1. An Appearance of Freedom

My reflections on technique and technology lead me to sometimes unexpected places. Recently, I wrote about a situation I observed when reflecting on our technical and technological uses: we are free, in appearance, but certain options become impractical or unthinkable without however being prohibited.

For example, and to take up the topic of the moment, we are free to use generative artificial intelligence just as we are free to send our mail by post rather than by email. Nothing forbids resorting to postal mail and no law prescribes email. No sanction strikes the person who sends a paper letter rather than an email. And yet, the freedom to use postal mail is very often reduced to nothing: the option exists but it is no longer reasonable. Nobody has really removed the choice or the possibility. They have simply made the alternative absurd.

I wanted to understand why we had come to this point. The easy answer is that (technical) progress leads us to use email rather than postal mail. That AI saves us time and that it is therefore “logical” to use it. This simple observation actually reflects a situation that our usual concepts struggle to capture. I chose to name this situation an imperceptive. This neologism results from the contraction of imperative and imperceptible.

2. The Inadequacy of Available Concepts

I used the imperceptive to describe the “quasi-imperative” we had to use technology in our activities (both private and professional). I preferred to coin a new concept because “quasi-imperative” was imperfect and, upon examination, does not allow the situation to be precisely qualified. I had, by reflex, used the term “quasi-imperative” but it is a misuse of language. I realise there is something more that does not seem to be captured by existing concepts.

At this stage, I would therefore indicate that the imperceptive designates

the constraint exerted on the framework of decision. Its specificity lies in the fact that it neutralises the relevance of choices without suppressing their formal possibility.

Before getting to the heart of the matter, I would first like to confront my idea with concepts that resemble it. The objective is to verify that what I call the imperceptive is not named differently by others. There are indeed concepts that, without being identical, could resemble the imperceptive.

The imperceptive resembles other concepts such as the nudge, governmentality, social norms or radical monopoly. There are surely others that I have not yet discovered. These concepts are too close to be ignored in this verification step. Although the imperceptive resembles them, I believe this resemblance is insufficient.

2.1. The Quasi-Imperative

The adverb “quasi” is an adverb used as a prefix meaning “almost” or “approximately” (according to the Larousse definition). According to CNRTL, it “serves to mark a similarity, an approximation or a qualitative assimilation”. Yet what I am trying to describe is neither an approximation (an “almost”) nor an assimilation, nor a similarity of the imperative.

Furthermore, an imperative is a prescription. This prescription can come from several horizons (moral, legal, social, …). Behind this concept of imperative, we therefore find the notion of hierarchy. The imperative is an obligation and it is very generally transparent in its statement. We identify a sender and a recipient. The law is an imperative. Its traceability is clear: there is a sender (legislative power) and a recipient (the legal subject (citizen or company)). When I speak of the imperceptive, there is no precise sender or recipient. As I wrote in the aforementioned note, it is a form of (social) pressure that “compels” a behaviour.

With the imperceptive, one then acts on the motivation to adopt a behaviour without however becoming an obligation as such. The imperceptive thus joins incentive in this sense: it makes an option more attractive, more appealing in a certain way.

2.2. The Nudge

If the imperceptive joins incentive, one might think it is a nudge.

The nudge is the deliberate intervention on choice architecture to orient behaviour without formally prohibiting or constraining. Thaler and Sunstein proposed a definition:

The nudge, the term we will use, is an aspect of choice architecture that alters people’s behaviour in a predictable way without forbidding any options or significantly changing their economic motivations. To count as a mere “nudge,” the intervention must be simple and easy to avoid. “Nudges” are not mandates. Putting the evidence directly before people’s eyes is considered a nudge. Merely prohibiting what should not be done or chosen does not work.

The example one could take to illustrate would be placing vegetables or fruits before chips in the school canteen (what an idea to put chips in the canteen anyway!). This arrangement is intentionally designed, with an objective (eating better), and the designers “know” what they are doing or at least make a conscious decision about the objective to be achieved.

