the universal induction

18 december 2020

hunter, stepan ryabchenko.[0]

this paper was presented at a conference; you can view a recording of my talk here.


The array of social functions which comprise “big data” and their subsumption of various activities of private life into the realm of abstract human labor (colonization)[1] – i.e., value-bearing physiological expenditures of energy – have presented a formidable contradiction to the concept of the juridical individual whose original property is their labor, eroding the legal-contractual relations which historically define worker-capitalist dynamics.

I intend to develop an analysis to this point first by establishing an operative Marxist concept of the category of labor [i], to be employed in a Marxist problematic for data production [ii]; this will then continue into a discussion of the broader erosion of the legal-political standing of the worker via the expansion of the data production labor process [iii]. These strands will then form a more cohesive thread with a discussion of literature and conclusions towards the questions raised, thus introducing a Marxist critique of the contemporary environment of critical data studies [iv].


The necessary connectedness of contemporary capitalist societies makes interaction with the broad domain of “online” functionally compulsory. Scrolling on Facebook, the movements of your cursor, those things which you digitally manipulate, your objective, online existence – these minuscule mechanical actions constitute engagement, and are persistently logged in near real-time, instantiated in microscopic data points aggregated and assimilated in macro, sold as business insights (their utility being reconfiguration of the marketing process, seizing upon market shares) in an array of processes which Kitchin and Lauriault seminally know as data assemblages,[2] though which might be better known as big data,[3] a composition of functions, processes more than a concrete thing.[4]

Big data, following the precept laid out by Adorno that “… concepts ‘which are basically short-hand for process’ elude verbal definition,”[5] is difficult to pin down, and typically nebulous in its usage; nonetheless, data mined and sold in aggregate tends to be of massive market value, Google AdSense in 2019 reporting USD$134.81 billion in revenue. Imperatively, though, AdSense only presents in itself a general capital outlay; it comprises an array of digital surveillance apparatuses, an online infrastructure which tracks qualitative phenomena – e.g., your multi-variably determined level of interest in an advertisement – converting it into discrete, cardinal variables. AdSense as such might be the means of producing data, but it is not autonomously forming data points; machinery does not produce new value:[6] so what produces data?

This paper intends to confront with a Marxist theoretical framework the question of the human production of data, and the broader conclusions to be drawn from the expansion of digital data production outwards, implicating, simply, all of us – disconnection from a connected society being functionally impossible.[7] The induced intention to produce data (i.e., the intention of the capitalist to produce data through the worker as proxy) is universal, totalizing: no offline.

i. Marx, on the historical category of labor

Labor, or the capacity for it, plainly might be understood by Marx as “… the aggregate of those mental and physical capabilities existing in the (…) human being (…) set in motion whenever he produces a use-value of any kind”[8] (underscore mine). The broad concept of labor here is one which is firstly abstract, as such – it refers not directly to the physiological expenditure of energy by the concrete person, but is instead an abstraction from that mechanical energy as realized in the use-value (abstract utility) of the products of labor, i.e., commodities.[9] Rather, abstract human labor is not actualized in the pure act of mechanical energy expended when manipulating an object – it emerges principally in the form of the commodity, which contains a use-value and an exchange-value (equivalence in a determinate quantity of other commodities). The mechanical energy of a person expended towards manipulating an object emerges as abstract human labor with the instilling of use-value in its object, and thus exchange-value – i.e., in becoming a commodity.

These precepts of the concept of abstract human labor suggest an understanding of labor not as a definite, trans-historical concept, but as a historically variable/conditioned category, whose particulars emerge as exercises of mechanical energy intended towards the production of commodities. The category of abstract human labor, and its subsumption of mechanical energy as a value-bearing thing, requires a degree of capitalist relations to have emerged – an outlay of functions which constitute a market, a place where the commodities have realized circulation – the emergence of the commodity form, and as such the basal category of abstract human labor necessitates at minimum primitively developed relations of capital, primitive accumulation.[10] “… the product of labor is an object of utility in all states of society; but it is only a historically specific epoch of development which presents the labor expended (…) as its [the object’s] value.”[11]

To illustrate: a particular expenditure of mechanical energy subsumed by this category of abstract human labor, then, might be most apparent in the workplace: e.g., you enter the workplace, and past a determinate threshold your (contractual) job is the production of tennis shoes – there is a largely fixed point of production. You are expending mechanical energy: either ensuring the rubber enters the mold correctly, or affixing the soles, or checking and then packaging the tennis shoes. Your work represents a labor process towards a definite, manipulated product with a definite use-value crystallized in the utility of the shoes; it is the historical conditioning of capitalism which endows it with exchange-value. Marx here writes: ‘… all labor is, speaking physiologically, an expenditure of human labor power, and in its character of identical abstract human labor, it creates and forms the value of commodities,” which continues, then, with a phrase particularly difficult for problematizing big data: “… all labor is an expenditure of human labor-power in a particular form and with a definite aim.”[12]

ii. A Marxist problematic for data production

“… if data is a commodity it has value, and if it has value, it is necessarily the product of labor,”[13] Rogan posits in “The Universal Factory,”[14] introducing a model for understanding the data-commodity as proceeding from Thatcher, O’Sullivan, and Mahmoudi.[15] Forming a Marxist concept for the production of data, given the notion of labor as a historical category containing temporally conditioned particulars, seems at surface to be simple: an expansion of the well-defined labor process “category” to include a new dimension of particulars. A “novel” array of mechanical energy expenditures acquire determinate use-values, a process of data colonialism.[16]

