Editorial
The article first appeared in a special issue of the ETAI Forum 2013, contributed to the 20th Anniversary of the Lexical.

Collocation Use in Writing Among Israeli Learners of English

Tina Waldman, Israel

Tina Waldman received her Ph.D. in Applied Linguistics from the University of Haifa, Israel. Her research interests are vocabulary and formulaic language, applied corpus linguistics and writing in an additional language - the subjects she teaches to trainee English teachers at Kibbutzim College of Education, Technology and the Arts where she also heads the department of English for Academic Purposes. E-mail: Tina_Wal@smkb.ac.il

Menu

Introduction
The study
Research questions
Corpus used in the study
Procedure
Results
Discussion
References

Introduction

There is little dispute that mastery of multi-word units is currently viewed as a necessary component of L2 lexical competence. Knowledge of phraseology makes learners come across as proficient and fluent (Boers et al, 2006), and distinguishes advanced learners from intermediate ones (Thornbury, 2002). Conversely, lack of this knowledge may impede the comprehensibility of learners' expression. Hill observes that learners often produce long winded and error ridden utterances because "they don't know the four or five most important collocations of a key word that is central to what they are writing about (1999: 5)."

The study described here (for the entire study see Waldman, 2009) investigates the use of collocations by Israeli EFL learners of three proficiency levels in free written production. Collocations do not have one simple and precise definition. Researchers seem to include under the term ‘collocations’ most multi-word units, including idioms, e.g. kick the bucket, fixed expressions, e.g. to and fro, leap year (Nation, 2001), and even functional expressions, e.g. excuse me, and proverbs e.g. let's make hay while the sun shines (Boers et a., 2006). In my study, I adopt the approach of earlier phraseologists e.g. Cowie (1981), so that I consider multi-word units such as throw a disk, pay money to be free combinations (words are replaceable following grammar rules), throw a party, pay attention to be collocations (restricted co-occurrence and semantic transparency), throw someone's weight around, pay lip service to be idioms (whose meaning is often opaque).

The study

Three factors motivated the study reported below: the belief in the importance of multi-word units in general and collocations in particular; the realization that collocations present a difficulty to language learners, including advanced learners; the conviction that learner language corpora can provide an invaluable source of data on language performance. Hence, the study is an analysis of collocations on the basis of a learner corpus. In this study, I investigated three groups of learners at three proficiency levels, and compared each group's performance to each other in order to trace any possible changes in group performance.

Research questions

The specific research questions were as follows:

1. Is there a relationship between accuracy in learners’ production of verb-noun collocations and their level of language proficiency? To answer this question, the following relationships were explored:
   1. between the number of well-formed verb-noun collocations and the learners' level of language proficiency
   2. between the number of deviant verb-noun collocations and the learners' level of language proficiency
2. What proportion of collocation errors are due to L1 influence at each level of proficiency?

Corpus used in the study

I collected the data for the learner corpus from learners in educational institutions all over Israel. The corpus consists of 759 argumentative and descriptive essays that were produced as part of the learners' course work at school, college and university. Care was taken to collect essays written only by speakers of Hebrew or Arabic as L1.The corpus contains one essay per learner.

I divided the corpus according to proficiency groups and called the level of essays written by 9th and 10th graders 'basic', the level of 11th and 12th graders 'intermediate', and the level of the college and university students 'advanced'. The 'basic' sub corpus contains 200 essays, 41,621 words. The 'intermediate' sub corpus is compiled from essays written by 252 learners and comprises 47,117 words. The 'advanced' sub corpus comprises essays written by 307 learners and contains 202,311 words. The average length of the essays varies in the sub corpora: in the 'basic' corpus it is 210 words, in the 'intermediate' it is 187 words and in the 'advanced' it is 504 words. The size of the corpus is small when compared to native speaker corpora since it consists of 291,049 words. However, it is large when compared to other learner corpora. For example the GeCLE (German Corpus of Learner English), which Nesselhauf (2005) used in her study of collocation production by German learners of English, comprised 154,191 words.

