In late 2010, the online services giant Google released a tool for public use called the Google Books Ngram Viewer. The Ngram viewer was a Web interface that enabled virtually anyone with an Internet connection to run frequency searches of character strings in a corpus of more than 5.2 million digitized books, or roughly 500 billion words, published between the sixteenth and twenty-first centuries. With these massive corpora at hand, a new branch of humanistic research has been established, dubbed “culturomics” (Michel et al., 2011). Culturomics enables investigators to examine changes over time reflected in print material associated with the lexicon being used and can offer evidence of social patterns and orientations within a language community. For example, one could use the Ngram viewer to chart how often the word “slavery” was printed over time and see that it increased dramatically in the years leading up to the US Civil War and had since declined substantially. Similarly, one could look at the appearance of technological innovations, such as the radio and the telephone, in print material and infer that the rate of cultural adoption of these technologies has increased over time. These kinds of capabilities are not contingent on the existence of the Ngram Viewer, but rather reflective of how analysts can look toward word usage in print materials from different times to understand changes within a community and culture (e.g., Lieberman et al., 2007).
What would such an approach tell us about the learning sciences (LS)? Granted, the field spans decades rather than centuries, and the corpus of associated text is modest even against the wide range of education research literature published during the same period of time. However, it is our suspicion that even for a field that is still emerging and establishing its academic identity, an examination of frequency of word use across different times could be informative for our understanding of the field. This chapter summarizes some of our efforts to do that work. Building upon an analysis presented in 2012 (Lee, Ye, & Recker, 2012), we provide here an elaborated comparative analysis of proceedings from the first and the most recent (at the time of this writing) LS conferences (Birnbaum, 1991; Polman et al., 2014), as well as a third conference, roughly collocated with the first LS conference (Gomez, Lyons, & Radinsky, 2010).