Matteo Pasquinelli, ‘Machinic Capitalism and Network Surplus Value: Notes on the Political Economy of the Turing Machine’, 2011
“The industrial modality appears when the source of information and the source of energy separate, namely when the Human Being is merely the source of information, and Nature is required to furnish the energy. The machine is different from the tool in that it is a relay: it has two different entry points, that of energy and that of information”. This insight by Gilbert Simondon on the second industrial revolution is not meant to underline a continuum between different technological age, to say that informationalism is the same of industrialism, but on the contrary to spot, as Gilles Deleuze and Felix Guattari would record, a bifurcation of the technological lineage, or machinic phylum.
Information can be found haunting also the instruments of the first industrial revolution: the Jacquard loom (invented in 1801) was in fact a mathematical device controlled by a punched card almost identical to the one standardised by IBM as data storage device in the 20th century.
In same years in which Simondon was sketching a counter-ontology of cybernetics, Romano Alquati introduced the concept of valorising information, that can be indeed adopted today as a conceptual bridge between the notions of information in cybernetics and value in Marxism.
In a long article on the ‘organic composition of capital’ at the Olivetti factory (an Italian company producing typewriters, mainframe computers and other automatised machines in Ivrea as early as 1950s), which was published in Quaderni Rossi in two parts in 1962 and 1963, Romano Alquati attempted one of the first Marxist analysis of cybernetics. Alquati frames the cybernetic apparatus (what we call today ‘digital networks’) as an extension of the internal bureaucracy of the factory, that monitors the production process by the means of control information [informazioni di controllo].
Bureaucracy descends into the bodies of the workers via the mediation of the circuits of cybernetics and machinery. Alquati introduces here the concept of valorising information [informazione valorizzante] as the ‘flow’ running along and feeding those circuits.
For the first time the modern reading of ‘information’ enters the essential definition of living labour and then of the Marxian surplus value itself, which is continuously absorbed by machinery and condensed into products in that way.
Easily the typical Marxist ‘organic’ distinction can be visualised here: living information is continuously produced by workers to be turned into dead information crystallised into machinery and the whole bureaucratic apparatus.
In fact, the important insight advanced by Alquati is the continuum merging bureaucracy, cybernetics and machinery: cybernetics unveils the machinic nature of bureaucracy and conversely the ‘bureaucratic’ role of machines, as they are feedback apparatuses that control workers and capture their know-how of the productive process.
Specifically it is the numerical dimension of cybernetics that is able to encode workers’ knowledge into bits and consequently transform bits into numbers for economic planning. In other words, operating as a numerical interface between the domain of knowledge and capital, the digital code transforms information into value.
At the beginning of the industrial age capitalism was exploiting human bodies for their mechanical energy, but soon it was realised that the series of creative acts, measurements and decisions that workers constantly have to take is the most important value that they produce. Alquati defines as information precisely all the innovative micro-decisions that workers have to take along the production process and that give form to the product but also give form to the machinic apparatus.
That a computational machine may occupy the space previously described by the division of labour was already a postulate shared by the early pioneers of cybernetics, such as Charles Babbage. Marx was already quoting him in The Poverty of Philosophy in 1847: “When, by the division of labour, each particular operation has been simplified to the use of a single instrument, the linking up of all these instruments, set in motion by a single engine, constitutes—a machine”.
Logically, in the first book Capital the chapter on machinery follows the chapter on the division of labour. Conversely, the division of labour in itself can be considered already a sort of abstract machine.
The important lesson we take from Marx here is precisely the refusal of technological determinism.
As much as the division of labour is shaped by social conflicts and workers’ resistance, so is the evolution of technology. The parts of the social ‘mechanism’ adjust themselves to the technical composition according to their degree of resistance and conflict. Machines are moulded by social forces and they evolve according to social forces.
If we endorse this political insight, that is looking at the social relations and conflicts replaced by information machines, we get finally a political methodology to understand the generic definitions of information society, knowledge society, network society and so on. As much as industrial machines were not just replacing workers’ horsepower but a whole set of relations that was developed under the regime of manufacturing, in the same way, information machines have replaced a set of cognitive relations already at work within the industrial factory.
