From Axioms to Irreversibility

Overview 1. The Impulsive Equation 2. Irreversible Divergence 3. The Truncation Fallacy 4. Global Permeation The Two Theorems Together

Overview

In Note 01, we laid down the four axiomatic pillars: the impulsive state-space, the bounded attractor, topological mixing, and the chaotic regime. These are the premises. This note derives the conclusions.

We prove two results that, taken together, constitute the mathematical core of Topoethics:

  1. Irreversible Divergence (Proposition 10): Any non-zero impulse creates a permanent separation between the actual trajectory and the "what-if" trajectory where the impulse never happened. The two histories never reconverge.
  2. Global Permeation (Proposition 12): This permanent difference is not confined to a local corner of the state space. Via topological mixing, the influence eventually reaches every region of the system.

Between these two results, we address the Truncation Fallacy: the mistaken belief that individual influence is "rounded to zero" by the vastness of the system.

The Logic of This Note

Note 01 asked: "What do we assume?"
This note asks: "What follows from those assumptions?"
The answer: your intervention permanently and globally rewrites the future—and this is a theorem, not a metaphor.

1. The Impulsive Equation

Combining Axiom 1 (State-Space Model) and Axiom 8 (Macroscopic Agency) from Note 01, we write the full impulsive differential equation governing the trajectory \( X(t) \):

The Impulsive Piecewise-Autonomous System

\[ \begin{cases} \displaystyle\frac{dX}{dt} = F(X), & t \neq \tau_i \\[10pt] X(\tau_i^{+}) = X(\tau_i^{-}) + I_i, & t = \tau_i \end{cases} \] where \( F \) is the locally Lipschitz autonomous vector field, \( \{\tau_i\} \) are the decision epochs, and \( I_i \neq 0 \) are admissible impulses (i.e., \( X(\tau_i^+) \in \mathcal{M} \)).

Piecewise-Autonomous: The Key Structural Insight

The vector field \( F \) is autonomous—it does not change over time. But the overall system is non-autonomous: the impulse times \( \{\tau_i\} \) and magnitudes \( \{I_i\} \) are not determined by the current state alone. They arrive from external encounters—chance meetings, unexpected events, new information. Each impulse discontinuously resets the state, which then serves as the initial condition for a new segment of the same autonomous flow.

The cumulative effect: life's trajectory is shaped by a sequence of externally imposed restarts, none of which were encoded in the original initial condition \( X(0) \).

Connection to CS: Interrupt-Driven Execution

In operating systems, the kernel executes a main loop until a hardware interrupt fires, saves the current register state, processes the interrupt, and resumes execution from the modified state. The impulsive ODE is structurally identical: \( F \) is the kernel, impulses are interrupts, and the post-impulse state \( X(\tau_i^+) \) is the new register file from which execution continues. The critical difference from a batch job is that the interrupt sequence \( \{(\tau_i, I_i)\} \) is not compiled in advance—it arrives at runtime from the external world.

2. Irreversible Divergence

We now state and prove the first core result. Consider two parallel universes: one in which an impulse \( I_i \) occurs at time \( \tau_i \), and one in which it does not. We ask: do the two trajectories ever reconverge?

Proposition 10: Irreversible Divergence

Under Axioms 1-3 and 8 (State-Space, Bounded Attractor, Chaotic Regime, and Macroscopic Agency), let \( X(t) \) be the trajectory with impulse \( I_i \neq 0 \) at \( t = \tau_i \), and let \( \widetilde{X}(t) \) be the null-intervention trajectory (i.e., \( I_i = 0 \)). Define the separation \( \delta(t) := \|X(t) - \widetilde{X}(t)\| \). Then:

  1. \( \delta(\tau_i^+) = \|I_i\| > 0 \).
  2. The asymptotic growth rate is strictly positive: \[ \limsup_{t \to \infty} \frac{1}{t - \tau_i} \ln \delta(t) > 0. \]
  3. In particular, \( \delta(t) \not\to 0 \): the two trajectories remain permanently distinct.
Proof Walkthrough:

Part (a) is immediate: at \( t = \tau_i^+ \), the impulse displaces one trajectory by exactly \( I_i \) while the other is unchanged, so \( \delta(\tau_i^+) = \|I_i\| > 0 \).

