|Institution:||University of Michigan|
|Keywords:||digital humanities; Information and Library Science; Social Sciences|
|Full text PDF:||http://hdl.handle.net/2027.42/111592|
This project systematically analyzes digital humanities blogs as an infrastructure for scholarly communication. This exploratory research maps the discourses of a scholarly community to understand the infrastructural dynamics of blogs and the Open Web. The text contents of 106,804 individual blog posts from a corpus of 396 blogs were analyzed using a mix of computational and qualitative methods. Analysis uses an experimental methodology (trace ethnography) combined with unsupervised machine learning (topic modeling), to perform an interpretive analysis at scale. Methodological findings show topic modeling can be integrated with qualitative and interpretive analysis. Special attention must be paid to data fitness, or the shape and re-shaping practices involved with preparing data for machine learning algorithms. Quantitative analysis of computationally generated topics indicates that while the community writes about diverse subject matter, individual scholars focus their attention on only a couple of topics. Four categories of informal scholarly communication emerged from the qualitative analysis: quasi-academic, para-academic, meta-academic, and extra-academic. The quasi and para-academic categories represent discourse with scholarly value within the digital humanities community, but do not necessarily have an obvious path into formal publication and preservation. A conceptual model, the (in)visible college, is introduced for situating scholarly communication on blogs and the Open Web. An (in)visible college is a kind of scholarly communication that is informal, yet visible at scale. This combination of factors opens up a new space for the study of scholarly communities and communication. While (in)invisible colleges are programmatically observable, care must be taken with any effort to count and measure knowledge work in these spaces. This is the first systematic, data driven analysis of the digital humanities and lays the groundwork for subsequent social studies of digital humanities.