Destructurations, Structurations and Assemblages through the Prism of Extratechnicity

We can thus specify our “destructurations” – artistic approaches considered in the central chapter – as gestures that affect less the technicity of the machine than its over-artificialized functioning. These gestures destructure the “extratechnical” level, allowing a surplus of coherence and harmony to surface in a system that, taken as a whole, can be conceived on the basis of Simondon as a superposition of reticulations: natural reticulation <-> technical reticulation¹ <-> psychosocial reticulation².

Structurations, for their part, are to be understood less as the extension of pre-established structural elements than as the emergence of new structures, but on the basis of regions of the technical reticulation that would have remained sheltered from the influence of the extratechnical layer, regions that would have retained a certain form of “purity.”

Opening a long but necessary parenthesis, we should ask ourselves here: what, more concretely, would this notion of the “purity” of technics consist of? Let us try to specify it by way of examples, while bringing it closer to our object of study. An algorithm in the C++ or Python language, for example, remains “pure” in its functioning when it is used to generate and modulate digital images³, the algorithm being understood here as the basic operative structure of computers, and the languages mentioned being very widespread, general-purpose and free languages. A digital image processing interface, for its part, contains additional conditionings with respect to the algorithm in the form of code, certainly, but not necessarily “impure” ones, so to speak. Indeed, the algorithm, in its primary operative function — gestural before being wrapped in the symbolic, concrete before presenting itself as abstract —, can always be mobilized by means of the interface. It is only when a new “feature” or design effect is added to the interface — one that, without improving the functioning of the tool, aims to increase the manufacturer’s economic margin — that this operation is diverted. Let us also think of an image processing script (whether for noise reduction, color correction or contrast adjustment): as long as the algorithm remains manipulable from the command line or in a minimal environment, it retains a certain purity of functioning. Yet when the same algorithm is integrated into software such as Adobe After Effects or into a plugin optimized for commercial purposes, it becomes an element of a dispositif carrying extratechnical constraints. The Open Source universe (for example, the exchanges on github.com), with open and free code, allowing different versions and free collective work, seems preserved. On the other hand, the addition of AI features – an operation under way today in most digital tools⁴ – is to be understood as a formidable thickening of the extratechnical layer. Finally, the mechanisms known as “filter bubbles” and “echo chambers,” which we discussed in the previous chapter in the light of Disnovation.org’s work Predictive Art Bot, and which affect certain social networks and certain search engines, introduce such a high level of extratechnicity into the technical element that they contribute (in combination with other algorithmic stratagems) to transforming it directly into a biopolitical dispositif.

However, to pick up the thread of the reflection, let us say that for a work to arise from a structuring approach within the digital, it is not enough for the artist to mobilize preserved zones (unaffected by extratechnicity) of the technological foundation. Indeed, the emergence of humanly unpredictable forms capable of structuring a digital work – let us think, for example, of the dynamic textures proposed by Toni Mitjanit by mobilizing thousands of autonomous agents –, on the basis of pre-existing logical frameworks (algorithms mobilized in their “normal” logic of functioning), can be due to two reasons:

• either the artist, by building their own algorithmic tools, manages to aim at unexplored territories of the – almost infinite, as we have seen⁵ – space-time of the digital image;
• or the artist, faced with an algorithm already built by someone else, manages to find, in the equally vast universe of possible data (that of the inputs), those that will compel the system (the algorithm), not to step out of its logic of design or functioning, but to aim, on the plane of the results of the computation (the outputs), at territories not humanly foreseen when the logic of the algorithm’s design was set up⁶.

Let us emphasize, still within the framework of structuring creative processes, that the possible construction of the algorithm by the artist, as well as the search for relevant input data, are processes based on an experimentation paced by successive human-machine exchanges, giving rise to countless adjustments⁷, a process grounded in the reception (perception) of transitional forms, in the process of being made.

As for the space-time of the results of the computation (the output, for us: the images produced by the algorithm), we may say that it includes regions that have already been “explored,” or else “unexplored” ones, depending on whether or not these images will have been the object of human perception – or rather, of human perceptions –, despite the fact that these regions are all contained within the framework of the formal logic comprised by the algorithm and the input data.

Assemblages, hybrids between structuring and destructuring processes, are potentially affected by the two distinct dynamics that drive them. Indeed, on the one hand, assemblages can produce forms through the activation of relatively “pure” regions of the technical foundation (still extending Simondon’s use of this term). On the other hand, these assemblages can act by dismantling the “impure” parts, affected by the introduction of extratechnical elements arising from strategies of domination and/or from the commodification of technics. However, as has already been suggested, through their functioning – which consists in linking elements that were not connected in the initial system –, these assemblages manage to shed light in a particular way on the metastable aspect of their initial domain: it is a “medium” charged with potentials, which need to be activated, triggered, by a gesture that introduces a surplus of connectivity within the initial reticulation. The deterritorialization and the reterritorialization that follows would thus have to be reinterpreted (ultimately) within this framework less as a displacement, or a journey toward an “elsewhere,” than as a significant restructuring of the local foundation that welcomes the artistic gesture, thanks to a process triggered by this very gesture.

1. Technics, for Simondon, is to be understood in the broad sense, as a mediator between the human and the world: “Technicity is first of all a matter of the relation between man and the world before being a matter of technical objects.” In Gilbert Simondon, Sur la technique (1953-1983), op. cit., p. 20.↩︎

2. The psychosocial, designating the psychic aspect of the human, emerges in Simondon within the living as a solution to new problems encountered in the course of evolution. This psychosocial is in fact, for the philosopher, a passage that opens the individual to the transindividual (that is, to the collective psychosocial).↩︎

3. Friedrich Kittler would not agree with this statement. Indeed, for the German media theorist, the last humans to have truly “written” are the engineers who, in the 1970s, spread out 12 square meters of paper to draw by hand the layout of the first integrated microprocessor, the Intel 4004. Subsequent processors were designed with the assistance of their predecessors, the machine becoming, for Kittler, its own agent. From that date on, for Kittler, no human has ever truly created anything. In Friedrich A. Kittler, Mode protégé, op. cit., p. 30.↩︎

4. The whole Adobe suite is affected by this transformation, and so are, beyond image processing, search engines, social networks…↩︎

5. On the basis of the works Every Icon (1997) by John F. Simon and Every Possible Photography (2013) by Jeffrey Thompson, considered in the central corpus, subchapter Structurations.↩︎

6. We explained such a situation on the basis of the principle of “logical omniscience,” discussed following Samson Abramsky in the “Structurations, Perspectives” section of chapter 2 (Why Do We Compute? A Matter of Finitude). This principle states that when a logical proposition is known, then all the logical consequences that follow from it would also be “known.” This principle is obviously false for the human, while the machine can only determine the structure of the outputs following the computation. This can be translated into the terms of the present research by saying that if the software and algorithms used by artists are “known” to others (for example, to the people who designed this software), just like the physical system on which these algorithms concretely run, this does not mean that all the digital images that the algorithm is able to produce are themselves known: even when artists design their own algorithms, they obviously do not know all the logical consequences — and a fortiori all the sensible manifestations — that can follow from the computation.↩︎

7. Such is the observation that emerges from the exchanges with the artists, as well as from personal practice.↩︎