This fascinating and important paper by Alice Marwick develops a theory of morally motivated network harassment to explain the pervasive dynamics of online abuse we can see on social platforms, beyond what she terms the dyadic harassment which can equally be seen in face-to-face settings and the normalised harassment in which certain forms of behaviour become common in particular online spaces. Her proposal is that we understand these distributed forms of harassment in terms of:
- An accusation that a norm has been violated, justifying a response which seeks to enforce it
- Amplifiers who increase the visibility of that accusation within particular online networks
- Swarming attacks from loosely connected individuals, aggregating into something deeply sinister, coalescing around ‘attack vectors’ in which characteristics of the target get drawn into the attacks (at least if these are characteristics which already generate attacks in wider society)
It captures something important about the responsibilities of amplifiers. To what extent should they be held responsible if their actions contribute to harassment without encouraging it in any explicit sense? This can be seen quite straightforwardly if you scroll down the responses which, say, a political scientist pushing back against ‘woke culture’ in a quote tweet (without writing anything inherently offensive) might generate. The superficially mild quote tweet can stimulate a thread with outright racism and harassment. It doesn’t quite seem right to me to attribute direct responsibility to the quote tweet but conversely to deny any responsibility seems obviously untenable:
Because amplifiers (highly followed nodes, in network terms) have so many viewers, they are able, consciously or not, to direct harassing behavior. It is often not the amplifiers themselves but those who follow them who engage in such behavior, making it difficult to ascertain responsibility
In some cases the amplifiers can be organisations. I watched a tweet from someone close to me about an event at a graduation ceremony be included in a newspaper story about the social media controversy this event generated, before the story and the tweet were reproduced in the newspapers around the world. This generated hundreds of misogynistic, occasionally racist, tweets from a distributed group across a couple of weeks, including one discussion in a men’s right YouTube channel and a number of threats. The story was reproduced because it was potent clickbait and the original tweet circulated because it resonated with so many as a reading of the event at the graduation ceremony. But this chain of reaction meant that it circulated far beyond its original network (where the underlying norm motivating it was widely agreed with) and instead found itself received by those who understood this norm as itself violating their norms. Hence the network harassment with the newspapers and websites as amplifiers.
I’d like to understand the dynamics of normative dissensus underpinning network harassment. There’s a risk that doing so obscures how such harassment intersects with existing forms of oppression and Marwick’s concept of ‘actor vectors’ is important for avoiding this. But there’s something deep here about how normative conflict becomes a routine part of social life through the mediation of social platforms which simultaneously encourage norm-enforcement through their visibility incentives (i.e. morally loaded content is more likely to generate a reaction) which perpetually facilitating boundary encounters between groups with radically different normative outlooks of a sort which are relatively rare as face-to-face encounters.