A multilayer exploration of the cognitive structure of publications in history

Citation networks among journal articles are perhaps the most common object of investigation in bibliometrics. For example, citation networks are widely used for science mapping as a way to explore the cognitive structure of scientific fields. Within this framework, the disciplines traditionally part of the humanities fare differently. Their main trait being the interplay of a broader array of publication typologies – monographs, edited volumes, journal articles – with a richer set of cited objects, including primary evidence. Consequently, when considered from a science mapping perspective, a community, field or specialism in the humanities might be represented as a multilayer network. We consider here a specialism in history, the history of Venice, and represent it using a set of publications including both books (edited and monographs) and journal articles. This set of publications is interconnected using three similarity measures: bibliographic coupling over references to books, bibliographic coupling over references to primary sources and textual similarity. The result is a multi-relation network with three distinct dimensions (that we will call layers), one per similarity measure, connecting the same publications. Given this representation, we proceed to analyse the different communities emerging from the three layers, to qualify them and consider to what extent they overlap or instead provide for orthogonal conceptual spaces.