corpus analysis

Academic essays entail taking a stance on the truth value of propositions. Epistemic adverbs deal with the speaker's assessment of the truth value of propositions. Employing a corpus-based approach with descriptive statistics and qualitative description, this study explored the use of epistemic stance adverbs in academic essays written by native English speakers and Iranian EFL learners. Following Biber et al.'s (1999) framework of stance adverbials, the researchers employed a corpus of 62077 words taken from class assignments written by Iranian EFL learners and a corpus of 65268 words taken from British Academic Written English (BAWE) to investigate the use of epistemic stance adverbs. Antconc software 3.4.3 version was used to search the most frequent stance adverbs. Frequency counts for each of the adverbs were extracted and normalized per 1000 words; then, Chi-square was run to pinpoint any differences between the two groups. The findings revealed both similarities and differences in the use of stance adverbs between the two groups. For example, EFL writers used more confident adverbs to show their authorial presence while native speakers used more maybe adverbs which are less authority-oriented. The findings may have implications for second/foreign language learners and writing instruction.

Despite the abundance of research investigating general and academic vocabularies and developing dozens of word lists, few studies have compared academic vocabulary with general service word lists such as conversation vocabulary. Many EAP researchers assume that university students need to know all the words in West’s (1953) General Service List (GSL) as a prerequisite to academic words (e.g., Coxhead’s, 2000) and teachers at language institutes recommend conversation students to learn words in Coxhead’s Academic Word List (AWL) as a follow-up to the GSL. The present study compared the academic and conversation vocabularies by exploring frequency and coverage of words in academic and conversation corpora. The GSL and AWL words were investigated in a conversation corpus and an academic corpus, each containing around 12 million running words. The analysis revealed that 1200 GSL word families were highly frequent in both corpora and 645 GSL word families were highly frequent in the conversation corpus but of low frequency in the academic texts. Also, a new academic word list of 700 word families was developed, which proved to be much more rigorous than Coxhead’s AWL. Further analysis indicated that the abovementioned 645 GSL words had a very low coverage of academic texts (0.7%), while they covered 4.05% of the conversation corpus. The new academic word list covered only 1.6% of the conversation corpus, whereas it had a high coverage of the academic texts (9.1%), much higher than that of the AWL (7.5%). The analysis of some other academic corpora revealed identical results.

In this study, corpus-based techniques were utilized in order to investigate the quantitative and qualitative features of adjectives across literary and technical texts. The corpus was composed of five literary novels and five engineering academic books in English. Seventy paragraphs were randomly drawn from each corpus and the frequency distribution of adjectives with respect to their position (attributive, predicative) and syntactic functions (descriptive, verbal, numeral, etc.) were tallied and summed. The results revealed that there is a significant difference in the frequency use of adjectives across the two corpora. From a register perspective, the high frequently use of adjectives in technical texts in comparison to literary texts (67.3% and 32.7%, respectively) may be due to the fact that almost all technical texts employ “expository” linguistic features which have a generally “informational purpose”, while most novels employ “narrative’’ linguistic features which have a direct functional association with the communicative purpose of telling a story of events which have occurred in the past. From a genre perspective, academic texts in engineering tend to be highly informational, non-narrative, and characterized by an impersonal style, whereas novels generally share the same primary communicative purpose of narrating a story whose purpose is to entertain.

The aim of this study was threefold: it aimed to develop a field-specific academic word list for accounting, to find the degree of coincidence between the word list and Coxhead's academic word list (AWL), and also to compare the occurrences of the most frequently used words in the list with six available word lists in different disciplines. A large corpus of accounting research articles was compiled and analyzed. We recognized 658 academic word families with the highest frequency in the corpus which we calledAccounting Academic Word List (AAWL ) . These 658-word families accounted for 10.16 % of the whole corpus. Further analysis indicated that out of these high-frequency word families we identified, only 354 coincided with those listed in AWL. Moreover, 50 most frequently used words in the list accounted for 3.98 % of the whole corpus. These words appeared in six available word lists in different disciplines with different degrees of occurrences which is a starting point for the development of a composite word list. Generally, this study confirmed the significance of subject-specificity of corpus-based word lists. The findings of this study suggest that AAWL can be used as a reference for the accounting community.

Stance has been examined extensively in the past few decades. However, the majority of the studies have been synchronic, with the few exceptions being snapshot diachronic studies without trend-related inferential statistics. This study adopted a trend-related inferential statistical measure and a more continuous diachronic dataset to examine the changes in using stance in the research articles of English applied linguistics from 2000 to 2020. To this end, 416 articles were randomly selected from 10 applied linguistics journals indexed in the first quartile of Scopus and probed using LancsBox for normalized frequency of Hyland’s (2005b) list of stance markers. Results were then analyzed for trends using the Jonckheere-Terpstra test. The results showed significant decreases for the overall stance markers (J=37168, z= -2.85, d= -.28, p= .004)), hedges (J=37014, z= -2.96, d= -.293, p=0.003), boosters (J=36298, z= -3.47, d= -.345, p=.001), and attitude markers (J=36647, z= -3.22, d= -.32, p=0.001), while self-mention markers were found to have experienced a slight, non-significant increase (J=42527.5, z= .94, d= .096, p=0.349). The functional analysis of the selected excerpts showed that the quantitative decrease in the use of stance markers has been compensated for, with an increase in the modification range and strength of the used stance markers as the two ways we could discover. After discussing the findings, the paper ends with some suggestions for further research.