this paper was published in the second volume of the johns hopkins richard macksey journal; you can access it here.
Critical data studies present a new endeavor, consisting in an object and a methodology for comprehending that object. The methodology is then a critical, theoretical prism – the object is the massive collection or accumulation, and assemblage of human data: big data. (Iliadis and Russo, 2016)
Big data, following the precept laid out by Adorno that “… concepts ‘which are basically shorthand for process’ elude verbal definition,” (Adorno, 1989) is then difficult to situate, an array of functions and relations more than a particular thing. (Puschmann and Burgess, 2014) The array here then characterizes not just the online or virtual: social media, digital engagement with various platforms – but as well, your measurable existence everywhere, your (unwitting) engagement with a nexus of surveillance apparatuses, perceptive things presiding over spaces. (Leszczynski, 2015) Big data affects you in the form of cameras on the street corner, the geospatial topographies enabled by cellular connectivity, the point-to-point inputs at determinate thresholds which facilitate contact tracing. Should you remove yourself from these watched spaces altogether (a demonstrable infeasibility), your sufficient proximity to someone unremoved involves you via inferencing. (Leszczynski, 2015)
Allow us to anatomize – big data is defined first by you: you perform some physiological expenditures of an indefinite quality in some space. These are initially distinct, differentiated: the expenditures are of different natures, heterogeneous. These expenditures then provide inputs for some perceptive thing: a camera records you, Google AdSense quantifies your interest (Google, n.d.), Facebook pixel tracks the granular motions of your cursor (Hootsuite, n.d.). These recorded, minuscule, and qualitatively differentiated expenditures are thus instantiated as quantitatively undifferentiated data, discrete and operable, stored in aggregate, and typified therefrom. (Puschmann and Burgess, 2014) Descending a level descriptively: you do something, it is measured by something, it becomes instantiated as data – this is the basic component of our object of study (Kitchin and Lauriault, 2014). As with any new study then, our first problem is proving the reality of the object, the import with which it is invested. The formulated question: what makes this [big data] meaningfully new?
I intend to demonstrate first that our main answers to this question, here arbitrarily distinguished as X and Y, are largely deficient – and therefrom posit Z, a Marxist formulation towards the question of data and the encounter: where the induced work on datasets is both new and analogous. Big data here constitutes an economic outgrowth precedented, and an ideological transformation of the particular mode of relations – with it, a defunction of the conceptual and relational tools of humanism. Our demonstration is then premised on an appendix to our original anatomy: sale – where the definition of big data is only coherent if it is formulated on the basis of data becoming a commodity, and thus a product of labor (Thatcher et al., 2016). Here, those physiological expenditures of our original anatomy are differentiated both in their qualitative character, and now as having no relation as equivalents on the market. Their undifferentiation as labors emerges through the prism of exchange, through sale: first in their transformation into discrete variables, homogeneous, though principally in their presence on the market as mutual equivalents, related to other objects and money. I should thus like to show where big data verifies our theoretical formulae towards these relations – and where it surpasses them.
The New Formulation
Position X would then hold that the newness of big data consists in a profound upheaval of the structures among which it is situated: big data reimagines the biopolitical, big data restructures the rhizome, big data is something other than capital, newer. As McKenzie Wark suggests (for the third time reiterated) in “Capital is Dead,” (2021) published by Verso Books: “… surely one has to at least entertain the thought experiment that this [big data] is no longer capitalism at all.” (p. 11) Here, the old formulation has been surpassed: capital is retained only as a residue on the new functions of big data.
Position Y therefore holds the contrary, parrying the question: the newness here is minimal, but useful. Y defers to the extant logic, intending through a backwards-reaching gesture to illustrate the veracity of the critical methodology through the temporally new but theoretically precedented prism of big data, data studies. Here, the same modes apply, the new engagements a neat development of a linear, mechanistic, and forward motion of capital. Capital is here one temporality drawing outwards, undulating, where big data is the newest (clarifying) crest. Big data is thus colonialism, really and residually; (Thatcher et al., 2016) big data is social physics, is capital.
