Vol. 4 Issue 6 – Research Paper: ‘Corpus-Based Approach Toward Teaching Collocation of Synonyms’ by Aliakbar Jafarpour, Mahmood Hashemian & Sepideh Alipour

Aliakbar Jafarpour, Mahmood Hashemian & Sepideh Alipour work with Sharekord University.

Abstract

The study was aimed at providing more insights on the application of the corpus-based approach. The objectives of the study were to compare its learning effects with the conventional teaching method’s effects on collocations of synonyms. The study was conducted with two groups of L2 learners. One group was randomly assigned to be the experimental group studying with the corpus-based approach whereas the other represented the comparison group studying with the conventional teaching method. The learners from both groups were matched in pairs according to their language proficiency and collocation knowledge on the pretest. During the study, the experimental group was trained through paper-based and hands-on activities to deal with the concordance information in the corpus whereas the comparison group was taught collocations of synonyms through conventional activities such as explicit teaching. The instruments for collecting data included the pretest, posttest, prewriting and postwriting.

Key words: collocation; synonym; corpus-based approach; conventional approach; writing

Introduction

Collocations are expressions consisting of two or more words that correspond to some conventional ways of saying things. In the words of Firth (1957, p. 181), “Collocations of a given word are statements of the habitual or customary places of that word.”

Since the advent of computerized corpora in the 1960s, research within corpus linguistics has demonstrated its potential not only for lexicography (Sinclair, 1987) and L2 research (Carter & McCarthy, 1999), but also as a resource in L2 teaching (Lewis, 1997). Concordancing software, or concordance, provides an easy and yet powerful means to study the multiple meanings and functions of a given word, compare usages and distribution of two or more words that are synonyms, and analyze vocabulary choices (Chan, 2002).

The ability to write effectively is becoming more and more important, and writing instruction is assuming an increasing role in L2 education (Weigle, 2002). However, writing is not a simple process, even in L1; it is very challenging for almost all L2 learners. Writing in English has always been a demanding task facing L2 learners. Even a skilled writer, who writes often does not necessarily find the writing process easy.

As far as the English writing performance of L2 learners is concerned, one of the salient problems of these learners is the application of collocations (Brashi, 2005). In writing, the use of right words is needed to make people understand the messages precisely. Erroneous utterances such as powerful tea are due to lack of collocational knowledge, not poor lexical or grammatical knowledge, as Crystal (1992) states that “collocations provide a major difficulty in mastering foreign languages” (p. 105). Zarei and Koosha (2002) found that Iranian L2 learners have problems with the production of English collocations.

The other problem that L2 learners encounter in their writings is the problem of synonyms, especially collocations of synonyms. Synonyms are lexical pairs, or sets, that have very similar cognitive or denotational meanings, but they usually vary in terms of collocations or register.  Partington (1998) says that synonyms can be distinguished in terms of their typical collocates. For example strong and powerful, tea can only be strong, but not powerful (Zarei & Gholami, 2007).

For these reasons, this study was an attempt to investigate L2 learners’ knowledge of English collocations, more specifically to examine the performance of advanced L2 learners with regard to the production and reception of collocations of synonyms.

The purpose of the present study was to investigate the comparative effects of the corpus-based method and the conventional approach. In line with what was said, the present study was an attempt to seek answers for the following questions:

1. Is there any difference between the effects of the concordance-based method and the conventional teaching method on L2 learners’ performance in collocation tests of synonymous words?

2. Can corpus-based approach improve L2 learners’ use of synonymous words and their collocations in writing?

Literature Review

Corpora and Language Teaching and Learning

The existence of technology has caused innovative changes in an L2. Corpora have changed views on language teaching and learning. The direct or data-driven application of corpora in the classroom implies that learners be exposed to authentic corpus material and be encouraged to discover things about language without any previous preconception about what they will find (Johns, 1991; Bernardini, 2002). The corpora can for instance be used to provide concordances or to select examples for learning activities. According to Granath (2000, cited in Bernardini, 2004) corpora can be an integral part of courses in grammar and in spoken and written proficiency and it can be used to create exercises, and to discuss near synonyms, and collocations. Penington (1996) argues that corpora increase the variety and diversity of learning opportunities. Tribble and Jones (1990) argue that concordances present language in a way that enables learners to discover word meaning and new knowledge about language for themselves.

Research on Concordancing, Collocation, and L2 Writing

Hsu (2007) investigated the effect of collocations and fluency in writing. He conducted an empirical study about the impact of lexical collocations on the writing of Taiwanese college English majors and non-English majors. He found a relation between collocations and writing fluency.

