-
31
영어교육 62권 4호 2007년 겨울
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge∗
Jaehak Chang (Kangwon National University)
Chang, Jaehak. (2007). The plausibility and generalizability of
Larsen-Freeman’s model of L2 knowledge. English Teaching, 62(4),
31-46.
This study examines the plausibility and generalizability of
Larsen-Freeman’s (1991) model of L2 knowledge, among others. The
main focus of the investigation, however, is placed on the adequacy
of the model as it relates to English relative clauses. The data
were collected from 685 students learning English in Japan, Korea,
Taiwan, and the United States and a relative clause test was
developed based on the accessibility hierarchy proposed by Keenan
and Comrie (1977). The issues addressed in this paper, thus,
include: (1) whether the same baseline model fits across different
groups (e.g., Korean vs. non-Korean; EFL vs. ESL), and, if so, to
what degree; (2) whether the constructs measured (i.e., form,
meaning, and pragmatic use of relative clauses) are similar across
different groups; and (3) the nature of the relationship between L2
grammatical and pragmatic knowledge (i.e., naturalness, formal and
informal register).
I. INTRODUCTION Different notions of what it means to know and
use a second language (L2) have
fundamental implications for L2 testing and teaching. The notion
of what constitutes L2 proficiency, for instance, provides a sound
theoretical basis for test design and development and serves as a
general framework of reference for teaching. Due to the importance
that an understanding of the nature of L2 proficiency bears on
language teaching and testing, a number of researchers have
proposed a variety of theoretical models of L2 proficiency (e.g.,
Bachman, 1990; Bachman & Palmer, 1996; Canale, 1983; Carroll,
1961; Chapelle, 1998; Harley, Cummins, Swain, & Allen, 1990;
Lado, 1961; Oller, 1979; Purpura, 2004; Spolsky, 1973). In short,
our conceptualization of this construct has evolved, and been
broadened, and
∗ This study was supported by 2007 Research Grant from Kangwon
National Univeristy.
-
Chang, Jaehak
32
reconceptualized (see Canale & Swain, 1980, for an excellent
review on earlier conceptions; also see Chalhoub-Deville, 1997;
McNamara, 1996; Shohamy, 1996, for in-depth discussions on models
of L2 proficiency that have been influential in the history of L2
assessment).
In the 1980’s, Canale and Swain’s (1980) theoretical framework
of communicative competence was considered “state-of-the-art” in
the field of applied linguistics. Rejecting Chomsky’s (1965) notion
of linguistic competence due to its disregard for the social and
functional aspects of language use, Canale and Swain (1980) not
only incorporated Hymes’ (1972) expanded notion of competence, but
also linked language teaching to testing. Among others, Hymes
(1972) argued that the social rules of language were of equal if
not of greater importance to linguistic rules in language learning.
In addition to Chomsky’s notion of linguistic competence, Hymes’
(1972) notion of communicative competence also meant knowledge of
language use and ability for use. For Hymes, language was used for
communication and as a form of social interaction.
By incorporating into their model then-current communicative
trends in L2 teaching as well as Hymes’ (1972) expanded notion of
competence, Canale and Swain (1980) were able to provide
researchers and educators with a sound theoretical framework of
communicative competence. As a result, this theoretical framework
has dominated the field for a decade (1) by serving as a working
hypothesis for later empirical studies (e.g., Allen, Cummins,
Mougeon, & Swain, 1983; Bachman & Palmer, 1982; Harley,
Allen, Cummins, & Swain, 1987; Harley, Cummins, Swain, &
Allen, 1990; Swain, 1985); and (2) by serving as a theoretical
foundation for later elaboration and refinement (e.g., Bachman,
1990; Bachman & Clark, 1987; Bachman & Palmer, 1996).
