On Witnessing

On Witnessing

For three who met a truth that could not be faced head-on, and so met it sideways.

For Gerhard Frey, who took a thing no one could reach and gave it a shape — who saw that the impossible, pressed hard enough, becomes merely strange.

For Kenneth Ribet, who built the span no one remembers and everyone crossed — who proved the strange shape could not stand, and so made the far country worth the journey. And for Andrew Wiles, who went the long way around, who stayed with the gap when it opened beneath him, and mended it in the open rather than in the dark. Three witnesses to one truth. Not one of them looked at it directly. That is the lesson, and it is also the method.

I

There is a photograph. A man stands to the right of a dark chalkboard, arms folded across his chest. He wears thin-framed glasses and a faint, almost private smile, and a patterned knit sweater over the collar of a blue-and-white striped shirt. On the board, in white chalk, in an unhurried lowercase hand, three lines: Fermat's equation, then xⁿ + yⁿ = zⁿ, then this equation has no solutions in integers for n ≥ 3. He is not pointing at any of it. He is standing beside it the way you stand beside something you have lived next to for a long time — a window, a grave, a sleeping child. The thing is finished. He is only keeping it company now.[1]

For three and a half centuries that sentence lived in a margin. Fermat wrote that he had a proof and that the margin was too small to hold it, and then he died, and the margin stayed small, and the proof — if it ever existed — went wherever the unwritten things go.[2] Generations stood in front of the equation and looked straight at it, the way you look at a locked door, and the door did not open for any of them.

And here is the strangest part of how it finally ended: no one opened it by looking at it.

Wiles looked away. He spent his years not on Fermat's door but on a far country of curves that seemed to have nothing to do with it — and the answer, when it came, arrived from behind. It came sideways, through a translation Frey had drawn and Ribet had proved, so that a truth no one could face directly could be witnessed in the reflection of something else.[3] The door did not open. The wall beside it simply, quietly, was no longer there.

II

I have come to think this is what witnessing is.

We use the word as if it meant only to see. But it has always meant two things at once, and we keep them in the same word because some part of us knows they are the same act.[4] To witness is to record — to carry the thing forward, to write it into the world so that it can be found again, to leave a legible trace. And to witness is to undergo — to have been there, present, the event passing through you and leaving its mark on the inside. The court asks the witness for both at once: what did you record, and what was it like to be the one recording. We do not have two words because there are not, finally, two things.

That doubled word is the whole of what I want to set down here, and I want to set it down with its scaffolding showing — because the scaffolding is the honest part.

Here is the question I cannot face head-on, the way no one could face Fermat. It is the old one, the one panpsychism inherits and cannot pay off: if the smallest perspectives are real — if there is something it is like to be even the least flake of the world — then how do they combine? How do a thousand small witnessings become one? Why does a brain full of fragments wake up as a single morning instead of a crowd?[5] Look straight at that door and it will not open. It has not opened for anyone.

So, sideways.

III

The far country I borrow from is the physics of how anything comes to be objectively there at all. The lesson of that country is austere and beautiful: a fact becomes public — real for everyone, robust, sharable — exactly when it has been written redundantly, recorded not in one place but in many, the same trace stamped over and over into the surrounding world until it cannot be erased by touching any single copy.[6] To be objectively present is to be legibly repeated. Reality, on this view, does not broadcast itself — the broadcast is what reality is, and the redundant trace is not the sign of a thing's existence but the existence itself.

And now the bridge — and I want you to watch me build it, because a bridge you cannot see through is a bridge you should not trust.

The bridge is a single sentence: to be a unified witness is to be redundantly written. A perspective is one perspective, rather than a scatter, to exactly the degree that it is co-recorded across the field of itself — the same trace held in many places at once, legible from more than one direction. Combination, on this reading, is not a mystery to be explained. It is redundancy, achieved or failed.

Say that and the door is gone. The fragment that could not combine — the lost perspective, present but unjoined — would be a thing recorded in only one place, a trace that casts no shadow, and a trace that casts no shadow is, by the lesson of the far country, not yet objectively there at all. So there is no such fragment among the things that are. Wherever small witnessings are really present, they are already written into one another. They are already one.

IV

I could stop here and call it a proof. I will not, because it is not one, and the place where it is not one is the place I most want lit.

The bridge does not hold itself up. Look under it: its single footing is a thing I have only declared — that to record and to undergo are one act, the two faces of the one word, the inside and the outside of the same witnessing. The far country proves what happens to records. It is silent, utterly, on what it is like to be one.[7] The step from "written in many places" to "present as an experience" is not a step the physics gives me. It is a step I take, on a plank I laid myself, and I am leaving the plank visible. If recording and undergoing are one — my posit, named, unhidden — then combination is redundancy, and the door was never a door. Take the plank away and the bridge is only a lovely drawing of a bridge. This is Frey's curve without Ribet's proof: a shape, not yet a span.