Another study conducted by Anthony (2006) investigated the role of concordancing in technical writing classrooms. The aim was for the learners to investigate for themselves the way language is used in target contexts.

Teaching collocations as claimed by Cowie (1981) is the step towards fluent writing. Another feature of advanced writing is the use of collocations, which can produce a highly influential writing. Mis-collocations are widely spread in foreign learners’ writing. This problem has to be solved by exposing the learner to highlighted word combinations, and raising his/ her attention to them.

Regarding the importance of writing, the role of collocation knowledge in writing, and the advent of concordances there are few studies on the application of concordances in L2 writing. The studies introduced below are among the few studies on concordancing, collocations, and L2 writing.

Yoon (2008) conducted a study on six L2 learners to determine the effect of corpus on L2 learners’ writings. The result reported was that corpus increased awareness of the importance of collocations among L2 learners and caused them to pay more attention to collocations in their writings. The other benefit of the study was that it helped learners solve their writing problems.

Synonyms

Despite their important role of synonyms, few studies have been concerned with synonyms, collocations, and L2 learning. L2 learners need to distinguish between words with similar meanings that are used in different contexts. L2 learners should find differences among synonyms in terms of frequency, their application in different registers, and collocations. Near-synonyms cannot substitute for the components of a collocation.

Yeh et al. (2007) exploited concordances and collocations to advise L2 learners in lexical choice. Through introducing the lexical semantics of synonyms they tried to reduce L2 learners’ confusion of collocations of synonyms. They said that collocation profile provides evidence for L2 learners in comparing and discriminating among synonyms.

Method

In this study, teaching collocations of synonyms and its effect on L2 learners’ production and test performance was investigated in terms of two approaches: 1) conventional approach and 2) corpus-based approach.

In the first stage, 90 male and female L2 learners, aged 20-23, took the Michigan Test of English Language Proficiency (MTELP). After that, to homogenize the participants in terms of collocational proficiency, they took a test of English collocations. The test was designed by the researcher and piloted for its reliability and validity. The researcher selected those whose scores were above the mean score of the collocation test. Among these 9o students six were excluded from the study as they were not comparable with others on the basis of their mean scores. Then at the last stage they were randomly divided into two groups: experimental and control.

After division the participants were asked to write on their favorite topic which served as their pre-writing. The purpose was to prepare materials for the training sessions total of 90 paragraphs were collected and their collocational errors of synonyms were spotted out. These errors were used as a guideline for selecting materials, tests, and tasks. This writing also served as the pretest for the second hypothesis.

Then a completion test on collocations of synonyms was given to the participants to determine their knowledge of collocations of synonyms. This test was given as the pretest and comprised 50 items with its reliability estimated to be 0.89 using Kr-21 formula. After that, students in both groups went under treatment.

After treatment the posttest was administered to determine the effect of instructions. Pre-test and post-test were the same in number and content, except that the items of the post-test were placed differently. Regarding the second hypothesis, the effect of instructions on the correct application of collocations in participants’ writings, the study entailed post writing after treatment. The experimental group was introduced collocations of synonyms through corpus-based approach. In this approach erroneous applications of synonyms made in participants’ writings along with their collocational patterns were taken from BNC and given to them in printouts for further study. The aim was for the students to explore language patterns.

But the control group was presented with conventional teaching materials. In other words they were taught collocations explicitly. It consisted of a concrete definition of collocations, examples of different types of exercises, along with follow up exercises. The exercises consisted of traditional exercises usually found in course books such as matching words with their collocations, multiple choice exercises that asked the participants to choose the correct collocation for each word. The materials for this group were taken from dictionary of collocations.

After the treatment, the posttest was given to compare their performance and again they were asked to write on the same topic as the pretest.

Result and Discussion

At first an independent t test was performed on the pretest scores of the participants to see whether they had any differences at the beginning of the study or not. The participants’ responses were counted as correct if they chose the correct choice. The result of t test showed that there were not significant differences between the two groups at the beginning of the study. What follows in Table 4.1 is the presentation and discussion of this result:

Table 4.1

Independent Samples Test
		Levene’s Test for Equality of Variances		t-test for Equality of Means
									95% Confidence Interval of the Difference
		F	Sig.	t	df	Sig. (2-tailed)	Mean Difference	Std. Error Difference	Lower	Upper
pre_ex_con	Equal variances assumed	.000	.990	.222	82	.825	.09524	.42957	-.75931	.94979
	Equal variances not assumed			.222	80.444	.825	.09524	.42957	-.75956	.95004

Table 4.1: result of pretest scores of both experimental and control groups

As it is evident from the table, the sig of the study is .990, meaning that the result of this test is not significant and there is no difference between the learners at the beginning of the study.

