The Limits of Network Analysis

Seattle.
The next AoIR 2011 speaker is Aristea Fotopoulou, whose interest is in digital networks. She focusses on the Feminism conference in London in 2009, using both ethnographic and Webcrawling methods. The conference is connected with the wider London Feminist Network, which not least engages with recent political changes in the UK. The network reframes views on violence against women, prostitution and pornography, but Aristea’s ethnographic work was able to trace a range of different versions of feminist identity.

Older divisions are reinvoked in networked conditions, and an imaginary perspective of feminism as a movement is evoked in the process; a continuing anxiety about catching up with digital technologies underpins some of these activities, too. In digital environments, doing network politics is being remediated through these imaginaries.

A second stage of Aristea’s work used IssueCrawler to examine the online connections of participants in the conference. She started with the conference as an issue itself; a compact network of strongly interconnected nodes emerged. She also examined whether links were generic (domain) links or deep links to specific pages; on violence against women, links appeared mainly to aid the circulation of specific documents, indeed – an issue network, in other words, but one which appeared to exclude dissident, alternative voices on the topic. This may also be a methodological problem, in fact.

Network mapping in this case does not reveal a presumed ‘real’ essence of a feminist network, but creates the conditions for meaning-making; non-human apparatuses like IssueCrawler are not neutral – it produces networks as meaningful political relationships. Issue network frameworks present an innovative shift in analysis, but the attention to single issues tells us little about how such issues come to be recognised as defined issues in the first place. Network analysis also overemphasises the flow of information, to the detriment of recognising specific local contexts of the process.