the word level and one at the phrase level. When

112

FIGURE 3.

A COMPARISON OF MISCUE TYPE AT THE WORD LEVEL

 

pm>ao>cfl uoc

Hamhm>mu

ScammflEo

cofiuummc«

Coausuflumnam

 

100

90
80

 

 

T

    

XX 4« x A]???
“Ki XX“ 7W KXA

x
AXAXXXXXX 4mm

  

 

   

Devine's study

Jensen's study

1%

I I: The present study

113

reversals occur in a text written in English, more often
than not, they either result in non—words or cause some
damage to the syntax. It is not the case with Chinese.

To recapitulate the finds so far: the data from
miscue types and structural levels reveal that the perform-
ance <3f Chinese readers is very similar to that of the
readers of English. For both groups of readers, the most
frequent types of miscues overall are substitution, then
insertion, then omission, and finally reversal. The data
also indicate that Chinese readers, like their English
counterparts, produce miscues which involve various levels
of language structures. The major difference between
these groups is that Chinese readers show a higher degree

of involvement at the bound morphemic level.

Semantic Acceptability

In the previous section, we have dealt with the cate-
gory of syntactic acceptability which concentrates on the
success with which the reader is coping with the structure
of the text sentence. In the present section, the focus
will be on semantic acceptability, which deals with the
question of how successful a reader is in producing
understandable structures. In determining the semantic
acceptability of a sentence, the major question is how well
the miscue fits the context semantically. According to the

short form of the RMI, there are two possibilities of

114

coding. If the sentence as finally produced by the reader
is semantically acceptable within the whole text it is coded
"yes." If the sentence as finally produced by the reader
is not semantically coherent with the total passage it is
coded "no." The following two examples illustrate how
this coding device works.

8M8 .4 m 3“ ii, 3

(1)Text: Hwa da mal jyau2 Sygu Shuan3 jin4 1e5

 

mother Hwalcall2 syau Shuan3 enter4 particle5 room6
inside7. (Motherlhnicalled Syau Shuan into the

inner room.)

8x48»: .1. Z? 381% €57 90

\ \ .. . \ o’ . } V ‘
Reader: Hwa da mal jyau2 Syau Shuan3 Jlnh 1e5 116 myan7 de8

8%

wudzg.

      
   

mother Hwal call2 Syau Shuan3 enter4 particles
inside6 side7 particle,3 roomg. (Mother Hwa called
Syau Chuan into the room inside.)

187737-b‘flft3 426788.82

(2)Text: Téi yang yez chu lai_ 134---- hOu myan jau6 jyan7
ding dz jye tou8 , shangll' gu ting
. /\ l”
a 37%, m7 481 we - 69 i

0 fl

\ . 75 6
'kOu'12 jei13 sz ge14 an dan15 de16 j n dzl7.

     
 

the sun1 also2 come out3 particle4.... behind5
shine6 see7 T-junction8 worn9 wooden tablet10 onll
'Ancient Pavilion'12 these13 four14 faded15
particle16 gold inscription17. (The sun had risen,
lighting up the worn tablet behind him at the
T-junction with its faded gold inscription:"Ancient
Pavilion.")

115

88.8883 88 p.88 1

Reader. Téi yang ye chfi 151 le.... hOu myan jéu jyan ding

34:51:51 #818 .1: 4.51:1?» 0 imfifl
dz jye tou bei byéh shang 'gu“ ting kOu' jéi $2 ge
“MB /k 89 $- ‘;9

8n dén de j1n dz.
In the first example, the reader substituted $.57 59/iv3"
1¥ myan de wfi dz 'room inside' for the expected response
éE3§§_wfi l! 'inner room.‘ This substitution is semantically
acceptable in the context of the story. In the second example,
the substitution ofifigbéi bya/n for the original text
wordsZ/ifipwb byaJn is not semantically acceptable, because
the miscueiig béi bya/n does not make any sense in the
story. Furthermore, in modern Chinese the character23£;_bei
has at least six different explanations 02g., quilt; to
cover; to spread; because of; a passive voice marker; a
Chinese family name, etc.). It is not clear to us which
of these definitions that the reader had in mind when he
produced the wordeZ“béi.

It is important to note at this point that one of the
very fundamental assumptions behind the Goodman taxononw'and
the RMI as well is that any sentence can be studied from
two perspectives. That is, it can be examined both for
its sentence structure and its sentence meaning. The
main reasoning behind this assumption is that a reader can
produce a sentence which is grammatically acceptable but

semantically nonsense. It is also possible for a reader to

116

change the grammatical structure of a sentence without
altering its semantic content. It is for this reason

that the semantic organization and syntactic organization
of a sentence are treated in two separate categories

in the taxonomy. Nevertheless, this separation of syntax
from semantics is more for practical convenience rather
than for theoretical consideration. The Goodmans and

other researchers are fully aware of the intricate relation—
ships between these two systems. As Montoro has stated,
"The syntactic and semantic components of a language are
virtually inseparable, and they are dependent upon each
other in the most profound ways. The Gooman Taxonomy does,
in fact, accommodate to this reality in several of its
categories. However, when we consider the acceptability

of a sentence and whatever changes the reader may have
made, it is most useful to make an arbitrary separation

of sentence grammar from sentence meaning" (Montoro,

1976, p. 34). One of the devices used to accommodate

to reality is to put a conditional requirement

when coding semantic acceptability. Semantic accept-
ability is never scored higher than syntactic acceptability
(K. Goodman, 1969, p. 27). In other words, the semantic
acceptability of a sentence is dependent upon or at least
is limited by the grammatical acceptability. This condition
puts the study of semantics within the context of syntax

and allows the researchers to treat meaning as a function

117

of both grammatical and semantic interrelation. The other
device suggested by the Goodmans and employed by many
researchers is to present the data of syntactic and
semantic acceptability together. When these two categories
are considered together, they provide a very clear picture
of how well the reader is handling syntactic and

semantic information available to him in the text.

Table 14 presents the data on syntacticzfiuisemantic
acceptability of the miscues for the readers of Chinese.
The data on acceptability point to one general observation.
All readers are more successful in using syntactic
than semantic information in the text. This is obvious
in the fact that the percentage of miscues totally
acceptable syntactically for all of the readers is higher
than the percentage of miscues totally acceptable semantic-
ally.

Table 15 summarizes the data on syntactic and semantic
acceptability of miscues for the readers of Blair's study.
Blair's data also indicates that readers are more likely
to score higher in syntactic acceptability than in
semantic acceptability.

Similar findings concerning the performance of
readers in the categories of syntactic and semantic
acceptability are reported in both the Devine and Jensen
studies (see Figure 4). A general observation which

emerges from the data is that all readers, regardless of

118

TABLE 14.

SYNTACTIC AND SEMANTIC ACCEPTABILITY IN Z
CHINESE ADULT READERS

 

 

SUBJECT SYNTACTIC SEMANTIC
l. 76 76
2. 88 76
3. 88 84
4. 84 68
5. 84 80
6. 88 80
7. 92 V 84
8. 96 84
9. 100 92

10. 80 72

MEAN 87.6 79.6

 

119

TABLE 15.

SYNTACTIC AND SEMANTIC ACCEPTABILITY IN Z

AMERICAN MATURE READERS

 

 

SUBJECT SYNTACTIC SEMANTIC
l. 96 60
2. 96 80
3. 96 72
4. 92 72
5. 88 76
6. 96 68
7. 84 80
8. 96 ' 68
9. 92 72

10. 96 72
11. 96 60

MEAN 93.5 70.9

 

120

FIGURE 4.

A COMPARISON OF SYNTACTIC AND SEMANTIC ACCEPTABILITY IN Z

100
90 fl

80 .7— F?

70

.0 g 8 8

¢ V ¢ %

50 ¢ / ¢

% / ¢

40 / / /

//’ //6 ///

3o 3;: /// jjj

20 ¢ / ¢

¢ ¢ /

10 ? ¢ ¢

¢ ¢ %

0 Z é a

 

 

 

 

 

 

 

 

 

 

 

 

Blair

Devine

D
.

Jensen The Present Study

Syntactic

Semantic

121

their language background, are more successful with syntac-
tic than with semantic information in their reading. A
comparison of the percentage of syntactic acceptability
with semantic acceptability of these four groups of’subjects
demonstrates this observation: 93.5% and 70.1% for Blair's
readers; 66% and 32% for Devine's group; 82% and 74.5%

for Jensen's; 87.6% and 79.6% for the present study. For
Jensen's and the present study, there is very little
difference between the two categories. Blair does not
provide information to explain this gap, therefore, no
comparison can be made here. One major reason for the gap
between syntax and semantics for Devine's readers is their
use of non-words. Non-words usually result in structures
which are grammatically acceptable but they must be consid-
ered as a total loss of meaning. In the present study,
very few instances of non-words exist. Instead, miscues
which cause a loss semantically are mostly involved with
bound morpheme substitutions. In most cases, the
substitutions end up with some types of new morphological
compositions which are also "real" words in the Chinese
language. They are coded as semantically unacceptable

in the present study because they violate some kinds of
dependency relations when miscue occur, as discussed in
the exampleflfl 397532?) I 25:- fi “413] ZI- Ta de ji'ng

yi ge béu shang. (See p. 84)

 
 

shen, syan dzai if

This kind of miscue not only causes loss in meaning, but

122

also damages the grammatical structure of the text
sentence. This seems to be the main reason that the
difference in percentage between the categories of syntac-
tic and semantic acceptability tends to be very low in the

present study.

Comprehending

The category of comprehending or meaning change is
another device used by researchers to discern a reader's
processing strategies in relation to semantics. The major
question to be answered in coding comprehending concerns
how much of the "intended meaning of the author is changed
by the reader" while he reads. To measure the degree of
meaning change the sentence is read as the reader finally
produced it. All corrected miscues are treated as finally
resolved by the reader. Table 16 presents the Compre-
hending Scores of the Chinese readers. The data indicate
22.8% of all miscues produced some changes in the intended
meaning of the text and 77.2% of all miscues produced no
change in meaning. The Comprehending Score presented in
the present study is obtained by combining together the
percentage of miscues which result in totally semantically
acceptable sentences and the percentage of miscues which
are corrected to become acceptable. Sentences which are
originally semantically acceptable take up 56.4%, and the

remaining 20.8% can be attributed to the result of

123

correction. Correction attempts made by a reader indicate
that he is getting meaning from the print. When reading
fails to make sense to him, he goes back or regresses and
corrects for meaning. 9

Table 17 summarizes the data on Comprehending Score
of Blair's American mature readers. The figures indicate
that 59.3% of all miscues are coded 'no' as having produced
no change in meaning and 40.7% of the total miscues as
having produced either some degree or major disruption in
meaning.

When these two sets of data are compared one finds
that except for the fact that the Chinese readers tend to
score higher in 'no' meaning change than their English
counterparts, there is no fundamental difference between
these two groups as far as the reading process is concerned.
Reading is primarily for meaning, as will become quiteeclear
if the figures in both the 'no'enui'minimal' categories
are combined. The new figures show that the majority
of miscues (84% for Blair's readers of English and 95.2%
for Chinese readers) preserve all or most of the intended
meaning of the text. Only a small number of miscues (16%
for Blair's and 4.8% for the present study) cause major

changes or loss in meaning.

Summary of the Findings

In the beginning of this chapter, the researcher has

124

TABLE 16 .

COMPREHENDING SCORE
CHINESE ADULT READERS

 

 

SUBJECT No. NO % NoglNIMALZ No. YES 2
1. 20 80 3 12 2 8
2. 20 80 5 20 0 o
3. 16 64 6 24 3 12
4. 22 88 1 4 2 8
5. 23 92 1 4 1 4
6. 14 56 10 4o 1 4
7. 20 80 4 16 1 4
8. 17 68 8 32 o 0
9. 21 84 4 l6 0 o
10. 20 80 3 12 2 8
TOTAL 193 45 12
MEAN 19.3 77.2 4.5 18 1.2 4.8

 

125

TABLE 17.

COMPREHENDING SCORES.

AMERICAN MATURE READERS

 

 

SUBJECT No. NO 2 NoglNlMAL% NO YES 2

1. 12 48 5 20 8. 32

2. 13 52 7 28 5 20

3. 13 52 9 36 3 12

4. 18 72 3 12 4 16

5. 17 68 6 24 2 8

6. 16 64 6 24 3 12

7. 13 52 8 32 4 16

8. 19 76 4 16 2 8

9. 11 44 8 32 6 24

1o. 16 64 7 28 2 8

11. 15 60 5 20 5 20
TOTAL 163 68 44

MEAN 14.8 59.3 6.1 24.7 4 16

 

126

posited two research questions to guide the investigation.

These two research questions are:

(1)

(2)

Are there any significant differences in the
oral reading strategies of the Chinese readers
and English readers in any of the following
categories: (a) graphic and sound similarity,
(b) grammatical function, (c) syntactic
acceptability, (d) semantic acceptability

and meaning change?

What is the relationship between meaning and
Chinese print? Can the readers get at the
meaning of a character even if they assign
an ”unconventional” sound to the character?

The results of the study can be summarized as follows:

(1)

(2)

(3)

(4)

There are no fundamental differences between
readers of English and Chinese as far as the
basic reading process is concerned. A11
readers regardless of their language back-
grounds are using the visual information
provided by the text. This is evident in

the fact that both the readers of English and
Chinese produce miscues which look like and
sound like the original text words.

The Chinese readers differ from the English
counterparts, both in the Blair study and

in other miscue studies reviewed, in that when
the Chinese readers miscue, their substitutions
show less graphic and sound similarity to the
text items than do those of English readers.
This is directly related to the differences

in their writing systems.

The miscues produced by readers of English
demonstrate that the higher the grapho-phonic
mean, the lower the semantic acceptability scores.
However, the miscues of Chinese readers with
high graphic similarity tend to show high
semantic acceptability scores. This disparity
between these two groups of readers is also
due to the differences in their language
systems.

The examination of the grammatical function
of the miscues produced by the Chinese
readers reveal that 89% of all substitution
miscues retain the same grammatical function
as the original text words. Only 5.7% of

(5)

(6)

127

the miscues result in a change of the
grammatical function. These figures are
very close to that of the Blair study.
87.6% of Blair's American readers' miscues
have the same grammatical function as the
expected responses and 9.8% of the miscues
have a different grammatical function.

The high percentage of same grammatical
function of the miscues produced by these
two groups of readers of very different
language backgrounds suggests that readers,
in general, are very sensitive to the syntactic
featurescaf their language, regardless of
what kinds of writing systems they are
exposed to. It is also a strong indication
that readers bring their syntactic knowledge
with them into reading.

The data on syntactic acceptability of the
Chinese readers' miscues indicate that 87.6%
of the sentences involved with miscues are
syntactically acceptable in the story. Blair
reports that 93.5% of his American readers'
miscues are syntactically acceptable. Blair's
figure is somewhat higher then that of
Devine's Spanish-speaking readers reading
English (66%) and Jensen's native speakers

of English reading English (82%). The
extremely high percentage of syntactic
acceptability found in the Blair study may

be attributable to the type of miscue
analyzed. Blair focused only on the
substitution miscues, while Devine, Jensen
and the present study analyzed all kinds of
miscues identified in the readers' oral reading.

The data from miscue types and structual
levels reveal that the performance of Chinese
readers is very similar to that of the readers
of English. For both groups of readers, the
most frequent types of miscues overall are
substitution, then insertion, then omission,
and finally reversal. The data also indicate
that Chinese readers, like their English
counterparts, produce miscues which involve
various levels of language structures. The
major difference between these two groups

of readers is that Chinese readers show a
higher degree of involvement at the bound
morphemic level.

128

(7) The data on syntactic and semantic accept-
ability of the miscues for the readers of
Chinese show that the Chinese readers are
more successful in using syntactic than
the semantic information in the text. This
is obvious in the fact that the percentage
of miscues totally acceptable syntactically
for all of the readers is higher than the
percentage of miscues totally acceptable
semantically (87.6% vs 79.6%). Similar
findings concerning readers' performance
in the categories of syntactic and semantic
acceptability are reported in both the
Blair and the Devine studies: 93.5% and 70.1%
for Blair's American readers: 66% and 32%
for Devine's Spanish group.

(8) Reading is primarily for meaning. Both the
Blair and the present study show that the
majority of readers' miscues (84% for Blair's
readers of English and 95.2% for Chinese
readers) preserve all or most of the intended
meaning of the text. Only a small number of
miscues (16% for the Blair and 4.8% for the
present study) cause meaning change or loss
in meaning.

As to the question "What is the relationship between
meaning and Chinese print?" the data in the present study
seem to suggest that Chinese readers are able to get
directly from the written symbols to meaning. This
conclusion is reached on the basis of the following
observations. First, miscues of the Chinese readers have
a tendency to show high "no similarity" scores graphically
and phonemically. Secondly, the majority of the Chinese
characters have a built-in meaning component, namely, the
radical. Readers in the present study appear to be able

to make use of this specific advantage facilitated by the

text language. It is clear from the evidence that miscues

129

with high graphic similarity also tend to show high
semantic acceptability. Thirdly, the readers of Chinese
are able to extract meaning from the printed forms even

if they fail to assign "conventional” pronunciation to the
words. Chinese readers produced a set of miscued words
which are high both in graphic similarity and semantic
acceptability, but very low in sound similarity. The
third factor seems to indicate strongly that Chinese
orthography can be mapped onto meaning at the level of
words rather than on phonemes. Consequently, it is not
necessary for a Chinese reader to translate the written
symbols into spoken language. Here, we find a significant
difference in possible silent reading strategies for
English and Chinese readers. Because the bulk of the
meaning lies in the signific, the readers of Chinese can
ignore the phonetic component when reading silently.
English readers, on the other hand, are easier to be
seduced into attending to grapho-phonic similarity first,

and then to translating spoken sound into meaning.

CHAPTER IV
THE TRANSACTIONAL THEORY OF READING

As Kenneth Goodman has said, miscue analysis provides
”windows" that enable us to see into the actual reading
process itself. Miscues are not simply random errors on
the part of an ignorant or careless reader, but highly
informative aspects of a reader‘s total language pattern
if we know how to make use of them. Furthermore, in
addition to the miscue analysis Goodman supplements the
procedure with the "re-telling.” In this way it could be
argued that the model is a complete picture of the reading
experience.

However, the retelling component has several
intrinsic problems which merit serious consideration and
which qualify to some degree any claim that the Goodman
model is a complete and final picture of the total reading
process in itself, not that Goodman has made such claims
himself. One of the major problems of the retelling was
pointed out by James Kalmbach (1980) when he said

Perhaps the most serious flaw of a story outline.

is that it is forced to treat each event and each

character more or less equally. . . . Consequently,

a retelling score does not yield insight into the

nature of the narrative the subject has created.

It only measures that narrative against the
original story (pp. 14-5).

130

131

Another reservation about the Goodman model was

expressed by Louise Rosenblatt (1978) in her The Reader,

 

the Text, the Poem. Rosenblatt repeated and expanded this

 

reservation in a 1981 article. She does agree with
Goodman in finding it useful to use the conception of
cue-systems to describe and analyze the reading process:

The process of reading a text, a process of
organizing meaning, can be understood as a
selective and synthesizing activity.

One seeks in the opening lines of a text syntactic
cues as to the sentence pattern to anticipate,

and semantic cues on which to base expectations
about what is forthcoming. . . . The reader
carries on a complex process of selection and
synthesis, the arousal and fulfillment (or frustra—
tion) of expectations, the construction of a
growing "meaning” until the final synthesis or
organization is achieved . . . (1981, p. 15).

But she finds the Goodman presentation handicapped by a
lack of awareness of the difference between reading a
literary document and reading a non-literary writing.
This general outline of the reading process is still
severely incomplete, however, precisely because it
does not provide for the important distinction
between the reading process that produces a
scientific report and the reading process from
which emerges a literary work of art (1981, p. 15).
In Rosenblatt's view, the Goodman's preoccupation with
information presented in the text in a sense hinders a
reader's fruitful appreciation of the reading experience
when that experience is concerned with poetry or fiction,
say, as opposed to scientific writing. Rosenblatt's

criticism is especially directed toward the technique

Goodman and other researchers use to elicit retelling.

132

According to Rosenblatt, giving instructions to a reader
before the actual reading starts to the effect that the
reader is to summarize when he completes the reading
amounts to forcing the reader to focus his attention

on certain kinds of information. In her 1981 article,
Rosenblatt uses a persuasive example to demonstrate her
point. She argues that a reader can read solely for
factual information if directed to do so. Suppose that
before presenting Yeats's line "I will arise and go now,
and go to Innisfree. And a small cabin build there, of
clay and wattles made," the researcher tells the reader
that he wants to know what information these two lines
offer concerning the material used in the building. In
Rosenblatt's opinion, the reader would then approach the
lines "with an efferent stance" and focus his attention
on the "referents" (1981, p. 17). In this situation” whether

the word in the poem is 'cabin" or "cottage" or ”dwelling"
is not a matter of importance because the reader's
concentration is on trying to find out what kinds of
material are used in the building; the aesthetic aspect is
lost.

Rosenblatt feels that one of the earliest and most
crucial steps in any reading experience is the "selection
of either a predominantly efferent, or a predominantly

aesthetic stance toward the transaction with the text”

(1981, p. 18). In this regard, the RMI model's special

133

emphasis on factual comprehension in the retelling is

a drawback, according to Rosenblatt, when it comes to the
reading of literary works of art as opposed to scientific
material. A

Rosenblatt's own reading theory, which she has
termed "transactional,” is an alternative which she believes
gives a fuller and more accurate account of the reading
process. Two basic tenets of her reading theory are
especially relevant to our discussion.

Central to Rosenblatt's theory of reading is the
view that reading is an "on-going" process which involves
both the literary text and the reader of the text. Al-
though this point of view is increasingly acceptable in
the early 1980's, she pioneered in such thinking as
long ago as 1938. In 1938, Rosenblatt published

Literature As Exploration, a book aimed at increasing the

 

teacher's awareness of the existence and importance of

the reader's contribution to the reading activity. The
main thesis of her reading theory, which she refined and
expanded over the next forty years, appeared for the first
time in Chapter Two: "The Literary Experience," In the

opening paragraph of this important chapter, she says

Terms such as the reader, the student, the literary
work, have appeared in the preceding pages. Actually,
these terms are somewhat misleading, though convenient,
fictions. There is no such thing as a generic reader
or a generic literary work; there are only potential
millions of individual readers of the potential
millions of individual literary works. A novel

or poem or play remain merely inkspots on paper until
a reader transforms them into a set of meaningful

134

symbols. The literary work exists in the

live circuit set up between reader and the

text: the reader infuses intellectual and
emotional meanings into the pattern of verbal
symbols, and those symbols channel his thoughts
and feelings. Out of this complex process emerges
a more or less organized imaginative experience

(p. 25).

