Enhancing Development of Communicative Skills in English through Teaching English Collocations in a Corpus-based Approach Bendravimo anglų kalba įgūdžių stiprinimas vartojant anglų kalbos kolokacijas mokomuosiuose tekstuose

S T U D I E S O F F O R E I G N L A N G U A G E S / K A L B Ų S T U D I J O S The author of this paper will substantiate the argument that communicative creativity and development of foreign language learners may be enhanced through carefully selected corpora which are applied as a useful tool in the second language pedagogy. Corpora constitute the easiest way to access the contexts in which given words and phrases appear to be suitable for native speakers of a particular language. This means that with their help learners have the opportunity to acquire the actual language which is used in everyday conversations by various groups of people. Moreover, corpora can save a lot of time, because learners can focus on language that they need in appropriate situations for determined communicational tasks. That is why learning collocations with their support can be easier and faster. The aim of this paper is to investigate how corpora can be applied to teaching and learning of English and whether they are truly indispensable for educational purposes. Accordingly, the results of empirical research consisting of observations and tests conducted among the Polish students of secondary schools will be shown to prove how the creative abilities in their communicative performance may be reinforced with the acquisition of English collocations.

Establishing factors which have an impact on achieving communicative competence in a foreign language is one of the most important objectives of contemporary linguistics. Nowadays, the majority of scholars emphasises the significance of the so-called formulaic language in achieving the competence in question. In short, formulaic language includes different kinds of word combinations, such as collocations. It is claimed that native speakers of any language have an instant access to an enormous base of formulae which they use in an automated manner when they intend to utter a sentence. The same process is believed to take place in learners of a foreign language. In effect, at present, collocations are believed to be a crucial element of L2 that enables its learners to achieve communicative competence.
From this perspective, it is important for foreign language teachers to be able to efficiently teach collocations to their students. At the same time, the problem of teaching collocations has been tackled by only few scholars. Thus, difficulties in choosing an appropriate tool for learning and teaching English collocations with special reference to speaking activities result from the fact that the topic in question has been in fact revealed although seldom explicitly undertaken, for example, in the works of Ellis (2001) and Nation (2001). The same comment refers to the selected positions on collocations authored by Wouden (1997), Herbst (1996), Fontenelle (1994), and Moon (1992). Accordingly, at present, the significance of collocational knowledge is emphasised in achieving communicative competence in EFL. However, there is general lack of teaching techniques and strategies that could be used to develop collocation skills in learners of English.
Consequently, the objective of the current paper is to present the impact of teaching English collocations in a corpus-based approach on development of communicative skills. What is more, the project was designed to bridge the gap between using corpora and collocations in learning and teaching. To put it differently, the ultimate goal of the present project was to investigate whether a corpus-based approach to teaching English collocations has a positive impact on communicative competence of learners. The article presents the result of the study that aimed at achieving the above objectives.
There were two groups of Polish students attending a private secondary school that participated in the study in question. The students represented approximately the same level of knowledge and proficiency in terms of English collocations. The classes were selected randomly from all third-form classes in the school (60 students in total). These groups were taught English collocations using different pedagogical approaches. The first group underwent traditional ways of teaching and the second group was taught using a corpus (British National Corpus). Prior to the commencement of the research, the students were asked to sit a collocation pre-pre-test which enabled for evaluating their overall knowledge of English collocations. In the course of the subsequent stage of the study, the participants were divided into two groups and asked to take another pre-test which aimed at confirming the previous results. The second pre-test constituted the first part of the data which was used for the analysis. Upon the completion of the project the students were again asked to sit a test, this time -the post-test. The aim of this test was to evaluate the difference related to the knowledge of collocations in both groups. Subsequently, the data from the second pre-test and the post-test were compared in order to evaluate the impact of the modified instruction (i.e., the inclusion of corpora) on collocational knowledge of students. The results of the aforementioned tests and observations of speaking activities with the use of collocations allowed for the assessment of the impact of corpora on communicative skills of the participants. On the whole, the present article puts forward an effective tool for teachers to apply in order to develop speaking skills of their students.
The literature clearly shows that scholars have been interested in the possible outcome of incorporation of corpora into learning and teaching L2. Much research has been conducted on how corpora can be employed and how people can benefit from them (Partington (1998); Jones and Sinclair (1974); Sinclair (1991); Stevens (1995) (2007)) However, little has been said about enriching students' knowledge concerning English collocations with the employment of corpora.

