Language use reflects scientific methodology: A corpus-based study of peer-reviewed journal articles

Abstract

Recently, philosophers of science have argued that the epistemological requirements of different scientific fields lead necessarily to differences in scientific method. In this paper, we examine possible variation in how language is used in peer-reviewed journal articles from various fields to see if features of such variation may help to elucidate and support claims of methodological variation among the sciences. We hypothesize that significant methodological differences will be reflected in related differences in scientists' language style. This paper reports a corpus-based study of peer-reviewed articles from twelve separate journals in six fields of experimental and historical sciences. Machine learning methods were applied to compare the discourse styles of articles in different fields, based on easily-extracted linguistic features of the text. Features included function word frequencies, as used often in computational stylistics, as well as lexical features based on systemic functional linguistics, which affords rich resources for comparative textual analysis. We found that indeed the style of writing in the historical sciences is readily distinguishable from that of the experimental sciences. Furthermore, the most significant linguistic features of these distinctive styles are directly related to the methodological differences posited by philosophers of science between historical and experimental sciences, lending empirical weight to their contentions.

Tracking the Footprints Puzzle: The problematic persistence of science‐as‐process in teaching the nature and culture of science

“…Appreciating variation among these communities precludes achieving a consensus view of the nature of science. Modern views tend to emphasize cultural, cognitive, epistemological, linguistic, methodological, and sociological features characteristic of diverse research enterprises and recognize that experiments often cannot provide answers to the questions that the historical styles of science, for example, ask (Argamon, Dodick, & Chase, 2008; Cleland, 2002; Dodick, Argamon, & Chase, 2009; Dodick & Orion, 2003; Frodeman, 1995; Gould, 1986; Kitcher, 1993; Kitts, 1977; Mayr, 1985; Orion & Ault, 2007; Schumm, 1991).…”

“…In brief and in regard to method of inquiry, philosophers and science education researchers simply do not agree on a unified nature of science (Alters, 1997; Argamon et al, 2008; Cartwright, 1999; Cleland 2001, 2002; Cooper 2002, 2005; Diamond 2002; Dodick et al, 2009; Feyerabend, 1993; Frodeman 1995; Gould 1986, 1989; Mayr 1985; Rudolph & Stewart 1998). Divides may be recognized between experimental and historical methodologies within science (even though both may hold prominence in the same field) or between pragmatist and realist philosophies of science.…”

Understanding Scientific Methodology in the Historical and Experimental Sciences via Language Analysis

“…Space permits discussion only of the high-level features here; a more detailed treatment has been given elsewhere (Argamon et al 2008).…”

Computational Register Analysis and Synthesis

“…It does, however, require a definition of register sufficiently precise that human annotators can label texts accordingly with high inter-rater reliability, which is not always easy to achieve. Register classification can comprise a stand-alone task (Stamatatos, Fakotakis, & Kokkinakis 2000, Biber & Conrad 2001, Argamon, Koppel, Fine, & Shimoni 2003, Finn & Kushmerick 2006, Santini 2006, Herring & Paolillo 2006, Abbasi & Chen 2007, Dong, Watters, Duffy, & Shepherd 2008, Sharoff, Wu, & Markert 2010 or may be used to derive insights into larger questions related to linguistic variation (e.g., Atkinson 1992, Argamon, Dodick, & Chase 2008, Eisenstein, Smith, & Xing 2011, Teich, Degaetano-Ortlieb, Kermes, & Lapshinova-Koltunski 2013, Clarke & Grieve 2017. Register labels, either manually or automatically assigned, can also be used to control for register in research on other text analysis methods (e.g., Carroll et al 1999, Giesbrecht & Evert 2009, Sharoff et al 2010; differences in register between training and testing data often affect outcomes for NLP tasks such as part-ofspeech tagging, parsing, or information extraction.…”