In Literature as Exploration Rosenblatt proposes that both

 

writer and reader are creative and that the literary experi-

ence that exists between reader and text is an interaction.

 

By 1954 she cametxabelieve that the relationship between

these two is a transaction, not an interaction. This is a

 

 

major change. The "interactional" terminology was rejected
by Louise Rosenblatt because it implies that reader and
text are two separate, independent entities acting on
one another. The concept "transaction," developed by
John Dewey and Arthur F. Bentley, seemed to her to be
more appropriate to designate those situations where two
entities are truly only separate aspects of a single
total situation in an ongoing process. A knower cannot
exist without a known. Nothing can be known if there is
not a knower. The relationship between a knower and a
known is transactional. The two exist only in relation-
ship to each other. In the same sense, there can be no
"reading" unless there is a reader and something to be
read. The term "transaction" is able to capture the
dynamic nature of the reading process, in which reader

and text are no longer two separate, distinct entities,

135

each self-contained, that interact, but rather two things
involved in a process that demands the existence of the
two simultaneously in a special relationship. Rosenblatt
describes this two-way, reciprocal relationship between a
reader and a text in the following fashion:
The poem, then, must be thought of an an event
in time. It is not an object or an ideal entity.
It happens during a coming-together, a compenetra-
tion, of a reader and a text. The reader brings to
the text his past experience and present personal-
ity. Under the magnetism of the ordered symbols
of the text, he marshals his resources and
crystallizes out from the stuff of memory, thought,
and feeling a new order, a new experience, which
he sees as the poem. This becomes part of the
ongoing stream of his life experience, to be
reflected on from any angle important to him as
a human being (1978, p. 12).
Thus, the text constrains the reader. On the other hand,
the reader brings to the text a special, unique, individual
set of experiences from life. The reader also offers
certain constraints on the process. Rosenblatt further
distinguishes the ”text" from the ”poem." The text consists
of ink marks on paper. The poem is created by the reader
in his mind through a transaction with the text. The
text has invited the reader to construct certain meanings
and to live through certain experiences; and the text
has also constrained the reader by its particular pattern
of verbal symbols. In Rosenblatt's view, the notion
that text is all-important, which she says is the

assumption of many teachers (including those called "New

Critics" in the 1950's), misses the reader's contribution.

136

Also, in her view, the notion, now widespread, that the
reader is all-important (she attributes this position to

David Bleich's Subjective Criticism, 1978), what she calls

 

the "psychological" view, misses the vital constraining
influence of the text. The position Louise Rosenblatt

sets forth clearly in her recent book The Reader, the

 

Text, the Poem is that reader and text are equally impor-

 

tant in the transaction that creates the poem.

The concept of the poem "as experience shaped by the
reader under the guidance of the text" bears very
significant implications for classroom teachers and
for those who are interested in the reading process.

If both reader and text are active elements in a reading
situation that produces a literary work of art, one does
not need to accept the idea that there is one and only

one "correct" reading of the text. However, she does
believe it is possible to decide in many cases which read-
ing is more relevant to the text, due to the guidance

and constraints provided by the text, a matter she takes
up under "validity of interpretation."

The second important premise of Rosenblatt's theory
of reading is her distinction between two "stances" that
a reader may transact with a text. One is the "aesthetic,"
the other is the "efferent." The reader actively elicits
meanings guided by the verbal stimuli of the text, no

matter whether he is reading a newspaper or the text of

137

Joyce's Ulysses. However, there are some qualitative
differences between these two kinds of reading. In an
aesthetic reading, the reader is reading a text for
literary experience. His attention is focused directly
on his "moment-to-moment alertness to what is being
activated in his consciousness” by the printed symbols
during his relationship with the text. In an efferent
reading, the reader is reading a text for information

he wishes to carry away aftertiuzreading event has taken
place. Efferent comes from effere, Rosenblatt says, a
verb meaning "to take or carry away." The distinction
between these two stances of reading derives from the
reader's choice of focus. Here Rosenblatt borrows
William James's notioncxf"selective attention" as her
theoretical basis for formulating literary experience.
Human consciousness, according to James "is always
interested more in one part of its object than in another,
and welcomes and rejects, or chooses, all the while it
thinks" (1890, pp. 284, 288). This "choosing activity”
is central to Rosenblatt's reading theory. First, it
provides theoretical support for the transactional stances
that a reader is to adopt, and secondly, it makes it
possible for a reader to choose from the multiple
possibilities provided by the text those elements which he

synthesizes and organizes into a literary experience.

138

Theoretically speaking, any text can be approached
either efferently or aesthetically, since "referents and
associations are all elements of the same stream of
thought" (Rosenblatt, 1981, p. 16). That is, a menu may
be read aesthetically by a gourmet, and a poem may be
read efferently by a would-be critic or examination-taker.
However, it is crucially important to make the distinction
between the reading process that results in a scientific
report and the reading process in which a poem is evoked.
In Rosenblatt's view, failure to recognize the difference
is to treat all kinds of reading as if they were the same.
One cannot look at the text and predict the response
that will emerge from a reading. Reading, according to
Rosenblatt, is always an event which involves a particular
reader with a particular text in a particular situation.
Furthermore, she says, the text of a poem, if the poem
is to be a poem, must be read aesthetically before it is
read efferently. If the would—be reader of poetry
does not value his own lived-through experience of
aesthetic reading, he will see the text of the poem as
some kind of puzzle he must solve. Without the pleasure
of the aesthetic experience to motivate him in future
poetry reading, he may simply turn away from poetry as
something he cannot make sense of. He will not value the
experience of reading poetry, only the process of gathering

information. As I have mentioned earlier in this section,

139

the main criticism Rosenblatt makes against the Goodmans'
RMI is that it fails to make the distinction between the
two stances that a reader may take toward a text, and
concentrates the reader's discussion on efferent reading.

Rosenblatt's emphasis on the lived-through experience
should not be taken to mean that she believes reading
is purely emotive. In fact, as she carefully points out,
even in great poems, which require the extreme aesthetic
stance, there are some ideational elements. While Rosen-
blatt does indeed emphasize the importance of aesthetic
experience in reading literary works, she is aware that
"although many readings may fall near the polar extreme,
many others may fall nearer the center of the continuum”
(1981, p. 17). In other words, many readers may initially
approach the text with an aesthetic stance, but after the
reading experience, the reader may find that he has
retained much of the informationiIISpite of the fact
that retention was not the primary concern of his reading.
Or some readers may select a "predominantly" efferent
stance, but still possess some degrees of awareness of the
qualitative and affective appreciation of the text.
Reading, as a form of a thinking process, is characterized
by a "continuing stream of choices" on the part of the
reader.

Rosenblatt followed up the idea of "reading as
transaction" by conducting experiments 1J1 reader response.

An example of the kind of research she has been carrying

140

out during the last forty years that centered on the use
of poetry in the classroom is given in Chapter Two in

The Reader, the Text, the Poem. Here she shows how a
classroom of students responded to a Robert Frost poem,
"It Bids Pretty Fair." The advantage her kind of research

has over the pioneering work of I. A. Richards in Practical

 

Criticism (1929) is immediately clear to the scholar who

 

knows both methods of research. She does not show I. A.
Richards' rather fixed notion of how meaning is built in
the reading process. Out of Richards' view came, eventually,

New Criticism with its emphasis on the text at the expense

 

of the role of the reader. She demonstrates in what way
the reader is active and in what way the text constrains.
In spite of Rosenblatt's fundamental and seminal ideas
about the reading process, the value and significance of
her theory is yet to be sufficiently recognized. It is
only recently that researchers interested in reading have
begun to understand the central importance of Rosenblatt's
point of view. Researchers at Indiana University, origi-
nally working within the tradition of miscue analysis, have
started investigating the nature of the act of retelling.
Jerome Harste and Robert Carey (1979) reported that a
retelling is the result of a "semantic transaction"
between the reader and the text, a creation of a new text
”as original and distinctive as the author's" (p. 17). Other

research which also supports Rosenblatt's position was

141

conducted by Sharon Smith. In her "Retelling as Measures
of Comprehension: A Perspective" S. Smith argues that

The analysis of the retelling, then, should not empha-

size recall in the form of repetitions or paraphrases,

although these will be of interest in the analysis.

The focus here is on the reader as author of his

own version of the content, which will indicate

the nature of his own active structuring processes.

The analysis, then, should attempt to reveal how the

reader is breaking meaning out of the text's language

and then reconstructing it in language of his own

(1979, p. 90).

From the passage, it is obvious that S. Smith is arguing
for the need to take both the contribution of the text
and of the reader into consideration when analyzing the
retelling.

The second part of this dissertation is a further
study of the Chinese readers already examined through the
techniques of the RMI but now making use of Rosenblatt's
transactional theory of the literary work. Instead of
lookingznzthe cues, we are now looking at the responses.

One of the very first and essential problems which
faces most of the researchers on literary response is the
problem of measurement devices. Different measurement
devices may channel the subjects' responses to certain
modes of responses, due to the kinds of questions asked,
the instructions given, the influence of the investigator
and the ways response statements are coded. The RMI,
for instance, may encourage more text-based responses.

As noted earlier in this chapter, Rosenblatt has

conducted a series of experiments in classrooms to find

142

empirical evidence to support her notion of ”reading as
transaction." The method she used to elicit responses
from her readers was simply to ask them to read the text
chosen and to write out whatever thoughts and feelings
had come to their minds. This technique would allow a
reader to select a reading stance on the basis of the
verbal stimuli of the text on the reader and the particular
expectations that reader might bring into the reading
situation. In Rosenblatt's view, the selection of a
dominant stance from which to transact with the text is
the earliest and most important decision in any reading
event. It can be argued that the device "let a reader
choose his own stance,"cn:to use Beach's term, "free-

association,‘ may bias a reader toward more subjective
responses. However, since Rosenblatt's method imposes no
prior restrictions on the reader, the reader has the
freedom to adopt any modes of reading which he feels most
appropriate. If reading is a problem-solving and decision-
making process, this very first step of deciding an appro—
priate reading stance should be considered as a part of the
whole reading process. Rosenblatt's method of eliciting
response statements from the readers appears to be the

most unbiased research technique that we have at the
present time.

Rosenblatt has provided a very useful way to collect

data if one wants to conduct research within the framework

143

of the transactional theory of reading. After the data was
collected, Rosenblatt analyzed the data in what she called
an "intuitive" study, which is especially relevant when
a small number of subjects are involved. When a larger-
scale investigation or a cross-cultural comparison is
intended, this kind of "intuitive" study may no longer
be sufficient. Although Rosenblatt did not develop
any coding system for this type of research, she does
make it clear what she thinks will be the possible
solution to this problem. In her 1977 paper delivered
at the Buffalo Conference on Researching Response to
Literature and the Teaching of Literature, Rosenblatt
made the following remarks: "Research designs for study
of response under my direction have included both system-
atic analysiscaf multiple elements present in the
responses of appropriately-selected groups, and also
intuitive study in depth of full individual responses"
(1977, p. 23). The research strength to be found in a
method that combines a content analysis with "intuitive"
study, or case study, is obvious. These two approaches
complement each other and enable researchers to study a
reader's reading behavior in breadth as well as in depth.
Louise Rosenblatt is not the only person who sees the
possibility of using these two analytical devices together
as a way of understanding the complex nature of the reading

process. Alan C. Purves, co-author of the Elements of

144

Writing About a Literary Work: A Study of Reaponse to

Literature, has made a similar suggestion: "The next

 

direction in research, thus, might well be exploration

into the complex system of literary responses. Such
exploration might well employ the case-study technique of
exploring many aspects of responses of a few individuals.
This technique should be combined with multivariate
analysis, multidimensional scaling, partition analysis, and
other more sophisticated treatment" (Purves and Beach,
1972, p. 37).

From the two quotations given above, it seems that
both Rosenblatt and Purves believed that the best possible
way to look into the dynamic nature of reading is by
joining content analysis methodology and case-study
techniquesinto a unified measurement device.. Since the
major purpose of this part of the present study is to
discover whether the Chinese adult readers approach a
literary text in the ways Louise Rosenblatt has described
in her transactional theory of reading, it has proved
best to follow the research designs recommended by her as

closely as possible.

Content Analysis

The inquiry into the literary "response” of a reader
is certainly not a recent interest. As early as 1929,

I. A. Richards pioneered an approach to the study of

145

reader's reactions to literary texts by analyzing the
readers' interpretations of thirteen poems. Richards'

method, set forth in his still very useful Practical

 

Criticism, is characterized by the "intuitive" judgment

 

of the researcher rather than by scientific control and
statistical analysis. Over a period of several years,
Richards issued printed sheets of poems with the authors'
names withheld to his Cambridge students. The readers were
asked to study poems at their leisure and then to make

and return comments on what they had read. Richards

found that to the majority of his readers these poems

were difficult, obscure and unmanageable. After intensive
analysis of a wide range of students' comments, Richards
identified ten chief sources of difficulties: (1) the
difficulty of making out the plain sense of poetry; (2)
difficulties of sensuous apprehension; (3) difficulties
that are connected with the place of imagery; (4) pervasive
influence of mnemonic irrelevances;(5) stock responses;

(6) sentimentality; (7) inhibition; (8) doctrinal
adhesions; (9) technical presuppositions; and (10) general
critical preconceptions. The theoretical assumption under-
lying these ten categories is clear. As pointed out by
Purves (1979), Richards' conclusions followed "from the
premise of hermeneutics that the literary text contained

a verifiable essence." Not only the meaning of the text

could be verified, so could imagery, feeling, tone and the

146

intention of the author. If one accepts Richards'
position, then the primary concern of research will be to
determine how close the reading of an individual is to a
hypothetical "correct" reading of the text.

In 1964, thirty-five years after Richards' first
research, James Squire undertook to study the responses of
fifty-two ninth and tenth grade students ranging from
age 14 to 16 while reading four short stories. The
rationale for Squire's research is that "the study of
literature must involve not only consideration of the
literary work itself but also concern for the way in
which students respond to a literary work" (1964, p. 1).
What makes this statement significant is the whole notion
of the reading process. Implicitly stated are the ideas
that meaning does not reside in the text alone, and the
reCognition of possible contributions from the reader
which help to shape the meaning. Perhaps the most important
contribution Squire made to the study of literary response
is the notion that reactions to a text should be obtained
during the act of reading. Squire believes that since most
of the studies of literary responses have been based almost
entirely on written responses obtained after the reading is
completed the facts gathered under this kind of condition
"may conceal as much as they reveal" (1964, p. 1). In
Squire's view, it is crucial to study "how the students

came to think and feel" while they read and responded to a

147

literary text. In order to capture the developmental sense
of the reading process, Squire developed a research design
which allowed him to gather data while readers were reading.
He divided each of the four stories selected into six divi-
sions. Each reader was asked to comment orally on what he
had read immediately after finishing reading one segment.
All of the subjects' responses were recorded. A total of
14,494 response statements were generated from the 52
readers' reading. Squire identified the following seven
categories of responses: Literary Judgments; Interpreta-
tional Responses; Narrational Reactions; Associational
Responses; Self-Involvement; Prescriptive Judgment and
Miscellaneous.

Purves began his research on literary response shortly
after Squire's work was published. The conceptual grounding
of Purves's work is very similar to that of Squire; however,
Purves believes that some of Squire's categories overlap and
are too broad. There is a need to refine these categories
and have a higher degree of discrimination. In 1968, Alan
Purves and Victoria Rippere published their book, Elements
of Writing about a Literary Work: A Study of Response to

Literature, in which they proposed a new system for content

 

analysis of responses to literature. They first organized
into four major categories the statements respondents

made about a literary text:

148

Engagement-Involvement, the first category,
defines the various ways by which the writer
indicates his surrender to the literary work,
by which he informs his reader of the ways in
which he has experienced the work or its
various aspects. '

Perception. . .encompasses the ways in which
a person looks at the work as an object distinct
from himself.... This conception (analogous

to "understanding") is analytic, synthetic, or
classificatory and deals with the work either in
isolation or as an historical fact needed to

be related to a context.

Interpretation is the attempt to find meaning

in the work, to generalize about it, to draw
inferences from it,tx)find analogies to it in

the universe that the writer inhabits. . .
Interpretation can be either of the form or of
the content. If it is a form, interpretation

is the drawing of inference from a formal aspect
of the work. . . . If the interpretation is of
the content, it can be as simple as the inferred
generalization that is character analysis, for
such generalization is based on knowledge or a
preconception about human nature. Generalization
leads to more complex interpretations: that of
seeing the work as imitative of the world, that of
seeing the work as distillation or abstraction
from the world, and that of seeing the work as a
medium of judgment or didacticism.

Evaluation, encompasses the statements about why

the writer thinks the work good or bad. His

judgment may be derived from either a personal or

an objective criterion (Purves, 1968, pp. 6-8).
In addition to these four main categories, Purves proposed
a fifth category, "miscellaneous." This category was
designed'for those statements which do not deal directly
with the work itself. Under "miscellaneous" appear two
elements often seen in literary response, references to

other writers on literature and comparison with other

works.

149

The four categories exist as the general framework of
the elements. They are further divided into 24 subcate-
gories, distributed as follows: 4 subcategories for
Engagement-Involvement; 9 for Perception; 6 for Interpre-
tation; 4 for Evaluation and l for Miscellaneous. Under
these 24 subcategories there are 115 elements. 10 of them
are distributed to the category of Engagement-Involvement;
49 to Perception; 34 to Interpretation, 16 to Evaluation
and 6 to Miscellaneous.

Purves's concern with readers and the roles they play
in the relationships with a literary text can be easily
seen in the fact that he devoted two of the four main
categories to report the response statements which are
more reader-based. The categories of Engagement-
Involvement and Evaluation can be regarded as being regu-
lated by the readers, while Perception and Interpretation
can be regarded as being regulated by the text. However,
when it comes down to the subcategories and elements, we
get a quite different picture. Out of 115 elements, only
26 can be identified as either Engagement-Involvement
(10) or Evaluation (16), yet there are 34 elements in
Interpretation and 49 in Perception. Since there are
almost three times as many in the categories of Perception
and Interpretation, the readers have more chances to be
coded either as "Interpreters" or "Perceivers." While

what has been pointed out does not necessarily mean this

150

analytical system is invalid, it suggests that there are
some limitations (Cooper & Michalak, 1981, p. 165).

In 1972, Purves and Beach presented a shorter version
of the original Purves coding system. ‘This new Purves-Beach
sCheme still maintains the four main categories, but each
is further divided into at least two subcategories:

Engagement-Involvement

(1) personal statement about the reader, an
"autobiographical digression"

(2) personal statement about the work, expressing
personal engagement with it

Perception

(3) descriptive statement or retelling part of
the work

(4) descriptive statement of aspects of the
work: language, characters, setting, etc.

Interpretation

(5) interpretative statement of parts of the
work

(6) interpretative statement of the whole work
Evaluation

(7) evaluative statement about the evocative-
ness of the work

(8) evaluative statement about the construction
of the work

(9) evaluative statement about the meaningful-
ness of the work

The newer Purves-Beach coding method has one major advantage
over the older one. As is pointed out by Odell and Cooper,

"the newer scheme enables us to capture information that would

151

be lost by using only the four broad categories. At the
same time, it allows a more convenient analysis than that
permitted by the use of either the twenty-four subcategories
or 115 elements of Purves' original system” (1976, p. 206).
Furthermore, this newer system is more neutral in the
sense that it has a more balanced distribution of reporting
components of the readers' response statements.

It is this Purves-Beach scheme which was adopted for

the present study.

Related Research

Knowledge of the nature of the reading process has
increased noticeably over the past fifteen years, largely
due to the great growth of interest in the reader and
improved research methodology. Advances in methodology
make it possible for researchers to do research with a
larger number of subjects and to deal with a wider range
of issues. For instance, since Squire's content analysis
was published in 1964, and followed by the Purves-Rippere
system in 1968, two large-scale investigations have been
completed, one of which involved ten countries, and several
dozen of doctoral dissertations at universities across the
United States have been completed. These studies have
covered a great range of interests, from the effects of
schooling (e.g., Weiss, 1968; Kirkpatrick, 1972; Heil, 1974;

Lucking, 1975), to age differences (Purves and Rippere,

152

1968; Mertz, 1972; Purves et al., 1973; Applebee, 1973),
sex differences (Purves et al., 1973; Applebee, 1973),
attitude and personality (Faggiani, 1971; Kuehn, 1973;
Cornaby, 1974), and the effects of different texts on

the readers (Cooper, 1969: Angelotti, 1972; Somers, 1972;
Haug, 1974 and many others). There is another group

of researchers whose main concern was with the ”process
of reading" itself. They addressed themselves to such
matters as how responses were initially formulated or the
differences between reactions made during small-group
discussions and private reading. Pollock (1972) and Beach
(1972) belong to this group.

The following examples of research on the reading pro-
cess were selected to demonstrate some of the common
traits found in the reading behavior of native speakers
of English reading in English. Since one of the major
purposes of the present study is to discover whether there
are any discernible patterns of Chinese readers' reading
behavior which are comparable to that of readers of English,
these findings from the discussion further serve as the
basis for comparison.

As previously mentioned, Squire explored the develop-
ment of responses of fifty-two ninth and tenth grade
students while reading four short stories. After intensive
analysis of the data according to the seven categories,
Squire reported that in spite ofthe discernible individual

variations in responses, "group tendencies are clearly

153

demonstrated in the development of interpretational,
literary judgment, self-involvement, and narrational
scores" (1964, p. 51). Among the categories of response,
Squire found Interpretational statements were the most
common (42.7%), followed by Narrational (21.36%), Literary-
judgment (14.85%) and Self-involvement (13.29% . Finally,
he saw that there were some obvious correspondences between
statements of Self-involvement and those of Literary -
judgment: "One of the more interesting findings is the
strong positive correlation between percentages of responses
labelled literary judgment and self—involvement, indicating
that readers who become extensively involved in stories

are also inclined to evaluate the literary qualities of a
selection. The two types of reactions were found to

covary during the average reader's process of responding

to an entire story" (1964, p. 51).

Squire also identified six sources of difficulty in
literary interpretation: failure to grasp meaning, reliance
on stock responses, happiness binding, critical predisposi-
tions, irrelevant associations, and unwillingness to
suspend judgment (1964, pp. 37-47), a format which clearly
indicates strong influence from I. A. Richards.

