Measuring the frequency of the academic formulas list across corpora: A case study based in TED talks and Yale lectures

Measuring lists of lexis across corpora is a well-established method in corpus linguistics. This article takes a novel approach and measures the frequency of occurrence of the Academic Formulas List (AFL; Simpson-Vlach and Ellis, 2010) across academic lectures (OYCLC) and an academic-adjacent corpus of TED talks (TTC). Frequency of occurrence is measured at three levels: overall inter- and intra-corpus variation; the composition of representation, to see which formulas are represented; and an investigation of the behaviour of formulas within texts. The corpora were found to be significantly different from each other in terms of overall representation with a medium effect size. The greatest difference concerned referential expressions and the smallest difference concerned stance expressions. In terms of intra-corpus variation the AFL was found to occur less often in the humanities and most often in the natural sciences for both corpora. The composition of coverage revealed Zipfian distributions for the AFL, with both corpora presenting a similar set of high frequency formulas within each group category. A combined ratio and minimum frequency measure identified salient formulas to each corpus. Concerning formula behaviour, differences were found between the corpora concerning the use of the same formulas. Pedagogic and methodological implications are discussed in the conclusion.