Drawing upon earlier works of communicative competence and
linguistic theory (e.g., Austin, 1962; Canale, 1983; Canale &
Swain, 1980; Halliday, 1976; Hymes, 1972; Searle, 1969; Widdowson,
1978), Bachman and Palmer’s (1996) model of L2 ability has often
been said to be the most comprehensive and inclusive model
available in the 1990’s (e.g., McNamara, 1996). However, recent
research findings have suggested that their model may still need
refinement. For instance, while Bachman and Palmer (1996) viewed
strategic competence as a set of metacognitive strategies, findings
by Purpura (1997, 1999) have suggested that cognitive strategy use
in addition to metacognitive strategy use may be the essence of
strategic competence. Moreover, while Bachman and Palmer (1996)
held a multi-componential view of metacognitive strategies, such a
componential view was not confirmed empirically (e.g., Purpura,
1997, 1999).
In addition, there have been competing views of grammatical
knowledge and unresolved issues related to it. Notions of
grammatical knowledge held by the researchers mentioned thus far
differ with respect to its level, its number of components, its
scope, and its divisibility. While Bachman and Palmer (1996) viewed
grammar as the structural aspect of language that operates at the
sentence level, a number of other researchers (e.g.,
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
33
Celce-Murcia, 1993; Celce-Murcia & Larsen-Freeman, 1999;
Givón, 1993; Halliday, 1994; Halliday & Hasan, 1985;
Larsen-Freeman, 1993; Purpura, 2004; Rea-Dickins, 1991) have argued
that grammar operates not only at the sentence level, but also at
the level of discourse. While Bachman and Palmer (1996) held a
uni-componential view of grammatical knowledge, a number of other
researchers (e.g., Larsen-Freeman, 1991; Purpura, 2004;
Rea-Dickins, 1987) have argued that grammatical knowledge may be a
multi-componential construct. Purpura (2004), for instance, states
that grammatical knowledge consists of two highly related, but
mostly separable components: grammatical form and meaning, whereas
Rea-Dickins (1987) argues that these two components may not be
distinct.
Larsen-Freeman (1991), on the other hand, contends that in
addition to knowledge of grammatical form and meaning, knowledge of
pragmatic use constitutes grammatical knowledge. She further argues
that each linguistic structure consists of those three
inter-related but (to a greater or lesser degree) distinguishable
components. She suggests that those three dimensions be taught in
order to enable L2 learners to use linguistic forms accurately,
meaningfully, and appropriately. Although Larsen-Freeman’s (1991)
notion of pedagogical grammar is intuitively appealing, a number of
other researchers (e.g., Bachman & Palmer, 1996; Canale &
Swain, 1980; Purpura, 2004) have viewed the social and functional
features of language use as pragmatics.
What these competing views of grammatical knowledge suggest is
that there may not be one correct or easy answer to what it means
to know and use an L2. If that is the case, it is possible that
Larsen-Freeman’s (1991) or Purpura’s (2004) model would work as
well, or even better than Bachman and Palmer’s (1996) or Canale and
Swain’s (1980) model. However, no studies to date have examined the
construct validity of the competing views addressed above. Nor has
the adequacy of the aforementioned models been examined and
compared empirically in a single study. Finally, the plausibility
of these models has yet to be examined as they relate to a
particular linguistic structure such as relative clauses that can
be viewed from these various perspectives.
This study examines the plausibility and generalizability of
several competing models of second language (L2) knowledge proposed
in the L2 assessment and L2 acquisition literatures. The rationale
behind examining several models of L2 knowledge is to narrow the
number of plausible models and to better understand the nature of
relationships among the theoretical constructs posited in the
models and their hypothesized components. The models under study
include Canale and Swain’s (1980) theory of communicative
competence, Larsen-Freeman’s (1991) model for teaching grammar,
Purpura’s (2004) model of grammatical ability, and Rea-Dickins’
(1987) model of grammatical ability, among others. The main focus
of the investigation, however, is placed on the adequacy of the
models as they relate to English relative clauses.
-
Chang, Jaehak
34
In this paper, I will first summarize the main findings of
single group analyses, which secure the preliminary construct
validity for the models in question in the present study. Next, I
will discuss the results of multigroup SEM for Korean and
non-Korean groups. After that, I will briefly discuss theoretical
and pedagogical implications of the findings of the present
study.