Using the Squire system of coding, Wilson (1966)
studied the responses to three novels written by fifty-four
college freshmen before and after class discussion. Wilson's

research structure was very similar to that of Squire, but

154

with a more in-depth investigation of nine of his subjects.
Comparing the response patterns for the two set of proto-
cols (before and after), he found a decrease in the number
of literary judgments and self-involvements as a result of
class discussion and an increase in interpretational reac-
tions. The results of statistical analysis repeated
Squire's finding, namely that Interpretation was the domi-
nant mode of students' responses, account for 65.6 percent
of the total responses. Wilson's more important findings,
however, were derived from the in-depth studies. There
were very few examples of the six sources of difficulty
noted by Squire, a finding which could be attributed to the
result of careful analysis of individual protocols. Wilson
pointed out that although some responses "were unSOphisti-
cated or evasive . . . fragmentary or partial . . . it
would have been difficult to demonstrate that these inter-
pretations were misinterpretations based on misreading or
on stock responses." Wilson's finding was not necessarily
contrary to Squire's; however, it suggested that statisti-
cal analysis can be supplemented profitably with in-depth
study of a limited number of subjects.

Another example of Wilson's conscious attempt to test
some of Squire's findings was found in his exploration of
the relationship between self-involvement and literary-
judgment. Squire reported that these two kinds of responses

tend to appear in the same reader's responses, but he did

155

not explain why they seemed to go together. On the basis
of the statistical results and individual study of certain
selected protocols, Wilson speculated that effective
interpretation might require an initial self-involvement:
"Perhaps most subjects can begin to concern themselves only
with questions which have personal importance. That is to
say, interpretation may be a secondary predictive process,
impossible without initial self-involvement" (Wilson, 1966,
p. 38).

Angelotti (1972) looked at the patterns of responses
to dissimilar works by comparing sixty-six eighth graders's
responses to a junior novel and to an adult novel. By
following Squire's lead, Angelotti divided each of the two
novels into four segments representing the exposition,
development, climax and the conclusion. The reactions
of the readers to each division of the junior novel

Tuned (hit were compared to the corresponding division of

 

the adult novel A Separate Peace and analyzed by using

 

Purves' (1968) coding system. Angelotti's major findings
include:

(1) The reaction to the more easily understood
novel was most often interpretative, while
the more difficult adult novel generated
more perceptual and factual responses.

(2) The ability of students to understand the content
of literature is the main factor in determining
their mode of responses.

(3) There were some discernible patterns of
developmental process across sections of
the stories. Involvement increased as the

156

story progressed, reaching a high point at the
climax, whereas evaluation declined to its
lowest point at the moment of climax. For

the simpler novel, statements of perception
were highest at exposition and lowest at the
climax, while for the harder novel they were
lowest during the exposition, and highest at
the climax. Interpretative responses to the
junior novel rose from a low during exposition
to a peak at the climax before falling sharply
in response to the conclusion. In response to
the adult novel interpretations were at their
highest during exposition and decreased to

the low point at the conclusion.

In 1972,Pbllock explored the responses moshort stories
made by seventy-three students selected from ninth graders,
eleventh graders and college freshmen. Among these seventy-
three readers, twenty of them were college students.
Pollock's college group was close to the present study in
terms of age and scope, therefore, the findings reported
in Pollock's study concerning the reading behavior of
college freshmen were used as one basis for comparison.

Content analysis was performed on the response state-
ments using Purves' major five categories and subcategories.
Pollock's data reveal for the college group that interpre-
tation was largest for the categories at 34.8%, with
perception next at 26%, ahead of engagement (19.1%) and
evaluation (18.2%). The ninth graders produced the largest
percentages of perception and evaluation while the dominant
mode of the eleventh graders was engagement. One of the
major conclusions Pollock drew from his data was "Maturity

and additional schooling were associated with more total

response, increased attention to interpretation, and

157

decreased attention to perception and evaluation in free

responses to stories" (Pollock, 1972, p. v).

CHAPTER V

RESULTS OF THE STUDY II:

FIVE READERS' EXPERIENCES OF A LITERARY TEXT

This chapter is divided into three major parts. The
first part describes the design and procedures employed
in this study (Study II). The second part presents and
discusses the data for the overall group. Data are
presented in terms of mean percentage in each of the
four Purves categories. The third part of this chapter
contains a qualitative analysis of the responses made

by five individual readers.

Design of the Study and Analytic Instrument

Sample and Data Collection

The ten readers who did the oral reading for the RMI
in Study I were also the same readers for this Study II.
As mentioned in Chapter III, these readers were chosen
under two conditions. First, they had to have come from
Taiwan and to hold a bachelor's degree from a Taiwanese
university. This condition was intended to insure that the
readers would have shared a similar educational background
and have achieved a high degree of reading proficiency.

Second, the subjects could not be literature majors. This

158

159

criterion was set up because the primary concern of this
study is to discover some general reading behavior patterns
of Chinese readers. A literature major may respond
differently from non-majors due to hiS/her special training.
To have included one or two literature majors in the

group might have complicated the issue by adding unneces-
sary variables.

These ten subjects were asked to read one contempo-
rary short story written by Lu’Syun. This story was
selected on the basis of a set of criteria originally
suggested by Squire (1964). The story ought to be

(1) interesting enough to be able to evoke responses
from different readers

(2) unfamilar

(3) rich and complex so that it will generate a wide
range of reactions from the readers

(4) of appropriate length

These ten readers were asked to read the story "Medi-
cine" in Chinese. Since "the lived-through experience" is
so important in Rosenblatt's transactional theory of reading,
the researcher felt it was necessary to find some way to
look into a reader's responses to a literary work while the
reader is actually in the process of reading. Following the
leads of Squire and Angelotti, the method used for this
purpose in the present study was to ask the reader to
comment on what he had just read when he completed reading

a certain part of the story. The story chosen for this

160

study contains four natural divisions made by the author.
Therefore, it was not necessary for the researcher to
make any arbitrary interruptions. The following form
would be presented to the subjects before assignment of
the actual reading:
This story consists of four parts. When you finish
reading each part, please stop and tell me any
feelings or thoughts that have come to your mind.
When you complete the reading of the whole story,
you may add any additional feelings or thoughts
about the story as a whole.
The responses made by the ten readers in Chinese were
taped and later transcribed into print. The protocols of
five readers were further translated into English to be

presented and discussed in the section of qualitative

analysis of individual reader's reading.

Instrumentation and Data Analysis

The transcribed data were analyzed and classified
according to Purves' 1972 version of content analysis
described on page 150. A fifth category "miscellaneous"
was added for those statements which were either not
relevant in the context or references to other literary
works. Two independent readers were used to analyze
the data to ensure the reliability of the coding. The
degree of agreement between these two independent analysts
is 88 percent.

The basic unit used for scoring an essay is called a

response "statement," which Purves described as "anything

161

that is set off by its own terminal punctuation, including
sentence fragments and epithets" (1968, p. 68). According
to this definition, a statement does not necessarily
closely correspond to a "sentence" in the purely linguistic
sense. A response statement can be a simple, complex, or
compound sentence, or simply a fragment. The choice that
a researcher makes to divide readers' responses into
statements affects the number of responses reported. Those
researchers who choose to treat "a single theme or idea" as
a scoring unit will tend to divide protocols into more
and smaller units. Wilson, for instance,used this device
and coded the following utterance as four units: "What a
way to live. No money or food or job" (1966, p. 34).
Pollock's study was used as one of the sources for
comparison with the findings of the present study,
especially in the area of quantitative analysis. In order
to stay as close as possible to Pollock's research, the
present study adopted the definition of ”statement” used
by POllock and treated ”the entire sentence" as the basic
coding unit. The following is a protocol collected in the
present study, which is meant to illustrate how response

statements are marked and coded:

162

Sample Protocol

”Medicine"

Reactions to Division I

 

Many words in this section are used metaphorically./l
For example, "road" and "lantern" seem to be used in
accordance with emotion./2 He,[§he author] says, "His
[Lau Chuanflg] hope lies in front of him;" later when he
takes the steamed bread mixed with blood home he is again
hopefu1./3 The metaphors used here are obvious./4 I feel
that these episodes are very vividly described./5

Reactions to Division II

 

This section is entirely about how they treat the
steamed bread before their coughing child eats it./6 He
does not use any obvious analogy./7 To write about the
procedure he uses only a simple undisguised technique./8

Reactions to Division III

 

I think that this section in comparison is slightly
disorderly./9 If it is not read from beginning to end,
it is difficult to understand in many areas how the ‘
characters come to be related./10 But after all, L6 Syun's
writing style I believe has always been of this kind./1l
It is his customary manner to use highly concise language
to express the things he wants to express./12

Reactions to Division IV

 

I feel that this story is very sorrowful /13 The
parents try many means to carry out their hopes for their
son./l4 They use all their energy to get the medicine
that can cure their son's illness; nonetheless, their son
died in the end a d only a grave remains./15 I think in
his early years Lu Syun also wrote of ignorant people who
are always tender towards their own offspring./l6 I find
this kind of story very distressing because there is nothing
they can do about it./l7 This is a story full of metaphors

163

and suspenses./18 Though short, the proceedings hold the
reader's attention./l9 More than anything else, I perceive
an extremely strong sense of hopelessness in this story./20

Not only is the main character ignorant, but the other
characters around him are all ignorant as well./21 Even
after the sons were dead, they continue to pray that a
crow appear and fly in the right direction indicating
that "it is the son's divine manifestation."/22 Though
they are ignorant people, their affection for their
children is by no means less than that of educated people
for their children./23

Sample Scoring

Paragraph Statement Subcategory Category

1 l 4 descriptive of Descriptive
aspects of the statement
work

1 2 4 descriptive of Descriptive
aspects of the statement
work

1 3 3 retelling part Descriptive
of the work statement

1 4 4 descriptive of Descriptive
aspects of the statement
work

1 5 7 about the evoc- Evaluative
ativeness of the statement
work

2 6 3 retelling part Descriptive
of the work statement

2 7 4 descriptive of Descriptive
aspects of the statement
work

2 8 8 about the con- Evaluative
struction of statement

the work

Paragraph
3

Statement

9

10

ll

12

l3

14

15

16

17

18

19

20

21

22

23

164

Subcategory

8

about the con-
struction of
the work

about the reader

descriptive of
aspects of the
work

descriptive of
aspects of the
work

about the reader

retelling part
of the work

retelling part
of the work

of the whole
work

about the work

descriptive of
aspects of the
work

about the evoc-
ativeness of the
work

of the whole
work

of the whole
work

retelling part
of the work

of the whole
work

Category
Evaluative
statement
Personal
statement
Descriptive

statement

Descriptive
statement

Personal
statement

Descriptive
statement

Descriptive
statement

Interpretative
statement

Personal
statement

Descriptive
statement

Evaluative
statement

Interpretative
statement

Interpretative
statement

Descriptive
statement

Interpretative
statement

165

Presentation and Discussion of Results:
Quantitative Analysis of the Coding Data

Following the procedures outlined in the previous
section, a total of 436 response statements obtained from
the ten readers were coded. Table 18 presents the results
of this statistical analysis. As the data indicate, the
category with the largest mean percentage is Perception
(37.84%), followed by Interpretation (24.31%), Engagement-
Involvement (22.9%), Evaluation (9.86%), and finally,
Miscellaneous (5%).

This finding is somewhat different from Pollock's.
Pollock analyzed 497 response statements to six short
stories by twenty freshmen college students (native
English speakers) and found the category of Interpretation
to be the dominant form of responses (34.8%), with
Perception next at 26%, ahead of Self-Engagement 19.1%
and Evaluation 18.2%. Miscellaneous responses account
for 1.67 percent (Pollock, 1972, pp. 136-140).

The difference may be attributed to the difficulty
of the text chosen for the present study. In fact, almost
half of the readers complained at some point that the text
was obscure and hard to grasp. In this regard, Angelotti
(1972) in his research on the effects of the difficulty of
the text on the patterns of responses has pointed out that

the ability of students to understand the content of

166

TABLE 18

 

MEAN SCORES OF RESPONSES BY

CHINESE ADULT READERS

 

RESPONSE

 

CATEGORY NO 7°
I. ENGAGEMENT 100 22.935
1. about the reader 32 7.339
2. about the work 68 15.596
II. PERCEPTION 165 37.843
3. retelling part of the work 90 20.642
4. descriptive of aspects of
the work 75 17.201
III. INTERPRETATION 106 24.311
5. of parts of the work 73’ 16.743
6. of the whole work 33 7.568
IV. EVALUATION 43 9.862
7. about the evocativeness
of the work 15 3.44
8. about the construction of
the work 22 5.045
9. about the meaningfulness
of the work 6 1.376
V. MISCELLANEOUS 22 5.045

 

167

literature is a major factor in determining the mode of
response. When a reader dealt with a text of more

complex form and content, he had a tendency to produce
more factual responses, while the leading response pattern
to a less sophisticated story was interpretational. Studies
of the age factors on the patterns of responses also report-
ed that interpretational responses increased at older ages,
which also suggests that the cognitive load contained in

a text might have some effects on the performance of the
reader. The differences in story genre may be another
possible explanation for this discrepancy. The story
selected for the present study adopts a structure which

is commonly found in the genre of "mystery" story.
Generally speaking, a story of suspense requires close
attention to the textual clues from the readers. In
addition, the fact that the borderline between the
categories of Perception and Interpretation is not always
very clear might have influenced the outcome of the
statistical analysis. Alan Purves might have anticipated
this potential problem because in many places in his
Elements of Writing About A Literary Work he stresses the
distinction of these two (cf. pp. 7, 21-22, 27, 31-33).

It is possible for a researcher to move across categories;
thus,in small-scale studies such as Pollock's and the
present ones, a few ”crossing-over" response statements

would be enough to change the picture.

168

Examination of the subcategory reveals that Chinese
readers are most sensitive to the factual information
contained in the text. Narrational and retelling part
of the work led in percentage (20.6%), with descriptive
aspects of the work, such as language, characters, setting,
etc. next (17.2%), followed by interpretative statements of
parts of the work (16.7%). These three subcategories
account for 55 percent of the total response statements.
Pollock's study reported that four subcategories--reaction
to content, content itself, general interpretation, and
interpretation of content together-~constituted about
60 percent of all responses (Pollock, 1972, pp 139-140).
The fact that these subcategories make up more than half of
the readers' responses in both studies suggests that the
linguistic symbols play a very important role in the
readers' process of reading. The subcategories such as
personal statements about the reader, or about the work,
or about the evaluative statements about the construction
of the work, which highlight more personal and idiosyncratic
reactions to literature, take up about 38 percent. The
figure is very significant because more than one-third of
all of the responses bear some traits of creation from the
individual readers. They cannot be directly related to any
overt verbal symbols of the text.

A number of generalizations emerge from our data.

First, the data indicate that the overall patterns of

169

responses made by the Chinese adult readers are not dras-
tically different from that of Pollock's American college
readers. It is true that Chinese readers seem to favor

the perceptional information offered by the text, whereas
Pollock's readers tend to produce more statements which

are interpretative; nevertheless, the differences are marked
only by a very slight margin, as shown in Figure 5. The
second observation, which appears to be even more valuable,
is that we are able to identify and classify from the
Chinese readers' reactions to a literary text response
elements similar to those discussed by Purves and many
other researchers. This second observation clearly
demonstrates that readers, regardless of their diverse
language background, seem to approach the reading of litera-
ture in a very similar fashion. They are sensitive to the
verbal symbols and are able to respond to them. They try

to make sense out of the text by using various sources of
information, such as clues from the text and the reader's
own prior experience and knowledge of the real world.
Readers make statements about the factual inputs contained
in the print, about actions in the plot, and also offer
their explanations of certain events. They make evaluation-
al remarks about the construction or meaningfulness of the
work which they are reading. They also make statements
about their personal reading experience with the text. We
see those general patterns of responses not only in Pollock's

study but also in the studies of Squire, and of Wilson, and

170

FIGURE 5.

A COMPARISON OF PATTERNS OF RESPONSE OF POLLOCK'S AMERICAN
READERS AND THE CHINESE READERS OF THE PRESENT STUDY

 

 

 

 

 

 

 

 

 

 

 

 

 

8 m
U C1 H £3 23
C: O U 0 0
q) .,...| co «4 Q)
E u u u C
0) Cu 0) ‘0 N
3° 8 '61. 3 :1
60 H H (U (D
C: Q) (D > U
O) Q: U Q) 0)
E E
100
90
80
70
6o
50
4o
30
/
20 Z
1° é é 2
0 Q A 4 4 i

 

 

 

 

 

 

 

 

 

 

Z23 Pollock's study

‘—

 

 

l The present study

171

of other researchers.

At this point, it seems appropriate to discuss the
findings of the present study in the context of Louise
Rosenblatt's theory of reading. Rosenblatt has repeatedly
claimed that reader and text are equally important in the
transaction that creates the poem. It is true that quanti-
tative analysis can only provideaigeneral picture concerning
what kinds of information a reader is likely to use in the
act of reading and gives very little help in answering the
question of how a reader and a text come together to evoke
a poem. Nevertheless, the results obtained from the content
analysis do confirm the notion that reader and text are
mutually essential in the transaction. Our data indicate
about 60% of the responses are more or less text-based and
about 40% of the reactions are reader-oriented.

The third observation evolved from the data concerns
the development of responses as readers move from one
division of the story to another. If reading is an on-going
process which takes place over time, as asserted by Rosen-
blatt, we should be able to find some indications which
reflect this process from our data. Table 19 presents the
numbers of response statements made by each reader in each
of the four story division. It is clear from the data that
readers produced more response statements in Division IV
than in the other three sections. Division IV alone

generated 219 units of the total 436, but this fact does

172

maomcmHawosz n z

6>Num=Hw>m u m

6>NumuouauwucH n H

.hwaumlommu Haw comoao who: mHOUOHOHn 660:3 mumvmou 6:6 cum 6
Hmcomuom I m "mucuEuHMum mo vcowwa

o>66aHHowon I a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6.6 N.6H 6.6N N.6N 6.6H 6.N 6.H6 N.66 6.6N 6 6.HH N.N6 6.NN 6.H 6.6 6.6H 6.66 N.6N HN6 666:
HN 6N N6 66 66 6 6 6N 6N 6N 6 N 6 N6 6H H N HH 66 NN H6666
N6.6H N6.6H N66 N6.6H N66H N66H N6.66 N6.66 6H 6666666
6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 H 6 N 6
N6.6 N6.66 N66 N66 N66 N6N N6H N66 N66 N66 NNN N66 6 6666666
6 H 6 6 6 6 6 6 6 N 6 H 6 6 6 6 6 6 N 6
N6 N66 N6.H6 N6.N6 N66 N6.NH N6.66 N6.66 N66 N66 6 666666
6 N 6 6H N 6 6 6 6 H 6 6 6 6 N 6 6 6 N N
N6.6 N6 N6H N66 N66 N6.6H N6.66 N66 N6.6 N6N N6.66 N6.66 N6.6N Now 6 666666
H N 6 N 6H 6 H 6 N 6 6 H 6 6 6 6 N 6 H 6
N66 N6.H6 N6.6 N6.66 N6.66 N6N N6.N6 N6.NH N66 N66 6 666666
6 6 6 6 H 6 6 N H 6 6 6 N 6 H 6 6 6 H H
N66 N6N N6H N6H NN N6.NH N66 N6.N6 N6.6N NN6 N6H N6N N66 N6H 6 6666666
6H HH 6 6 H 6 6 H 6 N 6 6 N 6 H 6 6 6 6H 6
NN.66 N6N N6H N6N N6N N66 N6N N66H N66 N6N N66 6 6666666
6 6 6 6 6 6 6 H N H 6 6 6 N 6 6 N 6 H N
NNN N6.66 N.HH NHH N6.66 N6.66 N66H N6.66 N6.66 6 666666
6 N 6 H H 6 N H 6 6 6 6 6 N 6 6 H 6 N 6
NNN N6.66 NN.NN NHH N6N N6N N66H NHH NHH N66 N66 N 666666
6 6 N 6 N 6 6 6 6 N 6 6 6 6 6 H H 6 6 6
N6 N6.66 N6H N6.6N N6.NH N6.66 N6 N6.N6 N6.N6 N6.N6 N6N N66 NN.66 H 6666666
N 6 6H 6 6 6 6 HH H 6 6 6 6 6 6 6 6 6 6 6
z 6 H 6 6 z 6 H 6 6 z 6 H 6 6 z 6 H 6 6 66666666
mucum 6:6

>H

 

HHH

 

HH

 

 

mo flowuw>wn

 

rmOOMH<U =U¢m mow onmH>HQ mo<u mm9h< mn<£

mmmzommmm ho mmu<fizmommm nz< mammzbz

oH mam<w

 

173

not necessarily mean that all of the response statements of
the four categories arrive at the highest pointzh1percentage
simultaneously when the readers got to the final section.
Figure 6 is a graphic representation of the data summarized
in Table 19. Figure 6 reveals that:

(a) The dominant response is the descriptive one.
More than 40 percent of the responses to the first segment
of the story are descriptive, which strongly suggests
that readers responded to the verbal stimuli provided by
the text right from the very beginning of their reading.
The statements of perception increase in number and reach
a peak of occurrence at division II (62.68%), then
decrease through division III (34.2%) to a relative low
percentage of responses at the fourth division (29.2%). It
is interesting to note that at division II, nine out of the
ten readers scored more than 50 percent on the category of
Perception. Four of them produced nothing but factual
retelling from the text. The reason for the sharp increase
in perception from division I to II may have something to do
with the way in which the story is presented. When we get
to the analysis of the protocols of individual readers in
the following section, it will become more obvious how the
author manipulates the readers' curiosity. In fact, more
than half of the readers demonstrated their frustration
by complaining that the story was "obscure" and "scattered."

As a result, they did not understand "what is going on."

174

FIGURE 6

THE DEVELOPMENT OF RESPONSES

 

100
90
80
70
60 . °
50 . ‘ . °
40 ‘
30
20
10

 

 

 

 

Divisions of the story
——--- I. Personal statements
II. Descriptive statements
———— III. Interpretative statements

---- IV. Evaluative statements

175

To find out "what is going on" seems to become the central
concern of their reading. Furthermore, the writer of

the story seemed to have anticipated the readers' anxiety,
because in section II he did provide Some more obvious
clues to the mystery which he deliberately created in the
beginning part of the story. Probably that is the main
reason why factual retelling became the dominant mode of
responses. In other words, what we have learned from the
data is that the form and content of the text might have a
definite effect on the readers in terms of which mode of
response to take.

