The languages of science: A corpus-based study of experimental and historical science articles

Studying the communication patterns of scientists can give us insight into how science works in practice. We have been studying the language of peer-reviewed journal articles in several scientific fields, and argue that methodological differences between different scientific fields are reflected in differences in how scientists in these fields rhetorically organize their texts. Techniques of computational stylistics enable us to compare rhetorical styles between the historical and the experimental sciences, and thus relate differences in language use to methodological differences posited by philosophers of science. A previous study (Argamon & Dodick 2004) demonstrated a clear difference between writing style in paleontology and that in physical chemistry; this paper extends those results to peer-reviewed articles from twelve journals in six different fields. Our current results demonstrate more generally that historical science writing is readily distinguishable from that of experimental science. Moreover, the most significant linguistic features of these distinctive styles are directly related to differences posited by philosophers of science between the historical and experimental sciences