The so-called division of labour is, foremost, a separation of mechanical from intellectual organs.
If Marx opens the chapter on machinery of Capital by stating “the machine is a means for producing surplus value”, he afterward clarifies machinery precisely as means for the augmentation of surplus value (as in Marxian terms machines cannot produce surplus value as they cannot be exploited, only workers produce surplus value).
But in Alquati as much as in Marx the relation of the worker with the machine is a conflictive one. And the living information (or living knowledge) that feeds everyday the cybernetic machinery is a field of resistance and struggle. This border of the transformation of living knowledge into dead knowledge and between the individual brain and the social brain are the issues that set today’s debates on labour and information, and that should challenge also Deleuze and Guattari’s notion of machinic.
[Deleuze and Guattari’s] notion of machinic was inspired, specifically, by the mechanology introduced by Simondon in Du mode d’existence des objets techniques,19 that was itself a reaction to the rigid determinism of cybernetics, its ‘feedback system’ and its idea of information as a mathematical unit of measure.
In 1972, in Anti-Oedipus Deleuze and Guattari initiated the notion of desiring machine in order to constitute an immanent political economy // where ‘desire’ could be finally recognised ontologically (and economically) as a productive force and not just as a linguistic operator of the psychoanalytical theatre.
Deleuze and Guattari describe also a machinic surplus value. Eight years after, however A Thousand Plateaus appears to introduce a more postmodern reading that focuses on machinic assemblages and abstract machines.
The machinic assemblage is immanent and productive too, but it is clear here the shift to a more relational ontology. Due to this ambivalence, recently, the notion of machinic happened to be received and reduced just to this relational paradigm of assemblages that obliterates the very dimension of production in Deleuze and Guattari together with their Marxist background.
More precise dictionaries highlight specifically the ancient root mach- that means growth, augmentation, amplification of a force. The same root mach- surfaces, for example, both in the Latin magia (’magic’) and magnus (’great, big’). Similarly, in Old High German the word macht refers to power, skill, ability and wealth in a similar way to the Latin potentia. In other words, when Deleuze and Guattari were referring to machinic surplus value, they were just making the ancient root of the word ‘machine’ resonating again.
Machine then is more about surplus than assemblage.
In a note of the Anti-Oedipus Deleuze and Guattari show to know the chapter on machinery in Marx’s Grundrisse. Inspired by that reading, in the same page they try to introduce the concept of “machinic surplus value produced by constant capital”, “recognizing that machines too work or produce value, that they have always worked, and that they work more and more in proportion to man, who thus ceases to be a constituent part of the production process, in order to become adjacent to this process” (Deleuze and Guattari, 1972: 232, n76).
Curiously, the notion of ‘abstract machine’, that Deleuze and Guattari put at the centre of their ontology in A Thousand Plateaus, is inspired by the same term used in cybernetics, where an abstract machine is the project of an algorithm that subsequently can be implemented in a virtual machine (such as computer software) or in a material machine (computer hardware or any mechanical apparatus).
If the notion of machinic is superficially applied, it may result into an apolitical continuum, where everything would become ‘productive’ and where then it would be impossible to distinguish living labour and dead labour, variable capital and fixed capital, that is exploitation and autonomy.
Together with Antonio Negri and Maurizio Lazzarato, Paolo Virno was one of the first thinkers of operaismo to extract and liberate // resolutely the living knowledge from the greasy gears of the industrial machine and to make it ‘breath the city air’.
Indeed the whole debate on post-Fordism and its cultural industries can be condensed in the following question: can living knowledge/labour be autonomous? This is the original contribution that operaismo gave to contemporary political economy and at the same time the reason of many irrational attacks by those who still consider workers just horses performing muscular power. For sure, in this exodus from the factory, the old Marxian borders of fixed and variable capital no longer keep: a more precise notion of machinic has to be discussed in order to explore this threshold.