(b) Percentage of interpretative response tend to be
low at the first two divisions (14.8% and 11.9% respec-
tively), then to rise sharply in reaction to the third and
fourth divisions (31.57% and 28.3%). What is being sug-
gested here is that the sense of meaning and interpretation
develop over time. Readers first have to establish some
kind of relationship with the text before they can actually
"generalize about it, . . .draw inferences from it,
find analogues to it in the universe that the writer
inhabits" (Purves, 1968, p. 7).

(c) In his discussion of the development of responses,
Squire has pointed out that there is "a strong positive re-
lationship of self-involvement responses and literary judg—
ments" (1964, p. 35). No such correlation has been found
in the present data. However, except for readers 3 and 10

who produced no personal statements in the first section of

176

the story, our data seem to support Wilson's argument that
"an initial self-involvement is necessary for effective
interpretational processes" (1966, p. 38). As shown in
Table 19 and Figure 6, personal statements started at the
highest point at the beginning of the story (29.7%). The
fact that the statements of perception and personal engage-
ment take up about 74 percent of the total response in the
first section suggests that information from the text and
the personal knowledge and experience which a reader brings
into the reading situation might be the very first re-
sources that help him in making sense of what he reads.

(d) Evaluative responses remain relatively low for
all four divisions of the story. However, the percentages
of literary judgment appear to be a little higher in
reaction to the first division (9.5%) and to the final
division (13.24%) as compared to the second (2.9 %) and
third (7.8%). This Observation seems to correspond to the
one made by Squire: "Clearly the readers . . . tend to
judge stories in terms of literary values to a greater ex-
tent when they commence reading or complete reading than
they do as they become engrossed in the central experiences
depicted in the literature" (1964, p. 31). The evaluative
statements made after the reading of the first section
tend to be sweeping generalizations regarding such
stylistic matters as the writer's presentation of the
story, while the final judgments tend to evaluate the

work as a whole, its construction and meaningfulness.

177

It is evident from the data that reading is accomplished
in a temporal process in which readers reassess their
relationships with the literary text continually while in
the activity of reading. But it is not directly observable
from these figures how a reader modifies his expectations
or how he readjusts his strategies in the process of evoking
his responses. To understand the constructive operations
of the reader in the development or shaping of meaning,
we must examine the linkage between the source of given
information and the response made to it. One possible
way to look into this complex matter is to study the
actual responses made by the readers closely. To serve
this purpose, case-study reports of five individual readers

are presented in the following section.

Presentation and Discussion of Results:
Responses of Individuals

As pointed out in the previous section the statistical
analysis of data can only reveal in a very general way
how a reader approaches a literary text during the reading
process. In order to have a better understanding of how a
reader deals with the reading material, a closer and more
detailed study of the actual data becomes an indispensable
operation. For this reason, this section is devoted to the
discussion of individual readers' protocols. Due to the
nature and scope of this present study, the researcher

felt it necessary to eliminate as many variables as possible;

178

therefore, the producers of the five protocols are all male
readers. However, using the same sex also has limitations.
Female readers may not read in the very same way as male
readers. To reach a broader generalization, a comparative
study would be required.

Another point concerning methodology needs to be
mentioned here. In analyzing these data, two independent
readers other than the researchers herself were involved.
These two readers were asked to read the five subjects'
protocols, and then rank these five readers on a ten-point
scale with the highest aesthetic rating being ESE: and the
highest efferent rating being aaaa. The total points for
each reader are ten, which could be allocated to either
the efferent or to the aesthetic or a combination of the
two. For example, if a reader is assigned a aia on the
efferent scale, his rating on the aesthetic scale must
be faag. The agreement among the three independent readers
was remarkable, with only slight variation in point

assignments.

Responses of Case Study Subjects

The presentation of this section takes the forms of
(l) a "response profile" chart for each subject with the
percentage of responses in the five Purves categories and
(2) a discussion of each subject's protocol based on the
finding presented in the chart as well as the actual

response statements of the individual subject.

179

Lin's response profile

 

 

Personal Descrip- Interpre- Evalu- Miscel-
state- tive state- tative _ ative laneous
ments ments statements state-
ments
Lin: 9.2% 38.2% 19.7% 14.5% 18.4%
Group 22.9% 37.8% 24.3% 9.9% 5%

means:

Lin's response profile indicates that he is a reader
who is able to draw upon information from many levels simul-
taneously. A detailed analysis of his protocol confirmed
this initial observation. One of the two independent
readers who read Lin's transcribed response statements
pointed out that ”this reader is very sensitive to a lot of
what is going on in the story and read information into the
smallest details" but "this is also a person for whom
experience is rich and as sensuously as cognitively." Lin
demonstrated this unusual quality of reading style right
from the very beginning of his reading of the story:

I feel that this passage is the beginning of a plot
development. For instance, the atmosphere and the handling
of atmosphere point to plot development. The backdrop of
the story is midnight. Furthermore, there are dogs. Usually
if a dog barks it makes a person feel less nervous. But in
this story the dogs do not bark, which makes the atmosphere
even more mysterious. In addition, the colors are either
black or gray. The arrangement of details is on the T¥junc+
tion. The main character is Lgu Chfian. He suddenly stopped
at the T-junction and did not move. It means, of course,
that a new episode is going to take place. What follows
is that between the person in black and Lau Chuan "one
hand hands over the money, the other hand hands over the
goods at the same time." It makes a reader feel a sense
of urgency and you want to know what is in the parcel.

180

Although on the surface this is just a steamed bread, a
very common thing, it seems that inside the steamed bread,
there is something else. Furthermore, because the setting
is night time, the main character a senile old man, and

the business that he is doing with the other stranger, all
point to the parcel as being uncommon, not usual. Because
all of the descriptions are of how Léu Chuan got there, this
passage tells two things. One, is the setting: the time is
deep into the night and the location is the T-junction.
Second, is the crucial thing, the parcel, the steamed

bread.

It is evident from the very beginning that this reader made
use of the textual sources as the vehicle for construing the
meaning of the text. For example, he talked about 166
Chfian, the person in black, the T-junction, the dogs and

the steamed bread. However, he was not just trying to give
all of the facts that occurred in the story. This is a
reader who is alert to how the information is given in

ways other than by simply being told that first this
happened, then that happened. He seemed to be most
interested in colors, and in the atmosphere in which the
first section of the story takes place, because he repeat-
edly said "the backdrop of the story is midnight... the

dogs do not bark, which makes the atmosphere even more
mysterious." He associated characters in the story with
colors: the man in black, the senile old man, and everything
is "either black or gray." The way in which this reader
talked about the first section makes one feel that he was

actually living in the story, as if he were actually feeling

the sense of mystery and "urgency."

181

The quantitative analysis of Lin's reactions to the
first division of the story shows that about 60% of his
response statements are related to the textual stimuli.

It seems particularly revealing to study these two sets

of data together. On the basis of Lin's reading behavior
as evidenced in his reactions to the first part of the story,
it can be said that the amount of information from the
text a reader makes use of is not the only crucial

factor in determining whether the reader is to have a
literary experience or not. It is indeed true that a
reader must pay close attention to the verbal signs of the
text because "the literary work is always anchored in the
text,” as Rosenblatt states. But what is also very impor-
tant is that the reader has to pay ”an equal closeness of
attention to what that particular juxtaposition of words
stirs up within each reader” (Rosenblatt, 1978, p. 137).
From the ways that Lin recounted his experience with the
text, he seemed to possess a high degree of sensitivity to
verbal nuances, but at the same time, he also demonstrated
that he had a capacity for intellectual and emotional
involvement. Lin was not only alert to the ways in which
textual information was given, but was also fully conscious
of the effects they had produced on him. This is probably
the reason why he was able to talk about the sense of
urgency and strong desire of knowing what was happening.

In sum, what can be shown about Lin as a reader on the basis

182

of his responses to the first section of the story is
that he was an anticipating reader who was emotionally as
well as cognitively involved in evoking his responses.

From the reactions that he made to the reading of the
second division,there is evidence for the idea that this
reader is someone who is both intellectually and emotionally
involved.

In the second passage, it seems that the location
shifts to the characterxsown house. The time is the next
day, oh, the next morning. And the character brings back
the steamed bread. Oh, it seems to be some kind of medicine.
It seems that in this section another plot has been arranged.
What I mean is that the hunchback has arrived; although
the author hardly touched upon the arrival, it seems that
this character is;1foreshadowing for the next development
because the time of this arrival is the same time as the
making of the steamed bread. Maybe the author will explain
later why he suddenly appeared at this crucial moment. That
parcel looks like some kind of medicine because it is given
to 156 Chfian. Oh, possibly it is some kind of medicine to
cure coughing. The atmosphere in this passage is not as
urgent and tense as the first passage. What I mean is
that now it is daylight and you feel that everything has
become more mild!

The results from the quantitative analysis show that about
57% of the total responses to section II were coded as
descriptive, which indicates that this reader was able to
pick many cues from the text. What is not so directly
observable from the figure is that Lin's personal and
emotional investment is as great as his cognitive one.
However, a close reading of his reactions to the second
part of the text reveals his evocation and responses to it
are also a "blending and balancing of overtones, attitudes,
feelings and ideas." He says "the atmosphere . . . is not

as urgent and tense . . ." and "daylight . . . you feel

183

everything has become more mild." What these statements
reveal is that this person was emotionally living through
the story which he was reading because he talked in terms
of things being mild.‘ Furthermore, Lin was also a reader
who anticipated what was coming next when he mentioned that
the appearance of a certain character was "a foreshadowing
for the next development." This reader appears to have been
anticipating what might happen and setting up his own
hypothesis in line with what he thought the author's
intentions were. In this sense, this reader was creating
his own possible developments for the story on the basis of
the verbal cues provided by the print.

One major difference between this section
and the first one is that there is more factual information
given in the second section, whereas there are more
impressionistic accounts in the first. It may well be
that the reader tried to orient himself, or, in Rosenblatt's
terms, began ”to develop a tentative sense of framework
within which to place what will follow" (1978, p. 54)
when he began the opening part of the story. He tried to
develop certain kinds of expectations about the setting,
about the characters, about the ideas, about the themes and
about what kind of text that would be forthcoming. Since
he did not have enough context to build up the textual
meaning of the story, he could only guess on the basis of

what he had. Perhaps this is the primary reason that his

184

response statements in the opening section tend to be more
impressionistic. When he moved to the second section new
elements were introduced, and apparently he found it
necessary to develop new structuring principles. Consequent-
ly, he reproduced more factual details in this later part
which were not available to him when he first started his
reading. Indeed, this sudden increase in factual retelling
seems to bea general trend for almost all of the ten
subjects, as shown in the quantitative analysis.
Lin's responses to the third section are only three
in number; this part of his total response is short.
In this section the development of plot seems to move
forward; that is, Kéng Da Shfi shows up, and because
the people in the teahouse react to him in such a
submissive manner he seems to represent some kind
of power. What I mean is that all those people agree
with whatever he says.
The ways in which this reader chose to react to this most
puzzling section of the story is interesting. It is
surpriSing to see how consistent a reader could be in
terms of selecting cues to respond to. His concern for the
development of plot has been the organizing power which
helps him to put the story together, and his use of
characterization as the main source for organizing the
plot development has been very consistent up to the third

section. For example, in the first section, be mentioned

that LXh Chian, the man in black and many more characters

185

all came to the T—junction which means "the beginning of
plot develOpment." In the second section when he saw the
hunchback appear at the moment when the family was in the
process of preparing the steamed bread, Lin was speculating
that "this character isaaforeshadowing for the next
development." Now in the third section, among so many
incidents, he chose to deal with characters and their
relationships to the possible plot development.

In contrast to the short responses of section three,
Lin's reactions to the fourth section make up a relatively
long passage. It takes up almost half of Lin's total
response statements. In the first part of this section
of retelling, Lin began to connect up the beginning and
middle of the story with the ending of the story.

The setting for this paragraph is a graveyard. From

this we know that Syfiu Chuan died. H1s mother went

to the grave to pay her respect.
The fact that Syau Chfian died at the end seemed not to
surprise him. It is possible that he had figured out
at the very beginning of the story the death of this
young man to be inevitable. Otherwise, he would not have
picked up on the colors and atmosphere in the first
section and guessed what they were leading to. They

seemed to set the theme for him. Many other readers did

186

not feel the atmosphere was as dark and gray in the first
section of the story as he did.

In the fourth section, Lin also finally came up
with his answer to his own implicit question: what is the
message that the story is trying to convey? This question,
apparently, was central to Lin's reading, because he was
continually looking at the foreshadowing.

At the beginning of this paragraph there is an account
of the graveyard where one section is reserved for

the poor common folk and in the other part lies
executed criminals. The description is very
poignant.‘ "On the left of the road were burial mounds
of such criminals as had been executed or had died
from exposure in jail; on the right was a burying
ground for paupers. Both sides were already crowed
with rows and files of mounds like the ceremonial
cakes at a rich man's birthday party.” I feel those
who died are probably the victims of the vicious

power of pressure groups. . . . King Da Shfi

represents some sort of evil power. He exchanged
steamed bread for money. This action is an action

of immorality because Ka'ng Da Shfi does not care about
people's lives. He only cares for money.

Some of the readers did choose to read the story for its
poignancy, that is, for the pity and the sadness one feels
for the family itself. But Lin sought a larger sociological
context and treated the young man who died of tuberculosis
and the young revolutionary as victims of society. The
society is so sick that none of the characters are normal
people.

The entire setting is very gloomy; the colors are

always dark and gray. First there is night, then

later bright daylight. But even in daylight there
are a lot of people who are either old, sick or

handicapped, or who are soldiers. So in this story
we do not see one who can be considered normal.
And with the death of Sygu Chfian a great impact on

187

our psycheijscreated. The senile old man, the hunch-
back and the graveyard. . . everything is in decay.

Here is a reader who talks as if he was experiencing anger
and sadness as he went through the reading experience.

Lin's responses to the fourth section of the story
constitute half of his total responses. The distributions
of these statements to the five Purves categories are as
follows: Personal statements (2.4%), Descriptive statements
(19%), Interpretative statements (19%) and Miscellaneous
(33.3%). About 33 percent of his responses in this section
were identified as Miscellaneous because at the middle of
his retelling he departed from the text by bringing in the

French writer Maupassant.

The short story written by L6 Syun reminds me of the
French writer Ma...Ma...Maupassant, especially in their
ways of writing and treatment of details. They both
tend to focus on an event in great detail, and then
make a very sudden surprising ending. In a short
story by Maupassant there is an old man living in

the city. Every day he goes to the park and sits
there watching young couples driving around in
carriages. The old man is a lonely bachelor. This
story revolves around the old man's visits to the

park and going to Work. Whenever the old man deviates
from this pattern of life he is unable to function
properly. The author does not explain how the old
man died at the end of the story. But his corpse

is found hanging from a tree. I find this tec nique
of a sudden ending similar to the ending in L Syun's
story. I find this to beaatypical way of short

story writing. That is, you focus on a single plot,
and then develop it

The reader departed from the text at the end of the story
at a very crucial point where there are some mysteries that
can only be solved by the way in which the reader interprets

signs like "the flowers without roots" and ”the crow."

188

Instead of staying with the text and trying to figure out
what these signs might have meant in the context, he
backed off from the text by bringing in Maupassant

and his horror story. His separation from the text is
disappointing because he ended up by saying "plot becomes
not so crucial." On the other hand, this deviation is
still satisfying in the sense that his departure is a
departure into literature by another author from another
culture. He was making a cross-cultural comparison and
developing a world-like truth.

Lin's protocol reveals him to be a reader who
undertakes a "spontaneous transaction" with the text.
This is a reader who is able to pick many cues from
the text and who responds to them emotionally as well as
cognitively. He pays close attention to the verbal
symbols of the text while at the same time he is equally
aware of the "feelings, sensations, images, ideas that he
weaves between himself and the text.” In short, in Lin's
account we have foundaireading experience that can be
accurately described by using Rosenblatt's terms ”aesthet-
ic." On the ten-point scale, Lin is ranked with an average
of 8.5 at the top end of the aesthetic scale and with an

average of 1.7 on the efferent side by the three independent

readers of his protocol.

189

Gu's response profile

 

 

Personal Descrip- Interpre— Evalu- Miscel—
state- tive state— tative ative laneous
ments ments statements state-
ments
Gu: 9.1% 69.7% 9.1% 12.1% 0%
Group 22 9% 37.8% 24.3% 9.9% 5%

means:

 

Gu's approach to the text is very different from our
first reader Lin. Instead of responding widely to the visual
and sensuous possibilities afforded by the text, Gu selected
out just one incident from the first section to focus on
and ignored many other incidents.

This passage seems to describe a scene where a

prisoner is being beheaded in Ancient China. The

setting of the story is probably in the northern

part of China. There is some flavor of northern

China here, m m m. It is very compactly written.

This is my impression of the first passage.

How did he know that it is a scene in which a prisoner is
beheaded? None of the other readers were able to get this
point until the very last section. How did he sense that
"there is some flavor of northern China" in the story? Gu
was coming at the story much more personally. As it turned
out in the later part of his retelling Gu had actually been
to the northern part of China.

All of Gu's response statements to section two were

coded only as Descriptive, but the way in which he recounted

the story reveals interesting things about him as a reader.

190

I forgot to mention one point; that is, the plot of
the story centers on the son who has tuberculosis.
People in China formerly believed that bread made

with blood can cure tuberculosis. The father's name
is LXu Chuan” This story is about how the father bribed

the executioner into getting blood with which to make
the bread that can cure their son, hoping that their
son will be rid of his illness soon. Naturally, the
parents care a lot. The incidents of the story are
not explicitly presented of course. The family seems
to have opened a tea house, probably that is. Then
in the second passage the incidents became more clear
and easier to understand. It is about how the father
took the bread home and had it prepared by steaming
it or putting it in the oven for his son. It is not
shown clearly whether the son had known that the
bread he ate is made with blood or not. Probably
he will know it in the end. I am not sure. At this
point he simply ate it and went back to sleep.
This is a reader who when he does not understand something
will not speculate verbally about it, but who is still
seeking understanding by collecting information from what
he has read. In contrast, the first reader seemed to
understand a great deal without having put everything
together concretely the way this second reader seemed to
have done. Gu did not want to mention L56 Chuan's family
until he was absolutely certain what was going on. In fact,
in the second section he does mention that "the incidents
became more clear and easier to understand." It is very
interesting to look at this reader's reaction to the first
section and the second section together. At first, he
made some generalizations about the setting, then when he
got to the second portion of the story he said "I forgot

to mention one point..." and went back to comment on what

he thought had happened in the first section. The fact

191

that he digressed and tried to recapture some incidents

that occurred in the first passage tells us something not
only about Gu as a reader but also the reading process in
general. It provides us with a clear demonstration that

a reader does not process information in a linear order.
Readers not only look forward but also backward to modify
the tentative hypothesis they have set up as new
information enters their focus of attention. Readers
select, organize and synthesize ideas, feelings and attitudes
on the basis of what seems to be most interesting or
relevant to them at the particular moment of their
relationship with the text. That Gu failed to follow the
main line of the story, that is, 166 Chdan and the steamed
bread at the beginning of the story, but instead talked
about northern China, can be used to demonstrate this point.
The recognition of something familiar which reminds him of
his childhood experience might have created a greater
emotional impact on him than the story itself. He chose

the northern China theme as his center of attention and
pronounced the expression "there is some flavor of northern
China here." That a reader can reserve judgment on something
until an appropriate moment arises is not unique to Gu. In
the first reader's retelling, we find a similar example.

The reference to "soldiers” appears in the first section of
the story. However, Lin did not say anything about soldiers

until he was in the fourth section. In commenting on the

192

sick, decayed society Lin said that in this society we do
not see any normal healthy people, and those who we do see
are "either senile, sick, or handicapped, or...soldiers."
Gu's comments on the third passage are short. They
are mostly factual retelling from the text.
This passage is about the prisoner who was killed.
He seems to have been a revolutionary who was captured
and imprisoned by the China government. During that
time China was run by the aliens from Manchuria. The
jailer wants to make some money. He is very poor,
that is all.
It is clear from his retelling of the second and third
sections that this reader read mainly for information as
opposed to reading for pleasure. Except in the first
section in which he demonstrated some degree of emotional
involvement, we do not get any idea at all that this person
was having any emotional responses to the reading experience.
But this is a reader who is very careful and offers much
information. In his factual retelling, he demonstrated
that he had some prior knowledge about the Chinese people,
culture and history which the rest of the readers might not
have had. For example, he said "people in China formerly
believed that bread made with blood can cure tuberculosis,"
and "during that time China was run by the aliens from
Manchuria." This prior knowledge that he possessed and his
experience of actually being in the northern part of China
enabled him to put a rather obscure plot into a coherent
reading.

Gu's reactions to the fourth section cover a wider

193

range of response statements, although the dominant one

is still descriptive. The results of quantitative analysis
of this section show that the category of Perception leads
the other three with a percentage of 50, and followed by
Engagement, Interpretation and Evaluation, each of them
taking up 16.7%.

This passage is about two mothers who visit their
son's graves. It is already the time of the year
when the annual journey is made to ancestral graves.
Obviously, Hwa Da M5 is the mother of the man who
ate the bread made with blood, and yet who died;

and through implication only, it is probably that
the other woman is the mother of the prisoner who
was beheaded. The outward appearances of the

two graves are not unalike. The two mothers grieve
differently although, of course, they are both

very sad. And it turns out, one woman comforts the
other. At this point they see a crow on a tree
branch. One mother hoped that the crow would reveal
to her that her son is innocent. In addition, this
shows that Chinese people are still superstitious.

This story is good. The episodes are not explicitly
told; they are all implied. The author does not
force his viewpoint. Also, there is nothing super-
fluous. The author of the story is Lu Syun. He is
a northerner. Actually he was not born in the north.
Usually, Southerners do not write this way. I know
that the setting is in the north because I used to
live there myself, and I also know that some writers
use north China as background for their stories. I
know this is true because the vocabulary, tone, and
dialect of northerners and Southerners differ.

Although I am still making the point that this is a reader
who was reading for information (that is, as an "efferent"
reader) we must notice that he put information in a clear
and coherent way. He was able to get an important point
across, namely, that the two mothers grieved differently,
one woman comforting the other. The woman whose son died of

tuberculosis comforted the mother whose son had died as a

194

prisoner. The bare outline of the plot abstracted of any
colors is what this reader got, but what he got he got
constantly and well.