The physical act of browsing, e.g., Facebook and pausing on an advertisement long enough to “stimulate” Google AdSense towards the completion of the production of a data point[17] provides a microscopic crystal of value in that data point, comprised of the again, minuscule quantity of mechanical energy expended as labor (scrolling, perceiving), with the larger portion of instilled value derived from investment in the array of perceptive apparatuses (e.g., AdSense) as capital. Your mere existence on the internet is providing unwitting work towards the process of market share acquisition – constructing and being constructed by data assemblages in an iterative process of development, known to Ian Hacking as dynamic nominalism.[18] “… in this framework, [we] are not so much incorporated into the data assemblage as much as we work and are worked by it.”[19]

The problematic component of this model for understanding might be the question of definite aim, or intentionality in the data production process; the apprehension of mechanical energy by an array of perceptive apparatuses, crystallized in data does not present an apparent point of definite, intended production as an enclosed workspace might. From Marx: “… the use-value of every commodity contains useful labor, i.e., productive activity of a definite kind, carried on with a definite aim.”[20] Facebook does not present itself as an apparent factory – scrolling, perceiving does not present itself as a definite, intended form of production.

Though the definition laid out by Marx of abstract human labor and its exercise lends itself to interpretation as intention within the worker/self towards manipulation of their object, the social-historical functions of capital present a more induced intention: the intention of the worker producing, e.g., tennis shoes being the exercise of the intention of the capitalist by proxy. The labor exercised during a workday by the worker-producer might be towards the production of tennis shoes, intended as such, but this intention is the determination of the capitalist, exercised through the worker upon the contractual sale/purchase of their labor.[21]

This note of contractual is particularly imperative to the data production process: “… the purchaser of labor-power consumes it by setting the seller of it to work,”[22] Marx states, the language of purchase implying an apparent legal relation emergent between purchaser and vendor, i.e., capitalist and laborer, respectively. The second problematic component with the construction of a Marxist model for the production of data then is the notable absence of a substantial contractual relationship between the purchaser (i.e., the owner of that array of capital which apprehends via “perception”) and the vendor (the user, whose labor is in the first instance engagement). Inherent in the labor process as presented in data production is the lack of informed consent, as well as the capacity for it, analogous more so to a form of colonization than an open market sale: a dispossession, more than a purchase.[23] The question posed for us here is: what does this extra-legal labor process entail for the broader (legal) relations of workers and capital?

iii. The juridical individual and data

The particular characteristic of the worker among the productive relations of capitalism which presents them as a codified individual, i.e., an abstract, legal human, is therefore their ownership of “original property” as labor: “… every man has a property in his own person: this nobody has any right to but himself,” Locke posits – “… the labor of his body, and the work of his hands, we may say, are properly his.”[24] Though Locke’s definition of original property remains ungrounded in concrete relations, broadly the notion of (abstract human) labor as that property which renders the individual an abstract, legal human – juridical individual – proves useful in understanding the legal position of the worker.

Though the intentions of the worker are necessarily induced – i.e., the worker as proxy – the entry into a vendor-purchaser relationship on the labor market presents some degree of legal consent, regardless of the circumstantially compulsory nature of work. Data production (extraction), in contrast, presents no noteworthy capacity for consent, with no particular bounds or thresholds entered, besides existence in a connected society;[25] both the determinate point of production, and the apparent legal relation foregone.[26] Facebook does not present itself as a factory, scrolling does not present itself as apparent work, and AdSense does not extend the pertinent worker a contract, or wage, though the data-commodity (in aggregate) produced demands an outsized exchange-value.[27]

Thus, the presence of an apparent labor process here – the production of a data-commodity through a minuscule expenditure of mechanical energy, apprehended by an outlay of perceptive apparatuses (capital) – does not find a legal instantiation of itself, i.e., a codification. The unwitting extraction of the value of labor, then (or the extraction of a particular use-value from the mechanical energy of a person’s connectedness) presents itself as extra-legal. The laborer clearly retains some degree of property as labor which is alienated via data mining, i.e., a self-contained property, though this property is entirely legally undefined, thus becoming a sort of commons by dispossession.[28]

iv. The universal induction and regulatory infeasibility

The dual subsumption of the physiological expenditures of energy which comprise general existence, engagement with connected society by the historical category of abstract human labor – and the absence of a legal realization of the propertied individual as such presents significant questions both for privacy, and for the legal-political position of the worker. Leszczynski states directly: “… foregoing all forms of digitally mediated communication [is] an infeasibility of living in connected societies.”[29] Thus, all persons existing among capitalism, and all persons in connected society – myself, you, everyone you know – are all implicated in the instillation of use-value as business insights in our very existence, and the extraction of this use-value unwittingly, constantly, in near real-time.[30] All are induced towards the intention of data production, acting as proxies: a universal induction.

