Michel Foucault and the Enlightenment in Reverse

By Matthew Parish

Tuesday 1 September 2026

There is a fashionable sort of intellectual dismissal that consists not in arguing with a philosopher but in announcing that he was never really a philosopher at all. Michel Foucault has suffered this fate more than most. To his detractors he was a historian with an eccentric prose style, a Parisian celebrity, a political provocateur, a theorist of prisons and sexuality, perhaps even one of the founding fathers of the more extravagant forms of contemporary identity politics — but not, properly speaking, a philosopher. According to the most severe version of this criticism, Foucault contributed nothing to philosophy because he offered no systematic metaphysics, no theory of knowledge capable of being expressed in propositions and proofs, no ethics from which duties might straightforwardly be deduced and no political philosophy comparable with those of Hobbes, Locke, Mill or Rawls.

This is an extraordinarily narrow conception of philosophy. It would exclude Nietzsche almost as readily as Foucault and would make rather uncomfortable reading for Socrates. Philosophy does not consist only in constructing systems. Sometimes its most important task is discovering that the assumptions upon which our systems rest are not nearly as secure as we imagined.

Foucault’s great philosophical achievement was to make us suspicious of something that the modern West had come almost universally to assume: that the progress of reason and the progress of human freedom naturally accompany one another.

That proposition lies close to the heart of the Enlightenment project. Kant famously described Enlightenment as humanity’s emergence from intellectual immaturity — the achievement of the courage to use one’s own understanding rather than relying upon somebody else’s authority. The broader Enlightenment inherited an optimism that science, reason and the dismantling of superstition would progressively enlarge human freedom.

Foucault asked an unsettling question.

What if they do not? What if knowledge sometimes enlarges power faster than it enlarges liberty? And, still more disturbingly, what if the institutions created in the name of rationality eventually become mechanisms through which human beings surrender precisely the intellectual independence that the Enlightenment was supposed to confer upon them?

Seen from the third decade of the twenty-first century, this question looks considerably less eccentric than it did when Foucault was alive.

The peculiar Enlightenment of Michel Foucault

It is misleading to describe Foucault simply as an enemy of the Enlightenment. His relationship with it was considerably more complicated. Near the end of his life he returned explicitly to Kant’s 1784 essay What Is Enlightenment? Foucault interpreted Kant’s short work as something surprisingly modern: philosophy turning its attention towards the question of what is happening to us now. Philosophy was no longer merely asking what is true, what exists or how knowledge is possible. It was asking what sort of historical moment we inhabit and what that moment is doing to the people living within it.

Foucault therefore retained something profoundly Enlightenment-like: criticism. What he rejected was the comforting assumption that criticism necessarily produces a straight road towards emancipation. History, for Foucault, had no escalator. Humanity does not automatically ascend from superstition to reason, from tyranny to liberty and from ignorance to knowledge. Institutions change. Forms of knowledge change. Forms of domination change with them. Sometimes an apparent liberation merely replaces an obvious form of coercion with a subtler one.

That is the central insight of Discipline and Punish.

The medieval sovereign could torture a criminal publicly. Modern society regards this, correctly, as barbaric. It replaces the scaffold with the prison, the executioner with the administrator and spectacular violence with correction, observation, classification and rehabilitation. Surely this is progress. Foucault’s answer was: yes — but look more carefully.

The disappearance of spectacular brutality was accompanied by an enormous expansion in society’s capacity to observe, classify and regulate individuals. The prison became only one manifestation of a much broader disciplinary logic also visible in schools, hospitals, factories, barracks and bureaucracies. Modern institutions created what Foucault famously called “docile bodies”: people whose activities could be measured, organised, corrected and made useful.

This was not a conspiracy. That point is essential to understanding Foucault and is frequently lost in vulgarised versions of his thought. There need be no sinister committee sitting at the centre of society designing systems of oppression. Power may emerge from ordinary institutional practices. Teachers want pupils to learn. Doctors want patients to recover. Prison governors want inmates to behave. Employers want factories to operate efficiently. Governments want populations to be healthy.

Yet every one of these benign ambitions generates information. The pupil acquires a record. The patient acquires a medical history. The prisoner acquires a file. The worker acquires an assessment. The citizen acquires statistics. Knowledge makes administration possible and administration generates still more knowledge. Hence Foucault’s celebrated preoccupation with the relationship between power and knowledge.

The digital Panopticon

This is where Foucault begins to look prophetic. He died in 1984. There was no World Wide Web. There were no smartphones, social media platforms, ubiquitous CCTV networks equipped with facial recognition, consumer databases tracking billions of purchases, algorithms predicting political preferences or artificial intelligence systems capable of constructing extraordinarily sophisticated models of individual human behaviour. Yet Foucault had already identified the philosophical architecture within which these developments might be understood.

His favourite architectural metaphor was Jeremy Bentham’s Panopticon — a proposed prison arranged so that prisoners could potentially be observed from a central point without knowing whether they were actually being watched. The genius of the Panopticon was psychological. You did not need permanently to watch everyone. People who believed that they might be watched would begin regulating themselves.

Twenty-first-century humanity carries the Panopticon in its pocket. The smartphone records where we go. Search engines reveal what interests us. Online shops learn what we desire. Social networks discover our friendships, resentments, attractions and political enthusiasms. Employers increasingly possess the technical capacity to measure the behaviour of employees. Governments possess surveillance capacities that would have astonished the secret police of twentieth-century dictatorships. And much of this has happened voluntarily. We purchased the surveillance devices ourselves. We charge them every night. This is almost comically Foucauldian.

