The recent emergence of English-language publications by academics at Indonesian State Universities (PTKIN) underscores the need for corpus-based analysis to systematically examine lexical and terminological patterns in academic discourse. The purpose of this study is to investigate the linguistic characteristics, terminological patterns, and transliteration issues found in English-language academic articles in the field of religious studies published by PTKIN in Indonesia. This study uses a corpus-based approach with AntConc tools to analyse word frequency, collocations, and keyword patterns, with a particular focus on Arabic-derived terms absorbed into Indonesian and transliterated into English. The study found recurring linguistic errors, including Indonesian grammatical interference, literal translation, inconsistent use of articles and determiners, non-standard capitalization and punctuation, and redundancy in glossary strategies. Furthermore, significant orthographic variation and inconsistent transliteration practices are found for Arabic-rooted terms such as wudhu/wudu, aqidah/Aqidah, and sharia/syariah, reflecting both linguistic adaptation and identity-driven choices. Semantic shifts are also evident when culturally rooted Islamic concepts are translated into English, leading to a narrowing of meaning or a partial loss of meaning.