Zeno’s paradox, in its classical form, suggests an apparently fundamental difficulty: to reach a target, one must first traverse half the distance, then half of the remaining distance, and so on. This results in an infinite succession of intermediate steps that seem to need to be completed before the final goal is reached.
Generalized Definition of Zeno’s Paradox Link to heading
A Zeno structure is a tuple:
$$ \mathcal{Z} = (\Sigma, d, \sigma_0, \tau, \phi, \mathcal{I}) $$
where:
| Symbol | Type | Interpretation |
|---|---|---|
| $ \Sigma $ | Metric (or topological) space | The set of all possible states of the system |
| $ d : \Sigma \times \Sigma \to \mathbb{R}_{\ge 0} $ | Metric | Measures the distance between two states |
| $ \sigma_0 \in \Sigma $ | Initial state | The point from which the evolution starts |
| $ \tau \in \Sigma $ | Target state | The state that must be reached |
| $ \phi : \mathbb{N} \to \Sigma $ | Trajectory | $ \phi(n) $ represents the state after $ n $ steps |
| $ \mathcal{I} \subseteq \mathbb{N} $ | Index set | Subset of steps that must be performed |
Zeno Condition (General Form) Link to heading
We say that $ \mathcal{Z}$ is a valid Zeno structure if and only if all the following are simultaneously satisfied:
$$ \textbf{(Z1)} \quad \mathcal{I} = \mathbb{N} \quad \text{(infinite steps)} $$
$$ \textbf{(Z2)} \quad \forall n \in \mathbb{N}: \quad d(\phi(n), \tau) > 0 \quad \text{(no step reaches the target)} $$
$$ \textbf{(Z3)} \quad \lim_{n \to \infty} d(\phi(n), \tau) = 0 \quad \text{(steps converge to the target)} $$
$$ \textbf{(Z4)} \quad \sum_{n=0}^{\infty} d(\phi(n), \phi(n+1)) < \infty \quad \text{(total progress is finite)} $$
The Paradox — Formal Statement Link to heading
Zeno’s paradox (general form) consists of the tension between:
- (Z1) + (Z2): The agent never reaches $ \tau$ in any identifiable finite step
- (Z3) + (Z4): The process as a whole converges to $ \tau$ and consumes finite resources
Mathematics does not discover that the rabbit catches the tortoise. Mathematics builds a universe (the ZFC universe) in which the proposition the sum of a convergent series is its limit is true by axiomatic construction.
| Axiom | What do we buy? | Ontological cost |
|---|---|---|
| Extensionality ZFC1 | Uniqueness of the set of steps — {s₁, s₂, …} is a unique, well-defined object | We accept that an object's identity is completely determined by its elements |
| Empty Set ZFC2 | Starting point for constructing ℕ, 0 := ∅ exists as a Von Neumann object | We accept that nothing is a legitimate mathematical object |
| Pairing ZFC3 | Stepwise construction of ℕ — n+1 := n ∪ {n} requires {n} to exist | We accept that any two objects can be grouped into a collection |
| Union ZFC4 | Gluing steps into a coherent path — n ∪ {n} exists | We accept that a union of sets is itself a set |
| Power Set ZFC5 | Existence of ℝ as a set-theoretic object — Dedekind cuts live in 𝒫(ℚ); without this axiom, ℝ does not exist and convergence has nowhere to land | Cardinality jump: |𝒫(ℚ)| > |ℚ| We accept that the set of all subsets of an infinite object is itself an object |
| ⚡ Infinity ZFC6 | Right to speak of infinite steps — without it ℕ does not exist, the paradox cannot even be formulated or solved | We accept that actual infinity exists as an object, not just as a potential process |
| Separation ZFC7 | Construction of Dedekind cuts — A = {q ∈ ℚ | q < r} is a legitimate set; without it ℝ remains a slogan, not an object | We accept that any property expressible in ZFC carves out a real set — the property becomes ontologically productive |
| Replacement ZFC8 | Legitimacy of the sequence φ(n) = D/2ⁿ as an object — the image of a definable function on ℕ is itself a set | We accept that infinite functions are objects, not just calculation rules |
| Regularity ZFC9 | Induction without circularity — excludes pathological sets of type A ∈ A | We accept the well-founded hierarchy of the set-theoretic universe |
| Choice ZFC10 | Strong completeness of ℝ — every Cauchy sequence converges; without it the limit in (Z3) may not exist in ℝ | We accept selections without explicit algorithm — existence without constructibility |
Completeness of $ \Sigma $ is not a physical property — it is an axiomatic choice, grounded in ZFC through the Axiom of Infinity, Separation, and Power Set.