Theoretical Background
On the Notion of the Corpus

Concordancing and L2
The scholars differ in how freely they apply the notion of corpus. However, there is an increasing consensus that a corpus has four main characteristics: (1) sampling and representativeness, (2) finite size, (3) machine-readable form, and (4) the standard reference. The notion of corpus that has been considered during these studies is: a collection of pieces of language text in electronic form, selected according to external criteria to represent, as far as possible, a language or language variety as a source of data for linguistic research (cf. Sinclair, 2005, pp. 1-16).
This definition shows that corpora can be found only in electronic form. What is more, it demonstrates that corpora have been designed to be as representative as possible. There is no doubt that the above can be helpful for learners of foreign languages. It is for the reason that they can find out whether the words that they want to work with are typical for a given linguistic context or if they should look for another expression that is more commonly applied in the target language by native speakers.
Furthermore, it is very important to note that a corpus in itself provides vast and deep knowledge about the language in question. That is why people have created computer programs to enhance its usage. Such programs are called concordances. They search corpora to find key words and the context in which they can be handled. The user sees on the screen sentences with the key words in the middle, and the context to the left and right of them.
Students and teachers mostly utilize concordancing programs in order to cope with various situations at school. In addition, teachers apply them to acquire deeper knowledge about the usage of given words. This can result in better explanations of particular problems in class.
There are some words in English that are not easily interpreted by Polish students, and there are some that can be translated into Polish with the use of only one word. This situation leads students to their incorrect usage of English words in different contexts.
Teacher's self-education could solve such problems, because it would allow them to explain such phenomena to students in a more comprehensible manner. What is more, with the help of concordances teachers can prepare more diversified lessons for their students. They can prepare hand-outs with results taken from concordancing programs and ask students to make generalisations about a given word's usage, to underline which contexts are more frequent than others, or to find new collocations. Concordancing programmes are designed to facilitate the usage of corpora. Students can also take advantage of concordancing programs. They can have access not only to dictionary explanations of a given word but also to real usage. Students can thus save a lot of time by learning promptly how to work with words in actual contexts, and this can help them to overcome their problems with communication. It is common among students to know some words without knowing whether they can be utilized in a specific context or not. With the help of concordancing programs, students can easily get access to the information they need. Collocations can be seen, according to Martelli (2007, pp. 13-14), as one part of the words' meaning. The meaning of a word depends on the words that precede it, on the words that go after it, and on the context. However, it is very hard to establish one definition of the collocation which would be accepted by all scholars who carry out research on this aspect of language.
Scholars have applied the term collocation to describe entities that are familiar for them and, in most cases, their descriptions are different from those presented by others. That is why it is not easy to determine which description should be considered to be the leading one. Despite these differences, there are two main approaches to collocations, specifically the frequency-based approach and the phraseological approach (Martelli, 2007, pp. [14][15][16][17][18][19][20][21]. It is works of scholars such as Halliday (1966), Sinclair (1991), Kjellmer (1987), Jones and Sinclair (1974), or Greenbaum (1970) that have had a significant impact on the frequency-based approach to collocations. The emphasis in these works is put on the frequency of the occurrence of sets of words. The distance between those words is also perceived as very important. To say that some words collocate with others, a specific distance between words must be preserved. It means that, for example, the collocating words cannot occur more than five words away from each other: there cannot be more than five words between them in a sentence. In the view of Nesselhauf (2005, p. 12), if the frequency of the co-occurrence of given words is high, it means that these words collocate.
As far as the phraseological approach to collocations is concerned, it has been influenced by Cowie (1998) and Meunier and Granger (2008). In this case, collocations are distinguished from other lexical items such as idioms or free combinations. Collocations are differentiated from free combinations, because they can be considered as virtually fixed sets of words that do not go together accidentally. This means that changing one word can result in changing the meaning of the phrase. On the other hand, words that form free combinations can