II. PRELIMINARY FINDINGS In order to examine the adequacy of the
models in question, Chang (2004) conducted a
construct validation study. In this study, the data were
collected from 685 students learning English in Japan, Korea,
Taiwan, and the United States. The participants were secondary and
college students in academic settings. All participants were 14
years of age or older with a median age of 17. The participants
represented 19 different native languages, with the majority being
Chinese, Korean, Japanese, and Spanish. They were given a 44-item
relative clause test, which consisted of a total of eight measures
(i.e., test tasks). Five of the eight tasks sought to measure the
form and meaning of relative clauses in English and the remaining
three measures concerning the socially appropriate use of relative
clauses.
Where relative clauses are concerned, the results suggested that
Larsen-Freeman’s (1991) and Purpura’s (2004) models may be better
representations of the data than Canale and Swain’s (1980) and
Rea-Dickins’ (1987) models. More specifically, the results
suggested that (1) English relative clause structure consists of
three dimensions: form, meaning, and pragmatic use (Larsen-Freeman,
1991) and (2) grammatical form and meaning may be highly related,
but this does not necessarily mean that they are identical
(Purpura, 2004). In addition, a further examination of alternative
models suggested that grammatical knowledge substantially involves
knowledge of both grammatical form and meaning, but may not
encompass knowledge of pragmatic use.
III. PRESENT STUDY
1. Research Questions The present study further examines the
degree to which the preliminary findings
described hold true for Korean and non-Korean groups. More
specifically, the present study examines factorial invariance –
whether the hypothesized patterns of relationships among latent
constructs represented in the models in question are equivalent
across Korean and non-Korean groups, and, if so, to what
degree.
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
35
In short, the current study addresses the following research
questions: 1. To what extent do the models in question represent
the data for the Korean and
non-Korean groups? 2. To what extent are the constructs measured
(i.e., form, meaning, and pragmatic use of
relative clauses) similar for the Korean and non-Korean
groups?
2. Participants In the present study, 685 students in two groups
were participated. The Korean group
consists of 215 students learning English in Korea and the
non-Korean group consists of 470 students learning English in
Japan, Taiwan or the United States. Three hundred and fifty six of
them were males, representing 52% of the population and 329
participants were females, representing 48% of the population. The
mean age was approximately 19, with the youngest being 14 and the
eldest being 43. The median age was 17. Data from these two groups
are summarized in Table 1.
TABLE 1
Descriptive Statistics and Correlation Matrix Form Meaning Use
Variable 1 2 3 4 5 6 7 8
1.SC – .51 .45 ..50 .32 .19 .15 .27 F 2. WO .54 – .43 .50 .35
.27 .14 .35
3. ER .63 .54 – .46 .38 .26 .25 .31 M 4. WM .58 .59 .65 – .43
.31 .27 .41 5. MM ..40 .39 .43 .52 – .20 ..15 .17
6. IR .25 .27 .42 .43 .28 – .29 .29 U 7. FR -.10 -.05 -.17 -.20
-.10 .15 – .21
8. NL .20 .29 .16 .28 .21 .18 .14 – Korean M 4.36 1.74 4.86 4.26
3.67 2.79 1.89 1.37
SD 1.86 1.13 3.32 2.08 1.19 1.23 1.39 .72 Non-Korean M 5.13 2.51
6.26 5.80 3.87 3.18 2.66 1.67 SD 1.45 .78 3.01 1.42 1.10 1.09 1.50
.61
The Korean group is presented below diagonal; the non-Korean
group is presented above diagonal; F = Form, M= Meaning, U = Use;
SC = Sentence completion; WO=Word order; ER=Error recognition/
correction; WM=Word meaning; MM= Morphosyntactic meaning;
IR=Informal register; FR=Formal register; NL: Natural ness.
3. Measurement Instruments and Theoretical Constructs
Table 2 presents the description of the Relative Clause Test
used in the present study and
the theoretical constructs underlying it. As shown in the table,
language knowledge, as viewed in this study, is concerned with both
grammatical knowledge and pragmatic
-
Chang, Jaehak
36
knowledge. L2 grammatical knowledge refers to the mental
representation of what learners know about the form and meaning of
expressions of their target language (Purpura, 2004). L2 pragmatic
knowledge, on the other hand, is concerned with what learners know
about social and functional uses of their target language (Bachman
& Palmer, 1996).
