Students’ preferences for syntax usage in Turkish language using distributional linguistic theory

Communication is one of the most important aspects of social life. Social interactions have increased the necessity for communication and learning of language. Social needs which constitute the main goal of teaching activities fill the gap of learning language. Linguistic surveys have revealed an important finding on educational activities and outputs. One of them is distributional linguistic theory. The purpose of this work is to describe the connection between Polish speakers who learned Turkish as a foreign language having preference for syntax and Turkish native speakers having preference for syntax using distributional linguistic theory. In this study, survey model, a kind of quantitative research, was used. Quantitative research data were collected by document review. The survey was carried out with Polish and Turkish university students in 2013/2014 academic year in spring semester. The working group consists of 20 students who were chosen with stratified random sampling. The findings explain the meaning statistically, and consequences were tabulated. There are differences in the syntax and sentence length of the working group in written language.


INTRODUCTION
Meaning represents a concept which is greater than its components since it occurs in advanced and complex expressions. Meaning transformation activity which is the main purpose of communication takes place wit h words. Being the main components of sent ences, words are the main elements which enable meanings to take place in the determined direction. Furthermore, meaningful features, obtained with words in a sentence, ha ve determinant roles in shaping and diversifying meaning levels. Usages, which widen meaning limits of words during communication, and new meanings, whic h are obtained through syntagmatic roles within sentence, consist of the important part of the study area of linguistics.
The studies conducted abroad about syntax gained new levels with advancement of linguistic field. Two schools draw attention. First is American structural linguistic school and second is productive transformational linguistic school (Işkın, 1999). Among these schools, Bloomfield analyzed the grammar of language in details and conducted important studies on the subject of syntax analyses. Therefore, he became a muse for structural science people, which came aft er him, using his approach of linguistic analyses which was called "direct components analyses".
These studies of Bloomfield enabled the establishment E-mail: cemerdemm@gmail.com.
Authors agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License of "distributional linguistic" by Zellig Harris, who was not structural with subject of form analyses. They enabled the occurrence of productive trans formational grammar w hich was developed by Chomsky (Demirel, 2011). Distribution approach, whic h is shaped in this foundation, found meaning with the thought that syntax and meaning are distributed in a similar fashion (Rubenstein and Goodenough, 1965). Indeed, with this approach, it is assumed that structure could be understood as a whole with distributional representations of word meanings throughout a sentence (Landauer and Dumais, 1997).
On the other hand, Chomsky intended to s eparate t he sentences of a language, which are appropriate to grammar from the inappropriate ones and to reveal the structures of sentences, which are appropriate to its grammar, in linguistic analyses related to surface structure. Chomsky stated that there were mathematical formulas in terms of structure in sentence setting for all languages. The main purpose of linguistic analyses is to study which sentences are linguistically correct and which are not appropriate linguistically (Chomsky, 1965;Demirel, 2011, p.34).
