What does Jewish studies “look like”? This:
Let me explain.
Last year I set out to try to understand what we might mean by the academic field of “Jewish studies.” I had done some previous work on networks on citation analysis and so figured that this was as good a place as any to start. This was a project that I thought would be easier than it turned out to be, but now, just about a year later, I have some preliminary insights.
I started with the text of about 20,000 research articles over forty years found in about thirty Jewish studies journals. The text, supplied by JSTOR, was in JSON files. Working with a student, Gabriel Burstyn, we used an LLM running on a local machine to extract the citations, allowing us to map hundreds of thousands of citations. We cleaned the data and then began to analyze it.
The goal was to see how these citations clustered and the relationships between these clusters. We thus started by running an algorithm that identifies denser cluster networks and organizes them into communities. When I looked at the top authors in each of these machine-generated communities their themes became quite clear. Different analyses helped us to identify the relative importance of each of the major communities (as measured through the citation network) and their structural importance. In network parlance, that would be the “pagerank” and “betweenness centrality” scores. The latter is particularly interesting to me, as it helps us identify the critical domains through which knowledge flows.
In the diagram above, the size of each node is adjusted by its betweenness centrality score. The larger the node the higher the score the more important it is to the network as a whole. We can see that there are some areas with Jewish studies that are relatively self-contained; their citation patterns look inward, even if they have high pagerank scores. Some of these we might expect. New Testament and Early Christianity, for example, draws from Jewish studies scholars but their own work is relatively marginal to the Jewish studies network as a whole.
A different diagram allows us to better see how domains relate to each other:
The numbers on the axes are insignificant. Here we look simply at the clusterings. The Western Europe area is perhaps the most disconnected of all these major areas. The clustering in the bottom left shows how these areas are all still so rooted in the Medieval period.
A few broad observations:
- Jewish studies isn’t really a single field. It can be seen as a network or constellation of semi-independent scholarly worlds.
- History is the strongest organizing principle (perhaps reflecting also the organization of the journals that supplied the data).
- There is a cluster of circulation hubs (“engines”) of the field, but several areas operate more like islands.
- There are small hidden bridges that connect some of the relatively isolated areas.
- There are dense areas and quiet zones that suggest that the ideas, methods, and scholars travel unevenly through the field.
Does every field look in the humanities/social sciences like this? That would require a bit more research.
We can also, though, go much deeper into this network. We ran a community detection algorithm on the “Rabbinics” cluster, for example, to break the many citations down into further subcommunities, and then drilled down into deeper into one of those subcommunities. In future posts I’ll provide some updates on those experiments.
I’ll be presenting on this project at the Association for Jewish Studies Annual Conference on Sunday, December 14, 2025.