TABLE 2
Description of the Relative Clause Test (44 items)
Scales No Items Used PART I: Grammatical Knowledge Section A:
Grammatical Form 17 Trait Method Morphosyntactic form
with relatives MC-sentence completion 7 1 – 7
Morphosyntactic form with relatives
MC-word order 3 8 – 10
Morphosyntactic form with relatives
Grammaticality judgment (Error recognition/correction)
7 (11–17) +
Section B: Grammatical Meaning 12 Trait Method Word meaning with
relatives Fill-in-the-blanks
(Word list provided) 7 18 – 24
Morphosyntactic meaning with relatives
Semantic interpretation (True/false)
5 25 – 29
Scales No. Items Used PART II: Pragmatic Knowledge Section C:
Register Awareness (i.e., appropriateness) 12 Trait Method Informal
register
with relatives Appropriateness judgment
(6-point scale) 6 31, 32, 35,
37, 38, 39 Formal register
with relatives Appropriateness judgment
(6-point scale) 6 30, 33, 34,
36, 40, 41 Section D: Naturalness 3 Trait Method Sensitivity to
naturalness
with relatives Naturalness judgment
(More natural/less natural)3 42, 43, 44
MC = Multiple-Choice; Relatives = Relative clauses + The
parenthesis indicates the items that are scored for the two
criteria: error recognition and error correction.
L2 grammatical knowledge, as viewed in this study, is concerned
with both form and
meaning. Grammatical form is concerned with the correct
formation of words, phrases, and sentences, including phonological
and morphosyntactic form. Therefore, Section A of
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
37
the test aimed at measuring grammatical from and consisted of
three parts: sentence completion, word order, and error
recognition/correction and contained 17 items. Five of them were
limited-production items. These tasks were designed to measure the
test-takers’ ability to recognize or produce the correct formation
of relative clauses. Example items for each of the three tasks are
presented in Table 3.
TABLE 3
Example Items for Grammatical Form Circle the correct
answer.
Morphosyntactic form (Sentence completion)
A: I think we should get together more often. B: That’s exactly
_____ in mind.
(a) I had (b) I had it
(c) what I had (d) what I had it
Morphosyntactic form (Word order)
A: What do you think of Jack? B: Well, I wouldn’t trust _____.
He has such a big mouth.
(a) every word he says
(b) he says every word (c) every word says he (d) word every he
says
Decide if the underlined part of the sentence is correct or
incorrect. If you think the
underlined part is correct, circle C (Correct) and go on to the
next item. If you think it is incorrect, circle I (Incorrect) and
correct it by changing, adding, or deleting only ONE of its
elements. You can use the space provided if needed.
Morphosyntactic form (Error recognition / correction)
A: Do you have a pen I can borrow it? C I B: Here it is.
Grammatical meaning, on the other hand, refers to (1) the
literal meaning expressed by words, phrases, and sentences, and (2)
the meaning encoded in grammatical structures (Larsen-Freeman,
1991; Purpura, 2004), including word meaning and morphosyntactic
meaning. Thus, morphosyntactic meaning, as viewed in this study,
was defined as the context-independent, literal meaning of what is
said. In this sense, what is referred to as
-
Chang, Jaehak
38
morphosyntactic meaning in the present study is similar to what
Purpura (2004) terms as grammatical meaning in his model.
Accordingly, Section B was designed to measure test-takers’ ability
to recognize the lexical meanings of relative pronouns and adverbs,
and the morphosyntactic meanings (i.e., literal meanings) of
relative clauses. This section consisted of two parts:
fill-in-the-blanks and semantic interpretation (true/false tasks)
and contained 12 selected-response items. Example items for both of
the tasks used in this section are presented in Table 4.
TABLE 4
Example Items for Grammatical Meaning Use the words listed
below. Complete the sentences. You can use the same word as many
times as needed.
Word meaning (Fill-in-the-blanks)
what, who, which, how, why, where, when
A: Wow. You did it so fast. I’m not sure I can do it myself. B:
This time I’m going to show you _____ I did it step by step,
Okay?
Read the first sentence. Then, decide if the second sentence is
true or false. Circle T or F.