Therefore, the findings obtained through syntax analyses enable one t o define basic language data about the establishment of meaning units and ordering (Sahlgren, 2008). Likewise, Pantel (2005) stated that word syntax are prone to gain similar meanings in syntax. Firth (1957) added that it is important to know what word groups follow in the whole sentenc e. Generally, distributional structure argues that there is a close relationship between structure and meaning within a text (Sahlgren, 2008). Saussure (1985) argued that parole is individual and langue is a societ al concept. Therefore, langue should be analyzed. Saussure (1985) regards langue as a system in which the components, that create the langue, are in interaction with each other within specific rules (Demirci, 2009). According to Saussure (1985), paradigm is a plane on which we can make horizontal changes. On the other hand, syntagma is a plane on which we can make horizontal and vertical changes On the other hand, Zellig (1968) focus ed on the concept of verbal language (Friedman, Kra, and Rzhetsky, 2002). He considered words wit h their different distributions on usage. Distributional approac h means evaluation of independent components which created the meaning (Sahlgren, 2008). Through this context, it could be stated that distributional approach depends on the assumption that there is distribution and meaning similarities within the nature of language. After Zellig (1970) developed his studies, he understood the concept of distribution required of him to work infinite number of sentences than when he focus ed on the concept of "trans formation". Therefore, he tried to specify kernel sentences, since once kernel sentences could be found, the structure of language would be understood Erdem 55 with those, according to Zellig (1970), stated that this method could be applied to all languages (Vardar, 1999). Four different bits of information could be obtained about language based situations with distributional analyses. They are as follows: potential us age of syntax, communicational or situational content, phonologic al idea in syntax classifications and prosody (Chater and Finch, 1998).
According to Zellig (1970), syntax based and word based distribution give general information about the whole text. The whole text could be understood with t he distribution of concept (Piits, 2013). Redington et al. (1998) stated three syntax classifications in distributional approach: 1. Determining distribution of each word within the text 2. Comparing the distribution of word combinations in the text 3. Determining the words, which are similar and us ed together, in the text Leonard Bloomfield stated that the structures, which we ignore since we do not know their meanings, could be understood through their distribution in the following sections of the text (Harris, 1970). Therefore, he claimed that the word groups, which are sourc e of establishment of text components in linguistic manners, create t he linguistic meaning (Harris, 1968).
Learning a foreign language is acquisition of new linguistic system, which is composed of meanings and sounds; to learn speaking rules related to this system and different types of communication functions; and use them in an appropriate and correct ways. In such a learning process, one could only be possible with a language teaching approach, which focuses on a person to comprehend the language cognitively and him/her to develop communic ational competence (Matthews, 1997). Learners learn new system and its rules as utilizing language systems and their native language in order to develop language skills and they focus on competencies (Tura, 1983).
According to Redington et al. (1998), syntax distributions of words in sections of the text are defined as important sources in language acquisition and in defining the text (Redington et al., 1998). On the other hand, Turney and Pantel (2010) stated that specifying the syntax usage of individuals in statistic manner give answer to the question of which is understood. Again, in the biomedical studies, it is found out that distri butional approach is effective i n defining the natural usage of language. Then it is stated that the findings could be used with neuro-linguistic method (Weeds et al., 2005).
Not only the theories in the area (Sentaks) should be given, also the results of the research on the syntax should be included. In particular, the results of the studies emphasizing the relationship of syntax to writing skills