The Enlightenment turns around

Nevertheless the deepest contemporary relevance of Foucault lies somewhere beyond surveillance. The Enlightenment’s political ideal required a particular kind of human being: the autonomous individual capable of exercising judgement. Kant’s exhortation was Sapere aude — dare to know. Think for yourself.

The contemporary information environment increasingly invites us to do precisely the opposite. Algorithms select what we read. Recommendation systems determine what we watch. Search rankings decide which sources appear authoritative. Social-media dynamics reward emotional certainty more effectively than intellectual hesitation. Political discourse is compressed into slogans whose success is measured by circulation rather than truth.

Artificial intelligence may now perform the intellectual labour involved in answering questions that earlier generations had to answer through reading, reflection and argument. None of these technologies is inherently hostile to Enlightenment values. Quite the contrary. The internet represents the greatest democratisation of knowledge in human history. Artificial intelligence may become an extraordinary instrument for extending human intellectual capacities.

But Foucault would surely have insisted upon asking the second question. What kinds of dependence are being created simultaneously? That is his method. Whenever modernity announces a liberation, inspect the machinery accompanying it.

From knowing citizens to managed populations

Foucault’s later lectures developed the concept of governmentality — the techniques and rationalities through which populations come to be governed. Modern government does not merely issue commands. It accumulates knowledge about populations: births, deaths, illnesses, employment, crime, education, productivity and innumerable other variables. The population becomes something measurable and therefore something administratively manageable. Foucault’s lectures on governmentality and biopolitics explored precisely this conjunction between knowledge, administration and the modern state.

Again this is not necessarily sinister. Governments cannot build hospitals without knowing where people live. They cannot construct pensions without knowing how long people survive. They cannot manage epidemics without epidemiology. But administrative knowledge subtly transforms the relationship between citizen and state.

The Enlightenment imagined the citizen asking government: By what right do you govern me? The administrative state increasingly asks: Into which category do you belong? The distinction is profound. One conception begins with an autonomous person possessing reason and rights. The other begins with a data point.

Foucault’s great mistake — and his great strength

Foucault undoubtedly made mistakes. His political judgments were on occasions spectacularly poor. Like many French intellectuals of his generation he could romanticise revolutionary movements whose actual politics were distinctly less attractive than their rhetoric. His prose could be unnecessarily obscure. His historical claims have been disputed by specialists. His reluctance to articulate a clear normative foundation for his criticisms creates an obvious philosophical difficulty.

If power is everywhere, why should one arrangement of power be considered morally preferable to another? If concepts of normality, truth and reason are historically conditioned, upon what basis does the critic stand when criticising them? These are serious objections.

But they do not establish that Foucault contributed nothing to philosophy. They establish that he left philosophical problems unresolved. There is a difference. Indeed Foucault’s unwillingness to provide a universal theory may partly explain the durability of his work. Genealogy was designed to demonstrate that institutions and systems of thought we imagine to be necessary frequently arose through contingent historical developments. He wanted us to examine the intellectual machinery around us rather than merely accept its description of itself. That is recognisably philosophy.

The return of intellectual immaturity

There is therefore a remarkable historical irony emerging. The eighteenth-century Enlightenment sought to release human beings from authorities that told them what to think. The twenty-first century may be constructing authorities incomparably more sophisticated than anything the eighteenth century knew. They need not burn books. They merely rank them. They need not prohibit opinions. They merely decide which ones appear on the screen. They need not compel obedience. They can construct environments in which particular behaviours become easier, more profitable, more socially rewarded or simply more convenient than others.

This is power of precisely the sort Foucault wanted us to notice — decentralised, frequently benign in intention and embedded within ordinary practices rather than descending exclusively from sovereign command.

And this may be what it means to say that the Enlightenment project is going backwards. Not that science is disappearing. Science is advancing extraordinarily quickly. Not that knowledge is becoming scarcer. Humanity possesses more information than at any previous moment. The regression lies elsewhere. Knowledge is expanding while individual intellectual autonomy may be contracting. That possibility would have seemed paradoxical to many Enlightenment thinkers. It would not have surprised Foucault.

The unfinished Enlightenment

There is nevertheless no reason to read Foucault pessimistically. His mature understanding of Enlightenment was not simply a denunciation of reason. It was an invitation to practise criticism continuously. Scholarly treatments of his position emphasise precisely this ambivalence: modernity generates new freedoms while simultaneously producing new institutions and techniques of normalisation and control.

That is why Foucault may ultimately belong within the Enlightenment tradition rather than outside it. The naïve Enlightenment says: Use reason and you will become free.

Foucault replies: Use reason — and then examine what reason has built. Ask who defines rationality. Ask how institutions classify people. Ask what information they collect. Ask which behaviours their architecture encourages. Ask how scientific categories become administrative categories. Ask whether systems created to liberate people subsequently acquire capacities to manage them. And above all, never assume that yesterday’s liberation cannot become tomorrow’s discipline.

Foucault died at fifty-seven. His later work was moving increasingly towards questions of ethics, subjectivity, parrhesia— courageous truth-telling — and the ancient conception of philosophy as a way of life. His premature death prevented this project from reaching completion. That loss now seems particularly unfortunate, because the central philosophical problem of our age may turn out to be remarkably Foucauldian.

Humanity has constructed machines capable of knowing more and more about us. Governments and corporations possess unprecedented capacities to classify populations. Algorithms increasingly mediate our encounters with reality. Artificial intelligence promises simultaneously to magnify human reason and to relieve human beings of the necessity of exercising it. The Enlightenment asked how humanity might escape from intellectual tutelage.

Our century may have to ask how we avoid voluntarily returning to it. Michel Foucault did not solve that problem. But decades before the technologies that make the problem urgent existed, he taught us how to see it. That is not contributing nothing to philosophy. It is one of the things philosophy is for.