The literature here (and the internal logic of this theoretical framework) presents an apparent conclusion: connectivity is totalizing, eroding the defined point of labor/production incidence – and as such, this connectivity through perceptive apparatuses forming a capital outlay reflects the well-defined principle of capital as totalizing. Rogan presents the extraction of your data in the smart city as “… working upon the machinery which is the city”[31] – beyond the boundaries of technological urbanism in particular, surveillance is thus unwittingly, extra-legally working upon the machinery which (are) the productive functions which comprise capitalist society, from which there is no substantial self-removal: “… personal control over information is technically impractical.”[32]

What can then be done about this extra-legal extraction of labor? The more classical liberal refrain towards the question suggests well-defined property rights in the instance of data, i.e., the codification of determinate bounds on the unwitting extraction of use-values from the person would then resolve the colonial dynamic of data dispossession.[33] The infeasibility of imposing regulations of this type, though, (i.e., worker-side) should be readily apparent. The principal component of the lucrative market of data extraction is unawareness, i.e., sans consent – though even were there a clear threshold of “consent” to data extraction, the hieroglyphic working & subsuming character of data assemblages[34] renders consent – much less informed consent – functionally impossible.[35]

Inversely, the more so social-democratic (i.e., statist) model for remedying the extra-legality of data production tends to then suggest, as Carissa Véliz does in the recent (2020) “Privacy is Power,” a regulatory regime on the side of capital.[36] The extensiveness of this regime is variable, Véliz suggesting an altogether abolition of the data market – although this suggestion perhaps, as well, misapprehends the inherent functions and precepts of data production; it is a similar infeasibility. Though a Marxist analysis might tend one towards a similar conclusion (the necessary abolition of the exchange of data as a commodity), Adorno, following Hegel declares: never assign to the part that which belongs to the whole.[37] Data as a crystallization of labor might be understood as a particular within a historical category (abstract human labor). Thus, deconstruction of the data market would be more alike to a post-hoc regression of the development of an extant stage of capitalist productive relations, than it would be a foreclosure upon the conditions which necessarily gave rise to the stage.

Our model, then, does not demand in its conclusions the freeing of labor from the particular inductions of data extraction – it rather reflects more decisively on the broad character of labor, and the capital-driven degradation of the juridical-political position of the worker, a function of the historical contradictions between boss and worker.[38] The colonial subsumption of the mechanical energy of engagement under the category of abstract human labor, that is, makes apparent the broader want of capital towards control, and towards a regression of the distinct legal and political position of the worker. The conclusion, then, cannot be a reconstruction of the bounds of labor, or those particulars which comprise the labor-category as it is realized – it must be a disposal of the category and its preconditions altogether, which itself constitutes the conditions under which the data market arises: “… labor is free in all civilized countries; it is not a matter of freeing [abstract human] labor, but of abolishing it.”[39]


The abstraction of use-values from the physiological expenditures of energy by a human, crystallized in a commodity constitutes the labor process – and this process, in a more or less opaque form, presents itself in the broad array of functions which comprise the production of data. Data production subsumes that particular realm of minuscule mechanical human energy – typically in the form of engagement, i.e., interaction with the connected environment – into the category of abstract human labor, homogeneous and historically variable; working upon the capital outlay of surveillance, constituting a primitive accumulation.[40]

The general development of capitalism as such, then, gives rise to the worker whose original property consists in their own labor.[41] The retention of their original property then places them among the social-historical relations of capitalism, compelled towards participation in the labor market,[42] which, by dint of the contractual (i.e., legal) nature of the purchase and sale of labor renders the worker not only an abstract human laborer, but a human in the abstract as a juridical individual, whose rights via property are actualized in law, with determinate point(s) of work. The thresholds of the workspace, and then private life, even if not necessarily spatial, are historically well-defined.