In the second phase, to see whether the learners progressed toward the end of the study or not, one  paired-sample t test was run on the pretest and posttest scores of the experimental group and one on the scores of the control group. The results are provided in Table 4.2 and Table 4.3

Table 4.2

Paired Samples Test
		Paired Differences					t	df	Sig	Sig. (2-tailed)
		Mean	Std. Deviation	Std. Error Mean	95% Confidence Interval of the Difference
					Lower	Upper
Pair 1exp_pre_posttest		-1.59524	1.96381	.30302	-2.20721	-.98327	-5.264	41	.000	.000

Table 4.2: the result of t test on the pretest and posttest scores of the experimental group

As the significance of the study shows the study is significant at p<.05, so the experimental group progressed toward the end of the study.

Table 4.3

Paired Samples Test
		Paired Differences					t	df	Sig	Sig. (2-tailed)
		Mean	Std. Deviation	Std. Error Mean	95% Confidence Interval of the Difference
					Lower	Upper
Pair 1	CON – posttest_Con	-.38095	1.71012	.26388	-.91386	.15196	-1.444	41	.156	.156

Table 4.3: the result of t test on the pretest and posttest scores of the control group

The result of t test is not significant at this part, meaning that the control group did not progress toward the end of the study and the teaching of collocations through traditional approaches to control group is not effective.

As mentioned in the methodology section, the first research question focused on comparing the learning effects of the application of concordance-based method and the traditional teaching method on the comprehension of the learners. It was hypothesized that there will be no significant differences between the participants’ scores in experimental and control group on the measures of their collocation comprehension. To test this hypothesis the data were collected from one measure of collocational knowledge in each group. T test with the (p=.o5) was performed to examine the learning effects of the corpus-based method and the conventional teaching method on collocation learning. Table 4.3 clearly shows the result of t test carried out to show the direction of hypothesis.

Table 4.4

Independent Samples Test
		Levene’s Test for Equality of Variances		t-test for Equality of Means
									95% Confidence Interval of the Difference
		F	Sig.	t	df	Sig. (2-tailed)	Mean Difference	Std. Error Difference	Lower	Upper
post_ex_con	Equal variances assumed	.011	.040	3.561	82	.001	1.30952	.36775	.57796	2.04109
	Equal variances not assumed			3.561	81.742	.001	1.30952	.36775	.57792	2.04113

Table 4.4: result of posttes scores of both experimental and control groups

On testing this question, it was found that the p- value was significant at p<.05, indicating that the two methods had different learning effects on learners’ learning. The results of t test shows that the participants performed significantly better when they were taught through concordances than the time they were taught without concordances. There was a significant improvement of the experimental group from the mean score of 9 in the pretest to 12 in the posttest.

In a selection of prewritings collected for data analysis a total of 343 collocation errors of synonyms were collected and it was found that 180 (52.47%) were incorrect in the prewriting of the experimental group and 163 (47.53%) were incorrect in the prewriting of the control group. The data were analyzed against concordancing.

	Number	percentage
Experimental group	163	47.53
Control group	180	52.47

Table 4.4: collocation errors of synonyms in learners’ prewritings

From the prewriting it is concluded that there was a little difference between the experimental and control group in number of collocation errors of synonyms.

After treatment their postwritings were analyzed using concordance.  A total number of 139 collocation errors of synonyms were collected and it was found that the number of collocation errors of synonyms in the postwriting of the experimental group was 54(38.84 %), and that of the control group was 85 (61.16%). Results are provided in the Table below.

	Number	percentage
Experimental group	54	38.84
Control group	85	61.16

Table 4.5: collocation errors of synonyms in learners’ postwritings

As the Tables above illustrate, the comparison of the learners’ prewriting and postwriting shows that once learners learned the process they could use collocations in the correct way. Results show that the total number of collocational mistakes in post writings of the experimental group was less than those of that in prewriting. Upon studying the essays the researcher found that the learners made fewer collocation errors of synonyms when used the corpus for instruction.