As a synthesis, I should like to suggest position Z: big data does not rearrange the structure of structures, constituting some new assemblage of the whole inventory of standing social functions – but big data does not either cleanly make some extant, opaque universal clear through itself as an inert discursive tool. Big data is capital, though the relation here is defined by capital as an animating but distant logic to social functions, the particulars of which are historically determinate. Capital here expresses its logic as differentiated, developing relations – rather than as a subsuming and proximate trans-historical category apparent in its parts.
Big data is capital inasmuch as it consists in the expansion and development of the logic of capital, through an historical, material encounter between two parties or temporalities. (Althusser, 2006) The first temporality is the productive modes of capital; the other the plane of otherwise differentiated expenditures, a different productive mode. The encounter here presents an analogue to precedented encounters, the basal functions well accounted for in theory, against the main decrees of position X. Though against position Y, the real model of an encounter defies reduction to a linear strand of capital growing outwards, of which big data is supposedly illustrative. The ideological inversion whereby the encounter of big data surpasses the humanist language of the laborer as subject, developing and inverting the real humanist mode of relations, presents something really new – a real disanalogy for “new metaphors.” (Thatcher et al., 2016) To begin, let us illustrate the basic component of our study through a known metaphor: the production of data as a commodity.
The Data Commodity
“… If data is a commodity, it has value, and if it has value, it is necessarily the product of labor,” Rogan posits in “The Universal Factory” (2019). We might use the example of Google AdSense to illustrate the proposition here.
Our first premise: i) data mined and sold in aggregate – or subsumed as a component part of other commodities through marketing – tends to be of massive exchange value. Google AdSense in 2019 reported USD$134.81 billion in revenue (Statista, n.d.). ii) Google AdSense as such only constitutes a general capital outlay – an array of embedded digital surveillance mechanisms, an online infrastructure collecting your multi-variably determined e.g., level of interest in an advertisement (or types thereof.) iii) AdSense is then the mode of data production, the mechanism through which production of business insights happens, though the outlay does not autonomously create data points. The data would be useless, and moreover unprofitable if it did. Machines instill a portion of their fixed value in their outputs, but no new value is added (Marx, 1990). Therefore: something is engaging the machinery, capital to produce the data commodity – this something is you, performing microscopic, qualitatively differentiated physiological expenditures then instantiated in data.
How do you (or I) produce data, then? To elucidate, let us return to the former portion of our original anatomy: you perform some physiological expenditure of an indefinite quality in some space. Standing outside, sitting on a bench; scrolling on your phone, clicking a mouse – these constitute physiological expenditures, differentiated, with unique qualities or intentions each: you want to stand, you intend to sit, you want to scroll. The expenditures here contained thus present at first (before apprehension) a utility, an intention; they are principally the attainment to some will and relate to the world and other expenditures as such (Marx, 1990, pp. 140 – 54).
The transformation of this relation of (deficiently defined as qualitatively) differentiated physiological expenditures into quantitative, undifferentiated labors then consists in the moment of our first premise here, the appendix to the next portion of our original anatomy: sale, following the expenditure constituting an input for some perceptive thing. Standing outside, sitting on a bench is watched by some surveillance camera; scrolling, clicking is watched by some digital infrastructure, e.g., Google AdSense, Facebook Pixel.
These minuscule expenditures are instantiated in data, as is well established. Though this data is then, in the act of sale, brought into relation with some other thing through which its value is expressed, into relation with its equivalent in the form of money or other objects. Here, the principal attribute of the relation is now exchange, and thus a mediated value. Your microscopic, differentiated expenditure is now an undifferentiated labor, a value-bearing thing as objectified in data related to other value-bearing things, and relates to the world and other labors as such (Marx, 1990). These expenditures enter the logic of the labor process of capital.
Therefore: the physiological expenditures, differentiated, are then transformed unwittingly into undifferentiated labors instantiated in data as a commodity (or as the component part of a commodity e.g., business insights), through a dispossessive perception (Thatcher et al., 2016). The relation in its basic productive logic, though confounded by the nexus of surveillance apparatuses and the opacity of their presence, is then fundamentally analogous to (or a metaphor of) the labors done in relation to more apparent capital – assembling tennis shoes, digging at lithium, plowing a field. The general relation here is well-contoured, theoretically. Differentiated mechanical energy is exerted and crystallized in some object; the object, a commodity, is related to other objects as its equivalent on the market. The initial energy is now an undifferentiated thing, known through the backwards-gesturing prism of the value it binds up in the commodity (Marx, 1990).