Data production, then, constitutes an erosion of the threshold of the workspace via the extraction of use-values from all general physiological expenditures occurring in the private sphere, apprehended by perceptive apparatuses: a colonization of private life from which there is no self-removal.[43] It presents, as such, an extra-legal labor process, whose vendors (the laborers) maintain no real codified status presently, and are broadly non-consenting – and entirely unwitting – to their performance of work and extraction of use-values therefrom, the use-value being business insights: i.e., working, unaware, upon the assemblages which in turn work you.[44]

Thus, the universal induction which data production constitutes: the implication of you and everyone you know in the realization of the intentions of capital, the boss by proxy via unwitting data production; and the infeasibility of regulating a market which presents itself as an effectively natural outwards development of capitalism present, in conjunction, imperative conclusions on abstract human labor. Among them, that the nature of intentionality within the context of commodity production is necessarily induction, intention by proxy – and as such, that the tendency the labor-category is a self-expansion, a colonization of a broader range of particulars; i.e., the principal historical-political contradiction between the capital and worker (individual) arises as the desire on the part of the boss to degrade the historical-political subjectivity of the worker.[45] This line of inquiry and Marxist problematic for “big data” reveals the infeasibility of simply regulating the primitive accumulation which data production constitutes – the parts constitute a whole.

mao's silhouette, yu youhan.


[0] All images c. WikiArt.

[1] Jim Thatcher, David O’Sullivan, and Dillon Mahmoudi, “Data colonialism through accumulation by dispossession: New metaphors for daily data,” Environment and Planning D: Society and Space (2016), 1 – 17.

[2] Rob Kitchin and Tracey Lauriault, “Towards Critical Data Studies: Charting and Unpacking Assemblages and Their Work” (2014).

[3] Andrew Iliadis and Federica Russo, “Critical data studies: An introduction,” Big Data & Society, (2016): 1 – 7.

[4] Cornelius Puschmann and Jean Burgess, “Metaphors of Big Data,” International Journal of Communication 8, (2014): 1690 – 709.

[5] Theodor Adorno, “Society,” in Critical Theory and Society: A Reader, ed. Stephen Bronner and Douglas Kellner (New York: Routledge, 1989), 267.

[6] Karl Marx, "Capital: Volume I" (London: Penguin Group, 1990), 508 – 17.

[7] Agnieszka Leszczynski, “Spatial big data and anxieties of control,” Environment and Planning D: Society and Space (2015), 970.

[8] Marx, Capital: Volume I (1990), 270.

[9] Ibid., 131.

[10] Ibid.

[11] Ibid. 153 – 4.

[12] Ibid., 137.

[13] Ibid., 125 – 38.

[14] Kevin Rogan, “The Universal Factory: Data Production and Platforms,” Enquiry (ENQ) 16, no. 2 (2019), 23.

[15] Thatcher, O’Sullivan, and Mahmoud, “Data colonialism” (2016).

[16] Ibid., 1 – 2.

[17] “About Engagements Reporting,” Google.

[18] Kitchin and Lauriault, “Towards Critical Data Studies” (2014), 8.

[19] Rogan, “The Universal Factory” (2019), 23.

[20] Marx, “Capital: Volume I” (1990), 133.

[21] Ibid., 283 – 93.

[22] Ibid., 283.

[23] Thatcher, O’Sullivan, and Mahmoud, “Data colonialism,” (2016).

[24] John Locke, "Second Treatise of Government" (Indianapolis: Hackett Publishing Company, 1980), 11.

[25] Leszczynski, “Spatial big data and anxieties of control” (2015), 970.

[26] Thatcher, O’Sullivan, and Mahmoud, “Data colonialism” (2016).

[27] “Google: annual advertising revenue 2001 – 2019,” Statista.

[28] Thatcher, O’Sullivan, and Mahmoud, “Data colonialism” (2016).

[29] Leszczynski, “Spatial big data and anxieties of control” (2015).

[30] Thatcher, O’Sullivan, and Mahmoud, “Data colonialism,” (2016).

[31] Rogan, “The Universal Factory” (2019), 23.

[32] Leszczynski, “Spatial big data and anxieties of control” (2015), 971.

[33] Jeffrey Ritter and Anna Mayer, “Regulating Data as Property: A New Construct for Moving Forward” Duke Law & Technology Review 16, no. 1 (2018), 220 – 77.

[34] Kitchin and Lauriault, “Towards Critical Data Studies” (2014).

[35] Carissa Véliz, Privacy is Power: Why and How You Should Take Back Control of Your Data (London: Bantam Press, 2020).

[36] Ibid.

[37] Adorno, "Society" (1989).

[38] Karl Marx, "The Germany Ideology" (New York: International Publishers, 1996).

[39] Ibid.

[40] Marx, "Capital: Volume I" (1990), 873 – 877.

[41] Locke, "Second Treatise of Government" (1980).

[42] Marx, "Capital: Volume I" (1990), 272 – 273.

[43] Thatcher, O’Sullivan, and Mahmoud, “Data colonialism” (2016).

[44] Kitchin and Lauriault, “Towards Critical Data Studies” (2014).

[45] Marx, "German Ideology" (1996).