Transcript Document

A Corpus-based Analysis of
Errors in Parts of Speech in
Chinese Learner English
by Zhao xincheng
Huaqiao University
Fujian, PRC
I. Introduction
1. Definition
Errors in parts of speech (POS), defined as the misuse
of parts of speech but with meanings of words or
utterances sufficiently conveyed (Yang, 2005), are
among the most frequently committed errors in the use
of words found in compositions by Chinese learners of
English, accounting for 4.18% of the 21 different types
of errors with frequencies exceeding 1% in the Chinese
Learner English Corpus (CLEC) (Gui & Yang, 2003).
2. Objectives
1) patterns of errors in POS by Chinese learners of
English at elementary and advanced levels
2) possible correlation between error patterns and learner
proficiency levels
3) possible causes of errors in POS
4) implications for vocabulary teaching and learning
II. Subjects and Methods
1.
Subjects
ST2 and ST6 were chosen from the CLEC,
representing learners of English at elementary and
advanced levels respectively. The reason why these two
samples were chosen is that they are learners’ written
productions under non-examination conditions,
supposedly free from time constraints, anxiety, and
avoidance, as language production under test
conditions is more liable to deviations from the norms.
Moreover, ST2 and ST6 can be regarded as the two
poles on the continuum of Chinese English learners
with homogeneous social and cultural background
under formal instructional settings, thus a comparison
can be more justifiably made between them to
investigate similarities in their production errors, and
probable correlations between proficiency levels and
types of errors.
2. Tools and method
Learner errors are classified into 61 categories in the
CLEC. First, errors in POS (coded wd2) were collected
from ST2 and ST6 branch corpuses using ConcApp,
and their frequencies were tallied. Next, the errors were
further classified into subcategories and labeled
accordingly. See table 1.
Learner type
ST2
No
Types of errors and labels
Errors in POS
frequency
wd2
412
Percentage
%
ST6
frequency percentage
%
308
1
Adj as N
(adjAn)
65
15.78
60
19.48
2
N as Adj
(nAadj)
51
12.38
47
15.26
3
V as Adj
( vAadj)
13
3.16
8
2.60
4
V as N
41
9.95
42
13.64
5
Wrong inflections (Winf)
42
10.19
12
3.90
6
Adj as Adv
37
8.98
29
9.42
7
N as V
37
8.98
35
11.36
8
Adv as Adj
( advAadj)
32
7.77
21
6.82
9
Prep as V
(prepAv)
21
5.10
1
0.32
10
Adj as V
(adjAv)
14
3.40
10
3.25
11
misuse of Ns with same roots (nSR)
11
2.67
11
3.57
12
Verb forms
0
0.00
17
5.52
13
Coined words
4
0.97
12
3.90
14
Wrong pron
5
1.21
0
0.00
15
Others
37
9.47
3
0.97
(vAn)
(adjAadv)
(nAv)
(Vform)
(coinW)
(Mpron)
( Other)

Errors in POS were further classified into 15
subcategories, as shown in table 1. Type 1 to 14 can be
defined with certainty, but type 15 includes errors that
are not specifiable for the following reasons 1) it is
difficult to identify the writer’s intention because of the
gravity of the errors, especially in ST2; 2) errors with a
frequency of less than 2, which is statistically
insignificant; 3) errors that are difficult to identify in
CLEC.
