A Co-Citation Network for Philosophy

Kieran Healy

The graph below represents co-citation patterns based on all articles published between 1993 and 2013 in Nous, the Journal of Philosophy, the Philosophical Review, and Mind. These four were chosen because they are all high-impact, high-prestige, and self-consciously "generalist" journals. The goal of the analysis---apart from teaching myself a bit of D3---was to get a rough, descriptive sense of what the world of high-prestige, professional, academic, English-speaking Philosophy has been talking about for the past twenty years.

I collected all of the citations contained in the 2,262 articles published since 1993 in the four selected journals---about 34,000 citations altogether. The graph shows co-citation patterns for the 500 most-cited items---that is, it takes the books and articles that have been talked about most often over the past 20 years in these journals, and shows which items are talked about at the same time. In fact there are 520 items in the graph, so as not to arbitrarily exclude some items with the same number of citations as other, included items. The colors of the nodes represent the results of a community-detection algorithm applied to the co-citation matrix. The community colors are generated inductively, not assigned in advance.

Note again that the unit of analysis is cited items, not authors, so the same author may appear in different places in the graph for different books or papers. Each book or paper only appears once, however.

The graph is generated dynamically, so it will look a little different each time you load it. You can also click and drag particular nodes, to "pull" parts of the graph around. When you release the mouse the graph should reconfigure itself. If you do this, let it finish moving around before you try moving another node. The idea is for the dynamic aspects of the visualization allow you a better view of different parts of the graph, while maintaining the global structure of relations between the parts. It is best viewed on a high-resolution monitor.

See the main blog post for more details and further discussion, including information on changes and corrections to the graph data..