What differentiates the nudge from the imperceptive is its structure. The nudge is implemented by an author (an administration, a company, a designer) who has deliberately arranged the choice architecture with a declared objective (or not).

We all know nudge techniques.

They can be public: steered by public authorities with an assumed purpose (e.g.: tax incentives for installing photovoltaic panels or using low-emission vehicles). They can be private, implemented by companies to orient consumer choices to their advantage.

In both cases, there is an actor who decides what they are doing and why they are doing it.

The nudge thus fits within a logic of paternalism boosted by a certain utilitarianism dear to our capitalist society. This model was already criticised by Milton Friedman when used at the state level because of its impact on individual freedom (I will return to this below). The nudge is a local and targeted technique for orienting choice.

There is an intentionality that is incompatible with the imperceptive.

Moreover, the definition of the nudge explicitly provides for “ease of avoidance,” which, to take the example of postal mail vs email, cannot be considered as I will explain below regarding the impact of the imperceptive on our freedom.

2.3. Governmentality

The quasi-imperative is approximate. The nudge is intentional. One might then think of Foucault’s governmentality. It designates the set of techniques, procedures and rationalities by which a population is governed. It is the production of norms, knowledge and subjectivities. The governed individual internalises the norms and conducts themselves according to them.

For Foucault:

It is about a certain type of state control over populations, a certain mode of exercising power where to govern is “to have towards the inhabitants, the wealth, the conduct of all and each a form of surveillance, of control no less attentive than that of the father of the family over the household and its goods.”

There is first a difference in mechanism: governmentality produces subjects (it acts on identity and subjectivity). For me, the imperceptive acts on the framework of decision without necessarily reshaping the subject’s identity (although this point would deserve a reflection in its own right. Whether our choices shape our identity or merely express it?). The subject remains: they simply think their choices are rational because they do not feel they are conforming to a conduct. They adopt it naturally, considering it rational.

There is also a form of paternalism in governmentality that is close to the public nudge. The direction of conduct is given by the authority to adopt a behaviour, a conduct. What distinguishes governmentality from the imperceptive is its scope. Foucault focuses on institutions and explicitly political power relations. Governmentality presupposes a political reason of government. The imperceptive, as I conceive it, does not imply a rationality of government because it impacts the conditions of what appears as rational.

While governmentality aims at the conduct of conducts, the imperceptive touches on the conditions of possibility of conduct itself. The imperceptive is the regime of rationality that configures upstream the criteria of the reasonable, such that certain options, although formally possible, cease to appear as rational choices.

One could then consider that the imperceptive plays a role upstream of governmentality as well, in that it configures and structures the possibilities of governmental rationality.

2.4. Social Norms

Contrary to what I propose above, one could also think that the imperceptive is a “product” of governmentality. This capacity to configure the possibilities of the rational could then be a consequence of decisions about conduct that produce notably “social” norms. These norms come to structure conduct. This fabric inherited from the past, shaped throughout our existences, comes to guide populations.

Here again, I identify a strong distinction with the imperceptive. The social norm is known and the recipient of this “social norm” is targeted by it. For example, we know that we must dress a certain way for a job interview. We know there are rules of politeness, ways of being that are expected depending on the context (for example, at a funeral or a birthday). This knowledge is precisely what allows conforming to the social norm or choosing to renounce it for a host of reasons that I will not address here.

There is therefore a distinction to be made on this first aspect: nobody says that one must use one technical device rather than another. It is more subtle because the imperceptive touches the regime of rationality of choices.

Then, a distinction can be made on the object. The social norm points to a specific behaviour in a specific situation. There is a more important contextualisation when examining the social norm compared to the imperceptive. The imperceptive is much more transversal. There is no identifiable prescription or determined action in the imperceptive.

Finally, we can also note a distinction regarding the sanction for non-compliance with the social norm compared to that of the imperceptive. Jump the queue and you will be scolded by the people you passed, show off on a train or bus and you will be called to order by other passengers. For the imperceptive, nobody will sanction you if you use outdated or obsolete technical means. At worst, you will be perceived as the “old school” or “vintage” member of the group, you will be disconnected from certain digital services, but this (possible) exclusion is not comparable to the transgression of a social norm.