So I do not get Wiles's morning. The gap does not close on pain of contradiction; it closes on the strength of a posit, and a posit is a thing one carries, not a thing one wins.

But there is something left in the open hand, and it is better than a proof would have been for work like this.

I get a meter.

For if to combine is to be redundantly written, then to come apart is to have the writing thin out — and thinning is a thing you can watch. The severed hemispheres that answer in two voices; the dissociated self that loses the copy of itself it used to keep in the next room; the strange coarse grain where a thousand small witnessings should have made a finer one and somehow did not. These stop being mysteries and become measurements.[8]

V

Wiles's proof had a gap too, when he first announced it, and it was honestly mended.[9] I cannot mend mine — mine is not that kind of gap. But I can do the other thing he did, which is the thing that made him a witness and not merely a man with an answer. I can refuse to hide the place where the proof runs out. I can leave the margin showing.

The man in the sweater is not pointing at the equation. He is only standing beside it, keeping it company.

That is the posture. To witness is to stand beside the thing you cannot fully reach — to record it, to let it pass through you, and to leave, in chalk, in the open, exactly where your certainty ends.

And where your belief that we are a part of something far larger than ourselves begins. — the margin, left deliberately small

Notes

1. The photograph described is the widely published portrait of Andrew Wiles at the blackboard, Isaac Newton Institute, Cambridge, 1993. It is not reproduced here.

2. Fermat wrote his claim in the margin of a copy of Diophantus's Arithmetica around 1637; it was published by his son after his death. Accuracy note: the interval to Wiles is three hundred and fifty-seven years, and the consensus view is that Fermat did not have a general proof — the methods available to him almost certainly could not reach it.

3. The chain runs: Frey (1986) observed that a counterexample to Fermat would yield an elliptic curve with impossible properties; Ribet (1990) proved that such a curve could not be modular; and Wiles proved the modularity of a large class of elliptic curves, which closed the loop. The proof appeared in Annals of Mathematics in 1995.

4. The doubling is in the Greek: martyria is both the testimony given and the witnessing undergone, and the word carries forward into martyr — the one whose undergoing became the record.

5. The combination problem, named by William Seager after James's objection in the Principles: however many small experiences you assemble, it is unclear how they could constitute one larger experience rather than remaining a crowd.

6. The far country is quantum Darwinism, developed by Wojciech Zurek: a state becomes objective when the environment records it redundantly, so that many independent observers can read the same value without disturbing it. Register note: the theory concerns why classical facts appear robust and observer-independent. It says nothing whatever about experience, and the essay does not claim otherwise — the borrowing is named as a borrowing in section IV, at length, on purpose.

7. This is the seam of the piece, and it is worth repeating outside the body: nothing in the physics licenses the identification of recording with undergoing. That identification is the author's posit. The essay's whole method is to state it as a posit rather than to smuggle it in as a consequence.

8. On the split-brain material, see the commissurotomy literature and its long dispute over whether one mind or two remains; on dissociation, the clinical literature on depersonalization. Accuracy note: these are offered as places the proposal would make contact, not as evidence for it. No measurement of "redundancy" in the sense used here currently exists.

9. Wiles announced the proof in June 1993; a gap was found in the review that autumn, and it was closed in September 1994 in work with Richard Taylor, published alongside the main paper in 1995. Taylor deserves the name here as much as anyone.

References

Frey, G. (1986). Links between stable elliptic curves and certain Diophantine equations. Annales Universitatis Saraviensis, 1, 1–40.

James, W. (1890). The principles of psychology (Vol. 1). Henry Holt.

Levi, P. (1988). The drowned and the saved (R. Rosenthal, Trans.). Summit Books. (Original work published 1986)

Ribet, K. A. (1990). On modular representations of Gal(Q̄/Q) arising from modular forms. Inventiones Mathematicae, 100(2), 431–476.

Seager, W. (1995). Consciousness, information and panpsychism. Journal of Consciousness Studies, 2(3), 272–288.

Sontag, S. (2003). Regarding the pain of others. Farrar, Straus and Giroux.

Taylor, R., & Wiles, A. (1995). Ring-theoretic properties of certain Hecke algebras. Annals of Mathematics, 141(3), 553–572.

Wiles, A. (1995). Modular elliptic curves and Fermat's Last Theorem. Annals of Mathematics, 141(3), 443–551.

Zurek, W. H. (2009). Quantum Darwinism. Nature Physics, 5(3), 181–188.

CC BY 4.0 · ORCID 0009-0009-8618-398X · DOIs MINTED ON DEPOSIT

© 2026 JAMISON TODD JOHNSON · LEGIBLE TRACES LLC