Bilimsel yayın yapma sürecinde Türk akademisyenlerin yaşadığı dilsel sorunlar

Many academic institutions worldwide have compelled academicians to publish their works in high-impact international journals as a part of institutional hiring, promotion and reward. Besides, at many universities around the world publishing in international indexed journals has even become a prerequisite for doctoral students to obtain their degree. In this sense, the number of academic members and research students, willing to secure publication in international arena which is overwhelmingly dominated by English, are increasing. However, many scholars experience a number of troubles and difficulties in writing for publication processes. Therefore, in order to help them improve both the quantity and quality of their scientific outputs, it is very important to identify the problems scholars face in their publication process. This study aims to investigate the linguistic problems experienced by Turkish scholars in their writing for publication process. For this reason, The Multidisciplinary Corpus of Writing for Publication was compiled as a part of this study. It consists of 216 unedited research articles written by Turkish scholars from five disciplines: communication, economics, education, engineering, and medicine. In addition, a reference corpus, consisting of 163 published research articles written by Native Speakers of English was compiled for the comparison of the Turkish corpus. Three linguistic dimensions were examined: lexical diversity, syntactic complexity, and errors they made in their manuscripts. Lexical diversity was investigated using three different measures: type/token ratio, standardized type/token ratio, and moving average type/ token ratio. In the analysis of syntactic complexity, Lu's (2010) L2 Syntactic Complexity Analyzer (L2SCA) was used. Each article in both the study and the reference corpus was investigated 14 measures in 5 different categories. Finally, grammatical errors were investigated using Louvain Error Tagging Taxonomy in which grammatical errors were analyzed in 8 broad categories which were broken down into 54 subdomains. The results of the type/token ratio analyses revealed that scholars in both the MCWP and the reference corpus used a wide variety of vocabulary in their manuscripts even though the MCWP had slightly less lexical variation than the reference corpus. The results of syntactic complexity demonstrated that native scholars used significantly more syntactically complex structures in all of the indices except for two (CP/C and CP/T) in which Turkish scholars used comparable amounts. In this regard, Turkish scholars were found to have remarkable similarities with upper intermediate Japanese and Chinese learners in terms of syntactic complexity. The computer-aided error analysis showed that Turkish scholars made 20.08 errors per 1000 tokens with an average of 110.15 errors per text. It was also found that the most problematic three categories were Grammar, Punctuation, and Register, respectively. In addition, the examination of the errors revealed that Turkish scholars experienced difficulty, especially in articles, verbs, punctuation, and using the correct lexical item. It was concluded as a result of this study that the first language of the writers is a significant factor in syntactic complexity and Turkish scholars suffered from sufficient knowledge of the rhetorical aspect of writing a research article in English, which lead to more errors in their manuscripts.