III. Main Findings
1. Frequency patterns of errors in POS
1)The 4 most frequent types of POS errors made by ST2
and ST6 are:
adjAn > nAadj > vAn > nAv;and ST2 and ST6
demonstrated similar tendency in these 4 types of POS
errors in the use of lexical words;
2)ST2 committed more errors than ST6 in the following
categories:
Winf > prepAv > others
3)ST6 made more errors than ST2 in the following
categories:
adjAn > nAadj > vAn > nAv > Vform > coinW
4) ST2 and ST6 had similar frequencies of errors in
comparison between ST2 and ST6's errors in POS
25
percentage
20
15
types
ST2
ST6
10
5
0
1
2
3
4
5
6
7
8
9
10 11
types of errors
Chart 1: Patterns of POS errors by ST2 and ST6
12 13
14 15
2. Distribution of errors in POS
As table 1 and Chart 1 indicate, overwhelming majority
of the errors in POS are related with lexical words,
amounting to 343 by ST2 (83.3%), and 293 by ST6
(95%). The most frequent errors are adjAn, followed by
nAadj, nAv, vAn, adjAadv, advAadj. Other types of
errors are much less frequent, and there is a noticeable
difference between ST2 and St6, too.
distribution of POS errors by ST2 and ST6 (%)
40.00
35.00
30.00
25.00
ST6
ST2
20.00
15.00
10.00
5.00
types of errors
Chart 2: distribution of POS errors by ST2 & ST6
vf
or
m
co
in
W
Mp
ro
n
ot
he
rs
nS
R
ad
vA
ad
j
pr
ep
Av
ad
jA
v
nA
v
Wi
n
ad f
jA
ad
v
vA
n
ad
jA
n
nA
ad
j
vA
ad
j
0.00
Chart 2 shows that the percentages of errors in the use
of lexical words by ST2 and ST6 are close to each other,
while the use of prepositions as verbs and wrong
pronouns only occurred in ST2, and errors in verb
forms only occurred in ST6, who also used coined
words three times as much as ST2.
3. Some tentative conclusions
(1) Errors in POS do not reflect learners proficiency
levels; whatever their level is, they commit frequent
POS errors in the use of lexical words which might
be an indication of L1 transfer. This supports Sun
Zhimian (2001), who collected tons of typical errors
made by Chinese learners of English during a decade
of his English teaching career, all of which can be
traced to Chinese L1 thinking patterns and linguistic
habits.
(2) there are some observable differences between some
types of POS errors, which might be developmental
errors, caused by learners’ incomplete or incorrect use
of L2 rules, the use of learning or communicative
strategies( the more frequent errors in Vform, coined
words by ST6), or induced errors due to lack of
attention to constraints on the use of linguistic rules
(overgeneralization).
IV. Discussion
1. Transfer of Chinese word class features into
learners’ interlangauge
Why is it that ST2 and ST6, the two poles on the
proficiency continuum, demonstrate similar patterns
of errors in lexical words?
For Chinese learners of English, parts of speech are
treated as basic grammatical concepts .
Classroom observations suggest that Chinese learners
of English, especially high proficiency learners, have a
strong meta-linguistic awareness of word classes and
their respective syntactic roles.
According to empirical studies by Liu (2006:131) word
class is that part of the word knowledge which is
acquired early, and “ invariably the best demonstrated
type of word knowledge” on Chinese learners’ English
L2 development route, either receptively or productively.
All factors considered, the only possible cause of their errors in
POS in lexical words is transfer of word class features in L1
Chinese. A comparison between English and Chinese
word classes may shed light on Chinese L1 transfer in
the domain of word class in English.
2. Linguistic markedness, English and Chinese word
classes
(1) Definition
According to Larsen-Freeman and Long (2000:101),
linguistic markedness is usually defined in terms of
complexity, relative infrequency of use or departure
from something that is more basic, typical in a language.
It is generally accepted in markedness theory that linguistically
unmarked features of L1 will tend to transfer, but linguistically
marked features will not. The relation between linguistic
markedness and transfer can be summarized by the
following table.