The alignment of big data here with the known concepts, formulae, then contradicts position X, and partially validates Y. We discover our first difficulty in situating Z here, where big data is (supposedly) easily defined as a new particular of the well-defined category and extant logic of commodities, commodity production. To defy the reduction of big data to an inert descriptive tool emergent from a linear development, while straying from the misguided introductions of a new system altogether, we then might look first to the prehistory of big data. The labor gains a value in its instantiation as data related to other things, but how does this relation emerge? The formulation here presupposes an array, an outlay of capital, a productive moment in motion: what is before the array?
Big Data and Primary Accumulation
The preceding illustration of data production descriptively employed the moment before a physiological expenditure is perceived, apprehended – the point zero of the sequence which constituted our original anatomy. Functionally and really, this moment is nonexistent as described. The very situation of your physiological expenditures among perceptive things, their output data related to equivalents, integrates the expenditure as labor. Assuming the perception happens, the expenditure was always labor, a potential realized with the same motion it is evaporated (Marx, 1990). If the perception does not happen – assuming the outlay, this is mostly a matter of spatial location – the question of point zero is irrelevant. The potential is functionally nonexistent having not engaged something to realize both the potential’s existence, and the objective output of the potential (Marx, 1990). We are then met with a difficulty: not where, as detailed here, but when are or were these expenditures not labors – anywhere? We look then to the moment of accumulation here emerging: the prehistory of big data.
First: position Y to this point brings out the language of subsumption (Thatcher et al., 2016). Here, Y presents a residue of an eschatology – one which transforms the critical methodology into a critical teleology. Subsumption posits an analogy between the logic of big data and the logic of colonialism, premised on a formulation of colonialism whereby the accumulation of capital – the broadening of markets – is a subsumption of active things. The subsumption is an integration of active things as inert components of a category immanent about them, proximate and expressive. The relation here is for Y a simple one: not yet capital become capital, mechanical energies become labor, subsumed by a category to which they were already related, forming an identity outside of themselves as capital.
Colonialism here, and thus the logic of primitive or primary accumulation, is the mechanistic outward development of a linear strand of capital realizing its purpose, telos. Supposedly, markets will of necessity expand such that capital will, as a category or identity, encompass the whole of physiological expenditures, mechanical energies – capital is searching always for new spaces of accumulation. Therefore, the exhaustion of these finite spaces is the subject of an eschatology – it is an eventuality, the arrival of which will bring an evaporation of capital altogether, defuncted, the world to come become at last (Haider, 2020). Big data is then simply slotted away, a developmental moment in the broader strand or sequence. It is the newest space subsumed by the category – it is capital, it is Revelation, it is the fourth bowl’s austere sunlight spilling outwards (Revelations 16:8 – 9).
Presenting the unique dimension of big data among the context of the productive relations of capital then requires first a basic rejection of the linear model of history. It is a model whereby capital is presently universal being, (a mechanical motor of history) everything else becoming. Our presentation is premised on the basic denial of the universalized capital, everywhere becoming or attaining to capital in the same capacity that a planted bulb in proper conditions is becoming – and functionally is – an onion, though still apparently a bulb. The rocks and trees and relations of people do not shout out loud: capital!
The basic rejection of the linear model should then bring with it a rejection of the critical eschatology, which renders big data a simple inert component of an effectively theological proof. We ought to turn instead to the model of the encounter, of primitive or primary accumulation, and therefore of different and distant histories and logics synchronized (Haider, 2020). Here, the prehistories of capital constitute separate temporalities altogether, whose tendencies and logic are interpolated and thus synchronized by an encounter with another. It is an encounter which attains only to the contained logic of either party to the encounter, rather than an outside tendency or spirit. The encounter of physiological expenditures in space by the outlay of perceptive apparatuses does not constitute an identity outside of itself as capital – it is the developing logic of the former becoming in the latter, the constant constitution of the social relations of capital (Althusser, 2006, pp. 164 – 207).
The prehistory of big data, a separate temporality, is then well elucidated through a rectification of the extant formulae for colonialism, for primary accumulation. Plainly: the encounter is not the realization of a telos in identity. It is instead an historical strand or temporality – surveillance capital, here ¬– interpolating, introducing its logic and rhythm to another, e.g., the temporality of physiological expenditures unperceived. Marx here writes (1990, pp. 153 – 4): “[the] product of [mechanical] 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” (underscore mine.) The historically specific epoch of development is then the moment emergent from the encounter, a transformation constituting: “… the confrontation of, and the contact between, two very different kinds of commodity owners [capitalist and laborer.]” (Marx, 1990, p. 874)
The precept here has its veracity embodied in the migrant yeoman driven by the hukou relocation program from steppe villages to Guangzhou, staffing warehouses and factories (Chuang editorial board, 2020); in the Adivasi dislocated, peddling ersatz things on the outskirts of Kolkata (Woodman, 2021). The encounters are not a telos realized, some clean subsumption, but an aleatory and scattered interpolation of logics and times, capital positing something outside through which it again constitutes and develops its social relations, inside. Productive modes, temporalities are everywhere misaligned; the encounter is a piecemeal alignment.
Google AdSense emerges first as the party to our encounter in 2003, the object of its encounter plainly you: the user (Google, n.d.). The digital space constituted by physiological expenditures as engagement: scrolling, clicking – was here underdeveloped towards capital accumulation, a misaligned productive mode. The expropriation of Adivasi farmland throughout Jharkhand presents an alignment of temporalities acre by acre, an infrastructure constructed like a grid of points forming a nexus constituting a productive mode (Woodman, 2021). Capital reaches outwards and around to encompass the outside, piece by piece. The feudal, subsistence temporality of rural West Bengal thus has its modes synchronized through an encounter with capital: the temporalities of New Delhi, Mumbai.
Google AdSense introduces an analogue here: the encounter of digital infrastructure and differentiated engagements is not the whole of a new mode, descending from on high and washing the residue away. Google AdSense instead constitutes a distribution of capital as nodes, platform by platform, perceptive apparatuses adopted by webmasters towards an array of cross-platform tracking, a veritable nexus of data accumulation (Google, n.d.). The nexus entails innumerable streams of data – streams of labor – stored in aggregate, assembled in a central dataset through which the nexus is then rebuilt, made efficient by curve-fitting and speculative inferencing (Kitchin and Lauriault, 2014). The heightened efficiency dissipates the other temporalities. The dissipation is an installation: an installation of more nodes throughout digital space, developing the productive logic and relational modes of capital here through an encounter, encounters. The online nexus of AdSense here now constitutes 38.3 million websites, nodes (Built With, n.d.)
Plainly: physiological expenditures in space – scrolling, walking, clicking, sitting; these constitute (among the prehistory of big data) a moment. This moment is a temporality of a unique logic and productive mode, distant but adjacent the temporality of capital. The encounter of the infrastructure of Google AdSense, of Facebook Pixel, of a watchful array and the temporality of your expenditures constitutes an interpolation. It is an alignment through a nexus of capitals distributed. These expenditures become labors, extricated from the person and objectified, instantiated in a thing related to other things, data. The linear model of history which Y presents is here surpassed. We are disinterested in the teleology of the encounter, where the mechanisms of capital are immanent in the vacuity of big data, proximately animated and determined. The camera being placed was not of an historical necessity. No spirit compelled Facebook Pixel’s conception. No spirit is realized in its function. We instead deal with facticity: the encounter happened, transforming the logic of the outside and developing that of the inside – thereby presenting some really new development, more alike to a lateral rhizome from which ginger bursts than a bulb become onion, a development of a relation by distance and difference.
Though rejecting the modes through which Y renders big data an inert tool, we are again met with a difficulty with the analogy or metaphor of big data and primary accumulation, the encounter. The question: how do we distinguish big data as new – is it not just an analogue of primary accumulation, a well-defined process? The development provides an illustrative mode of rectifying extant formulae, though what of itself is meaningfully new? Here we turn to the relational developments of ideology, reintroducing our note of big data and intelligibility.
Against (Data) Humanism
The anatomy of big data here has dealt principally with the basic, economic or concrete functions of capital. Though this anatomy should be understood descriptively – entangled in and really inseparable from the motion of that basic activity is the superstructure, the nebula of civic being, of ideology (Althusser, 2014, pp. 174 – 5). The development of the base is really that of the superstructure, a monism more than a dualism. The ideology referenced here should then be known as an affective thing: a real relation and a functional determinant of itself. Towards the ideology of commodity fetishism in particular, Marx writes in Capital (1990): “… to the producers, therefore, the social relations between their private labors appear as what they are.” (pp. 165 – 6) The ideological relation by which producers apprehend the world through the prism of self-active commodities, persons related through the mediation of objects – i.e., not persons related directly but related through their extensions in the form of labor and money – is a real relation: it appears as what it is. The ideals, epistemic frameworks through which political economy conceives markets as consisting in relations of objects instead of persons (Marx, 1990, pp. 163 – 77) – these denote real, material relations to the world. These are relations where the appearances of the ideological superstructure are the actual, functional determinations of the economic infrastructure. The hieroglyphic is its own objective reality.
The ideology of humanism then as well denotes a real relation to the world. It is a relation through a juridical instantiation, intelligibility of the abstract human through the basic, original property of their labor (Althusser, 2005, pp. 221 – 3). Here, we turn again to the formulation of primary accumulation: the assimilation of the differentiated physiological expenditures among the undifferentiated realm of equivalences then brings an ostensible original freedom: the freedom to sell your own labor, a freedom unafforded to the serf whose labor is inextricable from their mode of subsistence (Marx, 1996). The subjectivity constituted by capital is of an abstract humanity grounded in the basic, concrete structure of ownership, the first instance of ownership your own labor. The person is legally and politically instantiated as person, intelligible, through the contracts to which they relate in the presence of their goods (or labor) on the market, among equivalents. Plainly: the choice of the sale of labor (or property otherwise), the engagement with a plane of contractual relations of sale through the original ownership of some property then relates the person to the nebula of legality. They there become the abstract human, a subject with all of the endowments there entailed. The relational reality of humanism emerges as such.
Here position X attains to some validation: the concept of a contract seems infeasible in the context of big data, regulation through extant modes a nonstarter. The language and relational functions of humanism are surpassed, X introducing this such that the defunction of humanism is the defunction of capital. The first premise of the position here is then valid. The big data distribution of capital as nodes forms a nexus, both a productive mode and an encompassing whole of streams of data. The profundity of this nexus and the magnitude of the data collected and its capacity for drawing insights about the subject, then, makes the basic concept of sufficiently informed consent for contracts a difficulty (Véliz, 2020). The presence of a difficulty here might be anecdotally verified through asking: have I ever fully read the terms & conditions before clicking agree? The legal relational functions of humanism through contractual mediates seem here at least partially contradicted by big data.
Although, an inert, simple difficulty might not constitute a real contradiction to contractual relations; informed consent to a contract is hardly contingent upon the determined, particular competence of the signatory towards the work. The liability of learning to a large extent is yours. An ineffective regulatory mode (here, i.e., codifying legal-contractual thresholds on the sale of data) is still yet a regulatory mode through the relational prism of humanism; the basic vernacular is basically unsurpassed. What constitutes the ideological inversion, then, originally noted is not the reductive confounding of relations through the opacity of data accumulation, the contingency of particular competency. The inversion then consists in an inversion of the social dimension of the concrete relation to which it, humanism, corresponds, entangled. A particular contract of labor sale as instantiated in data is here not really particular; the concrete mode has changed. The proximity of the signing subject to others constitutes a third-party sale of their labor, as well. Big data presents an unwitting intelligibility of you and others through proximity to you and others, proximity to the nexus of perceptive things (Leszczynski, 2015)
The threshold of consent to a big data labor contract here involves all those to whom you are proximate, made intelligible through the typification and distribution mapping you enable by this original consent (Véliz, 2020). You do not here go to work and sell your labor; you go to work and bring all of your colleagues and neighbors unknowingly, whose labor is sold (or expropriated) as well. The mode of becoming a subject, a mode familiar to the relational mode of humanism, is here surpassed. The particularity of a relation as mediated through an object sold by contract is inverted. Position X here thus reemerges, already drafting a pitch to Verso; though capital has not been surpassed, as we have demonstrated through analogy of our extant concepts to big data. But the ideology of humanism and the temporally determinate relational mode has. The person here is not intelligible through the particular prism of a contract towards property: they are streams of data, streams of labor, related by their distance and difference from others, part of a distribution populated by inferences.
