[CP02] Mathieu O’Neil on online field theory

CP02: Mathieu O’Neil on online field theory

The game creates the boundary: field/force, actor-networks, and online research

From an Actor-Network Theory (ANT) perspective, deploying an analytical framework such as ‘field theory’ amounts to adding a layer of complexity above social actors which is arbitrary and unnecessary; all that is needed, goes this argument, is to trace the associations created by the actions of the actors involved in a controversy. In contrast I aim to show why, when it comes to analysing online activist spaces, a field-theoretical approach constitutes a valuable improvement to Actor-Network Theory.

Fields are created by the constant reciprocal adjustment of elements in relation to one another: field theory initially emerged in the physical sciences through ‘various attempts to comprehend how one thing could affect another without some substantive medium’ (Levi Martin & Gregg 2013: 40). Examples include gravity, electricity and magnetism. In the social sciences, the most well-known conceptualisation came from Pierre Bourdieu, who focused on the relationships of power that constitute and shape social fields. In Bourdieu’s sociology, society is differentiated into a number of semi-autonomous, internally coherent microcosms, governed by their own ‘game rules’, yet with similar basic oppositions – between economic and cultural capital for example – and general structures. These spaces are both fields of force (power is unequally distributed) and fields of struggle (people try to maintain or modify these power relationships). A number of objections have been levelled at field theory in general and at Bourdieu’s conception in particular. They can be summarised as follows: field theory is overly deterministic; it purports to exclusively unveil social domination; and it is simply adding an unnecessary analytical level to reality. For reasons of space, this post focuses on the last of these critiques.*

The notion that field theory adds an unnecessary macro-level structure above reality is central to ANT, which is currently being used by several analysts of online networks (see for example Marres & Moats 2015, Rogers et al. 2015). It is worth quoting ANT’s most famous proponent at length: ‘Instead of trying to simulate and predict the social orders, we wish to acknowledge the limitations of the simulation approach for collective systems and prefer letting the agents produce a dynamic and collect the traces that their actions leave as they unfold so as to produce a rich data set…’ (Latour et al. 2012: 605). Yet rich data sets appear to have distinct shapes and properties, which can be described in field-theoretical terms. This point can be illustrated by referring to boundaries, actors, and connections.

If we observe a ‘game’ or ‘contest’ (sporting or otherwise) and note that team players are wearing identifiable attire, acting in concert for a common purpose, and observing similar rules of behaviour, all of which are different to that of onlookers, then we can legitimately argue that they constitute a distinct social space. They are of course connected to the family members, friends and acquaintances watching them, to other players in other games, and so on, but it is safe to say that these ties will have limited bearing on the outcome of the particular game. A similar case can be made for online spaces (though their boundaries are more fluid than that of a sporting team): participants are there for a reason, which has nothing to do with the researcher’s construction of an object, or with a web surfer’s travels: no matter from which portal one enters, or the manner in which one interacts with other online actors, the collective purpose of the participants remains their own. The game creates the boundary. Suggesting, as Latour (2005) does, that local networks can be stretched indefinitely to cover the global, macro-perspective, provided that the material traces are accounted for, appears difficult to achieve in practice. ANT’s concern not to unduly project overarching structures onto local interactions is clearly valid. However in the online context boundary-making is justified by the socio-technical affordances of the Internet: for example the use of certain hashtags effectively circumscribes online fields.

In terms of players, an oft-mentioned contribution of ANT is to incorporate non-human actors into the network of connections and translations (Latour 2005). To return to the sports field example: no doubt the quality of the pitch and ball play a significant role; granted, the game would be vastly different without the referee’s whistle or the goal posts. Yet there is a fundamental difference in terms of agency between non-human and human actors: the whistle cannot blow itself; the ball cannot score a goal of its own volition. Both are dependent on human intervention. Similarly online there is a need to distinguish between actors who can autonomously make connections (such as people and, arguably, ‘bots’) and those who, though enabling important affordances in the diffusion of activism and controversies, rely on others to connect (to) them, such as Twitter hashtags. Nor are the algorithms that orient interactions across networks ‘just another actor’: non-algorithmic actors have no choice as to how algorithms affect them, so it makes better sense to define algorithms such as Google’s PageRank as governance institutions, whose influence stretches over the whole field.

Finally, not all connections are equal. It is far easier to create connections in Web 2.0 (by retweeting or liking a post, for example) than on Web 1.0, where hyperlinks have to be written into website link pages. Connections should accordingly be interpreted differently. Assuming that connections occur smoothly and naturally also overlooks a key fact, which is that actors may choose not to connect to one another. I have previously argued that in the online environment, where creating connections is at once costless and public, the absence of ties is highly significant: specifically, my co-author and I showed how activist actors who connected to similar issue-frames were not hyperlinking to each other’s websites, and we interpreted this absence of connection as indicating that there were major divisions in this field (Ackland & O’Neil 2011). In sum, like ANT, the version of field theory defined here acknowledges the desirability of staying at ground level. This is why our central concept of ‘field/force’ refers to the highly contingent capacity of actors to attract connections. But we also contend that this capacity or skill is inscribed in a local context: force operate in specific fields.


Ackland, R. & O’Neil, M. (2011). Online collective identity: the case of the environmental movement. Social Networks, 33(3), 177-190.

Latour, B. (2005). Reassembling the social: An introduction to Actor-Network Theory. Oxford, UK: Oxford University Press.

Latour, B., Jensen, P., Venturini, T., Grauwin, S., Boullier, D. (2012). ‘The whole is always smaller than its parts’ – a digital test of Gabriel Tardes’ monads. British Journal of Sociology, 63(4), 590-615.

Levi Martin, J. & Gregg, F. (2015). Was Bourdieu a field theorist? In M. Hilgers and Eric Mangez (Eds.), Bourdieu’s theory of social fields: Concepts and applications (pp. 39- 61). Oxon, UK: Routledge.

Marres, N. and Moats, D. (2015). Mapping controversies with social media: The case for symmetry. Social Media + Society, 1, 1-17.

Rogers, R., Sánchez-Querubín, N., Kil, A. (2015). Issue mapping for an ageing Europe. Amsterdam: Amsterdam University Press.


*To find out how we characterise an online field empirically, we invite interested readers to look at O’Neil, M. & Ackland, R. (in press) Towards a theory of online field/force. In M. Allen, J. Hunsinger & L. Klastrup (Eds.), International Handbook of Internet Studies Vol.2. Amsterdam: Springer. (Accepted 8/2/2016)] from which parts of this post originate. A pre-print of the chapter can be read on SSRN [http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2769684].


[CP01] Antonio Casilli on digital labour

CP01: Antonio Casilli on digital labour

The production of value online is always collective

My research is situated in a wider field than the simple automation of productive labour. The most controversial question, which is generating a lot of discussion about digital labour, is the return of implicit work, of invisible work. It is often boiled down to this question: is posting on Facebook working? But this is really a quite cursory way of phrasing the issue which makes me uncomfortable, even though I know journalists like these sorts of shortcuts. Said like that, it is just a provocation which does not account for the central point in my thinking, which is that the digital economy is largely based on hidden work, undeclared work that is sometimes unpaid.

Digital platforms enable us to perform activities which bring great benefits to us, but they also lead us to engage in practices which bring great benefits to their owners and investors. The identification of such types of value extraction from society is not a new concept in the social sciences. Before the Internet even existed, the analysis of the invisible labour of women, for example, led to long cycles of struggles for the recognition of this work, which occurred from the 1960s to the 1980s.

Several research areas converge so as to render visible a new way of understanding a labour that does not tell its name. My understanding of digital labour is part of a theoretical tradition. I can cite Marie-Anne Dujarier and her consumer work, or the whole Italian post-operaismo school, which emphasised immaterial labour, as well as Dallas Walker Smythe’s analysis of audience labour. In the end my thinking on informal labour in the context of digital networks revisits and reinterprets, revitalises these great research traditions, which have sometimes been neglected.

Digital labour occurs in several sectors of the economy. Let’s look first at the sharing economy. With companies like Uber, the peril for workers is obvious from the get-go. Neither the French state nor French city officials have done what the government did in South Korea, that is to say create a ‘counter-Uber’ by encouraging companies like Kakaotaxi, but also by creating an infrastructure that aims to transform Seoul into a “sharing city” (not only for cars, but also for lodgings and objects). This is a counterpower represented by public cooperative platforms… in France, these tensions generated classical forms of conflicts, which were, in my opinion, misunderstood by the press. The media applied old categories, and presented the conflicts of June and September 2015 between Uber and taxis as a defence, by the latter, of the status quo, of an antiquated corporatism. Basically, their argument was: ‘look at these remains of the past which are resurfacing and preventing the emergence of innovation’. The news media milieu is highly influenced by industrial interests and by the neoliberal doxa. And also, they are using an outdated rhetoric: the grand narrative of the forces of obscurantism, resisting innovation. But they did not see that it was a case of one platform battling another platform! And traditional trade unions working closely with new, emerging trade unions. The taxi drivers who burnt a few tires and flipped a car, they did put on quite a show, but they work for a platform, too. It is called for example Taxis Bleus or G7… Journalists did not notice either that in June, taxis were demonstrating against Uber, but in September Uber drivers themselves were organising their own trade union, and rallying in the streets to protest against the San Francisco-based company.

And this is also where research on digital labour progressively departs from previous studies on Internet use, on online contributions, because these studies were focused on content: photos, texts and videos voluntarily published on the Internet. With digital labour, we are less focussing on content as such, and deal a lot more with the metadata hiding behind the content. If I publish a picture on a digital platform, of course the human eye will look at the subject of the image, but the algorithm (the machine’s eye, so to speak) will zero in on the metadata. For example, the date, the place, the timestamp, the camera, the IP address on which the photo was routed. So, there are two completely different ways to think about the possible uses of this photo: academics who are interested in online engagement will analyse the wealth of content, the drive to participation, generosity, etc, but those who analyse how data is exploited and how wealth is captured by platforms will concentrate on the economics of the transformation of data into commercial value…

The question, then, is that of the value produced by platform users and of how it should be identified. And I will start by saying that we should not fall into the trap of considering such value as produced by each individual independently of all others.

The main problem is that, too often, researchers play into the hands of platform owners by concentrating on individual production and individual contributions. However, these contributions are intertwined. They are always collective, social activities. My data, my content has no intrinsic value, it only acquires value inasmuch as it is connected to my social graph, as I am linked to other profiles, other pages, other entities: there is a network effect. So, value is never singular, it does not originate from a social void. In contrast, platforms insist that contributions should be evaluated individually. Why? Because obviously for them, this approach has the benefit of atomising, fragmenting the efforts for recognition of digital labourers who end up being remunerated according to discrete tasks, and mostly very poorly. If you read the scientific literature on the various aspects of digital labour, you will see that these studies are often led by researchers working (or partnering) with the very firms that dominate the tech world. Facebook, Google and large telco companies have been trying to figure out for a while at what cost users would be prepared to sell their data, just in case…


Translated from the French by Mathieu O’Neil

This post presents extracts from an interview given by Casilli to the French-language INA Global site. The original article is here: http://www.inaglobal.fr/numerique/article/le-digital-labor-une-question-de-societe-8763

Biographical note: Italian-born Antonio Casilli has emerged as one of the preeminent theoreticians of online sociality in the French speaking world. Casilli teaches at Telecom ParisTech (the Telecommunication College of the Paris Institute of Technology) and also runs a monthly seminar on Theoretical and practical approaches of digital cultures at the prestigious École des hautes études en sciences sociales (EHESS). In 2015 Casilli co-authored Qu’est-ce que le Digital Labor? (‘What is digital labour?’). He can be found online at http://www.casilli.fr/