Procedure

I began the procedure by extracting nouns from each of the three learner sub corpora using a baseline of 220 high frequency nouns retrieved from analysis of a native speaker corpus compiled from essays of NS of comparable age (see Laufer and Waldman, 2011). Some of the nouns are child, question, knowledge, school, life, baby, aim, opinion, end, idea, law, guilt, place, television. I analyzed the entire list with the Vocabulary Profile available at www.victoria.ac.nz/lals/staff/paul-nation/nation.aspx and at www.lextutor.ca. Most of the nouns (201) belonged to the first 2000 most frequent words (e.g. society, money), 10 to the third thousand (e.g. benefit, damage), 4 to the fourth (e.g. access, welfare, approach), and 1 to the fifth. Four words appeared in lower frequency lists (e.g. Sabbath). The high frequency meant that I could expect the learners to know not only the meaning of these words, but also their usage, including their collocations. The following step in the procedure was to retrieve the collocations in which these frequent nouns appeared in each of the learner sub corpora.

I created concordances for each of these nouns so that the verb-noun combination in which they occurred could be identified and extracted. Subsequently, the extracted verb-noun combinations were checked in two dictionaries: the Benson, Benson, and Ilson Dictionary of English Word Combinations (1997), and Hill and Morgan's Language Teaching Publications Dictionary of Selected Collocations (1997). If the verb-noun combination was listed as a collocation in either one of them, it was noted as a collocation. Advanced learners produced altogether 852 collocations involving 13,805 noun tokens, intermediate learners produced 162 collocations using 3,057 noun tokens, and the basic learners – 68 collocations with 553 noun tokens.

An additional feature of the analysis of the learner corpus was the identification of erroneous collocations. A verb-noun combination was defined as an incorrect collocation when the intended combination (according to the context of the essay in question), should have been a collocation, and one of its components, usually the verb, was judged to be incorrect by a native speaker, and was not found in the BNC and the dictionaries of collocations. Here are some examples of deviant collocations found in the learner corpus: get the aim, inflict arguments, bring babies to the world, use a chance, learn children, do a decision, solve the disease. As one of my objectives was to investigate L1-Hebrew influence on collocations, all the deviations were examined for interlingual influence reflected mainly in word-for- word translation.

Results

Table 1 displays the following data for each sub corpus, (relative frequencies per 40,000 words are shown in brackets): the number of well-formed collocations, the number of deviant (erroneous) collocations, and the percentage of deviant collocations and the number of all ‘collocation attempts’ (correct + deviant collocations). Table 1 also shows the number and percentages of errors exhibiting potential influence from Hebrew. It can be seen that the learners at the three levels of proficiency produced a fairly high number of deviant collocations, which accounted for about a third of all the collocations they attempted to produce.

Table 1: Well-formed and deviant collocations in the Israeli learner sub corpora

Advanced learners produced 400 deviant and 852 correct collocations, intermediate – 82 deviant and 162 correct, basic – 34 deviant and 68 correct. An additional comparison of number of errors to the number of all the words in each of the compared sub corpora revealed a relationship between learner proficiency and the number of deviant collocations (χ2 = 26.27, p < 0.0001, Cramer's V=0.01). These results show that the advanced learners and the intermediate learners produced significantly more deviant collocations than the basic learners. Hence, not only is there no decrease in the number of errors with a growth in proficiency, but there is an inverse relationship between proficiency and correctness of collocations.

Research question 2 addressed the issue of L1 based errors in the deviant collocations of learners. I compared the three proficiency groups on the number of deviant collocations that appeared to be affected by Hebrew influence by relating them to the number of all deviant collocations produced in the three learner sub corpora (The raw scores of these errors were, in the advanced sub-corpus 258, in the intermediate 52, in the basic 15). No significant relationship was found between learner proficiency and the number of collocations potentially reflecting L1 influence (χ2= 2.12, p = 0.35).