The residual concepts of the abstract human here still inform the broad (non-critical) methodologies for assimilating big data as a new market (Ritter and Mayer, 2018, pp. 220 – 77). Recent regulatory efforts consist in tries at well-defined contracts, e.g., platforms operating in the Eurozone being required to prompt the user with a transparent account of the usage of their data, an abridged terms & conditions functioning as a digital contract. These concepts denote a real ideology, a real mode of relation to the world; though their material correlate here developed presents a disanalogy. Big data constitutes a real historical development in this capacity, affective, contradicting Y – it is a development which leaves the language of humanism, an ideology now outmoded. The encounter here aligns temporalities to a basic productive relation, precedented theoretically; though the particularity of the mode of relation as represented in ideology has transformed.
The formulae of humanism, an appearance corresponding (descriptively, though really intertwined) to a functional, basic determination are here then defunct. The development of the productive modes and relations as presented by big data illustrates the deficiencies of the humanist formulation. The abstract human presents a real, relational mode – though it should not be taken for the essential being behind itself. There is no annihilating behind to the ideological developments, no transhistorical basic relation or telos of capital at which we might try. The conceptual defunction of humanism towards big data is the real development of capital, the real outgrowth of an encounter, and we deal with these realities in opacity, where a moment of clarification grants us no transformation in itself, apart from a transformation of our concepts (Marx, 1990, p. 167). Humanism is surpassed through basic modes with which it is entangled being developed. The ideological appearance has transformed because the concrete predicate from which it has no (except descriptive) differentiation has transformed, developed, inverted.
We return to the original question: what makes this [big data] meaningfully new? The analogy of big data to extant forms, the known metaphors of the commodity and primary accumulation here contradict the contentions of position X, eager to publish the beginning of the post-capital productive mode. The Y position through which big data becomes an inert descriptive tool, through which big data is minimal in newness though maximal in utility towards validating our formulae – is then undermined. Y is here contradicted by the real disanalogy of the modal particularity of the basic and therefore superstructural relations brought with the encounter of big data: the surpassed humanism.
The sale of data as a commodity emerges from an encounter, though it is one neither denoting a post-capital temporality nor a static metaphor, a constituent part of a confounded proof of our concepts and their applicability to a linear history. Big data, through the encounter of a plane of physiological expenditures differentiated, and the synchronizing logic of capital through the infrastructure of perceptive apparatuses, nodes – constitutes a real outward development of capital by encounter, primary accumulation. It is the assimilation of the former temporality into the modes, logic of the latter. Though no telos is realized; the encounter did not have to happen. It is not a ruse of reason from whose parts an identity attaining to an external logic is formed. The alignment of temporalities, of productive modes in space is the constant situation and transformative development of the social relations of capital. The encounter and alignment of big data is a development for which our concepts have metaphors, though through which the relational mode, the ideological and conceptual mode of humanism is then outmoded.
Position Z is then validated in maintaining the continuity of capital, against X, and the inversion or transformation of its temporally determinate modes of relation, against Y – though we are met with a concluding difficulty. Now, a position W emerges with a reactionary hue, asking – even should big data constitute an outward development of these concepts, a lateral development differentiated from them, how can we deny the backwards gestures? The encounter of big data is surely the assimilation of physiological expenditures as labor, unwittingly, an expropriation. No contract seems capable, with the conceptual framework here given, of properly informing consent – of properly consulting all involved parties. The backwards gestures, W continues indignantly, would then begin to suppose through the clear social domination entailed by big data that all labor – emergent through similar processes, encounters, and subject to similar contract mediates – is subject to a form of social domination. It would, frankly, suppose that the very nature of all of these encounters is an imposed ascendance, leveling a critique at all labor. The conclusion: this [these premises] cannot be right.
The premises presented by position W are contradicted by their conclusion. Though the premises are correct; the social domination of big data makes that of past encounters clearer. The Revelations’ fourth bowl spilling outwards was the eschatological predicate of a world to come, sunlight encompassing the earth, burning up all humanity (Revelations, 16:8 – 9). Here the bowl has already been spilled. We wallow in the sunlight unknowing but known. Where is our world to come?