In other words, the rabbit must continue the race without any perturbation at each of these infinite samplings. Whether the physical universe honors this construction — that is a question for physics, not mathematics.
Let us now introduce a small variation.
Variant 1 — Constant risk at each step Link to heading
We assume that at each step of the argument there is a probability $ p > 0 $ that the rabbit does not continue the race (an obstacle, a distraction, a random event — anything). In order to catch the frog, the rabbit must successfully pass through all steps. The probability of success after $ n $ steps is $ (1-p)^n$, but the number of steps is infinite.
$$ \lim_{n \to \infty} (1-p)^n = 0 $$
Therefore, if there exists even a positive risk at each sampling, the rabbit almost surely does not reach the frog.
Variant 2 — Risk over time, not per step Link to heading
Let us now assume a more realistic model: risk occurs continuously over time, at a constant rate $ \lambda $.
In this case, the probability that the rabbit continues the race until time $ t $ is
$$ S(t) = e^{-\lambda t} $$
The total time required to catch the frog, resulting from the geometric series, is finite and we denote it by $ T $.
The probability of success becomes
$$ P(\text{capture}) = e^{-\lambda T} $$
The total time $ T$ is precisely the sum of the series from the classical paradox:
\begin{aligned}
P(\text{capture}) &= \prod_{n=1}^{\infty} e^{-\lambda \cdot t_n} \\
&= \prod_{n=1}^{\infty} e^{-\lambda \cdot \frac{D}{2^n \cdot v}} \\
&= e^{-\sum_{n=1}^{\infty} \lambda \cdot \frac{D}{2^n \cdot v}} \\
&= e^{-\frac{\lambda D}{v} \sum_{n=1}^{\infty} \frac{1}{2^n}} \\
&= e^{-\frac{\lambda D}{v} \cdot 1} \\
&= e^{-\frac{\lambda D}{v}} \\
&> 0
\end{aligned}
where $ D$ is the initial distance and $ v > 0$ the rabbit’s speed.
Variant 3 — Minimum Planck distance Link to heading
Let $ \ell_P \approx 1.616 \times 10^{-35}$ m be the Planck length — the limit below which the notion of spatial distance loses physical meaning. The Zeno decomposition stops at step $ N^*$, the first step for which:
$$ \frac{D}{2^n} < \ell_P $$
Resolving:
\begin{aligned}
N^* &= \left\lceil \log_2 \frac{D}{\ell_P} \right\rceil \\
\end{aligned}
Thus the number of steps is no longer $ \aleph_0$ — it is finite: 123.
Variant 3 dissolves the paradox at the physical level, not at the mathematical one:
| Description | Variant 1 | Variant 2 | Variant 3 |
|---|---|---|---|
| Steps | $ \aleph_0 $ | continuous on $ [0,T] $ | $ N^* < \infty $ |
| Saving mechanism | none | $ T $ finite | $ \ell_P $ truncates the series |
| $P(\text{capture})$ | 0 | $ e^{-\lambda D/v} > 0 $ | $ (1-p)^{N^*} > 0 $ |
| Is the Axiom of Infinity required? | Yes | Yes | No |
Variant 3 is the only case in which the Axiom of Infinity (ZFC6) is not invoked — the physical universe replaces it with a natural cutoff:
\begin{aligned}
\boxed{\ell_P < \frac{D}{2^n} \implies n < N^* < \infty \implies \text{ZFC6 becomes irrelevant}}
\end{aligned}
Physical setup Link to heading
In order to test the rabbit’s position, we send a quantum of light (a photon). This introduces an unavoidable perturbation — through the Heisenberg uncertainty principle:
$$ \Delta x \cdot \Delta p \geq \frac{\hbar}{2} $$
Each photon sent perturbs the rabbit with probability 0.5% 1 — constant at each step, since the perturbation comes from the act of measurement.
With a perturbation of $ 0.5%$ per photon and 123 real physical tests — the limit imposed by the Planck length — the rabbit catches the frog with probability only 53.98%: a simple coin toss.
Zeno’s paradox does not disappear through mathematics — it disappears through physics, and reappears through quantum mechanics.
Conclusion Link to heading
Zeno’s paradox, in its generalized canonical form, is a tension between local description (step by step, finite, inaccessible at the limit) and global description (convergent, finite as total resource, but realizable only if we accept that the state space is complete).
The generalization $ \mathcal{Z} = (\Sigma, d, \sigma_0, \tau, \phi, \mathcal{I})$ allows the same formal framework to be applied in probability, information theory, stochastic processes, and functional analysis — wherever the tension between infinite process and finite result appears.
-
Solvency II: $ p = 0.5%$ is exactly the annual risk level accepted in SCR calibration — ↩︎