Morphosyntactic meaning (Semantic Interpretation:
True/false)
The girl who speaks Greek is my cousin. My cousin speaks Greek.
T F
T = True F = False
In the word meaning part of Section B, the relative pronouns and
adverbs in the box
were given to test-takers. They were then asked to fill in the
blanks using one of the given words. In the morphosyntactic meaning
part of Section B, the test-takers were asked to decide if the
second statement was true or false given the first statement.
L2 pragmatic knowledge as operationalized in this study, on the
other hand, is concerned with sensitivity to register and
naturalness (Bachman & Palmer, 1996). Register, as viewed in
this study refers to linguistic choices we make that vary depending
on context of use, including formal and informal registers.
Section C was thus designed to measure test-takers’ knowledge of
when to use a particular linguistic form over another in a
particular context. This section contained 12 items and some
example items are presented in Table 5.
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
39
TABLE 5 Example Items for Register Awareness
Register awareness (Appropriateness judgment) Rate each item in
terms of what you would normally say in the
following situation, where zero means that you would be very
unlikely to use the expression in the given situation, and five
means that you would be very likely to use it. Note that all the
options in each dialogue may be appropriate to use under certain
situations. However, some expressions may be more appropriate to
use for informal conversations between friends.
Dialogue 5 (At the party, you are introducing Sam to your
friend.)
You: Hey, Joe. This is the guy ____________________. Joe: Sam?
You: That’s right. Joe, this is Sam. Sam, this is Joe.
very unlikely very likely 30. whom I was talking about 0 1 2 3 4
5 31. I was talking about 0 1 2 3 4 5 32. about whom I was talking
0 1 2 3 4 5 33. who I was talking about 0 1 2 3 4 5
In this section, the test-takers were asked to rate each item in
terms of what they would say normally, where zero meant that they
would be very unlikely to use the expression in the given
situation, and five meant that they would be very likely to use it.
The test-takers were also informed that (1) all options might be
appropriate to use in certain situations, and (2) some expressions
might be more appropriate for use in informal conversations between
friends.
Finally, Section D was designed to measure the sensitivity to
naturalness.. Naturalness (and its lack thereof), as viewed in this
study, refers to how similar a given utterance is to what a native
speaker would normally say. In this section, the test-takers were
given a pair of expressions. They were then asked to circle the
expression that they thought sounded more natural to a native
speaker. This section contained three items and an example item is
presented in Table 6.
-
Chang, Jaehak
40
TABLE 6 An Example Item for Naturalness
Naturalness (Naturalness Judgment) Read the dialogue. Of each
pair of expressions, circle the expression that would sound
more natural to native speakers. Dialogue 4 (You and your friend
Khan are getting near your house)
42-44. You: How much farther, Khan? Khan: We’re almost there.
Can you see the house ________ over there? You: Yeah. Is that your
house?
42. (a) with the red roof (b) whose roof is red
IV. ANALYSES AND FINDINGS In the present study, to establish a
baseline model that fit both Korean and non-Korean
groups, Larsen-Freeman’s (1991) model was hypothesized to
represent the relative clause data. Larsen-Freeman’s (1991) model,
as shown in Figure 1, was represented by the eight measured
variables concerning the form, meaning, and pragmatic use of
relative clauses. Specifically, this model postulated that: (1)
knowledge of grammatical form accounts for the relationship among
the three observed variables concerning the correct formation of
relative clauses; (2) knowledge of semantic meaning accounts for
the relationship between the two observed variables concerning the
meaning of relative clauses in expressions; and (3) knowledge of
pragmatic use accounts for the relationship among the three
observed variables concerning the socially appropriate use of
relative clauses.
Larsen-Freeman’s (1991) model was first examined separately for
each group and then simultaneously for both groups. As summarized
in Table 7, the results of these single-sample and multigroup
analyses suggested that Larsen-Freeman’s (1991) model represented
the data for each group reasonably well. Larsen-Freeman’s (1991)
model explained the data somewhat better for the Korean group than
for the non-Korean group. The overall fit of Larsen-Freeman’s
(1991) model evaluated across both samples was also generally
acceptable. Specifically, Larsen-Freeman’s (1991) model produced a
Satora-Bentler chi-square (S-Bχ2) statistic of 61.0756 with 33
degrees of freedom and a CFI of 0.976. This χ2/df ratio for the
model was within the recommended value of 3. A review of other fit
indices revealed that values for both the NFI and NNFI were above
0.95 and a value for the RMSEA was within the acceptable limit.
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
41
In short, the overall results supported the hypothesis that
Larsen-Freeman’s (1991) model serves as a baseline model that
applies equally well to both groups.
FIGURE 1
Simultaneous Multigroup CFA Results for Larsen-Freeman’ s (1991)
Model with respect to Relative Clauses and with Equality
Constraints on Factor Loadings
and with Standardized Parameter Estimates
TABLE 7 Goodness of Fit Summary
Goodness of fit summary Fit Indices Model χ2 S-Bχ2 df CFI NNFI
RMSEA Single-sample analyses Korean group only 26.283 24.9159 16
.983 .971 .051(.000
-
Chang, Jaehak
42
Given these findings, the degree to which the constructs
measured (i.e., form, meaning, and pragmatic use of relative
clauses) were similar for both groups was examined. To this end,
Larsen-Freeman’s (1991) model was examined again but with
cross-group equality constraints on the factor loadings not fixed
at 1.0 for model identification (five in total: two on the
grammatical form factor; one on the semantic meaning factor; two on
the pragmatic use factor). As summarized in Table 7, results of
this analysis suggested that the naturalness judgment task appeared
to measure knowledge of pragmatic use differentially across the two
groups. However, the results showed that the tasks of the relative
clause test largely measured three hypothesized constructs (i.e.,
form, meaning and pragmatic use of relative clauses) in comparable
ways for Korean students learning English in Korea and other ESL
and EFL learners. In sum, the results provided support for the
validity of inferences about knowledge of grammatical form and
meaning and knowledge of pragmatic use for these two groups of
students, based on these tasks.
V. CONCLUSION This study, utilizing structural equation
modeling, examined the plausibility of several
models of L2 knowledge proposed in the L2 assessment and L2
acquisition literatures. The models examined in the present study
had similarities and differences. What was universal in these
models was that all of them include knowledge of grammatical form,
semantic meaning, and pragmatic use. These models, however, differ
in terms of the number of components and the scope of grammatical
knowledge as it relates to relative clauses. For instance, one
group of models (i.e., Canale & Swain, 1980; Rea-Dickins, 1987)
predicts the eight measured variables concerning the form, meaning,
and use of relative clauses to display two distinctive patterns of
relationships (e.g., grammatical knowledge and sociolinguistic
knowledge), whereas another group of models (i.e.,
Larsen-Freeman,1991; Purpura, 2004) expects the emergence of three
distinctive patterns of relationships. In addition,
Larsen-Freeman’s (1991) model postulates that all the three
hypothesized components (i.e., grammatical form, semantic meaning,
and pragmatic use) constitute grammatical knowledge, whereas
Purpura’s (2004) model posits that only the first two components
(i.e., grammatical form and meaning) are part of grammatical
knowledge and that pragmatic use belongs to a separate area of
language knowledge (i.e., pragmatic knowledge).
The results of the present study, however, suggested that
English relative clause structure may consist of three identifiable
dimensions: grammatical form, semantic meaning, and pragmatic use.
Similar results were also obtained for the Korean and non-Korean
groups. Given these findings, the present study provided
researchers with a sound theoretical basis for assessing the
constructs of knowledge of grammatical form and meaning and of
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
43
pragmatic use with respect to relative clauses. In addition, the
present study also supplied valuable information about the degree
to which Larsen-Freeman’s (1991) model could be generalizable for
the Korean and non-Korean groups.
We do not know, however, how generalizable these findings are
with respect to linguistic structures other than relative clauses.
If linguistic structures other than relative clauses (e.g.,
question formation, negation, or modals) are selected for construct
validation, can grammatical form and grammatical meaning be
empirically separable? If the plausibility of the models in
question is examined as they relate to English modals, for example,
will the results reveal that L2 knowledge as measured by an English
Modals Test consists of three empirically identifiable dimensions
(i.e., form, meaning, and pragmatic use)? Subsequent validation
research of the models in question with respect to linguistic
structures other than relative clauses will thus give valuable
insights on the nature of L2 grammatical knowledge.
