10 fave papers

These are all paper that inspired me through my random walks in academia (with some emphasis on articles that deserve more attention). (It’s an updated version of a blog post from 2012.)

1. P Bearman, J Moody, R Faris (2002) Networks and history This paper blew my mind when I first read it. All of a sudden, network theory didn’t seem to have any limits. It is about checking the consistency of narratives by reconstructing their causal chains. It is also a case study of a book by the famous (and controversial) Swedish leftist Jan Myrdal. The book, Report from a Chinese village, is a pretty readable account of the Chinese civil war. (I should also say that I was traumatized by having to read his (I thought extremely boring) autobiography in high school.) BMF’s paper was also the direct inspiration to our Emergence of collective memories (even a bit . . hmm . . undercited paper)

2. JA Danowski, P Edison-Swift (1985) Crisis effects on intraorganizational computer-based communication This paper studies one year of e-mail communication within a governmental organization facing reorganization and funding cuts. As the tension grows, the e-mail networks (aggregated monthly) become sparser, this is accompanied by an increase of words relating to reorganization (in an analysis of the message contents). It has everything of today’s computational social science papers . . . but it’s from 1985. (OK, the one thing that would be different today would be the privacy discussion.)

3. L Edlund, E Korn (2002) A theory of prostitution Just the fact that there is a paper in an economics journal about the pricing of the services of prostitution says something about how cynical the sex industry is. This is an underlying theme of the paper, too, as they argue their main point—the prices of prostitute’s services equal the value of the forfeited married life. The terminology is quite comical at times (“marginal wife”). All-in-all, it’s a persuasive paper solving a problem I never thought of before—how the stigma of prostitution translates to the relatively high wages of prostitutes.

4. M Morris, M Kretzschmar (1995) Concurrent partnerships and transmission dynamics in networks This is my candidate for the first modern netsci paper. It studies how a dynamic system (HIV epidemics) is affected by the underlying network structure (assortativity by sexual behavior) by tuning the structure of the network in a computer simulation. Fairly well-cited but not in proportion to its legacy.

5. P Martin-Löf (1966) The definition of random sequences  Maybe not so obscure as a paper, still cool and good in many ways: It makes practical use of Andrey Kolmogorov’s complexity measure and explains it in an almost textbook-like way. It mixes in many hip theories, from number theory to the theory of computation, metamathematics, and stats. It must have been quite an adventure for the author, a Swedish guy, to pursue his Ph.D. studies in Brezhnev’s Russia (with Kolmogorov). I never met him, but I’m sure he has stories to tell.

6. HA Simon (1962) The architecture of complexity In the transition period, when (the too vague) cybernetics and (the too general) general systems theory was turning into something more useful—complexity theory—Herbert Simon was on fire. This essay is maybe not that obscure. Still, I’d say it is underread. Fifty-odd years later, the complex systems field is still revolving around the ideas here summarized by Simon.

7. A Tversky, D Kahneman (1967) Judgment under uncertainty: Heuristics and biases This is probably the most well-known paper on the list (over 40,000 citations and Kahneman being the second Nobel Prize winner on the list). I first read it as a final-year Ph.D. student, and it completely changed the way I thought about how to understand people and model social systems. It is also a good summary of (the still very readable) book of the same name.

8. DJ Watts, PS Dodds, MEJ Newman (2002) Identity and search in social networks I’ve always been a big fan of Watts and Newman, and this is their most imaginative work together. It really bears the mark of the anything-is-possible early days of network science. I remember reading this paper on the night train from Umeå to Uppsala and missing some sleep marveling at the insight that social spaces can be non-metric.

9. K Kaneko (2007) Evolution of robustness to noise and mutation in gene expression dynamics I really like Kaneko et al. ‘s way of solving concrete, experimentally testable problems with simulation modeling. This is probably my favorite paper. I was lucky enough to visit the group twice, but our projects didn’t get where we wanted them to go. (I’m still a bit vexed about it.) Besides, this was the first PLoS ONE paper I ever heard about.

10. WW Zachary (1977) An information flow model for conflict and fission in small groups We all know the data set as the E. coli of community detection. The social ties between members of a karate club were recorded, and two clusters of the social network foreboded a split of the group. But how many really read the paper? Zachary’s own analysis is impressive. . mainly that he was influenced by the Ford–Fulkerson algorithm (15 years old at the time) to find a minimum cut in the “information flow” between the two hubs. Just like the modern algorithms, he misclassifies only one node. Much of the Girvan-Newman algorithm is here already (although I don’t think G&N knew about Zachary when they started their work).