Furthermore, at the end of his retelling, this reader
made very interesting statements about the story's
vocabulary, tone and the dialect. These are elements which
create atmosphere. Thus, we know that he did see these
elements used by the author, but we also know that he did
not respond to them emotionally. He was looking at these
elements simply as information. Gu also mentioned that
the story was written in north China because he gathered
the evidence from the diction and the structure of the
language used. He knew that the setting of the story is
in the north because he actually at one time lived in the
north. But he did not seem to bring any personal feelings
into his responses. He was simply retelling what he
thought had happened. In terms of reading for textual
information, this reader is consistent, both in his choice
of clues to respond to and in the ways in which he made
use of his prior knowledge. He tends to focus only on what
he does understand and to dismiss what he does not.

In contrast with the first reader, Lin, who was
placed at the top side of the aesthetic scale, Gu was rated
as an efferent reader with an average 7.7 on the efferent

and 2.3 on the aesthetic.

195

Chen's response profile

 

 

Personal Descrip- Interpre- Evalu- Miscel-
state- tive state- tative ative laneous
ments ments statements state-
' ments
Chen: 32% 21.3% 40% 4% 2.7%
Group 22.9% 37.8% 24.3% 9.9% 5%

means:

Chen, like the first two readers, began by picking up
textual cues offered by the print. But the verbal signs
Chen chose to select are quite different from those chosen
by the other two readers. Whereas the first reader focused
on a relatively inclusive set of the literary aspects, and
the second one sought mainly the details that connected with
the development of the story, this third reader seems to be
particularly interested in characters.

After reading the first paragraph I began to speculate
on the possible ending of the story. I cannot guess
exactly what the end'ng will be, but at the present
time I know that Sya: Chfian is sick and that probably
the cure for his illness came from a human body gotten
by illegal means. The only remedy for Syau Chdan's
illness is something dripping with blood. ~

Then we learn the father Lau Chfian got some money

and asked someone to get the thing that he needs.
Although he has some reservations about whether or

not he will be able to get it, one thing is definite;
he loves his son. And we also know that Syau Chuan is
a very good son, a responsible son, which is hown in
the ways he tried to help his mother. And L u Chfian
suddenly felt young again because he was able to provide
the money which would make his son healthy again. Even
though a lot of money is required, to him this is a
form of relief. It brings him a sense of happiness.
That is why he feels young. But of course, this is
only my interpretation. And what about the red badge
on the numbered jackets? I think they are some kinds
of decoration on the uniform worn at the end of the
Ching Dynasty.

196

On the basis of his retelling here, Chen emerges as a
reader who was seeking to comprehend the story through
trying to understand the characters. ‘He fixed his attention
on the characters that in his mind he believed the author
created and tried to identify with. He demonstrated an
understanding of what the two parents in the story had been
going through. For instance, this reader noted that the
father, L56 Chfian "suddenly felt young again because he

was able to provide the money which would make his son
healthy again." This reader also tried to see the son,
Syau Chdan , through the eyes of the parents when he said
"Syau Chfian is a very good son, a responsible son." So
this is a reader who is very alert to characterization

and who pays close attention to the relationships of the
characters.

Another point made by this reader is especially
noteworthy. This reader was able to talk about what was
actually going through his mind while he was reading the
first section: "After reading the first paragraph, I
began to speculate on the possible ending of this story."
What Chen was doing confirmed Rosenblatt's statement that
"As he engages with the text, the reader. . . sets up
hypothetical frameworks, entertains expectations as to what
will follow, and uses these as guidelines for selecting out
from alternative responses” (1978, p. 137).

This paragraph makes the first paragraph more clear.

I mean that what I said in the first paragraph is
very close to what is presented in the second

197

paragraph. Now they talk about the two halves of the
steamed bread; therefore, we know that it is definitely
steamed bread. In addition to that, there is the blood
dripping scene. I guess the background is the end

of the Ching Dynasty and the beginning of the Republic
and that someone was executed. Someone took the steamed
bread and dipped it in to cure Syau Chdan's ailment.

I believe that all this confirms my original hypothesis
about the story. Of course, I cannot be sure whether
Sygu Chfian is cured.or not because I have not finished
reading the story yet. I can only say that these are
my responses to the reading, which confirms my previous
statements. Now a bunch of characters enter the story.
I did not receive any special impression from this part.

Responding to his own evocation which itself arose during
his reading of the first section of the story, this reader
was talking about whether the hypothesis he set up at the
beginning of his reading was correct or not: "I believe
that all this confirms my original hypothesis about the
story.” From these various retelling it would appear that
readers can be extremely conscious of ideas, sensations or
feelings that are stirred up in their minds in the trans-
action with the text. This reader as well as the second
reader could reveal confusion and uncertainty about what
things might mean and comment on the actual process of
trying to make sense of what was being read.
Now it is obvious. The setting is the end of the
Ching Dynasty and the beginning of the Republic. I
felt that the author was possibly using the common
fold's ignorant behavior to express people's general
attitude toward revolution. So human blood, bribery,
and the scene where the jailer hit the prisoner for
refusing to give him money, all point to society's
corrupt nature. It has become very obvious now that
the central theme of this story is that at the end

of the Ching Dynasty and the beginning of the Republic
the common people's concept of revolution is as of yet

198

undeveloped. Consequently, it is extremely difficult,

if not impossible, for revolutionaries to communicate

their revolutionary ideas.
One of the obvious characteristics of this reader's
strategy in approaching a text is that he always reads
ahead. He made predictions about what might happen, and
then commented on his own predictions. For example,
reflecting on what has been evoked in the second section,
Chen began his responses to the third section by saying
"Now it is obvious. The setting is the end of the Ching
Dynasty and the beginning of the Republic." Chen's reading
offers a concrete example which clearly demonstrates the
centralrmnfirmlof Rosenblatt's transactional theory of
reading. That is, the reader responds to the evocation he,
himself, makes under the guidance of the textual signs.

Up to this point, Chen's reading has followed the main
line of the story. His responses have covered most of the
central events that happened to the family. In the first
half of the third section, Chen proposed his interpretations
of what all those incidents meant to him: "I felt that the
author was possibly using the common folk's ignorant
behavior to express people's general attitude towards
revolution." However, in the middle of his retelling, he
changed his focus on the story and entered into a

discussion of democracy and saw this story as a piece of

propaganda.

199

If this story had been written by a modern Taiwanese
writer, I would say thatit depicts the first stage

of the development of the Taiwanese democracy. It
reminds me of an article written by Professor M6u,
Jfing-San. He said that democracy is something like

a skeleton. What I mean is that this skeleton becomes
alive only when everyone in this skeleton understands
democracy and cooperates. For instance, every meeting
has its set procedures. If the participants do not
follow the procedures then the skeleton collapes. In
other words, if we want democracy we need to heighten
the consciousness of the common people, then the common
people can operate the skeleton of democracy.

The main idea of this third paragraph is that there is
someone who wants to build the skeleton,lnn:unfortunate-
ly, the common people are not ready. Because the common
people do not understand the need for revolution they
said several times "Crazy! crazy!" They believe

that rebelling against the government in power is very
unwise. I think that the author is probably an

early advocate of democracy; in this story, it is

clear that he is expressing the concept of democracy,
and that is gradually spreading. I am not sure

but maybe this is a piece of propaganda writing.

For a person who was bent on understanding the story through
the characters in the first two sections of the story to
drop the main line of development suddenly and to move into
a discussion of democracy is striking. Chen had been very
consistent in his using characters as a path to the meaning
of the story. However, it is not clear until the third
section that he was not interested in the characters them-
selves as individuals or in their destiny,but rather in them
as a group that can fit into his political scheme. On the
basis of his retelling, there is no clear evidence to show
for certain that this reader had already sensed the political
overtones right from the beginning of his reading and had
anticipated that the story was going to develop into a

political writing. Therefore, it is quite possible that he

200

changed his original way of perceiving the story as the

new details presented in the third section entered his
consciousness. Of the three readers studied so far, this

is the only reader who shows great interest in this section.
Probably this is because in this section there is a prisoner
who might have died unjustly, which seems to be more
important to him than the fate of the other young man who
had a fatal disease that may or may not be related to his
poverty. It is true the third section contains a lot of
information with political implications. Maybe that is the
major reason why those readers who are less alert to the
political issues tended to ignore this third section or
found this section somewhat "disorderly." Apparently,

they did not find the political problems implied here
relevant to their understanding of the story. One of

the important things shown by these three readers' responses
concerns the matter of "selective attention" that Rosenblatt
has talked about. It is clear from these protocols that
each reader seemed to bring into the reading situation

his own special interest and intellectual strength, which

in turn seemed to affect his choice of what to focus on.

We have seen that each reader selected from "the multiple
possibilities” offered by the text those elements that can
be fitted into the relevant patterns of ideas, feelings, and

attitudes.

201

Now I start to wonder why they want to use steamed
bread mixed with blood to cure Sy u Chuan' s illness.

I think that the meaning of the blood may play a

very important role in this story. I remember a
story called "Dai Mai Ju Tswei Ji." In this story
there are some people who like to drink human blood.
When I first came across the steamed bread mixed with
blood I associated it with this. But it may mean
something else in the story. I do not want to dwell
on this. It is just a momentary association. I
think that maybe the steamed bread mixed with blood is
only a surface meaning.

I believe that the deep meaning of the story is that
the Ching Dynasty will fall. As far as those flowers,
the flowers without roots, it is obvious that they
were put there by someone. It indicates that there
are people who sympathize with the revolutionary
cause already. Then the crow. Especially the crow
mentioned in the last sentences. The crow flies

like an arrow into the distant sky. It gives

the impression of a lift- off. The impression is

very close to what Jwang Dz said' 'an eagle spreading
its wings" after long captivity and oppression to
enjoy newly discovered freedom. Perhaps I am looking
at this story from a political perspective. It may
be due to the fact that I have read a lot of articles
concerning politics. I have my own point of view on
the subject of democracy. I feel that sooner or later
democracy is an inevitable route.

Readers are found to talk more about the story itself in the
fourth section. Perhaps this is because there are some
obvious symbols in this section. It is interesting to see
how these readers reacted to these symbols that they chose
and to see how each of them manages to fit the symbols

into his own over-all structure of meaning. Chen, like

the two readers before him, responded to the symbols of
flowers and crow, but came up with a different interpreta-
tion. Instead of reading "the flowers without roots" and "the
crow" as a sign of superstition or as a sign of ignorance

of the Chinese people, as given by the first two readers

202

respectively, the third reader thought the flower
symbol was an unspoken sign of support of the revolutionary
activity. His idea about the crow is also tinged with
political color. He mentioned that the crow flying like an
arrow into the distant sky gave him the impression of
a lift-off, which reminded him of a famous phrase made by
an ancient Chinese philosopher, "an eagle spreads its wings
after long captivity and oppression to enjoy newly
discovered freedom." And he finished his retelling by
saying that democracy was bound to come sooner or later.
Analyzing Chen's responses has been an interesting
learning experience for the researcher. What can be learned
from his retelling is that a reader while reading the
story can actually talk about his confusion and doubts
about the meaning of the story and explain how and why he
came up with a certain kind of interpretation. Chen was
fully aware of why he approached the story from a
political point of view. He said that he was looking at the
story from a political perspective perhaps because he had
"read a lot of articles concerning politics." What he was
saying here is that his past experience was responsible for
the approach he took to the text.
From the retellings it can be seen that readers do
reach for information from their own background and try to

apply that information to the situations at hand. So readers

203

really bring their own understanding of life in general
to the body of the reading rather than staying strictly with
the verbal symbols provided by the text. The first of the
three readers brought in his knowledge of literature, the
second one his personal experience in Northern China, and
this third one his interest in politics, and as a result,
the three readers produced three different kinds of
readings out of the same set of verbal signs. These
protocols offer insight into how impossible it is for
readers to rely simply on the textual information to
figure out what is happening in the story. All three read-
ers went outside the text at certain points when they felt
confused to find ways of understanding what they were
reading. In going outside of the text, what matters is
where one seeks information and what one brings back to
the text.

Obviously, readers have different intentions
when they read, just as writers have different intentions
when they write. This third reader seems to be more
concerned with his own intention than that of
the author, so he apparently feels free to cast off the
information in the text and build up his idea of what the
story is about on the basis of limited details of the text.
This is a person who reads not for literary pleasure, but as
a means to an end. Rather than being immersed in the story,

he took the story in his own hands and used it as a tool

204

for his end, which was to speculate about establishing
democracy in Taiwan. As a reader, Chen was intensely
conscious of what was going on in his mind and was able to
articulate it. But he, like the second reader, was
reading for information. He was interested only in those
details which were related to a larger political context.
Chen's response statements reveal that he is more an
efferent reader than an aesthetic one. He scores 4.7

on the aesthetic scale and 5.7 on the efferent.

205

Dung's response profile

 

 

Personal Descrip- Interpre- Evalu- Miscel-
state- tive state- tative . ative laneous
ments ments statements state-
ments
Dung: 27.3% 24.3% 30.3% 18.2% 0%
Group 22.9% 37.8% 24.3% 9.9% 5%

means:

Except for the first reader who seemed to show no
frustration about the story's occasional obscurity; the other
readers demonstrated some degree of uncertainty at some
point in their reading. The fourth reader, Dung, bluntly
admitted that he did not have any clear idea of what was
going on in the first section of the story, partly due to
the fact that he was not accustomed to reading orally
while responding at the same time, and partly because
"the descriptions are scattered."

I really do not understand the purpose of the

descriptions in the paragraph. When I read this

section I felt that Lau Chfian wanted to do something,
but I am not sure what. Maybe I didn't understand
because I am not accustomed to reading orally and
simultaneously responding. I don't have any way of
connecting the author's thought to my own. Furthermore,

I feel that the descriptions are scattered. I have

no way of understanding why L86 Chfian went to the

T-junction to look at all the unusual people or

whatzthese people were doing. I do not have a distinct

idea of what is going on.
On the basis of his retelling, we may say that we are now
looking at a reader who does not have much confidence in
his ability to read. He felt frustrated because he did not

know how to pick clues from the text and put them together.

206

Unlike the third reader who, though also confused, yet
tried very hard to find out the possible direction of the
story, this fourth reader seemed to be willing to wait to
see what the story was going to tell him.

0k! This section is more clear. He says that L56

Chfian is going to get some steamed bread. I don't

know what kind of medicine is in the steamed bread.

It does not matter what kind of medicine it is.

The purpose of it is to cure his son's illness. This

is the main theme of this section.

In reading the second section, Dung began his responses
by saying "this section is more clear." He felt somewhat
relieved becauseluabegan to see the connection between
Lau Chuan, the group of strange people in T-junction

and the medicine for Lau Chfian's sick son. But he was
interested in plot. He paid no attention to such literary
details as colors or the implicit relationships between
husband and wife or father and son. All he wanted to
know was that there was a theme which connected section
one and two. From the response statements he made in
these two sections, it can be seen that he was a reader
whose primary concern was seeking textual information from
the printed material.

When he got to the third section, he still talked
mostly about what he had gained from the second section.
He mentioned the prisoner only in passing and went directly
to the slogan "The great Ching empire belongs to all of

US.

207

In this passage it is very clear that the medicine he
brought home is to cure his son's illness. No matter
what kind of lung disease, no matter what kind of
tuberculosis, this passage seems to suggest when
the medicine is taken the illness will go away--the
son is saved. In this passage, the author introduces
another plot. It is about Sya. Sya's son, namely,
the prisoner of A-Yi. Something needs to be noted here;
that is that he said ”The great Ching empire belongs to
all of us." This may be foreshadowing a future develop-
ment.
This third section is interesting because of the ways in
which the retelling is given. This reader has a lot more
capacity and capability than he used. Although this
reader seemed not to have much self-confidence, he actually
has a rather interesting style of retelling. The expres-
sions "No matter what kind of lung disease, no matter what

kind of tuberculosis. indicate that this reader has a
sense of poetry or a sense of individual creative force
which he was not consistently making much use of.

To be sure, Dung is not an experienced reader
in reading a literary text, but even so, his reading of the
story offers valuable insights. Once again the protocol
providesaiconcrete example which illustrates how the
process of reading a text is a process of the constant
shaping and revising of meaning. Dung began his reaction
to the fourth section with the sentences "In the passage,
I found my predictions to be incorrect. The son was not
saved; he died of tuberculosis." Obviously, this reader

had already formed his version of what the story should

be like, namely, that the medicine should be effective.

208

But when he moved into the fourth section, it turned out
that what he had been thinking was incorrect, and Dung
admitted that his predictions were wrong. As seen in
this example, Dung was responding simultaneously to the
verbal stimuli of the text and to his own responses
evoked from his previous reading. What Dung was doing
was very much like what the third reader did. Both of
them show involvement in the process Louise Rosenblatt
calls the ”second stream of response.” According to

Louise Rosenblatt the shaping of a sense of the
work demands a kind of embryonic critical testing of
tentative organization against the text itself. As he
engages with the text, the reader, we have seen, sets

up hypothetical frameworks, entertains expectations as to
what will follow, and uses these as guidelines for
selecting out from alternative responses. As the text
presents new elements, he may find it necessary to revise
earlier syntheses or to develop new structuring principles.
Thus an element of self-criticism vis-a-vis the text--a
concern with validity of interpretation--enters into

the very evocation of the work” (1978, pp. 137-8).

The second insight that can be gained from analyzing
Dung's retelling is that a reader can gain confidence over
time in the act of reading. Both the quality and quantity
of this reader's reading have progressively improved,

especially in the fourth section. In the first two sections

he was struggling with his confusion and was just able to

209

deal with the most obvious factual information. In the
third and fourth sections, he was beginning to show a
certain degree of sensitivity to the more delicate aspects
of textual clues. For example, he started to talk about
foreshadowing, and the subtle, delicate descriptions of

the tombs. The fact that he was actually able to recite the
exact wording cm? the text and to say that contrast was

very attractive to him strongly indicates that he was
emotionally involved in the part of the story.

In this passage, I found my predictions to be incorrect.
The son was not saved; he died of tuberculosis. I
felt that descriptions of the differences between the
tombs of the rich people and the poor people were

very subtle and delicate. He said, "The tombs of

the poor are piling up just like the ceremonial cakes
at the rich men's birthday party." This sort of
contrast attracts me very much. The rest of this
passage is a description of those people. However, we
do not know whether they died of tuberculosis as

did Syau Chuan.

There is no question about his being emotionally involved
because he ended up by saying that he did not like the
story. He felt that the story was unfinished.

I felt that this story was left pnfinished because
the author just touched upon A-Yi without giving

us any ex lanation. We do not know what happened to
him. A- Y mentioned what the prisoner had said,

that is, the ”Ch1ng belongs to us. I think that
this may be the main idea which L6 Syun is trying to
express in this story. This is his political
ideology. What surprised me is that the ending

of the story has nothing to do with politics, but
instead shows the country folk's ignorance. I felt
that this story was unfinished because there are
still several plots which remain to be developed.
What captures my greatest interest is the phrase
"The Ching is ours" and the vivid descriptions showing
the differences between the rich and the pauper,
especially the point that the tombstones are like
steamed bread on the tables of the rich. I do not

210

think that I like this story very much, probably

because of the way of describing ignorance about

medicine and the wrong belief that medicine can

alter fate. The reason that I do not like this

story is that it does not appeal to my senses.

I guess that the purpose of this story is to

show how when the country people are ill, they

rely on folk remedies. I think that the only

purpose of this story is to demonstrate the

ignorance of the country folk. If that is the

main idea that he is trying to convey, then he

has succeeded. The final scene at the cemetery

is another good indication of the country people's

ignorance.
The way this story is ended ”does not appeal to my senses."
This is a reader who had some particular interest when he
came into the reading situation. There are two places
in his retelling which illustrate this point. The first
example occurred at the place in the third section when
he first saw the expression "The great Ching empire belongs
to all of us,” and the second one took place at the
beginning of the fourth section when lne read the
descriptions of the graveyard. He was definitely
interested in political issues, as was the third reader.
However, there is a basic difference between them.
Chen, the third reader, used the reading occasion as a
springboard to express his own political ideology, while
Dung was more willing to wait to be told by the text.
Somehow he was not aware of the possibility of connecting
the political theme with the death of the prisoner and
with the flowers without roots. When he missed the

prisoner and his relationship to the other old woman at

the cemetery, his perception of the whole story changed.

211

The third observation derived from Dung's protocol
concerns the problem of aesthetic and efferent reading.
In her discussion of the distinction between aesthetic
and nonaesthetic reading, Louise Rosenblatt pointed out
”Actually, no hard-and-fast line separates efferent--
scientific or expository--reading on the one hand from
aesthetic reading on the other. It is more accurate to
think of a continuum, a series of gradations between the
nonaesthetic and aesthetic extremes. The reader's
stance toward the text. . . may vary in a multiplicity of
ways between the two poles" (1978, p. 35). In Dung's case,
it is clear that he is capable of taking either stance.
He was taking predominantly an efferent stance when he
dealt with the first two sections of the story. However,
as he was reading the first part of the third section and
the beginning of the fourth, he was apparently enthralled
by the content of the text, but he lost the intensity of
his interest and literary awareness when he found out that
the ending of the story was not what he had expected. He
went back to his former mode of responding, that of
retelling the factual details of the story. In short,
we find there is a back-and-forth movement of attention.
between the two stances of reading in this reader's reading

which clearly supports the position set forth by Rosenblatt.

Wong's response profile

 

Personal Descrip- Interpre- Evalu- Miscel-
state- tive state- tative ative laneous
ments ments statements state-
ments
Wong: 10.6% 40.5% 44.6% 4.2% 0%
Group 22.9% 37.8% 24.3% 9.9% 5%

means:

The first impression that one gets from reading Wong's
reactions to the first section is that he is a positive and
confident reader. Instead of making personal comments or
expressing frustration at not being able to figure out what
was going on in the story, he went right into the process
of transacting with the text and responded to the clues
that he chose to focus on by giving his interpretation.

I do not know what to say. Mm. . . , let me put it
this way. At the start, he he authdfl mentioned
that it was after midnight, t en later on he talked
of the sun rising. It seemed to me that he is saying
that there is some hope when he £156 Chfiafi] got the
steamed bread. Then he describe a group of people
standing there watching the soldiers. I think there
is an ironic, sarcastic element there because under
these kinds of conditions it should have been taken
very seriously. However, he described those people
in a very ridiculous way. For instance, he said that
their necks are stretched out very long like duck
necks. This sort of description, in fact, is L6 Syhn's
style. He thinks that when a group of people gather
together they become a mass without any individual.
thoughts.

Like the first reader, Lin, this reader is also able to pay
close attention to some of the more refined elements of the
story instead of producing a bare factual recounting. For

example, both of these readers are interested in details

213

such as background setting. However, it is interesting to
note that they picked up on different clues and consequently
came up with very different readings of the story. As
pointed out earlier in this section, the first reader
focused his attention on the colors of grey and black.