In this sense too, the imperceptive is insidious because it is more difficult to contest than a social norm. One can refuse a social norm by assuming the gaze of others or by ignoring it. One can hardly refuse the imperceptive in the same way: there is no gaze to face. There is simply a framework that progressively makes certain choices unreasonable.

2.5. Radical Monopoly

A notion proposed by Ivan Illich, it targets the place that certain institutions take in our societies. This monopoly would aim to modify, control and ultimately constrain populations to change their habits by restricting their choices and freedoms. In his essay “Tools for Conviviality,” Illich sees in radical monopoly:

a type of domination by a product that goes well beyond what is usually meant by the term.

For Illich, there is a radical monopoly when the programmed tool supplants the individual’s power to act. This domination establishes compulsory consumption and therefore restricts personal autonomy (p.80). It is therefore the antipode of individual autonomy.

In this framework,

Radical monopoly is established when people surrender their innate ability to do what they can for themselves and for each other, in exchange for something “better” that can only be produced for them by a dominant tool. Radical monopoly reflects the industrialisation of values. (p.84)

To take his examples,

That the car restricts the right to walk, and not that more people drive Chevrolets than Peugeots, that is radical monopoly. That people are forced to be transported and become powerless to move without a motor, that is radical monopoly (p.80).

The resemblance to the imperceptive is very strong but there remains, in my view, a very clear difference in what Illich was able to observe in his time.

Illich’s analysis is dated and essentially concerns consumption issues. The examples he takes to illustrate his point range from the automobile to education, from funeral homes to beverage manufacturers. He directs his critical observation at consumerist practices and targets the power taken by certain institutions. Radical monopoly is much more targeted. I see in the imperceptive a more structuring role, as I attempted to demonstrate for governmentality, which enables radical monopolies. Put differently, the imperceptive enables the creation of the radical monopolies proposed by Illich, as I will explain below.

3. The Imperceptible and the Promethean Gap

The detour we have taken has allowed us to trace the distinctions that the imperceptive could have with other concepts, focusing particularly on the notion of imperative.

Beyond the imperative, there is also the imperceptible. I believe that this absence of perception, this blindness, can only be considered from an internal point of view and this is what allows me to name it by attempting to step outside the framework that encircles us, in order to stigmatise it and attempt to criticise it.

This notion of imperceptible is not entirely new. It was found, in a certain form, in the concept of “Promethean gap” by Gunther Anders proposed in his book The Obsolescence of Man (Volume 1). This gap is defined by Anders as:

“the growing inadequacy between what humanity is capable of producing technically and what it is capable of imagining as consequences of this production”.

According to Anders, we build systems whose effects and consequences exceed our capacity for representation. This Promethean gap is one of the sources of the imperceptive. We are collectively incapable of imagining what our tools do to us and the constraint they exert on us cannot be perceived as such. It is in this sense that it is imperceptible, because we do not have the representations that would allow us to see it.

4. The Contours of the Imperceptive

4.1. The Relationship of the Imperceptive with Other Concepts

After having surveyed the related concepts, we can now identify a causal chain between certain concepts discussed. I see here the following chain: Promethean gap -> imperceptive -> radical monopoly.

The Promethean gap targets what humans create as technical objects whose effects exceed what they can imagine.

The imperceptive is the product of this gap. Because humans cannot imagine the effects of what they put in place, certain constraints settle in without being perceived as such. The imperceptive prohibits nothing: it neutralises alternatives by making them absurd, without this neutralisation being visible, deliberate or conscious.

Radical monopoly, in Illich’s sense, is one of the observable consequences of this process. When the imperceptive has reconfigured the rationality of humanity’s technical productions that Anders mentioned in 1956, certain tools or institutions end up occupying the entirety of available space for a given function. Radical monopoly is not the result of a monopolistic intention; it is the culmination of an imperceptive that has operated without encountering resistance, precisely because it was not perceived.