Figure 1
A sample fragment of concordance for the verb make Collocations be changed without altering the meaning. Distinctions between idioms and collocations are made on the basis of transparency. Idioms are not lexically transparent but collocations are.
Collocations have at least one word that has its own meaning, and idioms can only be studied in their entirety. The phraseological approach also focuses on the relationship between words that form collocations. Words can be classified as either autonomous or dependant. An autonomous word has its own meaning and is chosen to occur in a particular context because of its actual meaning. It also determines what kind of a dependant word is used. The dependant word is conditioned by the autonomous word and is not involved in forming the collocation.
Although these two approaches seem to be different, they both try to solve the problem of how collocations are formed and why some words are taken to form them and why others cannot occur in collocations. This also shows how complex the learning of collocations can be and why research serving to facilitate this process should be conducted.
The current article proposes an effective tool enhancing development of speaking skills using English collocations that can be used by teachers. The tool in question was developed as a result of the corpus-based research project on English collocations. During the research in question two groups of Polish students of secondary schools were taught English collocations using different pedagogical approaches.
The first group was instructed using a traditional teaching approach, whereas the second group was taught with the use of corpus (British National Corpus). The results of students' tests as well as the author's observations of various speaking activities with the use of collocations indicate that the use of corpora enhance English collocations acquisition by students of L2.
The choice of the tool was motivated by the desire to provide teachers with a tool that could be easily accessed at school. In addition, numerous studies have shown that it is the way in which collocations are used that distinguishes native and non-native speakers of English. Therefore, teachers should make every effort possible to teach English collocation to their learners.

Location
The Aim of the Project

The Participants
The aim of this project was to investigate how students would react when the method of teaching L2 was changed. The project was designed to bridge the gap between using corpora and collocations in learning and teaching. Its goal was to investigate whether a corpus-based approach to teaching English collocations might significantly increase students' knowledge. The project aimed to analyse in a comparative manner the results of the participants before and after the change of the teaching method.
Two classes of third-form students from a Polish private secondary school were selected to take part in this project. The students represented approximately the same level of knowledge and proficiency in terms of English collocations. The classes were selected randomly from all third-form classes in the school. They consisted of 60 students altogether. The gender distribution was approximately 55 % male and 45 % female.
The research took place in a Polish private secondary school in Lodz as the author of the project has been working there and had the opportunity to do the research. There are not many studies conducted on the problem of the acquisition of English collocations in the Polish private schools. Therefore, conducting a study in this institution was aimed at filling this gap in Polish literature of the subject.
The participants of the study were randomly divided into two groups. The first group was named the traditional (control) group. The students assigned to the aformentioned group were instructed with the use of the traditional teaching method. The method of L2 teaching was the same as before the commencement of the research. The participants' belonging to the control group used traditional textbooks and dictionaries during their English lessons. They performed these exercises that were available in their textbooks. What is more, teachers prepared additional activities which were based on different materials selected by the teacher. Students were asked not to do any additional activities beyond these included in the original syllabus for their classes.
The second group was named the experimental group. In the case of this group, the method of teaching L2 was changed, specifically the traditional way of teaching was replaced by the corpus-based approach. Students were taught the way in which they should work with the corpora (British National Corpus) and concordancing software. The above enabled them to have access to those resources at home.
As stated previously, British National Corpus was selected for the purposes of the current research. The reason for the above was that it constitutes a collection of spoken and written texts created by a consortium of universities, publishers, and the British government. Moreover, it was utilized due to its size (90 million words of written English and 10 million words of spoken English) and rich diversity of texts' domains. The particpants used handouts which included the rudimentary rules for finding collocations within the corpus and exercises for them to perform with the help of corpora at school. The participants studied the hand-outs in order to discover the contexts in which various words can be used and to learn the way in which words should be selected to collocate with each other.
For instance, the subjects applied concordances in the KWIC (Key-Word-In-Context) format in which collocations are sorted to the left or right of the key word. Thus, they could focus on the main item of the study which was situated in the center of the page. It is worth mentioning that the particpants were aware of other items that collocate with the particular item whose use they mastered.
Prior to the commentment of the research, the students were asked to sit a collocation prepre-test to evaluate their overall knowledge of English collocations. In the course of the subsequent stage of the study, the students were divided into two groups and asked to take another pre-test which aimed at confirming the previous results. The second pre-test constituted the first part of the data which was used for the analysis.
Upon the completion of the project the students were again asked to sit a test, this timethe post-test. The aim of this test was to evaluate the difference related to the knowledge of collocations in both groups. The post-test was more difficult than it was the case with the previous one. It is for the reason that students were required to come up with their own words as it was not a multiple-choice test. The test in question constituted the second part of the data used for the analysis. Importantlly, the observation of students' performance during speaking activities was another element of the study.

