باهم‌آیی‌ها در پیکرۀ نوشتاری فارسی‌آموزان خارجی: مطالعه‌ای بر پایۀ رویکرد عبارت‌شناختی

The present study is a mixed method (quantitative-qualitative) and cross-sectional research that was conducted to evaluate how to use collocations by foreign Persian learners at six levels from introductory to advanced. The statistical population included 1112 texts from the tests taken at the Saadi Foundation (Linguistic corpus with 142,608 tokens), which were sampled by stratified random sampling method, 300 written texts with balanced volume in each Level. Various linguistic approaches were studied to analyze the sample corpus and identify collocations; Then, by examining and matching the criteria of the Phraseological approach and Farth's approach with the researcher's observations in the sample, a comprehensive definition of coincidence and criteria for identifying it in the productions of non-Iranian Persian learners were obtained. Then, the sample body was coded and analyzed in the MAXQDA software according to a specific coding system, and the research hypotheses were verified by interpreting the descriptive and inferential statistics obtained from the SPSS software. The study found that Persian learners, regardless of their proficiency level, use a similar number of collocations in their writing so the collocations are important from the very beginning of Persian learning. However, more common collocations were observed in the lower-level Persian learners’ writing, and a list of them is provided in the article. Higher-level learners use a greater variety of collocations. These findings suggest that collocations are important from the beginning of language learning and should be incorporated into language instruction and assessment.
Extended Abstract:
Introduction
The effective use of collocations is crucial for accurate and fluent language production (Wray, 2000). collocations, as essential elements of pragmatic competence (Lewis, 2000), enable learners to act diverse communication contexts effectively. the concept of collocation has been explored decades ago by linguists like Palmer (1933) and Firth (1957). Later, the phraseological approach, with its categorization of lexical combinations into terms "idioms", "collocations" and "free combinations", offers a more precise criteria for distinguishing collocations from other lexical groups (Eisenstadt, 1979; Covey, 1981; Nottinger & DeCarrico, 1992; Howarth, 1996). This approach provides a robust theoretical foundation for the current study. Jalili and Sahrai's (2012) corpus-based study, while measuring vocabulary diversity, did not specifically address collocation usage. A comprehensive corpus-based examination of Persian collocation usage in non-Iranian learner texts remains absent from the literature. While studies on English learner productions, such as those by Cowie (1992), Howarth (1996), Moon (1998), Jackendoff (1997), Altenberg (1998), Men (2018), and Lu (2016), have been more extensive, a dedicated focus on Persian collocations in non-Iranian learner texts is still needed. The primary objective of this study is to establish a comprehensive definition and precise criteria for identifying collocations in the written productions of non-native Persian learners. Additionally, the study aims to investigate the frequency and diversity of collocations employed by Persian learners at various proficiency levels. To address these objectives, the following research questions and hypotheses are proposed:
     Research Question 1: Is there a statistically significant difference in the frequency of various Persian collocation types employed by learners at six distinct proficiency levels in their written productions?
     Research Question 2: Is there a statistically significant difference in the use of frequent and infrequent collocations by Persian learners at six distinct language proficiency levels in their written productions?
     Hypothesis 1: There is a statistically significant difference in the frequency of various Persian collocation types employed by learners at six distinct language proficiency levels: introductory, pre-intermediate, intermediate, upper-intermediate, advanced, and proficient.
     Hypothesis 2: There is a statistically significant difference in the use of frequent and infrequent collocations by Persian learners at six distinct language proficiency levels.
Methodology
     This study employed a mixed-methods (quantitative-qualitative) cross-sectional design. The statistical population comprised 1,112 texts from the Saadi Foundation (a corpus with 142,608 tokens). Using stratified random sampling, 300 written texts with a balanced volume in each category were sampled. To establish a comprehensive definition of collocations and clear identification criteria for identifying collocations from other word groups, a thorough review of linguistic approaches was conducted. The work of researchers such as Eisenstadt (1979, 1981), Van Roey (1990), Howarth (1996), Nesselhauf (2005), Laufer and Waldman (2011), and Men (2018) was considered. By adapting the definitions and criteria from phraseological and Firth’s approaches, along with insights from the sample corpus, a robust framework was developed. The sample corpus was then input into MaxQDA software and subjected to multiple coding iterations using a coding system based on syntactic patterns. This process refined the definition and identification criteria.The entire sample corpus was subsequently coded and analyzed according to the established criteria. Descriptive and inferential statistics, processed using SPSS software, were employed to verify the research hypotheses. To facilitate coding and analysis, lexical combinations were classified based on syntactic patterns, drawing upon the work of Hausman (1989) and Benson et al. (2010). A tailored classification of syntactic patterns was developed to suit the specific context of this research.
Results
     One of the most important achievements of this research is the development of a comprehensive definition of Persian collocations and the establishment of clear criteria for recognizing various linguistic combinations. This framework, visually represented in Figures 1 and 2, facilitates the identification of collocations produced by Persian learners across different proficiency levels. The definition and criteria were derived through rigorous analysis of a substantial corpus. Collocations are defined as combinations of two or more words closely co-located in a text, relative semantic clarity, and at least one of the following characteristics: (1) limited co-occurrence range for one or more constituent elements, or (2) high frequency of co-occurrence in native speaker texts, coupled with a close syntactic relationship. Based on this definition, the following criteria were established for identifying collocations in the written work of foreign Persian learners:

