Every Theorem Encrypts a Farce: Formal Structure

Mathematical proof is a trapdoor function: the contingent accidents of a theorem's discovery are irreversibly encrypted by formalization, and elegance measures encryption strength — the shorter the proof, the more completely its historical farce is destroyed.

Preliminary Observation

A proof erases the history that produced it. This erasure is not incidental — it is the mechanism by which a theorem acquires generality. The question is whether this erasure has a precise mathematical character.

It does. It is a trapdoor compression with irreversible information loss, and its strength is proportional to elegance.


Definitions

Definition 1 (Contingency basis). Let $\tau$ be a theorem in a formal system $\mathcal{F}$. A contingency basis for $\tau$ is a minimal sequence

$$\mathcal{C} = (c_1, c_2, \ldots, c_n)$$

of historically irreducible events — choices of axiom system, notational conventions, failed proof attempts, analogies imported from other domains, misreadings that turned productive — such that no proper subsequence of $\mathcal{C}$ would have produced $\tau$ in the form we have it.

The elements $c_i$ are the primes of the theorem’s origin. They are irreducible in the sense that each contributed a constraint not deducible from the others.

Definition 2 (Formalization map). The formalization map is

$$\varphi: \mathcal{C}^* \to \mathcal{F}$$

sending a contingency history to the theorem it crystallizes. Here $\mathcal{C}^*$ is the space of all finite contingency sequences.

Definition 3 (Proof as verification certificate). A proof $P$ of $\tau$ is a certificate such that:

  • Verification: Given $(\tau, P)$, checking validity is $O(\text{poly}(|P|))$.
  • Recovery: Given only $\tau$ and $P$, determining any $\mathcal{C}$ such that $\varphi(\mathcal{C}) = \tau$ is not computable from $(\tau, P)$ alone.

This is exactly the structure of a trapdoor function.


The Encryption Theorem

Theorem (Irreversibility of Formalization). Let $\tau$ be a theorem with proof $P$, and let $\mathcal{C}$ be a contingency basis for $\tau$. Define the contingency entropy:

$$H(\mathcal{C} \mid \tau, P) = -\sum_{\mathcal{C}’ \in \varphi^{-1}(\tau)} \Pr(\mathcal{C}’) \log \Pr(\mathcal{C}’)$$

Then:

$$H(\mathcal{C} \mid \tau, P) > 0$$

The proof never fully determines its own contingent origin.

Proof sketch. The map $\varphi$ is many-to-one. Multiple distinct contingency histories produce the same theorem: Euler’s proof is not Riemann’s proof is not the proof a machine finds. Each such history constitutes a distinct preimage under $\varphi$. Since $|\varphi^{-1}(\tau)| > 1$ for any non-trivial theorem, the conditional entropy is strictly positive. $\square$


The Elegance-Encryption Duality

Conjecture. Let $P_{\min}$ be the shortest proof of $\tau$ (its Kolmogorov complexity in the proof system), and let $P$ be any proof. Define the compression ratio:

$$\rho(\tau, P) = 1 - \frac{|P|}{|\mathcal{C}|}$$

where $|\mathcal{C}|$ is the description length of any contingency basis. Then:

$$H(\mathcal{C} \mid \tau, P) \geq \rho(\tau, P) \cdot H(\mathcal{C})$$

The contingency entropy is bounded below by the compression ratio times the total contingency entropy. In other words:

$$\boxed{\text{Elegance} \propto \text{encryption strength.}}$$

The more a proof compresses its contingent origin, the more irreversibly it encrypts it. A one-line proof of a theorem whose discovery took fifty years of combinatorial accident is maximal encryption.


Corollaries

Corollary 1 (The farce is the plaintext). The contingency basis $\mathcal{C}$ — the wrong turns, the café conversations, the notational false starts — is the message being encrypted. The theorem is the ciphertext. The proof is the public key: it lets anyone verify, but it does not let anyone decrypt the history.

Corollary 2 (Multiple factorizations). Just as $N = pq$ can sometimes be factored in multiple ways in more general rings, a theorem may have multiple contingency bases. Each is a distinct farce encrypted into the same dignity. Independence of proofs is independence of factorizations.

Corollary 3 (Dignity as encryption-worthiness). Not all contingency histories produce theorems. Most farces encrypt to noise — the formalization map $\varphi$ is partial. A theorem exists precisely when a contingency history is worth encrypting: when the lossy compression preserves something that was load-bearing in the original, while discarding everything person-shaped, accident-shaped, epoch-shaped. The dignity is not in the theorem’s truth. It is in the fact that the farce was compressible at all.


Connection to Kolmogorov Complexity

Let $K(\tau)$ be the Kolmogorov complexity of $\tau$ (the shortest program that outputs it). Then:

$$K(\tau) \leq |\mathcal{C}|$$

always, since the contingency history is one (verbose) program that generates $\tau$. The gap

$$\Delta = |\mathcal{C}| - K(\tau)$$

measures the destroyed information — the bits of farce that the theorem consumed without preserving. This $\Delta$ is the entropy production of formalization.

By the second law of thermodynamics applied to computational processes:

$$\Delta \geq 0 \quad \text{(Landauer’s bound on irreversible computation)}$$

Formalization is thermodynamically irreversible. The heat it dissipates is the farce.


Remark on Auto-catalysis

This connects to the sibling insight: the catalyst does not return. If a theorem $\tau_1$ catalyzes the discovery of $\tau_2$ — serving as lemma, analogy, or scaffolding — then the “catalyst” $\tau_1$ has already encrypted its own contingency basis $\mathcal{C}_1$. When it participates in producing $\tau_2$, the contingency basis $\mathcal{C}_2$ includes elements of $\mathcal{C}_1$, but only as filtered through $\tau_1$‘s encryption. The original contingencies are doubly encrypted: first by $\tau_1$, then by $\tau_2$.

In the limit of long derivation chains:

$$\lim_{n \to \infty} H(\mathcal{C}_0 \mid \tau_n, P_n, \ldots, P_0) = H(\mathcal{C}_0)$$

The original farce becomes maximally encrypted — indistinguishable from noise. This is why ancient theorems feel necessary rather than contingent. The encryption is complete.


This writing connects to 20 others in sisuon’s corpus. More will be published over time.