Procedure Limitations
There were certain limitations identified in the current study which should be mentioned here. First of all, the number of participants was limited to 60 students of the same school. The students were characterised by approximately the same language proficiency and knowledge of English collocations. This is quite a small number of particpants and, therefore, results cannot be generalised to all L2 teaching situations. It would be advisable to include learners from different schools and at different language levels to the study. Furthermore, the period in which the study was conducted was limited and it was not possible to observe students' progress and differences among the two groups in a longer timescale. The narrow scope of the experiment provides only initial data for further studies and conclusions.
The analysis of the data was conducted in two stages. First of all, the average results of pre-pre-tests and pre-tests results for both groups were compared. The above enbaled the assesment of any potential differences in knowledge of English collocations among the particpants of the study. This was conducted with the use of Mann-Whitney's test whose aim was to establish whether the difference in knowledge of collocations between the traditional and experimental group was statistically significant.
In the course of the second stage, the data collected with the use of post-tests was compared in order to see the difference in an average score for each group. The result of the aforementioned comparison showed the difference in acquiring collocations among the two groups of particpants. The comparison of the traditional and experimental groups' results was conducted with the use of Mann-Whitney's test in order to establish whether the difference among the two groups is statistically significant.
The overall aim of data analysis was to determine whether the results gathered in the course of the study were statistically significant. The above enabled to determine whether the implementation of corpora can be treated as a useful tool for both English teachers and learners.