As a result, his view of the story is bleak and gloomy.

This reader, on the other hand, chose to see the bright

side of the story: " after midnight, . . . the sun
rising. It seemed to me that . . . there is some hope
when he got the steamed bread." It appears to be a clear

example demonstrating how a reader's perception can
affect the meaning of a story.

That the prior knowledge and personal experience of a
reader play a critical role in the act of reading has
constituted a recurring theme in the present study. As we
have seen in the discussion of the previous four readers,
each reader brought some characteristics of his unique
understanding of the world into his reading of the story,
which obviously has further influenced his perception
of the written text. Knowing the author and prior
acquaintance with his works is what the fifth reader,
WOng, has brought into his encounter with the text. It~
is apparent from his retelling that he has read some other
works by the same author because he said "this sort of
description, in fact, is L6 Syun's style." That Wong sees

the story as a satire distinguishes his reading from the

214

others. Perhaps he had read other works by the same writer
which are indeed satirical and so was putting this story
into a category already established in his mind.

Wong is definitely a reader who responded to the
descriptive elements as well as to factual events of the
story. In the second section we have further evidence to
support this observation.

From the first paragraph to the second paragraph he
did not explain clearly what the steamed bread, or

we may say, the thing dripping with red blood, is.

In other words, he deliberately tried to keep the
object obscure to the reader in order to suspend the
reader's interest. Then came the hunchback who

asked what it was. Then the suspense is heightened.
And the author moved on to describe the son, but

he did not spend much time describing his personality.
But he did mention the way he ate the steamed bread.
From this very brief description I felt that he was

a very selfish person. What I meant is, he is eating
something and I think that he also wants to know what
it is; but he did not ask. It took him only a minute
to eat all of it. It seems that it is for his own
illness or something. It is not explicitly stated.
In this paragraph we also learn something about his
family's background. They own a small teahouse. It
does not look like they have a good business, a lot
of customers. I know this because there are a lot

of patches on their clothes. It shows that their life
is not a good one.

In this section Wong talked about the sense of suspense he
felt in reading the first two sections of the story because
the author refused to explain explicitly what the thing
dripping with red blood is. A similar comment was made

by the first reader when he said that he felt a sense

of "urgency and wanted to know what is in the parcel."

In fact, another reader (whose protocol was not analyzed)

215

made a similar observation. She used the word "mysterious”

to express the feeling and sensation that she had experienced
in reading the beginning part of the story. In viewing
these responses of these three readers, we see that we must
question any proposal that treats reading as a completely
subjective matter in which the text has no part to play.
Obviously, reading cannot be purely and only subjective,
because sometimes different readers react to a text in

what appears to be the same way. For example, different
readers sometimes select the same verbal symbols or images
to respond to. In the present study, almost all of the
readers chose to comment on phrases such as "steamed

bread mixed with blood," "the flowers without roots,"

"the crow" and the character Lab Chuan and his sick son.
All of the readers undoubtedly have been affected by some
of the verbal stimuli of the text. Although our data
affirm the essential role of the text in the literary
transaction, the presentation of the readers' protocols in
the preceding section has also shown us how illusory

it is to see the text as all-important and to treat the
reader "as a tabula rasa, receiving the imprint of 'the
poem'" (Rosenblatt, 1978. p. 4). As Edward Sapir has point-
ed out, most of the words have an "associated feeling-tone"
and it "varies remarkably from individual to individual
according to personal association of each" (Sapir, 1921,

p. 41). The protocols discussed so far have illustrated

216

the wide range of referential and affective responses
that might be activated by the verbal symbols.

In addition to responding to the descriptive elements
of the work, Wong also made use of the factual information
as a resource for making sense of the story. The statisti-
cal analysis of his second section data indicates that
50% of his response statements are coded descriptive
statements. Wong demonstrated a unique way of using the
factual cues from the text. Most of the readers in the
present study presented their retelling in an order
of actions, but Wong concerned himself not only with the
actions but also with the contexts in which these actions
might take place, and in this way managed to put the
actions into meaningful contexts. For instance, Wong
felt that the family was poor because he noted that there
were patches all over their clothes. The way in which he
talked about the incident of Syau Chfian's eating the
steamed bread is another example. Wong pointed out
that "it took him [Syau Chfiéfi] only a minute to eat all
of it" and "from this brief description" that he felt
that Syau Chfian was a "very selfish person." wong's
approach to the story is by picking up on clues presented
in the text and then putting them all together through
his own interpretation of those clues. In other words,
he gave the text new dimensions which might not exist in

’ the text, because, given all of the same cues, the other

217

readers either did not notice them or else came up with
a different view. From Wong's reading we are able to see
how a reader's responses to the text become integrated
into the work as perceived.

Wong's reactions to the reading of the third section
reinforced the statements that I have made concerning
his particular modes of responding to a literary work.
Wong is a reader who was getting a developing sense
of the story as much, if not more, from the author's
style and the atmosphere of the story as from the actual
details of the story.

He used the people's conversation in the group to
tell the readers what the steamed bread is used for
and what kind of illness Syau Chdan has. Then from
the conversation we learn the blood on the steamed
bread was from Syé's son. And as usual Ld'Syun
satirizes the ignorance of this group of people.
Through these people and their self-conceited
conversational tones we see that Sya's son is a
revolutionist. I like this kind of technique because
it is more convincing. If he presented it in a

very direct way, you would feel that he was preaching
or pushing some kind of slogan. It is very obvious
that he is making use of the speech of the common
people to show their ignorance. But there is

one thing that is not quite clear to me, t t is,
what is the relationship between Sya San Y and Sya
82 Y ? Are they brothers? Oh yes, they are in

the same family. And again we can see the selfish-
ness of Sya and the ignorance of the common people.

Wong mentioned how some of the incidents were presented in
this section and how this specific technique employed by
writer has affected him as a reader "I like this kind of
technique because it is more convincing. If he presented

it in a very direct way, you would feel that he was

218

preaching or pushing some kind of slogan." In this

reader, we find a wholly different way of looking at
retelling. He was interested in the author's style

and intent and talked in terms of whether they were accept-
able to him or not. This is definitely a kind of deep
emotional involvement which has to come down on the side

of aesthetic appreciation rather than on the side of
strictly cognitive acquisition.

It may be appropriate to rehearse here the matter
of aesthetic and efferent readings again. In dealing
with the protocol of our fourth reader, Dung, we noted
that there is a back-and-forth movement of aesthetic and
non-aesthetic transactions with the text during his
reading activity. His response statements are fused with
information obtained from the text and the qualitative
literary experience aroused in him. In fact, if we
examine the reactions of individual readers closely,
we will find most of the reading is a mixture of these
two modes of experience. The distinction between the
aesthetic and the efferent is usually a question of
degree, rather than of an absolute. Table 20 presents
the points assigned by the three independent readers
to the five readers on the aesthetic/efferent scale

based on each reader's protocol.

219

TABLE 20

Five Readers on Aesthetic/Efferent Rank

 

 

 

 

 

 

 

Reader I Reader II Reader III Average

A E A E A E A E
Lin 8 2 8 2 9 1 8.3 1.7
Cu 3 7 2 8 2 8 2.3 7.7
Chen 5 5 4 6 4 6 4.3 5.7
Dung 5 5 6 4 5 5 5 3 4 7
Wong 7 3 7 3 7 3 7 3

 

 

 

 

 

 

If we rank these five readers on a scale with the
aesthetic at one end and the efferent at the other, we

will get a continuum like the following:

Aesthetic Efferent
9 8 7 6 5 4 3 2 l

T. T .11 OT.

Lin Wong Dung Chen Gu

 

As the diagram shows, most of the responses fall somewhere
between the aesthetic and efferent poles, which is ‘
exactly what Louise Rosenblatt has claimed. A mixture

of these two modes of experience is, in fact, quite

220

characteristic of our reading behavior; it is rare to
find either in a pure form. Yet though they often mixed,
it is crucially important to make a clear distinction
between the two kinds of reading because they are
fundamentally different.

Wong has maintained a fairly high degree of
consistency both in his style of responding and in
the selections ofcues to respond to. As we have
noticed all along in our discussion of his protocol, he
demonstrated a clear pattern of responses from the very
beginning of his reading. That is, he stayed very close to
the text, yet at the same time he managed to arrive at a
meaningful explanation for the particular situation
that interested him. With respect to the selections of
cues and the construing of their meaning, Wong is also
very consistent. From the outset of his reading of the
text of this story he was predisposed to disliking the
characters and found them somewhat comical. For example,
he sees Syau Chuan as "selfish" and the group of people
as "a mass without any individual thinking.” The way in
which he perceives the opening scene seems to set the
tempo that determines his future relationship with the
text. His attention was directed to those elements which
bear the negative qualities of human nature. In the
third section, he repeatedly stressed the ignorance of

the common people. In the fourth section, while some

221

readers felt deeply for the two mothers involved, Wong
found them comical and ridiculous figures, who add "a
satirical dimension" to the story.

Obviously, eating the so-called steamed bread

mixed with blood doesn't help. Here, Lu Syun adds

a satirical dimension. He satirizes superstition,
including Mother Sya and Mother Hwa. He arranges
for these two women to appear at the same time in
this paragraph. One is the mother of the
revolutionary. The other one is the mother of the
child who ate the steamed bread mixed with the
revolutionary's blood. However, so far as I can see,
there isn't much difference between these two -
women. It is apparent that Mother Sya loves her son
although she does not understand him and is even

so ashamed of his behavior that she feels ashamed

to go to the cemetery. She is afraid to let other
people know what kind of a son she had. She said
that all her relatives have cut relations with her
and she did not understand why flowers had been put
on her son's grave. I think that during this period
not only this old woman, but many people, were igno-
rant and superstitious. One thing is certain: she
loves her son. There is another part that I find
ridiculous; Mother Hwa came to her son's grave.

She came to the cemetery to mourn her son. However,
it seems that she is more interested in what is going
on at the cemetery. I think it is pretty funny.

How can she be so easily distracted? I am

wondering whether she really felt sad or not.

Almost all of the details that he chose to comment on fit
nicely into the over-all theme that he has established for
the story, namely, that many Chinese people were "ignorant
and superstitious." However, the two most mysterious images,
the flowers without roots and the crow seemed to give him
some difficulty.

Finally, the crow is mentioned. I think that the

author may have a specific idea to convey, but I

cannot figure it out yet. However, it created a

very strange atmosphere. The other thing that I want

to bring up is the flowers on the revolutionary's
grave. From this incident it may be said that

revolutionists are not lonely.

222

He responded to the crow image by saying "the author may
have a specific idea to convey, but I have not figured it
out yet. However, it created a very strange atmosphere.”
It is uncharacteristic of him to pick up on something that
he could not explain. From a researcher's point of view,
this part of the protocol provides us with a precious
opportunity to look into the thinking process of this
reader in his shaping of meaning. Apparently, he had
vaguely sensed that there was something deeper in the
story than the interpretation that he had so far reached.
His reactions to the symbols of red flowers on the

grave of Sya's son seem to confirm this suspicion.

Twice Wong had mentioned the flower image. At first,

he saw the incident of the red flowers without roots

as an indication of ignorance and superstition of the
common people. However, after the crow image, he went
back and commented on the flowers incident again: "
revolutionists are not lonely." It is not clear to us in
the present context whether this final evocation was
activated by the reinforcement of these two symbols, or
whether Wong, too, like the third and fourth readers
perceived the political overtones long before he made the
concluding statement. Whether he would read this story

as a political writing as did the third reader is a
question that can only be answered by further investigation,
such as the study of re-reading. It would be interesting

to see how a reader alters his interpretation of a story

 

 

223

during subsequent reading experiences.

Our detailed study of Wong's protocol reveals that
Perception and Interpretation are his favorite modes of
response. This observation seems to agree with the finding
obtained from the quantitative analysis of his response
statements, which yield 40 percent for descriptive

statements and 44 percent for interpretative statements.

Summary of the Findings

In this chapter, two types of evaluative devices were
adopted to analyze the collected data. One dimension of
the investigation was quantitative, a statistical analysis
of five categories of responses. Another dimension of
analysis was qualitative, an exploration of the readers'

experience with a literary text.

The statistical analysis

 

(1) The category with the largest mean percentage is
Perception (37.8%) followed by Interpretation
(24.3%), Self-involvement (22.9%), Evaluation
(9.9%) and finally Miscellaneous (5%). These
figures are somewhat different from those of
a comparable study conducted by John C. Pollock
(1972). Pollock reported that the dominant
mode of response made by American readers is
the category of Interpretation (34. %) with
Perception next at 26%, followed by Self—
involvement 19.1% and Evaluation 18.2%.
Miscellaneous responses account for 1.8%.

(2) Examination of the subcategory reveals that
Chinese adult readers are most sensitive to the
factual information contained in the text.

(3)

224

Narrational and retelling parts of the story led
in percentage (20. 6%) with descriptive aspects
of the work, such as language, characters,

setting next at a percentage of 17.2, followed
by Interpretative statements of parts of the work
(16.7%). These three subcategories constitute
about 55% of the total response statements.
Pollock's study found that four subcategories--
reaction to content, content, general inter-
pretation and interpretation of content together--
take up about 60% of all responses. The fact
that these more text-based subcategories make

up more than half of the readers' response
statements in both studies suggest that the
linguistic symbols as well as information
designated by the verbal signs jplay an essential
part in the process of reading. The sub-
categories, such as personal statements about

the reader, about the work or evaluative
statements about the construction of the work,
which pertain to individual characteristic

of the readers account for about 38% of the total
responses. The fact that more than one-third

of the entire responses cannot be directly
related to any overt symbols of the text is a
strong evidence that reading is not simply an
automatic reaction to the print. In sum, the
results obtained from content analysis of the
subcategory support the notion proposed by
Louise Rosenblatt that reader and the text are
mutually essential to the reading process.

Analysis of the response statements made after
each division for each category indicate that
there are some discernible patterns of develop—
mental process across sections of the story

(See Figure 7). Descriptive statements are

the leading mode of response. More than 40% of
the responses to the first section of the story
are descriptive, which suggests that readers
respond to the verbal stimuli provided by the
text right from the very beginning of their
reading. The statements of description increase
and reach to the highest in number at division II
(63.2%), then decrease through division III to a
relative low 29.2% in the fourth division.
Percentages of the category of Interpretation
tend to be low at the first two divisions

(14.8% and 11.9% respectively), then to rise

"l..I-‘A".§:£"uflf I. .|* 9'.

225

sharply in reaction to the third and fourth
divisions (31.6% and 28.3%). Evaluative

responses remain relatively low for all of

the four divisions. Nevertheless, the percent-
ages of this category appear to be a little higher
in reaction to the first division (9.5%) and

to the third (7.8%). This finding seems to

agree with James Squire's (1964, p. 31).

The results of the qualitative analysis are, in general,

congruent with the findings of the content analysis of

response statements. However, the additional dimension of

qualitative exploration of the readers'protocols illuminates

the complex relationship that exists between a reader and a

literary text.

Qualitative Analysis

 

(l)

(2)

(3)

Reading a literary text is an unique experience
for each individual reader. The five protocols
under investigation manifest five totally
different kinds of reading experiences. However,
although it is obvious that each experience is
personal (how could it be otherwise?), it is
also clear that each of these unique personal
reading experiences is activated by the symbols
of the one single text.

Reading is an "on-going" process during which the
reader continually reassesses his relationship
with the text. The reader responds not only to
verbal signs offered by the text but also to his
own evocation activated by the verbal stimuli.
The reader is intensely conscious of what goes on
in his mind and is able to articulate it.

One of the most important things that we have
learned from reading the five protocols concerns
the problem of ”selective attention" that
Rosenblatt talks about. It is clear from these
protocols that each reader does bring into his
reading situation his own special interests and
intellectual strengths, which in turn affect his
choice of what to focus on. Readers select from

(4)

(5)

(6)

226

"the multiple possibilities" provided by the
text those elements that can be fitted into the
relevant pattern of ideas, feelings and attitudes.

Readers do reach for information from their own
background and try to apply the situations at
hand. People go outside of the text for inter-
pretation. It is impossible for readers to stick
within the text until they are completely through
a story to figure out what is happening. All

of the readers we have studied go outside of the
text at certain points when the text is confusing,
to find ways of understanding it.

Readers do not process information in a linear
fashion. The reader looks not only forward but
also backward to modifytfluatentative hypothesis
that he has set up for the text which he is
reading as new information enters his focus of
attention. The reader selects, organizes

and synthesizes ideas, feelings and attitudes on
the basis of what seems to be most relevant or
interesting to him at the particular moment of his
relationship with the text.

The distinction between the aesthetic and the
efferent is usually a question of degree, rather
than an absolute difference. In dealing with

the protocols of the five readers, we notice

in almost every case that there is a back and
forth movement of aesthetic and non-aesthetic
transactions with the text during the reading
activity. Most of the reading contains a mixture
of these two modes of experience.

CHAPTER VI,
CONCLUSIONS

The purpose of this dissertation has been to explore
the nature of the reading process by using Chinese adult
readers as subjects. Two empirical research projects
were set up for this purpose. The first study, conducted
within the framework of the RMI, concentrated on the
linguistic aspect of the reading process, namely, on how
a reader made use of the phonological, syntactic and
semantic knowledge of his language systems in his
encounter with a written text. The second study, in
contrast, focused on the experiential aspect of the
reading process while a reader attempted to arrive at a
meaningful and coherent reading of the text. This second
study consists of two parts: a quantitative analysis of
the readers' response statements by making use of the
Purves-Beach coding device and a detailed exploration of
five individual readers' protocols. Louise Rosenblatt's
transactional theory of reading served as a guiding
principle for explaining and interpreting the readers'
behavior. The results obtained from these two studies
were presented in Chapters III and V. In this concluding

chapter, rather than rehearse the detailed findings which

227

228

have already been highlighted in the chapter summaries
of these two chapters, it seems more appropriate to

discuss the relationship beteen a reader and a text as
reflected in our data, and also consider some possible

implications for future research.

The Relationship Between a Reader and a Text

From the five protocols presented in the preceding
section, we have seen five entirely different readings of
the sgmg text. We recognize that each reader carried on a
dynamic and unique personal activity. Each reader brought
different expectations, attitudes, moods and prior experi-
ence to the reading activity, which inevitably affected
their focus of attention and consequently resulted in
differences in responses. In view of these unique
experiences, we find the idea that "the text is merely an
object of paper and ink until some reader responds to the
work on the page as verbal symbols. . . and to bring a
poem into being requires always a reader" (Rosenblatt, 1978,
pp. 12-5) not merely acceptable, but demonstrably true and
useful. The concept of the poem as an experience shaped by
a reader allows us to account for the wide range of divers-
ity which we have discovered in each individual reader.

However, we can say with equal justice that of course

a reader cannot have a literary experience without a text.

 

229

Despite the obvious diversity of readers' literary experi-
ences, we observe that there are significant common
features and operations which can only be attributed to
the influence coming from the text.

At the beginning of our reading, we find ourselves
confronted with paper covered with written or printed
signs. Thus the first thing we experience is the visual
perception of these signs. The study of oral miscues
produced by Chinese adult readers provided the first
evidence in our study that a reader was under the guidance
of a text. As demonstrated in Chapter III, the visual and
sound properties of a language constitute an indispensable
element in the reading process. Readers in the present
study, in fact in all of the RMI studies reviewed, have
shown some degree of dependency on the graphic and sound
information. Readers produced miscues that either lggk
like or EQEEQ like the original text words. The following
is an example demonstrating how Chinese readers made use
of the visual information afforded by the text. Several
readers miscued the character {fii d_i_r_1g for the character
,gfii'jég. The physical resemblance between these two
characters is quite obvious. Readers also produced miscues
which retained the grammatical functions of the text words.
The data on syntactic acceptability of the miscues of the
Chinese readers show that 87.6% of the sentences involved

with miscues are syntactically acceptable in the story.

230

These figures do not include all instances in which the
readers appeared to use syntactic information in their
predictions which are acceptable only with the prior
portion of the sentence. If we combine the totally
acceptable predictions with these which are acceptable with
the prior portion, we see that 96% of the observed responses
preserve at least partially the grammatical structures

of the original text. It is a strong indication of the
readers' awareness of sentence structures offered by the
text and their ability to make use of them. If reading
were a totally subjective behavior on the part of the
reader in which the text had no role to play, it would be
difficult to explain why readers produced sounds, words
and grammatical structures which are so closely related to
the original symbols of the text. In fact, the whole
notion of miscue analysis is built upon the assumption
that a reading activity engages both a reader and a text.
"Reading is a meaningful interaction between the language
of the reader and the language of the author. . . . All
readers do deviate from the text, and these deviations

in oral reading are called miscues" (Y. Goodman and C.
Burke, 1972, p. 5). The importance of the written symbols
in an act of reading is unmistakably assumed here in this
statement. But the Goodmans feel it is equally essential
to make it clear that reading is EQE an exact process of

letter or word identification, but rather a process of

231

predicting, selecting and sampling of cues offered by the
text. Based on this assumption, the Goodmans have evolved
an elaborate analytical system to examine how a reader
makes use of his three cueing systems while reading. Louise
Rosenblatt also recognizes the necessity for a reader to
refer to the linguistic elements of the text while reading:

As one decodes the opening lines or sentences

and pages of a text, one begins to develop a

tentative sense of a framework within which

to place what will follow. . . . One evolves

certain expectations about the dictions, the

subject, the ideas, the themes, the kinds of

text, that will be forthcoming. Each sentence,

each phrase, each word, will signal certain

possibilities and exclude others, thus

limiting the arc of expectations (Rosenblatt, 1978,

p. 54).

Our analysis of the Chinese adult readers' oral miscues
confirms both the claim of the Goodmans and the claim of
Louise Rosenblatt that readers respond to the verbal cues
right from the outset of a reading activity. The fact that
the miscues of the Chinese readers do relate to the original
words in one way or anotherijsstrong evidence that readers
rely on the textual cues offered by the text.

The study of the readers' response statements is
another source which provides us with examples demonstrating
that readers indeed must use the textual information to‘
organize their literary experience. As shown in the
protocols, all of the five readers under detailed investi-

gation, responded to certain details of the story. They all

chose to react to the clues such as "the steamed bread mixed

232

with blood," "the flowers without roots” and Lau Chuan
and his family. The fact that readers all chose to
respond to those incidents indicates that the text did
serve as a basis for the readers to construct the meaning
of the text.

