LAUSR

Typical studies on the history of science, or particularly of biology, have been focused on a particular scientist or book, but this selection has a risk of being arbitrary. To find a more objective way of studying history of biology, we applied a statistical method. First, we downloaded from the PubMed database all available titles and abstracts of 934,807 articles in 32 selected journals from 1965 to 2014, and extracted most frequently used 322 terms by text mining. Clustering of these terms according to the annual frequency of usage resulted in three main clusters: Cluster 1 represented terms that were no longer used frequently, Cluster 3 included terms that became abundantly used recently, and Cluster 2 contained terms constantly used. Three phases were delineated in the history of biology over the past 50 years, with transitions in 1987 and 1997. In contrast with our tacit understanding that "function" is a key notion in biological thinking, the results suggest that function-oriented discourses are a new habit of biologists in the genomic era after 1997, in which biological researches focus on identifying a link between a molecule or a structure with its function. We hypothesize that, in spite of repeated warnings, function-related discourses have a teleological connotation, which is easily misunderstood by general audience and, with emphatic expressions such as "important" and "essential", fit to the need for justification of researches as part of researcher's responsibility for public funding.