Last but not least, the present study also provided
pedagogically useful information about how well L2 learners know
about the form, meaning, and use of relative clauses. Korean
students learning English in Korea (i.e., the Korean group), for
instance, were not fully aware that not all utterances that are
grammatically correct are appropriate for use in informal talks
between close friends. More specifically, Korean students who had
high scores on the tasks on the correct formation of relative
clauses tended to have low scores on the tasks on the socially
appropriate use of relative clauses. Some of them even found formal
utterances (e.g., this is a guy about whom I was talking) to be
more appropriate for use in casual talks between close friends than
informal utterances (e.g., this is a guy I was talking about). By
contrast, such a negative relationship between grammatical accuracy
and social appropriateness was not observed in the non-Korean group
(i.e., students learning English in Japan, Taiwan and the United
States). Given these findings, the differing degrees of formality
of relative clauses in expressions should perhaps be one of
instructional priorities for the Korean group of students in order
to enable them to use this linguistic structure accurately,
meaningfully, and appropriately.
REFERENCES
Allen, P., Cummins, J., Mougeon, R., & Swain, M. (1983).
Development of bilingual proficiency: Second year report. Toronto:
The Ontario Institute for Studies in Education.
Austin, J. (1962). How to do things with words. Oxford:
Clarendon Press. Bachman, L. (1990). Fundamental considerations in
language testing. Oxford: Oxford
University Press. Bachman, L., & Clark, J. (1987). The
measurement of foreign/second language proficiency.
Annals of the American Academy of Political and Social Science,
480, 20-33.
-
Chang, Jaehak
44
Bachman, L. & Palmer, A. (1982). The construct validation of
some components of communicative proficiency. TESOL Quarterly, 16,
449-65.
Bachman, L., & Palmer, A. (1996). Language testing in
practice. Oxford: Oxford Univesity Press.
Canale, M. (1983). On some dimensions of language proficiency.
In J. Oller (Ed.), Issues in language testing research (pp.
333-342). Rowley, MA: Newbury House.
Canale, M., & Swain, M. (1980). Theoretical bases of
communicative approaches to second language teaching and testing.
Applied Linguistics, 1, 1-47.
Carroll, J. (1961). Fundamental considerations in testing for
English language proficiency of foreign students. In Center for
Applied Linguistics (Ed.), Testing the English proficiency of
foreign students (pp. 31-40). Washington, DC: Center for Applied
Linguistics.
Celce-Murcia, M. (1993). A nonhierarchical relationship between
grammar and communication Part II: Insights from discourse
analysis. In J. Alatis (Ed.), Georgetown University Round Table on
Languages and Linguistics 1992: Language, communication, and social
meaning (pp. 166-178). Washington, DC: Georgetown University
Press.
Celce-Murcia, M., & Larsen-Freeman, D. (1999). The grammar
book. Boston: Hinele & Hinele.
Chalhoub-Deville, M. (1997). Theoretical models, assessment
frameworks and test construction. Language Testing, 14, 3-22.
Chang, J. (2004). Examining models of second language knowledge
with specific reference to relative clauses: A model-comparison
approach. Unpublished dissertation. Teachers College, Columbia
University.
Chapelle, C. (1998). Construct definition and validity inquiry
in SLA research. In, L. Bachman & A. Cohen (Eds.), Interfaces
between second language acquisition and language testing research
(pp. 32-70). Cambridge: Cambridge University Press.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge,
MA: MIT Press. Givón, T. (1993). English grammar: A function-based
introduction II. Philadelphia: John
Benjamins. Halliday, M. (1976). The form of a functional
grammar. In G. Kress (Ed.), Halliday:
System and Function in Language. Oxford: Oxford University
Press. Halliday, M. (1994). An introduction to functional grammar.
London: Edward Arnold. Halliday, M. & Hasan, R. (1985).
Language, context, and text: Aspects of language in a
socio-semiotic perspective. Oxford: Oxford University Press.
