| چکیده انگلیسی مقاله |
Lexical bundle research has recently come to the forefront of corpus-driven studies. Previous corpus studies have documented conflicting results regarding the frequency and function of lexical bundles (LBs) in academic prose. To date, however, no study has exclusively investigated LBs in the "discussion" sections of research articles generated by professional native English authors. The current study addressed this gap by examining the frequency, structure, and function of the most frequent four-word LBs. The corpus was composed of the discussions of published research papers authored by native (L1) writers. The data were extracted from five reputable international journals in the field of applied linguistics, consisting of over 300,000 words. Using AntConc, all the lexical sequences were retrieved with a frequency of 10 and a range of 5. The results revealed that LBs were predominantly used by English writers. Structurally, it was found that phrasal bundles were the most frequent in our corpus. The findings also demonstrated that functionally, referential bundles were extensively employed. In addition, stance bundles and discourse organizing bundles were the most prevalent after referential bundles. Finally, the findings are discussed in terms of the implications for non-native writers regarding the use of LBs in academic prose.
1. Introduction
Since research articles (RAs) are an indispensable part of academia, writing a highly qualified paper entails the competent deployment of linguistic features. The current study investigated a particular type of morphological feature dubbed “lexical bundles” (LBs), which refer to frequently-occurring word combinations. With the growing interest in this area, some corpus-driven studies have examined LBs across different academic genres (Biber, Conrad, & Cortes, 2004), academic registers (Biber, & Barbieri, 2007), disciplines (Cortes, 2006; Durrant, 2017), expertise levels (Staples, Egbert, Biber, & McClair, 2013), L1 versus L2 writing (Ädel & Erman, 2012; Esfandiari, & Barbary, 2017), and rhetorical moves (Alamri, 2020). The findings of prior research on L1 and L2 writing have illustrated inconclusive results concerning the function and frequency of LBs. For instance, Ädel and Erman (2012) observed that native English writers relied on LBs to a greater extent than non-native writers. However, there have been corpus-based studies indicating that non-native writers utilized LBs with a higher frequency than their English counterparts (Bychkovska & Lee, 2017; Pan, Reppen, & Biber, 2016). By the same token, the frequency of functional patterns of LBs has been found to vary in a number of previous corpus studies (e.g., Ädel & Erman, 2012; Bychkovska & Lee, 2017). This study was set out to contribute to this path of inquiry by investigating the frequency, structure, and function of the most frequent four-word LBs in a corpus of 'discussion' sections of RAs written by native English academic writers in applied linguistics.
1.1. Research Questions
1. What are the frequently used four-word lexical bundles in research articles' discussions written by native English academic authors in applied linguistics?
2. What are the structural and functional properties of these frequently used four-word lexical bundles?
2. Literature Review
2.1 Frequency of LBs
Frequency is the most basic attribute of LBs since a multi-word sequence ought to have the requisite frequency threshold to be considered as a bundle. Depending on the size of a corpus, the frequency threshold might vary from 10 (Biber et al., 1999) to 40 times per million words (pmw) (Pan et al., 2016). A variety of occurrences have been identified to be used by authors in preceding bundle studies in L1 and L2 writing. As an example, Esfandiary and Barbary (2017) observed that English academic authors used significantly more LBs than Persian writers. Conversely, Bychkovska and Lee (2017) found that Chinese undergraduate students used more LBs in their essays than English students did.
2.2 Range of LBs
Range or dispersion is another criterion for identifying LBs. Similar to frequency, the range threshold varies depending on the corpus size. For instance, Adel and Erman (2012) set the low dispersion threshold of 'three' thanks to the size of the corpus, while for a corpus of 176 texts, the range threshold was set at 20 by Biber and Barbieri (2007).
2.3 Structure and Function of LBs
LBs fall into different structural and functional patterns. Following Biber et al.'s (2004) functional and structural taxonomies, LBs were structurally classified into three categories: NP/PP based bundles (phrasal bundles), VP-based bundles, and Dependent clause bundles. Functionally, they serve three primary functions, namely stance bundles, discourse organizing bundles, and referential bundles. Previous research has shown varying results regarding the frequency of stance expressions and discourse organizers in L1 writing.
3. Methodology
3.1 Corpus
The present study used a corpus of research article discussions produced by native English academic writers in applied linguistics. The RAs were extracted from five highly-ranked international journals (Language Learning, Applied Linguistics, TESOL Quarterly, Studies in Second Language Acquisition, and Second Language Writing). The corpus was composed of 243 discussion sections published between 2005 and 2019.
3.2 Bundle identification procedure
In the initial stage, the discussions were removed from all non-textual content (i.e. plain texts). Using AntConc (3.5.8.0), a list of four-word LBs with a frequency of 10 and a range of 5 were retrieved. Then, the LBs were structurally and functionally analyzed based on Biber et al.'s (2004) structural and functional taxonomy.
4. Results
After retrieval, 142 types and 2,637 tokens of LBs were found to be used in the discussions, suggesting the prevalence of LBs in the academic prose of native English writers. The most frequent LBs found in the corpus were in the present study, in the current study, in the case of, it is possible that, the results of the, and on the other hand, which occurred over 50 times across the corpus. Structurally, most LBs were phrasal bundles consisting of NP-based and PP-based bundles. The functional analysis revealed that referential bundles accounted for 60.6% of all LBs.
|