Shelsea van Ornum and Professor Doris Dant, Linguistics and English Language
Main Text
Writers and editors are constantly faced with decisions about usage (the actual use of language), as opposed to grammar (the theoretical structure of language). To help writers and editors make these decisions, various organizations and language authorities have compiled manuals and handbooks that include advice on and examples of various usage items. However, actual usage may not always reflect what language experts claim is the correct choice. Furthermore, language experts often disagree with each other in their usage recommendations.
Linguists can add a level of objectivity to subjective opinions and disagreements about usage by collecting empirical data on usage. One method that has greatly enhanced the ability of linguists conducting usage research is the development and use of corpora—large, searchable electronic databases of texts. These corpora enable researchers to empirically investigate the possible discrepancy between prescribed usage in manuals and actual usage reflected in corpus samples.
I investigated the possible discrepancy between prescribed usage from two style manuals and usage represented in a recently released corpus—the Corpus of Contemporary American English (COCA).1 My focus was on academic science, technology, and medicine (STM) writing. The two style manuals I selected for this study were The Chicago Manual of Style 15th edition and The CSE Manual for Authors, Editors and Publishers 7th edition. I matched these two style manuals with two subgenres of the academic register that are featured in COCA: “Science/Technology/Agriculture” and “Medicine.”
My specific purposes were (1) to evaluate how close usage recommendations in the two style manuals correspond to STM usage reflected in corpus samples and (2) to see whether STM usage aligns closer with the recommendations of one manual over the other. Thus I selected ten usage items mentioned in both style manuals, compared the usage entries in the manuals for these items, and used corpus data to empirically investigate the areas where the two style manuals differ in their usage recommendations.
This comparison of usage entries between style manuals showed that in several instances usage recommendations differ between manuals. In most cases, Chicago and the CSE differ in how strongly they advocate a traditional usage rule or emphasize the possibility or need for deviation from the traditional rule. In other entries, one style manual may bring up a topic not mentioned by the other style manual (e.g., Chicago mentioning the need to sometimes use differently than); and in some entries, one or both manuals may approach the topic from a vastly different perspective (e.g., Chicago states that presently is ambiguous and CSE restricts the meaning of presently to “soon” or “shortly”). Thus STM authors and editors may derive different impressions on recommended usage depending on which style manual they consult.
COCA data showed that academic STM usage aligns with the recommendations of at least one of the two style manual in nine of the ten items I researched. At the same time, however, STM usage does not consistently align with one manual over the other. Of the 10 usage items, 4 (farther/further; flammable/inflammable; impact; presently) had patterns of use in corpus samples that aligned closer with Chicago’s recommendations than with the CSE’s recommendations. Three items (after/following; although/while; healthy/healthful) aligned closer with the CSE’s recommendations than with Chicago’s recommendations. Two items (different from/different than; due to/because of) followed the advice of both style manuals with some minor deviations. And one item (people/persons) aligned with neither manual. Thus although academic STM writing appears to follow traditional usage rules (which vary between style manuals), STM writing actually wavers between one set of prescriptions and another.
In most cases, the usage in the Sci/Tech and Medicine parallel each other; however, in a few instances there are intriguing differences (after/following; people/persons). These differences confirm that usage differences can exist even among closely related subgenres.
Perhaps this usage variance is not a bad sign. STM authors and editors may be adapting general usage rules to the needs of their discipline and genre. My research focused on corpus results for the Sci/Tech and Medicine subgenres. Because I do not compare results from these two subgenres with the other subgenres in the COCA academic sample, I cannot prove that discernable usage patterns in my samples are unique to STM writing. Also, my investigation of only ten usage items is extremely limited. However, future corpus research may further elucidate not just the difference between usage in corpus samples and usage recommendations in various manuals but also differences in usage that may be specific to an academic discipline or subgenre.
Results from this study confirm that corpus research is valuable in gathering descriptive information about usage, even on the very limited scope that I employed. Because of changes in usage, discipline- and register-specific usage patterns, and outright deviations from prescriptive recommendations, authors of these style manuals may need to reevaluate their prescriptive advice. Although style manuals may not wish to simply report what this and other descriptive studies suggest is most common usage, these manuals should at least consider descriptive evidence on usage as they construct their prescriptive recommendations.
STM writers and editors should be wary in arbitrarily applying all usage advice from Chicago and the CSE to their work since corpus data reveal minor to significant deviations from prescribed recommendations. Accessing corpus data through the publicly available COCA interface may help authors temper prescriptive advice and their own usage inclinations with information on overall usage patterns and specific examples from published texts in their field. Corpus data provides information about what editors and authors seem to consider acceptable usage—information that is typically more current than information in style manuals. Thus in addition to usage recommendations from style manuals, corpus data and research can be useful to not only linguists and usage gurus but also to the everyday editor and author trying to decide whether to use farther or further.
This research was presented at the Council of Science Editors 2010 Annual Meeting and an abstract of this study was published in the Science Editor, July-August 2010.
References
- http://www.americancorpus.org/