Discussion

The results of the study are based on an analysis of a relatively large corpus of Israeli learner English that provided data for comparison of learners at three levels of language proficiency. My results showed a significant growth in the occurrence of verb-noun collocations only in the advanced learner sub corpus when compared to the sub corpora of the basic and intermediate levels. This suggests that the development of collocation use is slow and uneven. It was also found that learners at the three proficiency levels produced a fairly high number of deviant collocations, about a third of all the collocations they attempted to produce. Here, my results are similar to the results of other studies, e.g. Nesselhauf (2005). However, if I consider the actual number of errors in the sub corpora in relation to all the words produced, the advanced learners and the intermediate learners produce significantly more deviant collocations than the basic learners. The two groups attempt to use more collocations than the basic learners, probably due to a higher degree of confidence, but with flawed success. Since a third of the attempts results in error, this means that learners who attempt to produce more collocations are likely to err more often. To put it differently, not only is there no decrease in the number of errors with a growth in proficiency, but as proficiency increases, the frequency of errors increases. L1 influence appears in about half of the erroneous collocations at all levels of proficiency, and does not decrease with time.

This study shows that collocation use constitutes a problem even for advanced learners. I attribute this problem to the inherent nature of collocations, and the nature of some teaching practices that stress input-based learning. Collocations are usually semantically transparent, e.g. make a decision, send a message, offer help, submit an application, hand in a paper, etc., since they are constructed from frequent individual words. Therefore, when encountered in the input, they may not be noticed by learners and teachers as problematic. Similarly, production is often difficult since “equivalent” collocations in L1 may often include at least one word that is different from L2. For example, English break the law is l'aavor al ha-chok 'pass the law‘ in Hebrew, give examples is lehavi dugmaot ‘bring examples'.

My results showed that collocations are problematic even for advanced learners, most of whom in the present study have been taught by communicative techniques. I suggest that in order to raise learners' awareness of collocations and the difficulties they present, communicative, task-based teaching should be supplemented by form-focused instruction involving pre-planned activities which single out the target items and has learners practice them out of an authentic, communicative context. These activities could follow two principles: emphasis on production, and cross-linguistic comparison. The first principle is motivated by the nature of collocations. Many of them are transparent in meaning, and therefore easily understood. Teaching efforts should concentrate on eliciting the collocations in exercises requiring matching the appropriate verbs or adjectives to nouns, selecting the missing part of a collocation from semantically similar options, completing parts of collocations without given options. The second principle is rooted in the influence that L1 has on the learners' collocations and the persistence of L1-based errors at advanced levels of learning, as shown by my results. Empirical evidence is available that shows that a brief explanation of L1-L2 differences in specific collocations and translation practice of these collocations proves more effective than other teaching methods that ignore the cross-linguistic differences (Laufer & Girsai, 2008). I hope that further research will explore the development of collocations over an extended period of time in additional learner populations, and will suggest instructional practices that can improve collocation learning.

References

Benson, M., Benson, E., & Ilson, R. (1997). The BBI dictionary of English word combinations (Rev. ed.). Amsterdam: Benjamins. Amsterdam: Benjamins.

Boers, F., Eyckmans, J. Kappel, H. Strengers, & Demecheleer, M. (2006). Formulaic sequences and perceived oral proficiency: Putting a lexical approach to the test. Language Teaching Research, 10, 245-261.

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

Hill, J. (1999). Collocational competence. English Teaching Professional, 11, 3-6.

Hill, J. & Morgan, L. (1997). LTP Dictionary of Selected Collocations. Hove: LTP.

Laufer, B. & Girsai, N. (2008). Form-focused instruction in second language vocabulary learning: A case for contrastive analysis and translation. Applied Linguistics, 29, 694-716.

Laufer, B. & Waldman, T. (2011). Verb-noun colloctions in second language writing: a corpus analysis of learner English. Language Learning, 61, 647-672.

Nation, I.S. P. (2001). Learning vocabulary in another language. Cambridge: Cambridge University Press.

Nesselhauf, N. (2005). Collocations in a learner corpus. Amsterdam: John Benjamins.

Waldman, T. (2009). A corpus based study of collocation production by Israeli learners of English as a foreign language. Unpublished doctoral dissertation, University of Haifa University, Haifa.

Please check the Methodology and Language for Secondary Teachers course at Pilgrims website.
Please check the Teaching Advanced Students course at Pilgrims website.
Please check the How to be a Teacher Trainer course at Pilgrims website.
Please check the Methodology for Teaching Spoken Grammar and English course at Pilgrims website.