Readers are sensitive not only to the verbal signs
of the text but also to the ways in which the story is
presented. The structure of a text constitutes another
type of constraint. In the present study, we find
readers demonstrated their sensitivities to the structure
of the story by either saying the story is "disordered,"
"obscured," scattered," or "mysterious." Although
readers used different terms to describe the impression
they got from reading the first section of the story,
nevertheless, they shared one thing in common. That is,
they all reacted to the technical devices employed in
structuring the text. Rather than following a general
scheme of a short story with events ordered and introduced
according to a casual, temporal relation, the story
"Medicine" adopts a device which is commonly found in
the writings of a "mystery"story. The story starts with
several events in juxtaposition, and none of them is
explicitly describedixldetail, only hinted at. As soon
as a reader began to read, he immediately found himself
surrounded by obscure details which he was expected to put

together. Probably this is the main reason that readers

233

felt frustrated and complained about the text being unclear.
However, as the readers moved into the second section, they
sensed that the story was told in a much more straight-
forward fashion. They showed their awareness of this
change either by saying "it becomes much more clear," or
simply by trying to explain what they thought had happened
in the story. The quantitative analysis of response
statements also reflects this trend. Nine of the ten
readers devoted more than 50% of their responses in the
second section of the story to factual retelling. In other
words, readers are generally sensitive to the technical
devices used to organize the text and are able to readjust
their strategies in their transaction with the text.
Research on the relationship between discourse
structures and the comprehension activities of the
readers appears to support this observation. G. Olson,
R. Mack and S. Duffy, for example,state in their 1981
report that
Our main claim about these readers is that they know,
either explicitly or tacitly, a number of conventions
for how texts of these types are written, and they
use this knowledge actively during their comprehen-
sion of the text. They have strong expectations
about what they will find in the text, and use these
expectations to guide their understanding. When ,
their expectations are violated. . ., their verbal
protocols reveal confusion and their silent reading
times are slowed down considerably. When their
expectations are met. . . they appear to be able

to profit from the application of their knowledge
of genre conventions to the process of understanding

(p. 39).

234

The preceding are the three most obvious kinds of
textual constraints found in our data. It should be suffi-
cient to illustrate the point that the text is a "necessary"
condition if a reading activity is to take place. The
second question is how to specify the degree of constraints
coming from the text and how to account for the readers'
experience of it. In view of the readers' miscue data,
evidence existstx>demonstrate that the readers are
responding to cues set1forth in the text while reading.
Difficulties arise, however, when one seeks to specify how
a reader and text actually come together to evoke a poem.
The analysis of the three cueing systems provides very
little information as to how a text is translated into
a reader's mind. It is important to note, however, that
the RMI contains an area of investigation which may provide
valuable insights into this process if proper research
questions are asked, namely, the study of the readers'
correction strategies. According to the Goodmans, when a
reader becomes aware that his prediction does not fit the
information, he has two options. One of these options is
to digress and make a new attempt. An attempt to correct

is an indication that the reader feels that his prediction

is not acceptable in some way. The readers' corrections
of oral miscues offer evidence that it is the readers who
formulate and modify the expectations, not the text; in

addition, the readers' correction attempt allows the

235

researcher to look into the reading process in action.

The study of the readers' response statements, on
the other hand, yields some valuable information which
may help us to picture the relationShip between the text
and the reader in the dynamics of the reading process. As
pointed out in Chapter V, the results of the quantitative
analysis show that about 55% of the total responses are
related to either linguistic symbols of the text or
information designated by those symbols, while about 40%
of the reactions bear no direct relationship to any overt
verbal symbols of the text. John Pollock studied the
responses made by American college freshmen and found very
similar results. These statistical figures are signifi-
cant in the sense that they present a clear picture
concerning what kinds of information a reader is likely
to use in the act of reading. However, they still give
very little help in answering the question of how a reader
and a text actually work together to re-create a literary
work of art. In this regard, the exploration of the
individual reader's protocols proves to be most useful.

It is evident from.our data that a literary work of
art is not manifested in the printed text, nor produced.
solely by the reader's imagination, but rather is the
result of the transaction between the two. Throughout
the process of our five readers, we find that there is

continual interplay between expectations and modified

236

expectation. A reader looks not only forward but also
backward to modify the tentative hypothesis which he him—
self has set up as he obtains new information. As you

may recall, the fourth reader, Dung, began his reaction

to the fourth section by saying "In this passage, I found
my predictions to be incorrect. The son was not saved;

he died of tuberculosis." It is clear from this state-
ment that Dung had already formulated his hypothesis about
the outcome of the story, namely, that the medicine should
be effective. But when he moved into the fourth section,
it turned out that what he had been thinking was incorrect,
and Dung admitted that his predictions were wrong. As

seen in this example, Dung was responding simultaneously

to the verbal stimuli of the text and to his own responses
evoked from his previous reading. So a reader responds

not only to the verbal signs afforded by the text but also
to his evocation activated by the verbal stimuli, as
Rosenblatt has proposed. The fact that a reader actually
reacts to his own evocation constitutes the vital link
between the text and the reader. The reader's reactions

to some aspects of the emerging work will become integrated
into the work as perceived. Thus a reader's communication
with the text possesses a continuous and cyclic character.
At each moment, the reader is in a process of reconstructing,
in which new information is picked up, modified and in turn

'directs further exploration.

237

Writing about the act of reading, Louise Rosenblatt
says:
The dynamics of the literary experience include,
then, first the dialogue of the reader with the
text as he creates the world of the work.
responding to cues, adopting a predominantly
efferent or aesthetic stance, developing anticipatory
framework, sensing, synthesizing, organizing and
reorganizing. Second, there is the concurrent
stream of reactions to the work being brought
forth: approval, disapproval, pleasure, shock;
acceptance or rejection of the world that is being
imagined; the supplying of rationale for what is
being lived through. There may be also awareness,

pleasant or unpleasant, of the technical traits
of the text (1978, p. 69).

This notion of concurrent stream of reactions" is crucial
to her transactional theory of reading and to our under-
standing of the reading process as well. It is this mechanism
which is responsible for the continuous and cyclic nature
of the readers' reading behavior found in this present study.
In fact, reading is not the only human experience that
depends on "spatial and temporal continuities." Perception,
attention and memory all have the same structure. Ulric
Neisser, the prominent cognitive psychologist, argues that
perception involves an active transaction between the
perceiver and the object perceived. Perception, according
to Neisser, is operated by "anticipatory Schemata" which
"accept information as it becomes available at sensory '
surfaces and is changed by that information; it directs

movements and exploratory activities that makes more

information available, by which it is further modified"

-'-———..-.

 

238

(1976, p. 54). Schemata direct our perception, but are
affected by what is perceived. So the perceiver and
what is perceived become part of a total situation.
Similarly, the text functions as a guide for "selecting,
rejecting, and ordering of what is being called forth"
(Rosenblatt, 1978, p. 11), but a physical text is trans-
formed into a literary one by virtue of a reader.

In sum, on the basis of the data in the study, the
researcher discovers the relationship between the reader
and the text is not the kind in which one side does
everything for the other. It is a relationship of
equality and reciprocity. The reader is expected to
imagine, to be sure. But the reader's imagination is
not unrestricted; it must play on a given body of material,

that which the text provides.

Suggestions for Further Study

Certain questions and concerns emerged during the
course of this study which indicate a need for further
research.

(1) Do female readers react to a literary work of
art the same ways as male readers do?

(2) Rosenblatt suggests that the act of reading a
scientific article is very different from the
act of reading a literary text.

How do these two modes of reading differ?

(3) Rosenblatt also suggests that any text can be
read either efferently or aesthetically.

239

(a) What makes a reader choose a particular
stance?

(b) What effects do various types of instructions
given before a reading activity takes place
have on a reader's relationship to a written
text?

(4) What is the relationship between oral miscues and
the quality of a reader's reading experience?

Are we able to tell whether a reader is reading

efferently or aesthetically by looking at the

oral miscues he produces?

(5) What is the relationship between oral responses
and written responses?

(6) What is the relationship between first reading
and re-reading? Does a reader transact with a
text in his second reading the same way as he
does in his first reading? If there are
differences, what are they? and why?

The conclusion reached from our discussion of the
functions of the text is significant in two respects. First,
it offers some theoretical justification for the claim that
THE READING PROCESS IN CHINESE AND THE READING PROCESS IN
ENGLISH ARE ESSENTIALLY THE SAME DESPITE THE DIFFERENCES
IN THE WRITING SYSTEMS, since our data clearly demonstrate
that the Goodmans' miscue approach and Rosenblatt's
transactional theory of reading do indeed hold for both
Chinese and English. Secondly, it suggests a methodology
for the study of reading experience. By examining the
oral miscues and response statements produced by the readers
together, we find a way to reconstruct the process of how a
reader makes use of all different levels of information and

his experience of them in the re-creation of a literary work

of art.

BIBLIOGRAPHY

240

BIBLIOGRAPHY

Allen, P. D. 1970. A Psycholinguistic Analysis of the
Substitution Miscues ofmSelected Oral Readers in
Grades Two, Four, and Six and the Relationship of
These Miscues to the Reading_Process: A Descrip-
tive Study. ’PH.D. dissertation, Wayne State
University.

 

 

 

 

 

Allen, P. D., & Watson, D. J., eds. 1976. Findings of
Research in Miscue Analysis: Classroom Implica-
tions. Urbana, 111.: ERICClearingHouse.

 

 

Angelotti, M. L. 1972. A Comparison of Elements in the
Written Free Regponse of Eighth Graders to Junior
NoveI and an Adult Novel. Ph.D. dissertation,
Florida State University.

 

 

 

Applebee A. N. 1973. The Spectator Role: Theoretical
and Developmental Studies of Ideas about and
Response to Literature, with Special Reference
to Four Age Levels. Ph.D. dissertatiOn,
University of LonHOn.

 

 

 

 

Beach, R. W. 1972. The literary Response Process of
College Students WhileiReading and Discussing
Three Poems. Ph.D. dissertation, University of
Illinois at Urbana-Champaign.

 

 

 

Bean, T. W. 1976. An Analysis of the Oral readinggMis-
cues of Hawaiian Islands Dialect Speakers in
Grades Four, Five and Six. Ph.DJ’dissertation,
Afizona State University.

 

Blair, H. L. 1977. A Psycholinguistic Description of the
Oral Reading Behavior of Mature Readers. Ph.D.
dissertation, Arizona State University.

 

 

Bleich, D. 1978. Subjective Criticism. Baltimore: The
Johns Hopkins University Press.

 

Cantril, H., & Livingston, W. K. 1963. "The Concept of
Transaction in Psychology and Neurology." Journal
of Individual Psychology 19: 3-16.

 

241

Carlson, K. L. 1971. A Psycholinguistic Description of
Selected Fourth GradECChildren Reading a Variety
ofiContextual Material. Ph.D. dissertation,
Wayne State University.

 

 

 

Chao, Y. R. 1968. Langugge and Symbolic Systems.
Cambridge: Cambridge University Press.

Cooper, C. R. 1969. Preferred MOdes of Literary Response:
The Characteristics of High School Juniors in
Relation to the Consistency of Their Reactions
to Three Dissimilar Short Storiés. Ph.D.
University of California at Berkeley.

 

 

 

 

Cooper, C. R. 1971. "Measuring Appreciation of Litera-
ture: A Review of Attempts." Research in the
Teaching of English 5: 5-23.

 

 

Cooper, C., & Michalak, O. 1981. "A Note on Determining
Response Styles in Research on Response to
Literature" Research in the Teaching of English

15: 163-169.

 

Cornaby, B. J. 1974. A Study of Influence of Form on
Responses of Twelfth-Grade Students in College-
Preparatory Classes to Dissimilar Novels, A
Short Story, and a Poem. Ph.D. dissertation,
University of”Washington. '

 

 

 

 

Cornelius, P. 1965. Languages in Seventeenth-and Early
Eighteenth Century Imaginary Voyages. Geneva.

 

 

Derrida, J. 1976. Of Grammatology. Translated by Gayatri
Chakravorty Spivak. Baltimore: The Johns HOpkins
University Press.

 

Devine, J. J. 1980. Developmental Patterns in Native and
Non-Native Reading Acquisition. Ph.D. dissertation,
Midfiigan State University.

 

 

Dewey, J. 1934. Art as Experience. New York: Minton
Balch.

 

Dewey, J., & Bentley, A. F. 1949. Knowing and the Known.
Boston: Beacon Press.

 

242

Faggiani, L. A. 1971. The Relationship of Attitude to
Response in the Reading of a Poem By Nineth’Grade
Students. Ph.D. dissertation, New YorkUniversity.

 

 

Fairbanks, J. K. 1976. China Perceived: Images and
Policies in Chinese-American Relations. New York:
Vintage Bodks.

 

 

Fenollosa, E. 1936. The Chinese written Character as a

Medium for PoEtry. San Francisco: City Lights
Book, Inc.

 

 

Gibson, E. J., & Levin, H. 1975. The Psychology of
Reading. Cambridge: The MIT Press.

 

Goodman, K. S. 1965. "A Linguistic Study of Cues of

 

 

Miscues in Reading." Elementary English

42: 639-643.

. 1967. "Reading: A Psycholinguistic Guessing
Game." Journal of the Reading Specialist

6: 126-135.

. 1969. "Analysis of Oral Reading Miscues:
Applied Linguistics.” Reading Research Quarterly
5: 9-30.

 

. 1970. "When a Child Reads: A Psycholinguistic
Analysis." Elementary English 47: 121-128.

 

1972. "The Reading Process: Theory and
Practice." In Language and Learning to Read:
What Teachers Should Know aBout Language. R.
Hodges and H. Rudorf (eds). Boston: Houghton

 

 

 

 

Mifflin.

1973a. "Miscues: Windows on the Reading
Process." In Miscues Analysis: Applications to
Reading Instruction. K. S. Goodman (edJ). Urbana,
111.: NCTE.

1973b. The Psycholinguistic Nature of the
Reading Process. Detroit: Wayne State university

 

 

Press.
1973c. "Psycholinguistic Uhiversals of the
Reading Process." In Psycholinguistics and Reading.

 

F. Smith (ed.). New York: Holt, Rinehart and
Winston.

243

Goodman, K. S., & Burke, C. 1968. Study of Children's
Behavior While Reading Orally. Washington, D. C.:
Department of Health, Education and Welfare.

 

 

1973. Theoretically Based Studies of Patterns
of Miscues in Oral Reading Performance. washington,
D.C.: Department of Health, Education and Welfare.

 

 

Goodman, Y. M. 1967. A Psycholinguistic Description of
Oral Reading Phenomena in Selected Young Beginning
Readers. Ph.D. dissertation, Wayne State Univer-
Slty.

 

 

1971. Longitudinal Study of Children's Oral
ReadingBehavior. Washington, D. C.: Department
of Health, Education and Welfare.

 

 

Goodman, Y. M. & Burke, C. 1972. Reading Miscue Inventory
Manual: Procedures for Diagnosis and Evaluation.
New York: Macmillan.

 

 

1980. Reading Strategies: Focus on Comprehen-
sion. New York: Hdlt, Rinehart and Winston.

 

Goodman, Y. M., & Burke, C., & Lindberg, M. 1974. Readrng
Miscue Inventory: (Evaluation Form) Evaluation
Questions, Coding and Summary Sheets.

 

 

Harste, J., & Carey, R. 1979. "Comprehension as Setting.”
In New Perspectives in Comprehension, J. Harste
and R. Carey (eds). ‘MOnographs in Teaching and
Learzigg, School of Education, Indiana University,
pp. - 2.

 

Haug, F. M. 1974. YoungChildren's Responses to Litera-
ture. Ph.D. dissertation,iUniversity of Minnesota.

 

Heil, C. 1974. A Description and Analysis of the Role of
the Teacher's Response While Teaching a Short
Stor . Ph.D. dissertation, Uhiversity of Pitts-
Burg.

 

 

Holland, N. M. 1975. "The New Paradigm: Subjective or
Transactive?" New Literary History 7.

 

Huey, E. B. 1968 (orig. publ. 1908). The Psychology and
Pedagogy of Reading. New York: Macmillan.

 

 

244

Ingarden, R. 1973. The Cognition of the Literary WOrk of
Art. Translated By R. A. Crowley and K. R. Olson.
Evanston: Northwestern University Press.

 

Iser, W. 1978. The Act of Reading: A Theory of
Aesthetic Response. Baitimore: The Johns
Hopkins University Press.

 

James, W. 1980. The Principles of Psychology. New York:
Henry Holt, pp. 284386.

 

Jensen, L. 1972. A Psycholinguistic Analysis of the
Oral Reading Behavior of Selected Proficient,
Average and WeakiReaders Reading the Same
Material. Ph.D. dissertation, Michigan State
University.

Kalmbach, J. R. 1980. The Evaluation of Narrative Retell-
ings by Sixth Grade Students. Ph.D. dissertation,
Michigan State university

 

 

 

 

Kenner, H. 1971. The Pound Era. Berkeley: University
of California Press.

 

Kirkpatrick, C. 1972. The College Literature Class:
Observation and Description of Class Sessions in
1The Scarlet Letter.‘ *Ph.D. dissertation,
Columbia University.

 

 

 

Knowlson, J. 1975. universal Language Schemes in England
and France 1600-1800. Toronto: University of
Toronto Press.

 

Kolers, P. 1970. "Three Stages of Reading." In Basic
Studies on Reading. H. Levin and J. P. Williams
(eds). New York: Basic Books, Inc.

 

Kratochvil, P. 1968. The Chinese Language Today.
London: Hutchinson University Library.

Kuehn, W. L. 1973. Self-Actualization and Engagement-
Involvement Response to’Literature amonngdoles-
cents. Ph.D. dissertation, University of

Mll’ll’léSOta .

Kuhn, T. 1962. The Structure of Scientific Revolutions.
Chicago: The University of Chicago Press.

245

Leibniz, Gottfried Wilhelm. 1949. New Essays Concerning
Human understanding. Translated by Alfred Grdoen
LangIey. Honolulu: University Press of Hawaii.

 

. 1957. The Preface to Leibniz ' NOVISSIMA SINICA.
Translated and edited by D.’Lach. Honolulu:
University Press of Hawaii.

1977. Discourse on the Natural Theology of
the Chinese. Translated by H. Rosemon, Jr., &
D. J. Cook. Honolulu: University Press of
Hawaii.

 

 

Lucking, R. 1975. A Study of the Effects of a
Hierarchically-Ordered Questioning Technique on
Adolescents1 Responses to Short Stories. Ph.D.
dissertatiOn, University 6f Nebraska at Lincoln.

 

 

Menosky, D. M. 1972. A Psycholinguistic Analysis of
Oral Reading Miscues GeneratediDuring the Reading
of Varying Portions of Text by Selected Readers
from Grades Two, Four, Six and Eight. Ph.D.
dihsertation, WayneiState university.

 

 

 

 

Mertz, M. P. 1972. Responses to Literature among
Adolescents, English Teachers, and College
Students: A Comparative Study. Ph.D. dissertation,
University of Minnesota. ‘

 

 

 

Miller, G. A., & Isard, S. 1963. "Some Perceptual Con-
sequences of Linguistic Rules." Journal of Verbal
Learning and Verbal Behavior 2: 217-228.

 

 

Mbntora, C. B. 1976. "Syntactic Acceptability, Syntactic
Change and Transformations.” In Findings of
Research in Miscue Analysis: Classroom Impli-
cations. P. D. Allen and D. J. Watson (Eds ).
UrBana, Ill.: ERIC Clearinghouse.

 

 

Morris, E. C. 1976. "Critique of a Short Story: an
Application of 'Elements of Writing about a
Literary WOrk.'" Research in the Teaching of
English 10: 157—175.

 

Neisser, U. 1976. Cognition and Reality: Principles and
and Implication ofiCognitive Psychology. San
Francisco: ‘W.iH. Freeman.

 

246

Neisser, U., & Beller, H. K. 1965. "Searching Through
WOrd Lists." British Journal Pyschology 56:
349-358.

 

Northrop, F. S. C. 1966. The Meeting of East and West.
New York: Macmillan.

 

Odell, L., & Cooper, C. R. 1976. "Describing Responses
to Works of Fiction." Research in the Teaching
of English 203-225.

 

 

Olson, G. M., Mack, R. L., & Duffy, 8. 1981. Cognitive
Aspects of Genre. Technical Report No. 11,
University of Michigan.

 

 

Pierce, J. R., & Karlin, J. E. 1957. "Reading Rate and
the Information Rate of a Human Channel." Bell
Systems Technical Journal 36: 497-517.

 

 

Pollock, J. C. 1973. A Study of Responses to Short Stories

 

by Selected Groups of Ninth-Graders, Eleventh-
Graders, and College Freshmen. Ph.D. dissertation,
UniverSity of Colorado.

 

 

Purves, A. C. 1979. "That Sunny Dome: Those Caves of
Ice: A MOdel for Research in Reader Response.”
College English 40: 802-812.

 

Purves, A. C. & Beach, R. 1972. Literature and the
Reader: Research in Response to Literature,
Reading Interests, and the Teaching of Literature.
Urbana, 111.: NCTE.

 

 

 

 

Purves, A. C.; Foshay, A. W., & Hansson. 1973. Litera-
ture Education in Ten Countries. Stockholm:
Almoqvist & Wiksell:

 

Purves, A. C., & Rippere, V. 1968. Elements of Writing
about a Literary Work: A Study of Response to
Literature. *Research‘Report No. 8. urhana, 111.:
NCTE.

 

 

Rawski, E. S. 1979. Education and P0pu1ar Literacy in
Ching China. Ann Arbor: The university of
Michigan Press.

 

 

247

Rene Wellek & Warren, A. 1956. Theory of Literature.
New York: Harcourt Brace andiWorld.

 

Richards, 1. A. 1929. Practical Criticism. New York:
Harcourt, Brace andiWorld.

 

Romatowski, J. A. 1973. A Psycholinguistic Description
of Miscues Generated by Selected Bilingual
Subjects During the Oral Reading of Instructional
Material as Presented—ih Polish and’English
Basal Readers. Ph. D. dissertation, Wayne State
University.

 

 

 

 

 

Rosenblatt, L. M. 1938. Literature as Exploration, New
York: Noble and Noble.

 

1964. "The Poem as Event.” College English

 

'26: 123-128.
. 1977. "The Transactional Theory of the Literary
WOrk: Implications for Research." Paper pre-

sented at the Buffalo Conference on Researching
Response to Literature and the Teaching of
Literature, October 27—29, State university of
New York at Buffalo.

1978. The Reader, the Text, the Poem: The
Transactional Theory Of"the Literary Work.
Carbondale, 111.: Southern Illinois uniVersity
Press.

 

 

1981. "Act I, Scene I: Enter the Reader."
Literature in Performance: A Journal of Literary
and Performing Act 2: 13-23.