However, not all imperceptives necessarily produce radical monopolies. Radical monopoly presupposes an infrastructural reconfiguration on a large scale. The imperceptive can also produce more diffuse, less institutionalised effects.

The relationship is therefore as follows: radical monopoly presupposes an imperceptive, but the imperceptive does not always lead to radical monopoly.

4.2. The Imperceptive as Choice Modelling

As already written, the imperceptive is an imperative that exerts itself without ever formulating itself as such. What distinguishes the imperceptive is that it operates on the framework within which the decision takes shape and not on the decision itself. This framework is the regime of rationality of which technology is, in my view, the privileged vector.

There is no invitation to do X or Y but our milieu is shaped in such a way that the framework of reflection “imposes” doing X or Y. It is a very perverse situation in that the imperceptive will model conceptions to orient choices. The subject does not perceive the imperceptive as an imperative. The subject simply thinks they are acting according to rational choices, as self-evidence or common-sense necessity given the framework and milieu in which they operate.

The imperceptive becomes imperceptible because it blends into the real and has been naturalised by the subject. The real is in fact only a representation of the subject, interpreted according to their own sensibility. On this point, I can refer to the conception proposed by Marcello Vitali Rosati:

there are only material models; the real is always the result of a modelling. The real is multiple and always mediated; there is no real outside of a mediation and all mediation is material. (…) The model is not a selection of the real, it is the real itself, as it results from a determined mediation; but outside of this mediation there is nothing at all.

On the basis of the above, I can refine my definition: the imperceptive is a model of rationality in which a dominant criterion, efficiency, imposes itself without being discussed, configuring upstream the conditions of the reasonable. Certain options remain formally possible but cease to appear as relevant.

4.2.1. The Example of the EU AI Regulation

Let us illustrate by taking the European regulation on AI.

Article 4 of the European Artificial Intelligence Regulation (hereafter AIR) illustrates this mechanism in a particularly striking manner. It provides that providers and deployers of AI systems must take measures to ensure a sufficient level of AI literacy for their personnel and the persons responsible for the operation and use of these systems.

There is no obligation to use AI. Article 4 of the AIR does not constrain anyone to deploy an AI system. It leaves the freedom to deployers (users) and companies to decide to use it.

However, if the company decides to use an AI system, then the regulation imposes obligations of means on users who, if they use AI, must take the necessary measures to (i) use it in a certain way and (ii) understand the tools used.

This mastery obligation therefore presupposes usage but also contributes to normalising this usage. This is how Article 4 reveals the imperceptive. The legislator did not decide that everyone should use AI. They organised the conditions of this usage. But in doing so, in anchoring in the very text the idea that adoption is a fait accompli that calls for regulation, the regulation contributes to naturalising this fact. The rule therefore has a performative effect in addition to being prescriptive: it normalises usage by treating it as self-evidence that precedes the norm.

Here is a legal imperative that, through its architecture, illustrates the imperceptive by treating as settled what was still a matter of choice. Article 4 does not say “you must use AI.” It says: “since you use AI, here is what you must do.” This presupposition is the moment when the imperceptive produces a visible imperative.

This analysis is reinforced by reading Recital 20, linked to this Article 4. It confirms this observation. It begins with the objective of “making the best possible use of AI systems.” This formulation presupposes usage before anything else. Adoption is not a hypothesis that the legislator examines: it is the starting point from which they reason.

The recital goes further by inscribing AI mastery in an explicit economic purpose: AI training is presented as a lever for improving working conditions and “consolidating the path to trustworthy AI innovation in the Union.” The legislator naturalises a causal link that is far from obvious: mastering AI would, by definition, mean working better. Usage becomes not only a fait accompli but a condition of progress.

And if you still had any doubt, reading Article 1 would finish convincing you. Article 1 of the Regulation sets out the purpose and states its aim to “promote the uptake of human-centric and trustworthy artificial intelligence (AI), while ensuring a high level of protection of health, safety, fundamental rights enshrined in the Charter, including democracy, the rule of law and environmental protection, against the harmful effects of AI systems in the Union, and to support innovation.”

End of part 1