Marazzi insists on the transposition of the machinic fixed capital into the human living body. “According to our hypothesis, aside from the traditional faculty of labour, the body of the labour force has to become the container of the function of fixed capital, that is machinery, ‘codified knowledge’ and ‘productive grammars’, in other words past labour”. This passage of Marazzi is radical: if according to Marx, capital is but a social relation, indeed there is no need of heavy actors such as machinery, industrial management and scientific research to describe contemporary production—the machinic source of profit can be externalised into the workers’ body itself.
These interventions by Marazzi into the grammar of political economy are crucial to underline once again that, when we talk about cognitive capitalism or the hegemony of immaterial labour, we do not refer to something ‘immaterial’ but to a very physical machinic intertwining of our own bodies and social relations.
To conclude: there is a machinic dimension of knowledge external to the industrial capital fixed into machinery. The collective dimension of machinic knowledge is called by Marx in the Grundrisse ‘general intellect’, ‘general scientific labour’, ‘general social knowledge’, etc. This collective dimension is productive in two ways: as physically embodied into industrial machinery, communication infrastructure and digital networks, but also as a mass intellectuality managing the division of labour of the social factory and producing new forms of life.
The individual dimension of the so-called immaterial labour can be the [sic] distinguished in itself into cognitive labour (creating new material, immaterial or social machines) and informational labour (operating in front of a machine and producing valorising information).
Of course the distinction between machinic knowledge and mass intellectuality, cognitive labour and informational labour is blurred. What is important to remark here is the primacy of living knowledge and living labour against any // fatalistic reading of machinery and new technologies as a perverse obstacle to the autonomy of the living.
Curiously all the metaphors employed to describe the machinic dimension of knowledge exiting machinery are still adopted from industrialism, see for instance ‘cultural industries’ or ‘edu-factory’.
Precisely if it is true that machines are moulded by social forces, we should recognises in the Turing machine the silhouette of living knowledge.
The term ‘digital code’ may refer to three different things: the binary digits encoding an analogue input into 0 and 1 impulses, the language in which software programs are written (such as C++, Perl, etc.), the script or textual source of the software programs (that embodies the logic form of an algorithm). In this text I propose to focus on the algorithm as the inner machinic logic of information machines and the so-called digital code.
And how it is clear in the case of videogames, an algorithm is not just a mathematical abstraction but it projects a very physical subjectivity. The algorithm exits the screen and play the operator that works in front of it.
The conceptual operation that I suggest here is to apply the notion of machinic to the algorithms of the digital code in order to recognise digital code and software programs as a form of machine in Marxian sense, that is a machine that is used to augment surplus value (even if we will have to discuss further the unit of measure, or better dismeasure of such a surplus value).
Algorithms are not autonomous objects, but they are shaped themselves by the pressure of external social forces.
But two kinds of information machines or algorithms must be distinguished then: algorithms to translate information into information and algorithms to accumulate // information and extract metadata, that is information about information.
It is in particular the scale of metadata extraction that discloses a new perspective on the economy and governance of the new means of production.
Economist as a very ‘industrial revolution of data’. If, as seen before, Simondon recognised the industrial machine as already an info-mechanical relay, today a further bifurcation of the machinic phylum can be introduced to recognise the information machine as a meta-informational relay, that handles information and metadata (or information about information). Metadata are the ‘measure’ of information, the computation of its social dimension and its transformation into value.
Thanks to this intuition by Alquati, Turing machines can be defined more generally as machines for the accumulation of information, extraction of metadata and implementation of machinic intelligence. The diagram of the Turing machines offers a pragmatic model to understand how living information is turned into machinic intelligence.
Briefly, here metadata are said to be used: 1) to measure the accumulation and value of social relations; // 2) to improve the design of machinic knowledge; 3) to monitor and forecast mass behaviours (dataveillance).
In conclusion Turing machines are defined as devices to accumulate valorising information, extract metadata, calculate network surplus value and feed machinic intelligence. To borrow few metaphors from Brian Holmes’s work on financial cybernetics, I guess it is time to move from the white cube of ‘digital creativity’ and dig deep into the black box of network surplus value and the algorithms designed for the capture of the common.