 

Sapir, E. 1921. Language. New York: Harcourt Brace
and WOrld.

Sims, R. 1976. "Structural Levels." In Findings of
Research in Miscue Analysis: Classroomilmplica-
tions. Urbana, 111.: ERIC Clearinghouse.

 

 

Slatoff, W. 1970. With Respect to Readers: Dimensions
of Literary Response. *Ithaca: ’Cornell University
Press.

 

 

248

Smith, F. 1969. "The Use of Featural Dependency Across
Letters in Visual Identification of Words."
Journal of Verbal Learning and Verbal Behavior

8: 215-218.

 

. 1971. Understanding Reading: A Psycholinguistic
Analysis of Reading. New York: Holt, Rinehart
and Winston.

 

 

1973. Psycholinguistics and Reading. New York:
Holt, Rinehart and’Winston.

 

1975. Comprehension and Learning. New York:
Holt, Rinehart and Winston.

 

Smith, M. G. 1978. A Comparison Between Oral Reading
Miscues of Readers with Normal Intelligence and
Educable Mentally Retarded Readers. Ph.D.

dissertation, the University ofCWisconsin,

 

 

 

Madison.
Smith, S. 1979. "Retellings As Measures of Compre-
hension: A Perspective." In New Perspectives

 

of Comprehension, J. Harste and R. Carey (eds.).
Monographs in Teaching and Learning, School of
Education, Indiana University, pp. 84-99.

 

Somers, A. B. 1972. Responses of Advanced and Average
Readers in Grades Seven, Nine, and Eleven to
TWo disSimilar Short Stories. thD. dissertation,
Florida State university.

 

 

 

Squire, J. R. 1964. The Responses of Adolescents
While Reading Four Short Stories. Urbana, Ill.:
NCTE

 

 

Weber, R. M. 1968. "The Study of Oral Reading Errors:
A Survey of the Literature." Reading Research
Quarter1y4: 96—119.

 

 

Weiss, J. D. 1968. The Relative Effects upon High School
Students of Inductive and Programmed Instruction
ih the Close Reading of Poetry. Ph.D. dissertation,
New York University.

 

 

 

Wilson, J. R. 1966. Responses of Colle e Freshmen to
Three Novels. Urbana, 111.: N E

 

APPENDIX A

READING MISCUE INVENTORY: (EVALUATION FORM)

 

249

READING MISCUE INVENTORY: EVALUATION

by Yetta Goodman and Carolyn Burke
and Margaret Lindberg

WORD LEVEL SUBSTITUTION 1N CONTEXT: EVALUATION

 

 

Evaluation of the following questions indicates whether
the student is making appropriate use of grammatical
function and of the graphophonic cueing systems.
Questions 1, 2, and 3 are answered for only word level
substitution miscues. under column headed IEEE list
the word which the reader substituted. Answer the fol-
lowing questions for each of these pairs of words. If
dialect is involved place a d_next to the reader's
substitution.

Q 1. Graphic similarity: How much do the two words
look alike?

high TWO of their three parts
are similar.

Beginning and middle.
Beginning and end.
Middle and end.
some ONE of their three parts
is similar.

Beginning or general con-
figuration.

Middle.
End.

none NONE of their three parts
is similar.

Q 2.

Q 3.

250

Sound similarity:

high

some

none

Grammatical Function:

S ame

questionable

different

How much do the two words
sound alike?

TWO of their three parts
are similar.

Beginning and middle.
Beginning and end.
Middle and end.

ONE of their three parts
is similar.

Beginning or general con-
figuration.

Middle.
End.

NONE of their three parts
is similar.

Is the grammatical function
of the reader's word the
same as the grammatical
function of the text word?
(To help answer this ques-
tion read the text sentence
with the reader's miscue in
it.

The reader's miscue is the
same grammatical func-
tion as the text word.

It is not possible to tell
whether the grammatical
function of the reader's
miscue is the same or dif-
ferent from the grammati-
cal function of the text.

The reader's miscue is a
different grammatical
function than the text
word.

251

LANGUAGE SENSE: EVALUATION

 

 

Evaluation of the following two questions indicates the
degree to which the reader is concerned with producing
acceptable language. Questions 4 and 5 are answered for
every sentence which contains one or more miscues. To
read for acceptability consider each sentence as the
reader finally produced it. All corrected miscues or
attempts at correction should be read as finally resolved
by the reader. ‘When there is no attempt at correction,
the miscues should be read as produced. Miscues which
are acceptable within the reader's dialect should be

considered acceptable.

Number each sentence in the text and place the number
for sentences containing miscues under the column headed
Sentence Number, Next to this, in the column headed
Number of Miscues, indicate the number of miscues con-

tained in each of the sentences.

Q 4. Syntactic Accept-
ability: Is the sentence involving
the miscue syntacti-
cally (grammatically)
acceptable in the story?
yes When the sentence is read

as finally produced by
the reader it is syntac-
tically acceptable in
the story.

252

no When the sentence is read
as finally produced by
the reader it is not
syntactically accept-
able in the story.

Q 5. Semantic Accept-
ability: Is the sentence involving
the miscues semantically
(meaning) acceptable in
the story?

yes When the sentence is read
as finally produced by
the reader it is seman-
tically acceptable in the
story. '

no When the sentence is read
finally produced by the
reader it is not seman-
tically acceptable in the
story.

COMPREHENDING: EVALUATION

 

 

Evaluation of this question indicates the degree to which
the reader changes the intended meaning of the author as
he reads. Question 6 is answered for every sentence
which contains one or more miscues. To determine the
degree of change the sentence is read as the reader
finally produced it. All corrected miscues or attempts
at correction should be read as finally resolved by the
reader. When there are no attempts at correction the

miscue should be read as produced.

Q 6.

Meaning Change:

no

minimal

yes

253

Is there a change in
meaning involved in the
sentence?

When the sentence is read
as finally produced by
the reader there is no
change in the intended
meaning of the story.

When the sentence is read
as finally produced by
the reader there is a
change, inconsistency
or loss to minor inci-
dents, characters or
sequences in the story.

When the sentence is read
as finally produced by
the reader there is
change, inconsistency
or loss to major inci-
dents, characters or
sequences in the story.

APPENDIX B
AN ENGLISH VERSION OF THE READING TEXT

Medicine

by Lu Syun (Translated by George A. Kennedy)

1

It was in the early hours of an autumn morning,
after the passing of the moon and before the coming
of daylight. Only a sheet of blue-black sky remained,
under which all lay asleep save the creatures of the
night.

Old Hua Shuan sat up suddenly. Striking a match,
he lighted a greasy oil lamp, the greenish glow from
which filtered through the two rooms of the tea-shop.

"Little Shfian's dad!" came an old woman's voice,
"Are you going now?" At the same time there was a
spell of coughing in the small inner room.

”Hm!" Old Shuan grunted in reply as he buttoned
up his clothing. He stretched out one hand. "Give it
to me!"

Mother Hda rummaged about beneath the pillow
for a long time, finally bringing to light a parcel
of dollars which she handed over to Old Shuan. He
took it from her with trembling hands and thrust it
into his pocket, giving it a pat to make sure it was
there. Then he lit a paper lantern, extinguished the
lamp, and walked into the inner room. There was a
rustling noise followed by a succession of coughs.
Old Shuan waited for it to quiet down, then spoke in
a low voice, "Little Shuan! . . . Don't get up. . .
The shop? . . . . Your mother will manage all right

There was no answer from his son, and thinking that
he had dropped off quietlytxasleep, 01d Shuan stepped
through the door and out on to the street. The streets
were dark and deserted. Nothing was visible but the
gray stretch of roadway and his two feet illuminated
by the lantern following one behind the other. From
time to time he met a few dogs, but not one of them
made a sound. The air was much colder than it had been
in the house. Old Shuan found it exhilarating. With
the coming of this dawn, Somehow, he was a youth again,

254

255

blessed with mystic insights, filled with life-
imparting power. His stride lengthened and the road,
as he walked, became clearer and the sky brighter.

Old Shfian's whole attention had been riveted on the
road, when he started back suddenly as he saw ahead
where it ended clearly at a cross street. He retraced
a few steps and, finding a little shop still tightly
shuttered, took shelter under the eaves and stood
leaning up against the door. Some time passed and he
realized that he was cold.

"Hey, an old fellow over there!"

It

"Seems cheerful enough .

Old Shuan started a second time as, staring ahead,
he saw a number of people passing by in front of him.
One of them turned his head to look at him. His figure
was indistinct, but he had the manner of a long-starved
wretch lighting upon food, his eyes glittering with
greedy desire.

The lantern had gone out, Old Shuan noticed. He
felt the pocket inside his clothing. It was still
there, hard and firm. He raised his head and looked
about him. Numbers of fantastic creatures were wander-
ing about like devils in groups of twos or threes.

He kept staring, but could discover nothing else that
seemed strange.

It was not long before he saw several soldiers moving
about, the great white circular design on the front and
back of their clothing being visible even from a distance.
Of those who passed nearer by he could make out in
addition the dark red trimming on the uniforms.

There was a rush of trampling feet, and in a
twinkling a large crowd had gone jostling by. The
scattered groups of twos and threes formed suddenly
into a mass, rolling forward like a tide. Reaching
the street crossing, the crowd came to an abrupt halt
and ranged itself in a semicircle. Old Shuan gazed
in their direction but saw only a multitude of human
backs. Necks were craned like those of so many ducks
being pulled upwards in the clutch of invisible hands.
After a period of deep silence, there seemed to come a

256

faint sound and then the crowd became agitated once
more. With a deep rumble it streamed back again, past
where Old Shuan was standing, sweeping him almost off
his feet.

"Hey Old Shuan! Pass over the money with one hand
and take your stuff in the other!”

In front of him stood a fellow dressed from head
to foot in black and with eyes like two swords that
bored into Old Shuan, shriveling him to half his usual
size. One of the man's huge hands was stretched out
toward him, while the other clutched a bright red roll
of steamed bread from which the redness was still
dripping drop by drop. Old Shuan fumbled hastily for
his money and held it out tremblingly but could not
bring himself to take the thing the other held.

"What are you scared of?” the man said impatiently,
"Why don't you take it?"

As Old Shuan still held back, the fellow in black
snatched up his lantern, ripped off the paper shade,
and wrapping it around the bread, thrust the parcel
in Old Shuan's hand. Then he snatched up the money,
tried the feel of it, turned, and was gone. "The old
fool . . ." he muttered.

"Who's the medicine for?” Old Shuan thought he
heard someone questioning him but he did not answer.
All his thought was concentrated on his package. It
seemed as precious as the last of ten generations of
only-begotten sons. All other thoughtswere subordinate
to this. He must now transplant into his home the fresh
life that his package held and reap its harvest of
happiness.

The sun had risen. In front of him it lit up a
broad street stretching straight to his house; behind him
it touched the faded gilt characters on the broken
signboard at the street crossing: "Site of the
Ancient . . . Pavilion."

2

When Old Shuan arrived home he found the shop
cleaned and set in order with its rows of smoothly
glistening tea tables. There were no customers, only
Little Shuan eating at one of the inner tables.

257

Great drops of sweat rolled down from his forehead and
his lined coat clung to his spine where the shoulder
blades stuck out like a raised number eight.* Old
Shuan's face clouded over at sight of him. His wife
came hastily from behind the cooking range and her
lips trembled somewhat as she peered at him.

"Did you get it?"
"Yes."

They went together to the back of the cooking range.
There was a moment's consultation; then Mother Hua went
out, to return shortly with an old lotus leaf Which she
laid on the table. Old Shuan unfolded the lantern
shade and wrapped the scarlet bread roll up again in the
lotus leaf. Little Shuan had finished his rice. His
mother called hurriedly to him:

"Little Shuan! Sit still! Don't come over here!"

She was stirring up the fire in the oven. Old
Shuan took the deep green parcel and the torn white
shade splotched with red and placed them together on
the coals. There was a dull red flare, and then the
shop filled with a strange fragrance.

"That's a f'ne smell! What are you_cooking?"
Hunchback Sao-y h the Fifth had come in. He invariably
spent the whole of every day in the tea shop, being the
first to arrive as well as the last to leave. He had
just come limping up to the table in the corner next to
the street, but no one paid any attention to his
question.

"Is it rice porridge?" Still no one answered him
although Old Shuan hurried out and poured him some tea.

"Come in here, Little Shuan!”

Mbther Hua led him into the inner room in the
middle of which she had placed a bench. When he was
seated she handed him a plate on which was something
round and very black. She said gently, "Eat this
up ... . . You're going to get all well again."

Little Shuan picked up the black object and looked

* The Chinese character for "eight":

258

at it with inexpressible wonder as though he were
holding his own life in his hands. He broke it open
carefully. From within the charred shell came a cloud
of white steam which drifted away to leave two halves
of a bread roll made from white flour.

In a short time he had finished it but he could
remember nothing of its taste. The empty plate was
before him, and beside him were his father and mother,
one on either side, with eyes fixed on him as though
they expected to infuse something into his body and to
withdraw something at the same time. This made his
heart beat quicker. He clutched his chest and burst out
in another fit of coughing.

”Sleep a little while . . . . You're going to be
all right."

Little Shuan went coughing to bed as his mother
directed. Mother Hua waited for his breathing to become
quieter before she gently covered him with the collection
of patches that made his quilt.

3

There were a great many people sitting in the shop.
Old Shuan was very busy carrying his great copper kettle
about to fill up the tea- -cups of his guests. His eyes
were ringed with dark circle.

"Is there something the matter with you, Old Shuan?
Are you sick?" asked a gray-bearded man.

IINO . I!

"No? I thought from your smiles that you didn't
look . . . ."
The gray-bearded man did not finish.

"He works too hard. Now if his son .

The Hunch-back's speech was interrupted by a sallow-faced
man who burst into the shop wearing a dark gown

thrown over his shoulders with the buttons all unfastened
and held carelessly around his waist by a loose dark
sash. As soon as he was through the door, he shouted

out to Old Shuan:

259

"Has he eaten it? Is he better? Old Shuan, I've
brought you luck! It's a good thing that I've a nose
for news.

Old Shuan was all smiles as he listened, one hand
holding the teakettle and the other lowered in a respect-
ful gesture. All the people in the room listened
respectfully, in fact. And Mother Hua, with darkly
ringed eyes, came smiling out to proffer a teacup and
tea leaves, adding an olive as an extra delicacy. Old
Shuan poured on the boiling water.

"This is a sure thing! This is quite exceptional.
Think of it!" The sallow-skinned man kept up boisterously.
"Brought home while warm and swallowed down still warm!"

"That' 5 right!” Mother Hua thanked him warmly.
"If it hadn't been for your kindly interest, Uncle K' ang,
how could we.

"Guaranteed, guaranteed! Taken hot like this--a
roll of bread with human blood--guaranteed to cure any
form of consumption!"

At the word "consumption" a change came over
Muther Hua' 3 face. She seemed displeased, but
immediately put on_ a smile and walked away mumbling to
herself. Uncle K' ang, however, had noted nothing.

He continuedix)shout at the tOp of his voice until
Little Shuan woke to keep him company with his coughing.

"To think," said the gray-bearded customer, "that
your Little Shuan should have such a stroke of good
fortune! Of course he will recover completely. No
wonder you smile all day long, Old Shuan.‘ Having
spoken, he stepped over to Uncle K' ang and said in a
low voice, "We have heard, Uncle K' ang, that the
criminal who met his end this morning was some child in
the Hsia household. Whose child was he really and what
was it about?"

”Who was he? Why, the son‘ of the widow Hsia.
The young scoundrel!" Uncle K' ang noted a general
pricking up of ears and grew highly animated. His
yellow skin blew itself out and his voice took up a
higher pitch. "That little wretch didn't deserve to
live. Well, that'esall right. But there was nothing

260

in it for me this time. They even stripped his clothes
off him and Red-eyed Ah Yi, the jailer, got it all.

uncle Shuan here was the luckiest; the next was Hsia's
third son who got twenty-five snow-white ounces of silver
to stick in his purse without having to spend a single
cash."

Little Shuan had come slowly out from the little
room clutching his chest with both hands and coughing
continuously. He filled a bowl with cold rice from the
range, poured hot water over it, and sat down to eat.
Mother Hau followed him. "Do you feel better, Little

Shuan?" she asked quietly. "Are you still as hungry
as ever?"

Uncle K'Eng glanced at Little Shuan. "Guaranteed!
Cure absolutely guaranteed!" Then he turned back to his

audience. "That third Hsia is a foxy one all right.

If he hadn't turned him in first, all his own property
would have been confiscated and the whole family
implicated. But look at him now! Silver!--That little
fellow was absolutely unbelievable. Even in jail he
tried to get his keeper to join the revolution!"

"Ai-ya! Imagine that!" A young fellow twenty or so
spoke up angrily from the back row.

"You see, Red-eyed Ah Y1 tried to find out from
him what the story was, but the kid gave him a political
lecture. He said the Great Ch'ing Empire by rights
belongs to all of us. Think of that! Does that make
any sense? Red-eye knew, of course, that he had only
an old mother at home. He hadn't realized, however,
that the child was too poor to be squeezed of the
tiniest bribe. Red-eye was bursting with anger over
that already, so when that kid with his politics
scratched the head of a tiger, old Red-eye let him have
a few cuffs for his trouble."

"Brother Y1 has a good fist!" spoke up the hunch-
back gleefully from his corner. "I'll bet those two
punches did the trick all right!"

"Nah, you couldn't frighten a worthless scoundrel
like him. He kept saying 'What a pity, what a pity!'"

261

"Pity?" said the graybeard. ”Who's to have pity
for a creature like that?"

Uncle K'ang looked him up and down with infinite
scorn. ”You misunderstood me. ‘What he meant apparently
was that Ah Y1 deserved pity."

The listeners suddenly dropped their eyes and a
silence fell. Little Shuan had finished eating his
rice. He was perspiring so freely that his head seemed
to be steaming.

"Pity for Ah Y1! What rot! He must have been
completely mad!" The graybeard blurted out his words
in a burst of sudden clarity.

”Absolutely mad!" agreed a man of twenty, also
suddenly clear about it.

Thereupon the customers in the shop began to liven
up and chat and laugh again. Little Shuan, taking
advantage of the commotion, coughed vigorously, but Uncle
K'ang stepped over to him and patted him on the shoulder.

"Guaranteed! Don't cough this way, Little Shuan,
I guarantee it!”

"He was mad!" said Hunchback Séo-ygh the Fifth
nodding his head.

4

Outside the west gate and lying close to the footof
the city wall was a piece of land that had always been
government property. Diagonally across it ran a narrow
crooked path, which, starting as a shortcut trodden out
by the shoes of passers-by, had ended by becoming a
natural boundary. On the left of the road were burial
mounds of such criminals as had been executed or had
died from exposure in jail; on the right was a buying
ground for paupers. Both sides were already crowded
with rows and files of mounds like the ceremonial cakes
at a rich man's birthday party.

It was unusually cold at the spring festival this
year. The buds on the poplars and willows were scarcely
half the size of a grain of rice. Not long after day-
break Mother Hua stood before a fresh grave on the right

262

hand side of the path. She set out four plates of food
and a bowl of rice, wept a while, and burned sacrificial
money. Then she sat mutely on the ground, apparently
waiting for something, though she could not herself
have told what it was she expected. A faint breeze
ruffled her thin har, whiter by far than it had been

the year before.

Along the path came another woman, also gray-
haired, with ragged coat and skirt. She carried a
round broken basket that had once been red lacquer and
a string of sacrificial paper money. She had to rest
every few steps. Suddenly she saw Mother Hua sitting
on the gound watching her. She hesitated and an
embarrassed flush spread over her pale thin face. Then,
summoning her courage, she made her way over to a
grave on the left-hand side and put down her basket.

This grave was in the same row as that of Little
Shuan, separated only by the narrow path. Mother Hua
watched the woman set out the four dishes of food and
the bowl of rice, saw her stand and weep and then ignite
the paper money. "It will be a son," she was thinking
to herself, "in that grave, too.”

The old woman had been moving about to survey things
when she gave a sudden start and, staggering backwards,
stood staring in terror. Mother Hua feared that the
woman was going insane from grief, and she felt con-
strained to rise, cross the path, and speak to her.

"Lao Ngi-nai," she said quietly, "don't take it so
hard. Let's go back home."

The woman nodded, but her eyes remained in a fixed
stare. "Look!" she stammered, ”Look, what's that?"

Mother Hua followed her pointing finger and her
eyes fell on the grave in front of them. It was ugly
enough, with patches of yellow dirt where the grass had
not grown. But a closer look produced a shudder of
amazement--for there,clearly encircling the rounded tip
of the grave, was a wreath of red and white flowers.

Even to their age-dimmed eyes the flowers were
quite distinct. There were not many of them; they
were somewhat wilted; but they were arranged very

263

regularly in a perfect circle. Mother Hua took a

hasty look at her son's grave and at the others, but
she saw nothing beyond a few bluish wildflowers braving
the cold here and there. She was possessed by a sudden
feeling of emptiness and was reluCtant to examine
further. The old woman went nearer to the grave and
regarded it carefully, muttering to herself.

"They have no roots. They don't seem to have grown
of themselves. But who could have come here?
Children wouldn't play here. . . . Relatives stopped
coming long ago. . . . What can have happened?"
Suddenly she broke out in a sobbing cry. "Oh, my child!
They all wronged you! And you can t forget it. Are you
still grieving? And is this a sign from the spirit
world you give me so that I will know?" Looking about
her she saw a black crow sitting on a leafless tree.
"I know now," she continued. "Oh, my child! They
murdered you. They'll get their punishment same day,
Heaven knows. Just close your eyes. . . . But if you
are really here, if you are listening to me, give me a
sign . . . make that crow fly on to your grave.”

The breeze had died away. The blades of dry grass
stood up stiffly like copper wire. The quivering
thread of sound echoed more and more faintly on the air
and was gone. All around was the stillness of death.
The two stood in the withered grass gazing up at the
black crow; and the crow, with head tucked in, sat on
his branch as though he were cast in iron.

A long time passed. The visitors to the graves
increased in number. Old and young passed in and out
amongst the grave mounds.

Mother Hua felt relieved of a great burden and
turned to go. "Come," she urged, "we had better go back."

The old woman sighed and listlessly gathered up
the dishes of food. She hesitated still a moment, then
walked slowly away. "What can it all mean?" she mumbled
to herself.

They had gone not more than twenty or thirty steps
when a loud caw sounded behind them. With a jerk, they
turned their heads to look. The black crow spread his
wings, gave a preliminary flap, and took flight like an
arrow into the distance.

"11111111111!"111115