the performativity of networks

Prompted in part by some conversations at the ASA meetings, in part by Gabriel’s discussion of the Social Structures author-meets-critics session, and in part by some gentle prodding from Cosma Shalizi, here’s a current draft of a paper of mine, The Performativity of Networks, that I’ve been sitting on for rather too long. Here’s the abstract:

The “performativity thesis” is the claim that parts of contemporary economics and finance, when carried out into the world by professionals and popularizers, reformat and reorganize the phenomena they purport to describe, in ways that bring the world into line with theory. Practical technologies, calculative devices and portable algorithms give actors tools to implement particular models of action. I argue that social network analysis is performative in the same sense as the cases studied in this literature. Social network analysis and finance theory are similar in key aspects of their development and effects. For the case of economics, evidence for weaker versions of the performativity thesis in quite good, and the strong formulation is circumstantially supported. Network theory easily meets the evidential threshold for the weaker versions; I offer empirical examples that support the strong (or “Barnesian”) formulation. Whether these parallels are a mark in favor of the thesis or a strike against it is an open question. I argue that the social network technologies and models now being “performed” build out systems of generalized reciprocity, connectivity, and commons-based production. This is in contrast both to an earlier network imagery that emphasized self-interest and entrepreneurial exploitation of structural opportunities, and to the model of action typically considered to be performed by economic technologies.

The usual disclaimers about work-in-progress apply.


Written by Kieran

August 26, 2011 at 10:59 am

8 Responses

Subscribe to comments with RSS.

  1. Awesome, thanks for sharing it Kieran!


    brayden king

    August 26, 2011 at 1:52 pm

  2. Cool, very glad you’ve posted this. Just yesterday I was saying to Viviana that in both this paper and your presentation at her panel, you’re taking a very introspective “physician, heal thyself” approach to econ soc.



    August 26, 2011 at 3:55 pm

  3. Fascinating paper, thanks for sharing! My only disappointment is that you didn’t include the most famous example of Google bombing (especially given the use of humorous anecdotes in the footnotes). Also, if you wanted to buff up the discussion of performativity outside of economics (which may not be worth doing), you could add in a bit about the 2010 Journal of Cultural Economy exchange between Butler and Callon. I think there is some mutual understanding, but it’s an interesting debate nonetheless.


    Dan Hirschman

    August 26, 2011 at 5:21 pm

  4. […] the performativity of networks – […]


  5. Agreed, this is really interesting. I knew something was making me slightly uneasy about some of the more recent work in network analysis drawing on electronic data sources, but I couldn’t put my finger on it until now. However, in terms of empirical demonstration of the comparability of network science and economics, it does seem that there is the lack of a smoking gun that MacKenzie provides via the direct historical connections he makes between the discipline of economics and practitioners in finance (e.g. the reference to Friedman’s discussion with Melamed). Tracing the variegated origins of the discipline back to the 19th century axiomatic debates doesn’t really get past this difference. One would need some evidence that Mark Zuckerberg took a soc or networks class or something of that sort (surely he took a networks class in CS). That is, the empirical lit on performativity tends to have more archival flesh tracing the intellectual history, and the connections between academics and engineers/practitioners (MacKenzie and Millo 2003) and a comparative case study would need some sort of empirical validation, which may demonstrate some differences between the two. Nonetheless, the methodological quandary for network analysts using virtual data, and the projection regarding future network models remain as the field coalesces into an interdisciplinary unit.



    August 26, 2011 at 6:03 pm

  6. The patent for Google PageRank has an interesting reference list, including:
    L. Katz, “A new status index derived from sociometric analysis,” 1953, Psychometricka, vol. 18, pp. 39-43.
    Mizruchi et al., “Techniques for disaggregating centrality scores in social networks,” 1996, Sociological Methodology, pp. 26-48.

    And many more computer science-y network references. That’s not a bad start for direct links.


    Dan Hirschman

    August 26, 2011 at 6:12 pm

  7. Does the old concept of double hermeneutic (Giddens) equal ‘weak’ performativity?

    Should the concept ‘performativity’ be limited to cases where the theory is self-fulfilling (e.g. claims made in network theory work better because agents know about network theory) and other types of cases where theory informs practice simply be identified as double hermeneutic?

    Some of it seems to revolve around the fact that network analysis is both ‘technology’ and ‘social science’ (often more of the former, no?). This might be a good way to summarize the context of the paper it up-front.



    August 27, 2011 at 10:40 pm

  8. […] is overlap with Kieran Healy's just-posted work-in-progress, The Performativity of Networks, {link} where he argues that network algorithms, when implemented in social media products, […]


Comments are closed.

%d bloggers like this: