There is a running debate in critical theory circles about the applicability of Marxian analysis to big data specifically, and to an economy dominated by immaterial goods, more generally (I have blogged about this periodically, circling primarily around the concept of primitive accumulation: see here and here). As part of working through that literature, here I want to lay out some of the broad outlines of the pro- side as I see it, and then offer some preliminary thoughts from Marx’s own work that address one specific objection.
Advocates of the applicability of Marx almost invariably (as far as I can tell so far; I don’t claim to have read anywhere near all of this literature yet) base their case on Italian autonomism. Autonomism is best known in the work of Antonio Negri (with or without Michael Hardt); important in this context are also Paolo Virno, Maurizio Lazzarto and Franco Berardi. Autonomism breaks with classical Marxism both insofar as it breaks with any sort of economic determinism (by shifting the base to class struggle, and not means of production), as well by advancing two more immediately applicable ideas. First, autonomism adopts Marx’s “Fragment on Machines,” an unpublished (by Marx) and fragmentary set of notes to the effect that capital will be relying increasingly on accumulated science and knowledge (Virno has what is probably the best synopsis of this account of the “general intellect”). Second, autonomist theory argues that capitalist relations have extended beyond the boundaries of the workplace to encompass all social relations. In Negri’s terms, we now face the complete subsumption of society by capital.
Combinations of these theses are at the heart of three papers central to the current discussions of Marx and big data. Because these papers are from the early 2000’s, they tend to reference Negri and a book in which Lazzarato, like Latour, thinks we have a lot to learn from the 19c sociologist Gabriel Tarde. Berardi’s Soul at Work (2009) and Lazzarato’s later work were written after the papers here. In the first, Tiziana Terranova (2000) essentially argues that the entire Internet functions as a machine of surplus value extraction: individuals voluntarily, and in their free time, contribute value to the Internet in the form of blog posts, web surfing, and so on (now of course also via social media). But they receive nothing in return, no pay: all the money goes to the companies that own the software and hardware of the Internet.
The second, Scott Lash’s “Power after Hegemony” (2007) pushes the argument that power is now immanent to society and social processes, rather than imposed from the outside. Hence cultural studies, he argues, needs to switch its dominant narrative away from one about hegemony in the Gramscian tradition, which conceptualizes power as transcendently imposed onto the social order. In its place needs to be the assumption that power is immanent to cultural processes, a point that Lash grounds in Negri’s reading of Spinoza’s distinction between potentia (power as ability) and potestas (juridical power). Lash acknowledges that one of the risks of this shift is that cultural critique will often become difficult to distinguish from the culture industries, as the immanence of power makes it difficult to adopt a point outside of it for critique (this territory has also been ably covered by Hardt and Negri’s critique of postmodernism in Empire). Lash similarly emphasizes the importance of networked communications, and the displacement of critique by performativity:
“At stake is not reproduction but the chronic production of society that is necessary in a fluctuating environment. Thus we can understand the contemporary importance of Carl Schmitt and his doctrine of the rise of the executive function and the ‘decision.’ We live in an age of such chronic decisionism: one in which legality as a mode of legitimation is displaced by performance. Hegemony – and the symbolic – are effective through meaning. Communications work through performativity. Legitimation is no longer separate from what it is meant to legitimate, it becomes automatic” (66)
So power is in the flows of information, and in the various structures that regulate and modify those flows. As a consequence, the loss of a standpoint for critique is replaced by the politicization of everything (as an aside, there’s no mention to Judith Butler’s earlier work on performativity and iteration, but I think she is saying some of the same things, and the way Butler talks about iterability I think improves in at least one way on Lash’s discussion: iterability builds in the notion of difference into every performance).
Finally, Nigel Thrift’s “Re-inventing invention” argues that capital is trying to increase innovation by fusing knowledge and life, a point he grounds in Virno. More specifically, capital is “seeking new ways to squeeze value by amplifying the rate of innovation through a general exteriorization of intelligence out from the corporation, in turn redefining what counts as value” (282). One will immediately recognize both the complete subsumption thesis, and the move to the “general intellect” from Marx’s “Fragment on Machines;” as Thrift puts it, capital will be “mobilizing forethought.” I have remarked elsewhere that the boundaries of corporations are likely to blur and shift as data becomes something more cheaply outsourced in a neoliberal economy. That is one way we can expect to see not just the extension of capital into society more broadly. Thrift identifies a broader process of an extensive effort to bring consumers into the process of consumption, both in terms of engaging their affects and in blurring the production/consumption line. As a result, the process of design and commodification becomes continuous: there will not be a finished industrial product at the end of the process, but rather a continuously modified and modulated commodity, the exact nature of which is continuously revised to adapt to consumer tastes and preferences.
Based on my admittedly unscientific reading of footnotes, these three papers are central to the emerging literature offering a Marxian account of big data and how it extracts value. What they share in common is their adoption of a heterodox account of Marx. Against this, John Michael Roberts (2016) argues that these arguments fail as Marxian both because they mischaracterize the results of prosumer activity as surplus value, and because they miss the Marxian connection to time (he is critiquing specific papers that came out after the ones summarized above, but they are both broadly autonomist). In the first case:
“These all might very well provide a service to productive capital through their surplus labor but they do not create surplus value because they do not add new value in the production process to an initial outlay of money-capital and productive capital. Only productive labor creates commodity-capital, which is then sold for a greater sum of money-capital than originally invested”
In short: prosumption does not add value to the production process, and so it cannot be considered productive labor. Since it is not productive labor, it cannot generate surplus value. I want to leave this point for the moment.
On the second point, efforts to tie labor to individuals miss that, for Marx, labor needs to be tied to an abstracted amount of time that it generally takes to produce a commodity. Individual efforts may fall either above or below this average, but the exchange value of a commodity is tied to the socially average amount of time it takes to produce it. The individualized, time-intensive activities of prosumers precisely fail here because there is no notion of socially average labor, only specific embodiment, and Roberts argues that the prosumer argument is much closer to Ricardo than to Marx, and was even the sort of theory that Marx specifically critiqued.
There are a couple of things one can say here, beyond a wariness about being too scholastic in applying Marxian theory to post-industrial economies. The first is to push the notion of time, and argue that post-industrial time is, phenomenologically, either faster than industrial time or irregular. Marx is pretty clear in Capital Vol. 1 that it is the steam engine and the production of physical objects that structure the temporality of life in industrial capital, so it’s not at all clear that that notion of time would import into a post-industrial economy.
Perhaps more interestingly, it seems to me that there is another Marxian answer, one that should also help in understanding how autonomist theories account for the creation of value. In volume 3 of Capital, Marx describes the M-M’ circuit of interest bearing capital (I’m quoting the Lawrence and Wishart edition. It’s chapter 21). In this, the lender treats money as a commodity, the use value of which is its ability to become capital and create surplus value. This feature distinguishes money from other forms of use valuel, however:
“In the case of the other commodities the use-value is ultimately consumed. Their substance disappears, and with it their value. In contrast, the commodity-capital is peculiar in that its value and use-value not only remains intact but also increase, through consumption of its use-value. It is this use-value of money as capital – this faculty of producing an average profit – which the money capitalist relinquished to the industrial capitalist for the period” (351)
The distinction is not quite the same as between rivalrous and non-rivalrous goods, since simultaneous uses of the money-capital are impossible, but there is a diachronic analogue: you can use the same money over and over again, and it doesn’t wear out. It just returns surplus value. From the point of view of the lender, the cycle is M-M’, where M’ is the original loan plus interest.
Importantly, Marx immediately compares this situation to “labor-power in its relation to the industrial capitalist. With the difference that the latter pays for the value of labor-power, whereas he simply pays back the value of the loaned capital.” In other words:
“The use-value of labor-power for the industrial capitalist is that labor-power creates more value (profit) in its consumption than it possesses itself, and than it costs. This additional value is use-value for the industrial capitalist. And in like manner the use-value of loaned capital appears as its faculty of begetting and increasing value” (351)
From the point of view of the owner of the commodity in question, then, two points. First, neither lending capital nor labor-power need be considered in relation to time. The important point is that both produce value above-and-beyond their initial state by being understood in terms of their use value. It may not be quite right to say they produce surplus value in the technical sense (although the M-M’ circuit requires that whoever invests the loaned money has to ultimately invest in labor to generate surplus value; much of the discussion hinges on one’s level of abstraction, something Marx is very careful with), but they both produce additional exchange value as a function of their use value.
Second, the argument helps to paint a picture of what the prosumer on his Facebook page does. Like loaned capital, he adds value to the site that, from the point of view of FB, spontaneously appears. From the point of view of Facebook, that is, the user functions like the use value of lending capital: a commodity that has the delightful feature of both generating increased value and not being used up in increasing that value. This is why social media is so attentive to increasing the stickiness of its sites, to keeping consumers invested in their participation in the sites, and so on: a user who quits the site is “used up” from the point of view of the site. Indeed, if anything, the prosumer becomes more valuable each time he returns to the site, because his affective attachment to it goes up (even if he doesn’t like it, he has more of his personality invested in it for something like a blog or FB. This would perhaps be the correct entry point for a discussion of surplus value creation: the investment of personality is always somewhat alienating in that it returns less value than the effort generates for the site. Because this requires the unusual move of reading Hegel and alienation into the account of surplus value, it's a topic for another day). So prosumption may be a poor fit for the industrial model of capital, but (a) the discussion in Capital Vol. 3 suggests that Marx does not mean for the description of factory production in Vol. 1 to be the final word on value in capital, and (b) the analogy – and that’s all I’ve established here, and that mostly at the proof-of-concept level – between finance and presumption suggests a hypothesis about the co-emergence of both in neoliberalism.