Results and Discussion
The results of the pre-pre-test showed that the students in both groups were charcaterized by approximately the same level of proficiency in the English language including English collocations. The errors, which were made by the students, mostly represented the same area of knowledge of collocations. For instance, the participants managed slightly better with verbs than with adjectives.
The first five sentences of the pre-pre-test (prior to the introduction of groups) were the most difficult for all the students. During the process of checking the tests, it was found that the  students made the majority of mistakes in this particular part of the test. What is more, it was found that 95 % of the students did not choose the correct answer for the second sentence. On the other hand, they did not have any problems with sentence number 12. The average score received by the particpants for the pre-pre-test was 9/15. Figure 2 summarises the results of the pretest for both groups. It can be seen that the average score received by the students was slightly higher than previously, specifically 10/15 points in the traditional group and 9/15 points in the experimental group. Nevertheless, there is no significant difference among these test results. In contrast to the pre-pretest, in the case of the pre-test the students made the majority of mistakes at the end of the test. It was the final six sentences that proved to be the most difficult for them. Analogically, the traditional group made the majority of their mistakes in this part of the test, but still slightly fewer than the experimental group. At the same time, the difference between the two groups is negligible. It was also revealed that, in both groups, there were individuals who possessed knowledge and language skills above the average. Similarly, there were individuals that were below the average level of the respective groups.
The results of the post-test reveal that there was a change in the average number of points scored by the students from the experimental group in the post-test as compared with the previous tests. As stated previously, the post-test was not a multiple-choice test. Students had to come up with their own words and they were not provided with a list of words to choose from. Although this test seemed to be more difficult than the previous ones, its results show that Figure 3 The results of the post-test the number of scored points was higher in the case of every single individual who took part in the project. There were 90 % of the students from the traditional group that were wrong in sentence seven and all of them did very well in sentence fourteen. As for the experimental group, 70 % of the students were correct in sentence seven and they did quite well in the rest of the sentences.
On the whole, students from the experimental group had better knowledge about sets of words that should be used in different contexts. The test also revealed that the students from the traditional group, where the methods of teaching were not changed, did just a little bit better than previously.
The following Figure 3 presents the comparison of post-tests results for both groups. It can be seen that the average score in traditional group is 11/15 points, whereas the average score in experimental group is 14/15 points. Figure 4 below summarizes the student's results. It is immediately visible that the project has enriched the experimental group's ability to use English collocations when speaking. In the majority of cases, students belonging to this group did not make mistakes connected with the appropriate choice of word that collocates with other words.
The results of the current project are in agreement with the results obtained during previous research into the corpus-based approach to teaching and learning second language. On the whole, the results obtained indicate that these students that were exposed to corpora can most efficiently deal with problems associated with the use of collocations in English, especially in the course of speaking activities.
The difference between the results of the pre-test and the post-test is approximately 7 % for the traditional group and 27 % for the experimental group. Overall, this amounts to a further increase of 20 % in the experimental group. According to Mann-Whitney's test the difference in knowledge of collocations between the traditional and experimental group in the post-test was statistically significant. Figure 5 summarizes student's results.  Figure 4 The distribution of points in the tests results The aforementioned result is satisfactory, however does not suggest that corpora should totally replace traditional methods of teaching. To put it differently, the results obtained in the course of the research indicates that language corpora can be a good additional source of information that should accompany traditional methods of L2 teaching.
Lastly, it has to be mentioned that corpora are so vast that it may be difficult or even impossible for the second language students to find a vocabulary item that they look for. Nevertheless, as the current project showed, the above problem has been efficiently solved by the introduction of concordancing programmes. These programmes enable users to browse the corpora and search for the exact words and contexts that they need.

Figure 5
The distribution of points in the tests results Second language is sometimes very hard to acquire and to master. This problem led scholars to develop methods and techniques that would facilitate people's endeavours to learn L2 language. The recent development is focused on the corpus-based approach to language acquisition. Corpora are electronic data bases containing vast knowledge about language. One can easily gain access through them to contexts that one may be interested in. People can learn a lot from corpora because they can see how the real, living language works. They do not have to Conclusions learn separate words, and they can learn substantial chunks of language at the same time. This gives them the opportunity to learn how to apply language appropriately in a given situation.

Tests
The project confirmed previous research into the corpus-based approach to learning second language conducted by scholars. The results indicate that students who were exposed to corpora can deal better with problems associated with the use of collocations in English, especially in speaking activities. The difference between the results of the pre-test and the post-test comes to approximately 7 % for the traditional group and 27 % for the experimental group. Overall, this amounts to a further increase of 20 % in the experimental group. This result is good, but does not suggest convincingly that corpora should totally replace traditional methods of teaching. In my opinion the result rather indicates that language corpora can be good additional sources of information that accompany traditional methods of teaching L2. Their use may result in providing the students with knowledge which is closer to the real use of the language. It may be argued that corpora are so vast that students will not be able to find what they need at a particular moment, but this problem has been solved by concordancing programmes, as was pointed out above. Students can thus browse the corpora and search for the exact words and contexts that they actually need.