Table 2: Markedness and transfer (Ellis, 1994)
Language 1(L1)
Target language (L2)
Interlanguage
1 unmarked
unmarked
unmarked
2 unmarked
marked
unmarked
3 marked
unmarked
unmarked
4 marked
marked
unmarked
(2) the differences between English and Chinese
word classes in terms of linguistic markedness
(a) most lexical words in English have overt forms
(deviational suffix) from which their parts of speech
can be inferred. For example, in the following list of
words taken from CLEC where errors occurred, the
parts of speech can be identified from the suffixes of
the words:
Happy / happiness, sleep / sleepy, pollute / pollution,
performance / perform, encourage/ encouragement,
hibernate/hibernation, minimize / minimum, emphasize /
emphasis, success/ succeed /successful, death / die,
jealous/jealousy, health / healthy, offend/ offence / offensive,
differ /different / difference, ornamental / ornament, help /
helpful, tradition / traditional, society / social, etc.
Take the following sentence as another example.
Statements (noun、plural) about crime are (plural) not
criminal (adjective)language(noun、singular); nor are
(plural) statements (noun、plural) about emotions
(plural) necessarily (adverbs) emotional (adjective)
language(noun、singular)
In the above sentence, the italicized words all have
word class markings. Even if one does not understand
the sentence as a whole, he can at least identify the
parts of speech of the words with the help of their
deviational endings.
(b) In contrast Chinese does not have overt formal
indication of word classes.
According to Otto Yespersen: most words in Chinese
have no deviations or inflections to mark word class
(quoted in Wu & Wang, 1994) and the parts of speech
of words can only be drawn by their syntactic functions
(that may account for the fact that in Chinese
dictionaries there are no parts of speech information).
According to Liu (1992:7): Covertness is a basic
characteristics in the Chinese language. By covertness it
is meant that words or utterances do not have formal
indications of their grammatical roles.
Look at the following example:
红旗漫卷西风。
她的脸上红一阵白一阵的。
他俩从来没有红过脸。
东方红,太阳升。
落红不是无情物,化作春泥更护花 。
In the above examples, the underlined character “红”
for red is used as an adjective, a verb, and a noun in
different sentences.
(3) In terms of learning difficulties from a psychological
viewpoint, Eckman’s Markedness Differential
Hypothesis (MDH) can best explain errors in POS by
Chinese learners of English. The MDH makes three
predictions (Eckman 1977, quoted in Larsen-Freeman
& Long, 2000:102) about where errors might occur in
terms of linguistic markedness:
(a)
(b)
(c)
Those areas of the L2 which differ from the L1, and
more marked than the L1 will be difficult.
The relative degree of difficulty of the areas of the
L2 which are more marked than the L1 will
correspond to the relative degree of markedness.
Those areas of the L2 which are different from the
L1, but are not more marked than the L1, will not be
difficult.
According to James (2001:183), the MDH offers a
better explanation about the relation between
markedness and L1 transfer.
3. Conditions of Chinese L1 transfer in word class in
CLEC
While discussing lexical transfer, Odlin (1989:79) made
the observations that
(1) Lexical transfer can occur when there is no
morphological similarity between words that appear
to be semantically equivalent;
(2) Transfer can also occur when the word forms are
not similar but the meanings are.
Our findings from the CLEC show that most POS
errors caused by L1 word class features occur to words
which have similar meanings in both Chinese and
English.
Look at the following :
Lines (1) to (4) are from ST2.
(1) gh[wd2,4-0]. Fail [wd2, 0-1][vAn] is succeed’s [wd2,0-1]
(2) ause it relations [wd2,2-4] [nAv] people’s healthy
[wd2,6(3) eyes. For our safe [wd2,3-], [adjAn] people don’t play
the fi
(4) nd black face likes [wd2, s-] [prepAv] my father’s very
much. Esp
Lines (5) to (10) are from ST6.
(5) to make a complain [wd2, 1-] [vAn] on the noise that
disturb
(6) you see a terminal [wd2, 1-] [adjAadv] ill person who
sream [fm
(7) ct. to our exciting [wd2, 1-1] [adjAn] the government is
now und
(8) feminists emphasis [wd2, s-] [nAv] those too much.
They think
(9) Since the industry [wd2, 2-1] [nAadj] revolution, in
order to m
(10) n help people broad [wd2, s-] [adjAn] their minds.
