The Building Nobody Built

Jeremy Bentham spent roughly two decades and a large part of his own money trying to get a prison constructed, and failed. Parliament eventually paid him compensation for the wasted effort. He died in 1832 with the thing unbuilt, which he considered one of the disappointments of his life, and which turns out to have been irrelevant, because the design did not need a building.

That is the useful fact about the panopticon and it is the one most explanations skip. It is not primarily a prison. It is a mechanism, and the mechanism works whether or not anything gets constructed around it.

What is a panopticon?

A ring of cells around a central tower. Each cell open to the tower and lit from behind, so the occupant is permanently visible in outline. The tower shrouded, with blinds and baffled passages, so the occupant can never see whether anyone is inside it.

Bentham published the design in 1791 as a series of letters written from Russia, and the title tells you it was never only about crime: Panopticon; or, The Inspection-House, applicable to prisons, workhouses, factories, madhouses, hospitals and schools. Wherever people needed to be kept under inspection. He thought he had found a principle, and he had.

The principle is that the tower does not need to be occupied. A prisoner who might be watched behaves like a prisoner who is being watched, because the cost of being wrong is too high to risk. Surveillance becomes, in Bentham’s own accounting, cheap. You are paying for the possibility rather than the act.

He described the result, with the confidence of a man who had solved something, as a new mode of obtaining power of mind over mind.

What Foucault did with it

For most of two centuries the panopticon was an architectural curiosity. Then in 1975 Michel Foucault made it the centrepiece of Discipline and Punish, and it stopped being about buildings.

Foucault’s argument was that Bentham had not designed a prison. He had drawn a diagram of how modern power works.

The older kind of power was spectacular and occasional. It waited for a transgression, then punished it publicly and horribly, and the book opens with several pages of a man being dismembered in a Paris square to make sure you feel the difference. That power was expensive, visible, and worked after the fact.

What replaced it was continuous, invisible, and worked in advance. It did not punish people for what they had done. It produced people who would not do it. The mechanism was visibility: arrange things so that the subject knows they may be observed at any moment and cannot verify when, and the subject takes over the work of discipline themselves.

Foucault’s formulation is that the person under this arrangement becomes the principle of their own subjection. They are not being coerced. They are doing it, and the doing feels like their own choice, which is the point. This is what he meant by discipline, and he traced the same structure through schools, hospitals, barracks, factories and examinations. The prison was the clearest instance rather than the only one.

The part that gets left out

Popular summaries stop there, which loses the more uncomfortable half.

Foucault was not describing a conspiracy of watchers. He was describing something closer to a technology, which anyone could pick up and which had no single owner. The panopticon is efficient precisely because it does not require a malevolent authority. It requires only an arrangement of visibility. The guard tower can be empty. It can be understaffed, indifferent, or manned by someone reading a newspaper. The effect holds.

He also thought it was productive rather than only repressive. Disciplinary power does not just prevent things. It manufactures a particular kind of subject: legible, comparable, ranked, self-monitoring. The examination, the file, the personnel record, the case history are all panoptic instruments, and what they produce is a person who can be managed, including by themselves.

Which means the question who is watching is usually the wrong one. Often nobody is, and it does not matter.

Does it actually work? The evidence

Foucault was a philosopher and historian rather than an experimentalist, and for a long time the panopticon’s psychological claim sat unmeasured. It has since been tested, and it holds up better than most theoretical constructs of its vintage.

The relevant literature is on chilling effects. When people believe they may be under surveillance, their behaviour changes even without any monitoring taking place.

The clearest natural experiment came after the 2013 Snowden disclosures. Traffic to privacy-sensitive Wikipedia articles dropped measurably after the revelations, and the drop persisted rather than rebounding, indicating both an immediate and a longer-term effect. Nobody had been told to stop reading those pages. Nobody was checking.

Controlled work found the same pattern. Making surveillance salient suppressed people’s intentions to engage in a range of online activities, with the effect extending to political behaviour, and holding across different populations and political contexts. The theoretical prediction and the empirical result agree, which is not something you can say about a lot of critical theory.

The most Benthamite finding is that it travels. Awareness of an online audience changes behaviour offline, in physical rooms where no device is present and the audience cannot see. Researchers called it the extended chilling effect. The watching does not stop at the edge of the platform, because the watcher was never really in the platform.

Why the metaphor is also wrong

Having made the case, the honest thing is to note the objection, because there is a serious one.

Surveillance scholars have argued for years that the panopticon is a poor fit for contemporary conditions and that its ubiquity as a metaphor has done real damage to the field. Bentham’s design was centralised, hierarchical, spatially bounded and aimed at a known population of inmates. Contemporary data collection is distributed, commercial, largely automated, and directed at populations who are not confined and often not identifiable. Nobody sits in the tower because there is no tower and no centre.

A counter-argument, which I find more persuasive, is that this reads Foucault too literally. His later work turned toward what he called technologies of the self, the ways people exercise power over themselves without any coercion at all. On that reading the panopticon was never mainly about who holds the tower. It was about what happens inside a person who assumes visibility, and that condition is more widespread now than in any prison Bentham imagined.

Both positions are defensible. What is not defensible is using the word as a synonym for surveillance is bad, which is most of its current usage and which throws away everything interesting about it.

What Bentham got right and did not intend

The design has one feature Bentham was proud of and modern readers usually miss. He wanted the inspector watched too. The tower was to be open to public visit, so that the watchers were themselves under inspection, and he thought this made the scheme democratic rather than tyrannical.

He was a reformer. The panopticon was intended as a humane alternative to hanging and transportation, and by the standards of 1791 penal practice it was. The horror of the thing is not that a cruel man designed a cruel machine. It is that a well-intentioned man designed an efficient one, and efficiency turned out to be the dangerous property.

That is the part worth carrying out of the concept. Not that someone is watching, which is often false and always beside the point. That an arrangement of visibility, set up by nobody in particular for reasons that seemed good at the time, will do the work of discipline on its own, and will keep doing it long after anyone has stopped paying attention.

The building was never built. It did not need to be.

FAQ

What is a panopticon?

A building design from 1791: a ring of cells around a central tower, each cell lit from behind so the occupant is permanently visible, the tower shrouded so the occupant can never tell whether anyone is inside. Jeremy Bentham proposed it for prisons, but also for workhouses, factories, hospitals, madhouses and schools. Its defining feature is that the tower does not need to be occupied. Uncertainty does the work.

What is the panopticon effect?

The behavioural change produced by believing you might be observed, whether or not anyone is observing. It has been measured rather than only theorised. After the 2013 surveillance disclosures, traffic to privacy-sensitive Wikipedia articles dropped and stayed down. Controlled studies find that making surveillance salient suppresses online activity, including political expression, across different populations. The effect does not require actual monitoring, only credible possibility.

What did Foucault say about the panopticon?

That Bentham had not designed a prison but drawn a diagram of modern power. In Discipline and Punish (1975) Foucault argued that older power was spectacular, occasional and punished after the fact, while modern power is continuous, invisible and produces compliance in advance. The panopticon shows the mechanism: arrange visibility so the subject assumes they may be watched, and the subject takes over the disciplining. He described the person under this arrangement as becoming the principle of their own subjection.

Was a panopticon prison ever built?

Not by Bentham. He spent around two decades and a great deal of his own money trying, and Parliament eventually compensated him for the failure. Several later prisons borrowed elements of the radial design, most famously the Presidio Modelo in Cuba, but Bentham’s scheme as specified was never realised. The design’s influence turned out not to depend on construction.