Conclusion

The main concern of the first research question was a comparison of the effects of corpus-based approach and traditional approach on learners’ collocation competence. Regarding this question it was hypothesized that there will be no significant differences between the two methods. It was found from the study that the corpus-based approach was more effective than the traditional approach in increasing learners’ collocation knowledge. As is evidenced from their scores, the average score of the experimental group (M= 12) is higher than that of the control group (M= 11) in the posttest, although there was no significant difference between their pretest scores (mean score of the experimental group in pretest= 9, control group= 10). The study supports previous studies that have found the advantage of corpus over traditional methods (Jafarpour & Koosha, 2006). The relatively high and positive correlation between the scores on the collocation test of the experimental group and the corpus-based approach suggests that learners’ collocational knowledge and the way collocations are learned are closely and positively related. In terms of teaching collocations, both group improved, though the improvement of the experimental group was more remarkable than the improvement of the control group.

Regarding the second research question, the corpuse had more significant impact on the overall performance of learners in postwriting in experimental group than the control group. With the corpus-based method L2 learners had access to the corpus and they could find their mistakes from the corpus lines. The reason is that these lines make collocations more apparent to learners, and it increases collocation retention. As Cobb (1999) stated, the data that is searched by the learners is more useful than the data provided to them by the teacher.

All in all, the results of the study show that corpus-based approach is a valid approach that plays a positive role in the students’ learning and writing processes and improves the quality of their writing. Potential differences in learners performance on collocation tests and their improvement in writing task is attributed to the corpus-based approach. Writing proficiency of learners improved as they had to discover the collocations of the words intended.

References

Anthony, L. (2006). Developing a freeware, multiplatform, corpus analysis toolkitfor the technical writing classroom. Ieee Transactions on Professional Communication, 49(3).

Bernardini, S. (2002). Exploring new directions for discovery learning. In B. Kettemannann and G. Marko (Eds.), Teaching and learning by doing corpus analysis (pp. 165-182). Amsterdam:Rodopi

Brashi, A. (2006). Collocability as a problem in L2 production. Reflection on English Language Teaching, 8(1), 21-34.

Carter, R., & McCarthy, M. (1988). Vocabulary and language teaching. London: Longman

Celce-Murcia, M. (2006). Teaching English as a second or foreign language.

Chan, M. K. M. (2002). Concordancers and concordances: Tools for Chinese language teaching and research. Journal of the Chinese Language Teachers Association, 37(2), 1-58.

Cobb, T. (1999). Breadth and depth of lexical acquisition with hands-on concordancing. Computer Assisted Language Learning, 12, 345-360.

Cowie, A. (1981). The treatment of collocations and idioms in learners’ dictionaries. Applied Linguistics, 2, 223-235.

Crystal, D. (1992). The Cambridge encyclopedia of language. Cambridge University Press.

Firth, J. R. (1957). A Synopsis of Linguistic Theory. In F. R. Palmer (Eds.) selected papers of J. R. Firth. London: Longmans.

Hsu, J. (2007). Lexical collocation and their relation to the online writing of Taiwanese of college English majors and non-English majors. Electronic Journal of Foreign Language Teaching, 4(2), 192-209.

Jafarpour, A., & Koosha, M. (2006). Data- driven learning and teaching collocation of prepositions: the case study of Iranian EFL adult learners. Asian EFL Journal, 8(4), 192-209.

Johns, T. (1991). Should you be persuaded? Two samples of data-driven learning materials. English Language Research Journal, 4, 1-16.

Lewis, M. (1997). Implementing the lexical approach: Putting theory into practice. Hove: Language Teaching Publications.

Parastuti, A. (2005). The negative transfer of English collocations written by the students of Gunadarma University.

Partington, A. (1998). Patterns and meaning: Using corpora for English language research and teaching. Philadelphia, PA: John Benjamins.

Tribble, C., & Jones, C. (1990). Concordancing in the classroom: A Resource Book for Teachers. London: Longman.

Weigle, S, C. (2002). Assessing Writing. Cambridge: Cambridge University Press.

Yeh, Y., Liou, H., & Li, Y. (2007). Online synonym material and concordancing for EFL college writing. Computer Assisted Language Learning, 20(2), 131-152.

Yoon, H. (2008). More than a linguistic reference: The influence of corpus technology on L2 academic writing. Language Learning & Technology 12(2): 31-48.

Zarei, A. A., & Golami, V. (2007). The effect of partial synonymy on L2 vocabulary learning. Iranian Journal of Language Studies, 1(2), 73-88.

Zarei, A, A., & Koosha, M. (2002). Patterns of Iranian advanced learners’ problems with English collocations. Iranian Journal of Applied Linguistics, 6(1), 137-169.