Television ca
2015/7/16
27
V. Conclusions
1. Patterns of POS errors
(1) A great majority of errors in parts of speech in
Chinese learner English occur with lexical words.
(2) Advanced learners may commit more errors in lexical
words than learners at elementary levels, because they
have to deal with more demanding writing tasks that
require the use of a wider range of vocabulary.
(3) Elementary learners commit more errors in functional
words than advanced learners, because these errors
are more of a developmental nature.
2. Relation between POS error types and
learners’ proficiency levels


The frequency of errors in parts of speech does not
reflect learners’ proficiency levels. It doesn’t follow that
advanced learners commit less errors in parts of speech
in lexical words. Actually they may committed more
errors in lexical words because their written tasks were
more demanding, requiring that they make more
attempts at expressing their ideas, so they may be more
willing to take risks at words in which they have less
confidence , and they may also have coined words
resulting from their incorrect use or overgeneralization
of deviational forms.
Advanced learners commit less errors in functional
words because these words are limited in number, so as
learners’ proficiency increases, their chances of making
mistakes decreases naturally.
3. Causes of errors in parts of speech
(1) Errors in lexical words are mainly caused by transfer
of Chinese word class features.
(2) Errors in functional words are likely to be
developmental in nature. They are part of the
learning process. As learners progress along their way
to a mastery of target language, they make less and
less parts of speech errors with functional words.
4. Implications for vocabulary teaching in the classroom
(1) More attention should be paid to the teaching of lexical
words at all stages of learning because they are larger in
number, and an indicator of learner proficiency.
(2) More systematic C-E contrastive studies should be made
in the classroom to raise learners’ awareness of the
differences between forms of word classes in the two
languages.
(3) Vocabulary should always be a priority on our teaching
agenda regardless of learners’ levels, but with different
foci at different stages.
(4) The use of CLEC should be introduced in the classroom
so that the teaching of vocabulary can be more targeted
at words which are more troublesome to Chinese
learners of English.
References
[1] Ellis, R. 1994. Understanding Second Language Acquisition. Shanghai Foreign
Languages Education Press.
[2] Gui, Shichun. & Yang, Huizhong. 2002. Chinese English Learner Corpus [M].
Shanghai Foreign Languages Education Press.
[3] Gui, Shichun. 2004. A Cognitive Model of Corpus-based Analysis of Chinese
Learners’ Errors of English [J]. Modern Foreign Languages, vol.27 N0.2.
[4] Gao , Yuan. 2001. Introductory remarks [A]. In James, C. Errors in Language
Learning and Use: Exploring Error Analysis [M], F30.
[5] James, C. 2001. Errors in Language Learning and Use: Exploring Error Analysis.
Shanghai Foreign Languages Education Press.
[6] Lott, D. 1983. Analyzing and counteracting interference errors. English Language
Teaching Journal 37:256-61.
[7] Larsen Freeman, D & Long M.H. (2000). An Introduction to Second Language
Acquisition Research [M]. Shanghai Foreign Languages Education Press.
[8] Liu, Shaolong. 2006. Researching L2 Word Knowledge Development [M]. Science
Press.
[9] Liu, Miqing. 1992. C-E and E-C Contrastive Studies. Jiang Xi Education Press.
[10] Odlin, T. 1989. Language Transfer: Cross-linguistic influence in language learning.
Shanghai Foreign Languages Education Press, 2001.
[11] Ren, Xueliang. 1981. A Chinese-English Comparative Grammar. China Social
Sciences Press.
[12] Sun, Zhimian. 2001. Chinese Environment and English Learning [M]. Shanghai
Foreign Languages Education Press.
[13] Wu, Jingrong & Wang, Jianzhi. 1994. On parts of speech in English and Chinese [A].
In Li R. Collection of Essays on English and Chinese Comparative Studies [C].
Shanghai Foreign Languages Education Press.