Syntagma
Paradigm My sister Tomorrow To London (she will) Fly Ali Today At us (he will) stay S/he İn the next week At Ankara -will be-I Tuesday -From-Germany (I will) leave should be included (Miller G, Walter C (1991). In this study, revealing the relationship between syntax preferences of students, whose native language is Polish, and regarding Turkish as a foreign language, and syntax preferences of students whose native language is Turkish aims to make syntax preferences of students visible (Table 1).

METHODOL OGY
Infor mation on research model, study group and data collection process is provided in this section.

Research purpose
The pur pose of the study is to reveal the relationship betw een syntax preferences of students, w hose native language is Polish, and regard Tur kish as a foreign language, and syntax preferences of students w hose native language is Turkish.

Research m odel
In this study, survey model w hich is a kind of quantitative research was used. This method intends to analyze research questions w ith different perspectives, and to find c learer answ ers for them as using both quantitative and qualitative data. The data in quantitative section are collected through document analysis (Yildir im and Şimşek, 2008).

Study group
The study group is composed of 20 students, w ho were selected w ith stratified random sampling method, and from Turkish Teaching Department of Gazi University in Turkey and Jagiellonian University, Warsaw University, and Adam Mickiew icz University in Poland, in spring term in 2013/2014 education year.

Data analysis
Content analysis method is used for data analysis in this research which is conducted through document analysis method. Content analysis is a systematic and repeatable method in w hich some words of the text are summar ized w ith s maller content categories, through coding depending on spec ified rules ( Büyüköztür k et al., 2008). For this purpose, essays obtained from the study group are analyzed according to distr ibutional approach analysis method. For quantitative analysis, dependent-independent groups t-test is used.
The students w ere asked to w rite a thought article. With a view to make it possible for students to express themselves in a better w ay no particular title w as identified. It w as explained that these articles did not have a purpose of evaluation as an exam and w ould not transform into exam marks but w ere made only to observe the situation; the students w ere asked to reflect the know ledge and skills regarding w riting composition w hich they had been obtaining so far in the articles. This process w as completed in 50 to 60 min.

Findings related to syntax
Item orders and sentence structures obtained from the students in Poland Turcology departments are analyzed as follows: Findings in Table 2 show that Poland students set sentences, which are c reat ed by at most 5 items. Additionally, it is found out that amount of sentences which are composed on single and 5 items are preferred at the smallest rate, in similar ways.
It is seen that Poland students use average of X =9.18 while Turkish students use average of X =10.20 words and also Poland students tend to write essays with average of X =15.7; while Turkish students with average of X =18.2 sentences (Tables 3 and 4).
Findings show that sentence structures, which are greater than t wo item -sentences, used relatively low amounts in students" written communication. Findings show that Polish students prefer using sentences which are created with at most 2 items for Turkish sentences.
Findings show that the distribution has common features up to sentenc es which are created with three items. Additionally, Polish students use sentences with 4 items relatively less than Turkish students. Demirel (2011) shares similar information regarding that writing focuses on thought and students can realize thoughts.
From the necessary assumptions to make ANOVA test come true, whether distribution shows normality or not was examined by "Kolmogorov-Simirnov test" (Table 5). Being between -1 and +1 of these values verifies the hypothesis that study group marks show normal distribution (Büyüköztürk, 2013;Cohen et al., 2000). Towards the findings ( Table 6) that were obtained, using    parametric tests in the analysis of data obtained from study group was approved and the execution was practiced accordingly (p>0.05). Tables 6 and 7 shows that in the sentences, which are used by Polish students in essays, meaningful differenc es could not be found in terms of word averages per sentence (p>0.05).

Findings related to item usage
In this section, part of the first 1400 words is considered in texts in order to eliminate the calculation errors which may be caused by the length of texts in the evaluations related to item amounts. In this purpose, a standard length is sustained in written communication of students. In the essays, which are created by Polish students in essays, meaningful differences could not be found in terms of item types (p>0.05).

Conclusion
It is found out that, S-V, O-V and A-V ordering in sentences with two items; S-A-V ordering in sentences with three items; A-O-A-V, S-A-A-V ordering in sentences with four items; and A-A-O-V ordering in sentences with five items are used frequently in sentence structures in essays of Polish students. On the other hand, it is found out that S-V in sentences with two items; S-O-V ordering in sentences with three items; S-IO-O-V ordering in sentences with four items are used frequently in sentence structures in essays of Turkish students. In this context, differenc es are established in sent ence structures of sentences, which are created with three and four items. It was found out Polish students use sentences, which are created with 2 and 3 items, relatively more frequently in written communication. It is found out that Polish students prefer using sentences, which are created with 4 items, less as compared to Turkish students. It could not find a meaningful difference in essays belonging to Poland and Turkish students in terms of item type frequency.
Adverbial item is used more frequently comparing to other items in sentences of Polish students.
It is seen that the sentences of Poland students have similarities with Turkish students" sentences in t erms of distribution of item amounts (Graphs 1, 2 and 3).
No meaningful difference in terms of word averages per sentences of students was found. It is seen that Polish students use average of X =9.18 while Turkish students use average of X =10.20 words Table 4. No meaningful differenc e in terms of sentence averages per essays of students was found. Therefore, Polish students tend to write essays with average of X =15.7; while Turkish students with average of X =18.2 sentences (Table 4).

Suggestions
Data diversification and interpretation studies, which might offer source for distributional approach could be widened to other skill areas. Materials related to students" Tura S (1983

Sentence types of Turkish Students in terms of Item amounts in texts
Amount of Sentence