Harley, B., Allen, J., Cummins, J., & Swain, M. (1987). The
development of second
language proficiency: Final report. Toronto: Modern Language
Centre. Harley, B., Cummins, J., Swain, M., & Allen, P. (1990).
The nature of language proficiency.
In B. Harley, J. Cummins, M. Swain, & P. Allen (Eds.), The
development of second language proficiency (pp. 7-25). Cambridge:
Cambridge University Press.
-
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 Knowledge
45
Hymes, D. (1972). On communicative competence. In J. Pride &
J. Holmes (Eds.), Sociolinguistics (pp. 269-293). Harmondsworth,
UK: Penguin.
Keenan, E., & Comrie, B. (1977). Noun phrase accessibility
and universal grammar. Linguistic Inquiry, 9, 63-99.
Keenan, E., & Comrie, B. (1979). Data on the noun phrase
accessibility hierarchy. Language Learning, 55, 333-351.
Lado, R. (1961). Language testing. New York: McGraw-Hill.
Larsen-Freeman, D. (1991). Teaching grammar. In M. Celce-Murcia
(Ed.), Teaching
English as a second or foreign language (pp. 279-296). Boston:
Heinle & Heinle. Larsen-Freeman, D. (1993). A nonhierarchical
relationship between grammar and
communication Part II: Theoretical and methodological
considerations. In J. Alatis (Ed.), Georgetown University Round
Table on Languages and Linguistics 1992: Language, communication,
and social meaning (pp. 155-165). Washington, DC: Georgetown
University Press.
McNamara, T. (1996). Measuring second language performance.
London: Longman. Oller, J. (1979). Language tests at school: A
pragmatic approach. London: Longman. Purpura, J. (1997). An
analysis of the relationships between test-takers’ cognitive
and
metacognitive strategy use and second language performance.
Language Learning, 47, 289-325.
Purpura, J. (1999). Learner strategy use and second language
test performance: A structural equation modeling approach.
Cambridge: Cambridge University Press.
Purpura, J. (2004). Assessing grammar. Cambridge: Cambridge
University Press. Rea-Dickins, P. (1987). The relationship between
grammatical abilities and aspects of
communicative competence: With specific reference to the testing
of grammar. Unpublished dissertation. University of Lancaster.
Rea-Dickins, P. (1991). What makes a grammar test communicative?
In J. Alderson & B. North (Eds.), Language testing in the
1990s: The communicative legacy (pp. 112-131). London:
MacMillan.
Searle, J. (1969). Speech Acts: An essays in the philosophy of
language. Cambridge: Cambridge University Press.
Shohamy, E. (1996). Competence and performance in language
testing. In G. Brown, K. Malmkjaer & J. Williams (Eds.),
Performance and competence in second language acquisition (pp.
138-151). Cambridge: Cambridge University Press.
Spolsky, B. (1973). What does it mean to know a language; or how
do you get someone to perform his competence? In J. Oller & C.
Richards (Eds.), Focus on the learner: Pragmatic perspectives for
the language teacher (pp. 164-176). Rowley, MA: Newbury House.
Swain, M. (1985). Communicative competence: Some roles of
comprehensible input and comprehensible output in its development.
In S. Susan & C. Madden (Eds.), Input
-
Chang, Jaehak
46
in second language acquisition (pp. 235-253). Rowley, MA:
Newbury House. Widdowson, H. (1978). Teaching language as
communication. Oxford: Oxford University
Press. Applicable levels: ESL and EFL learners at the secondary
and college level Key words: L2 proficiency, relative clause,
construct validation Jaehak Chang The Department of English
Education, Kangwon National University Kangwon National University
Road 1, Chunchon, Kangwon-Do, 200-701, Korea Email:
[email protected] Received in August, 2007 Reviewed in
September, 2007 Revised version received in November, 2007
The Plausibility and Generalizability of Larsen-Freeman’s Model
of L2 KnowledgeⅠ. INTRODUCTIONⅡ. PRELIMINARY FINDINGSⅢ. PRESENT
STUDYⅣ. ANALYSES AND FINDINGSⅤ. CONCLUSIONREFERENCES