Distinction between collocations and idioms: Collocations possess relative semantic clarity.
Distinction between collocations and free combinations: Collocations exhibit limited co-occurrence range and high frequency in native speaker texts.

   Data analysis revealed that collocation usage rates do not vary significantly across six language proficiency levels, indicating their importance from the outset of language learning.Descriptive statistics indicate that in the total sample size (including 47,388 Tokens), 9,055 collocation tokens were observed. The frequency of collocations used in the written productions of different language level learnrers shows a gradual increase of 2% from the introductory level (14%) to the advanced level (16%), which is insignificant.
Also, to examine the frequent and infrequent collocations in the productions of Persian learners of different language levels, only well-formed collocations of the “noun-verb” were analyzed, and it was observed that 4,388 collocation tokens and 1,232 collocation types were used in the entire corpus under study. This frequency of collocation types (1232) compared to the number of collocation tokens (4388) means that on average each collocation types is repeated about 3 or 4 times in the texts under study; however, the graph below shows that the distribution of collocation types in the frequency range of 1 to 5 times (bottom of the graph) is much denser than the top of the graph. In other words, a small number of highly frequent collocation types and, on the contrary, a very large number of infrequent collocation types (with a frequency of less than five times) were observed in the sample corpus.
Figure: Scatter plot of the frequency of "noun-verb" collocations in Persian learners' productions
 
Comparative analysis of collocation diversity in learner texts indicated that each proficiency level exhibits significantly greater diversity than the preceding level, aligning with the findings of Getsaki (1999), Zhang (1993), and Men (2018).
Conclusion
The production of collocation reinforcement tasks has consistently been a focus of interest for educators. However, there remains uncertainty regarding the necessity of explicitly teaching learners about the limitations of free word combination and whether a behaviorist approach, emphasizing Audio-lingual teaching method and repetition of word clusters, is sufficient for collocation instruction. This study aims to provide a comprehensive definition and precise criteria for identifying collocations in the written productions of non-native Persian learners, and to examine the representation and diversity of collocations in the Persian productions of learners at different language levels. Through analysis of the sample corpus, significant findings were obtained. Based on these findings, we conclude that explicit instruction on the limitations of collocation selection and informing language learners about collocation patterns can contribute to the development of their vocabulary knowledge. Therefore, it is essential to explore appropriate pedagogical methods and develop curricula and instructional materials for teaching collocations. The findings of this study have implications for the development of educational resources and proficiency assessment tools for second language learners, and can also serve as a foundation for identifying various word group types in the field of second language collocation research.
Acknowledgment
The authors would like to sincerely thank the experts in this field who collaborated in assessing the reliability of the coding of the learner corpus of this study.

Conflict of Interest
The data in this study were extracted from texts produced by Persian learners in an educational context and were used without mentioning the learners' names or personal information.