"Strengthening Connections with the Audience: Reformation and Exemplification in Mathematics Research Articles"
|
Kristy LesperanceKristy was in her third year of undergraduate studies at the University of British Columbia when this essay was originally written, studying Mathematics under the faculty of Arts. The paper was written for an upper-level, intensive research and scholarly writing course using corpus analysis to investigate discursive features of literature from the student’s chosen major. Contents |
MethodCorpus Selection A corpus of scholarly, peer-reviewed, online-access articles was constructed following the “select, analyze, select again” procedure outlined by Bauer & Aarts (2000, p.23). Journals were chosen within the top ranking American journals as listed on the SCImago (2007) website, in order to reflect the contrast between "hard pure" and "soft applied" sub-disciplines. Two journals were selected to represent “hard, pure” theoretical mathematics (hereafter simply "theory"): the Journal of the American Mathematical Society and SIAM Review (Society for Industrial and Applied Mathematics). Two journals were also selected to represent “soft, applied” mathematics education (hereafter "education"): the Journal for Research in Mathematics Education and Research in Mathematics Education. Individual articles were then selected which included geometry or geometric in the title, but restricted to those with the observed median length (common to both sub-disciplines) of 15 to 35 pages. This resulted in a corpus of 29 articles spanning nearly 40 years: 17 representing theory with a combined 192,200 words, and 12 representing education with a combined 103,279 words.
The list of code glosses provided by Hyland (2007) was used to search for target words within the corpus articles (see Table 1 below), as well as some additional target words that appeared frequently in the present corpus, but which were not discussed by Hyland. It was not feasible to access professional corpus analysis programs, therefore each article was examined using the PDF text-search feature. In the few articles for which this was not possible, text was searched by reading for target words. The search excluded code gloss use within titles, abstracts, footnotes, indices, tables and charts, quotes, acknowledgements, and references. However, usage found within subtitles was included, since it was observed that articles frequently used subtitles to indicate the function of the subsequent discussion. Each occurrence was evaluated to identify the target function of the term within the sentence in which it was found. Due to the complexity in the function of potential code gloss cues, some of the target words were more indicative of reformulation, but others were more ambiguous. Therefore, following Charles (2009), most target words were still incorporated into the analysis, despite both their clear or potentially ambiguous function; however, cues which were inarguably ambiguous to this author were excluded. Particularly, the term “which” was often more ambiguous than other terms in that “which” may be used to further describe or elaborate on attributes about a single concept, rather than to indicate a reformulation of that concept. To illustrate, “which” was excluded from the following statement: “…then by deleting those terms for which for each satisfying , a reduced program is obtained that is canonical” (Ecker, 1980).
Table 1: Frequently used code gloss markers (frequency per 100,000 words)
|