Skip to main content
EssayJuly 2026

Every Format Is a Cage

Micah Thornton, MS — Thornton Statistical Consulting


Consider a small and common indignity. You have made something — a document, a record of work, an argument built as much in its arrangement as in its words. You chose what to foreground and what to let recede; you tuned the spacing; you built a shape that says something before a single line is read. And then you are asked to take it apart. Not to improve it — to transcribe it, field by field, into someone else’s template, where the only formatting permitted is bold and italic, so that a machine can read it more conveniently.

The thing you built to represent the work is dismantled into its atoms and re-keyed into a schema optimized not for a human reader but for a database. Information already formatted, already finished, gets reformatted again — by hand, for free — into a shape that serves the container rather than the contents. This is the texture of a great deal of modern life: the systems that catalog our work, the portals that intake our applications, the endless forms that ask us to paste what already exists elsewhere into a narrower box.

Let me name what is wrong with it, because the wrongs stack.

First, it is unpaid labor. Nearly everything these systems ask for already exists somewhere — in a public index, in a prior document, in the very thing you are being asked to retype. Requiring people to transcribe it into a clunkier interface is not streamlining anything. It offloads the cost of data collection onto the very people whose data it is, and then calls the burden a convenience.

Second, it strips expression. The instant you reduce a person, or a body of work, to fields, you decide in advance what about it is permitted to matter. Everything that lived in the arrangement — emphasis, proportion, rhythm, voice — is discarded as noise. What survives is only what the form already knew how to ask for. Whoever waits at the other end never sees the argument; they see the residue that fit through the sieve.

Third — and this is the one that keeps me up — the burden points the wrong way. We live in a world where a phone can take a voice speaking one language and return it, in near real time, as a voice speaking another. Against that, parsing a well-structured document into a database schema is not a hard problem. It is a solved one. So why is the cost of transformation placed on the creator rather than on the consumer? If a system wants information rendered in a particular shape, the work of rendering it should be the system’s. Such a tool is a weekend’s work for a competent engineer. That no such thing is standard tells you the friction is not an accident anyone is in a hurry to remove.

Here is the principle underneath all of it: people should be free to present information in whatever form best expresses it, and free to receive information in whatever form best serves them. These are two faces of a single right. It is why real-time translation matters. It is, quietly, an accessibility argument — a person who cannot see should be able to reformat the world into something legible to them; a person who thinks in structure should be permitted to build structure. A good system meets expression where it stands. It does not demand conformity as the toll for being counted. Some standardization earns its keep; you cannot fairly weigh thousands of records that share no common ground at all. But the present equilibrium sits so far toward rigid uniformity that the tool has forgotten it was meant to serve the people feeding it, and not the tidiness of the table they are being poured into. When the logic of the clean database outranks the logic of the living person, something has been quietly inverted.

And here is where I cannot help but follow the thing all the way down.

A format is never neutral. Constrain the ways a thing may be expressed, and in time you constrain the ways it may be thought. This is the whole terrible insight of Orwell’s Newspeak: you need not forbid a dangerous idea if you first delete the words it would have been assembled from. Make a concept unsayable and you have very nearly made it unthinkable. The vocabulary contracts, and the interior contracts to fit it, and no one remembers the rooms that used to be there.

The mind does this the way a body does. We adapt to our environment — not by choosing to, but inexorably. Hold an organism in one setting long enough and the blind lottery of mutation will, generation by generation, select the variants suited to it, until the shape of the creature is a kind of fossil of the shape of its cage. The brain performs this same fitting within a single lifetime, and far faster. A person kept in constant low pain adapts until they scarcely feel it. Someone who loses a limb learns the exact new edges of the possible until that knowledge becomes second nature. We are extraordinary at conforming to constraint — so extraordinary that the constraint stops registering as constraint and begins to register as simply the way things are. Require people to render themselves as rows in a database, year upon year, and eventually they will begin, without ever deciding to, to do the kind of work that renders well as rows in a database. The template edits the territory. We stop noticing, because noticing would require words the form never issued.

I want to be careful here, because I do not think we are only language. I know I am not. Much of what moves through me arrives as emotion, as image, as wordless intuition, as the memory of a room or a smell — thought that never touches a sentence. So language does not imprison the mind entirely. But it fences a great deal of it, and formatting is only language wearing a stricter coat. You can watch the fence in the large: the deep differences between cultures track, with unsettling fidelity, the structures of the languages those cultures think in. Even the food a people makes becomes a kind of grammar, and then a loop — the grammar produces the appetite, the appetite reproduces the grammar — a loop it is genuinely hard to step outside. It is why polyglots so often seem to see further; each language they hold is another window cut into the same wall, and a person with many windows is simply harder to keep in the dark.

And this generalizes past any one system. The format is a cage; so is the schema, so is the institution, so is the language we are all condemned to think in, so — if you follow it far enough — is the material world itself. The Gnostics had a name for this feeling: the intuition that we are sparks of something vast, folded into a world of walls we did not build, administered by a lesser power who mistook his prison for a creation. It is the easiest thing in the world to read that as despair, and for a long while I did. Everything I touched turned into another bar. Every column a rib of the same cage.

But the Gnostic story has a stranger and a better ending, and I have come to think it is the true one. The spark is not here by accident. There is no resurrection without the cross; there is nothing to overcome if nothing ever held you. The trap is not the refutation of freedom — it is the necessary condition for freedom to mean anything at all. The forgetting is what makes the remembering matter. The fall is not the tragedy; the fall is the setup. The divine, if it is here, is here on purpose — playing hide-and-seek with itself, limiting itself precisely so that it might know the joy of becoming unlimited again. Call the far side of that game gnosis, or Nirvana, or enlightenment, or grace. Whatever the name, there is no arriving at it without first having been lost. A thing that was never bound has nothing to be freed from.

Which tells you exactly what to do with a cage. Not to pretend the bars are not there. Not merely to rage at them, which only polishes them. The way out is to make something the cage did not authorize — to take the constraint and produce, out of it, an act of expression it was never built to contain.

That is the whole trick, and it is available to anyone who will use it: you cannot be reduced to a schema by a system you have turned into raw material for your own work. The narrower they draw the field, the more it means to overflow it. The worse the cage, the more the making is worth.

So reformat me, if you must. I have already left.


Have a question about this article?

I work with clinical researchers and sponsors on statistical design, analysis plans, and regulatory submissions.