Part (b) uses the Chaotic Regime axiom. Since \( \lambda_{\max} > 0 \) (Axiom 3), there exists at least one direction along which infinitesimal separations grow exponentially on average. The impulse \( I_i \), being a finite vector in \( \mathbb{R}^n \), generically has a nonzero component along this maximally expanding direction, ensuring a strictly positive limsup. ("Generically" means for all directions except a measure-zero set that is exactly orthogonal to every expanding direction—a condition that a random perturbation satisfies with probability 1.)

Part (c) follows by contradiction: if \( \delta(t) \to 0 \), then \( \ln \delta(t) \to -\infty \), forcing the limsup in (b) to be \( \leq 0 \)—contradicting the strict positivity just established. Hence \( \delta(t) \not\to 0 \).

The Caveat: Infinitesimal vs. Macroscopic

The Lyapunov exponent is defined for infinitesimal perturbations in the tangent space \( T_X \mathcal{M} \). Applying it to a macroscopic impulse \( I_i \) involves an approximation: since \( \mathcal{M} \subseteq \mathbb{R}^n \), the finite displacement \( I_i \) can be identified with a tangent vector via the canonical inclusion \( T_X \mathcal{M} \hookrightarrow \mathbb{R}^n \), and the initial separation is dominated by \( \lambda_{\max} \) to leading order.

However, the proof requires only that \( \delta(t) \not\to 0 \)—not a precise growth rate. On a bounded attractor, the separation saturates at the diameter \( D \) of \( \mathcal{A} \), after which the two trajectories behave as uncorrelated points—entirely different histories.

What "Irreversible" Means Here

The claim is not that the separation grows forever. It is that the separation rapidly reaches system-scale magnitude, after which the two trajectories evolve as entirely different histories. The universe with your impulse and the universe without it are permanently, macroscopically distinct states.

Corollary 11: No Replaceable Individual

Under the axioms, removing any single impulse \( I_i \) from the sequence yields a trajectory that diverges irreversibly from the original. Therefore, no individual intervention is "replaceable." The absence of a single person is not a neutral event but a permanent alteration of the global state vector.

3. The Truncation Fallacy

Before proceeding to the second core result, we address a common objection: "Isn't individual influence eventually diluted to zero by the vastness of the system?"

We call this the Truncation Fallacy: it conflates the machine epsilon of numerical computation with the resolution of physical causality.

The Fallacy

A digital computer rounds values below its machine epsilon \( \varepsilon_{\text{mach}} \) to zero. This is a hardware constraint—finite-width registers cannot represent arbitrarily small values.

The universe possesses no analogous rounding operator for macroscopic events. Human interventions operate at classical scales (\( \sim 10^{-2} \) to \( 10^{0} \) m), exceeding the Planck length (\( \sim 10^{-35} \) m) by over thirty orders of magnitude. There exists no physical mechanism \( \operatorname{Fl}(\cdot) \) capable of mapping such massive structural changes to a true mathematical null vector.

Connection to CS: There Is No Fl() in the Universe

In IEEE 754 double-precision arithmetic, values below \( 2^{-1074} \) are flushed to zero because the 64-bit register cannot represent them. This is a property of the hardware, not of the numbers themselves. Concluding that individual influence is "below machine epsilon" is a category error: it applies a property of finite-width digital registers to analog physical reality, where no such width limit exists. The universe does not discretize its state into 64-bit words; it has no subnormal range, no flush-to-zero mode, and no rounding policy.

4. Global Permeation of Influence

Proposition 10 established that an intervention creates a permanent difference between the two trajectories. But permanence alone does not preclude the possibility that this difference remains confined to a small, inconsequential corner of the state space. The next result closes this gap: it shows that the difference permeates the entire system.

Proposition 12: Global Permeation of Individual Influence

Under Axioms 1-8 (requiring all axioms, including Topological Mixing), the influence of any impulse \( I_i \) eventually reaches every region of the state space in the following sense. Let \( U \) be any open neighborhood of \( X(\tau_i^+) \) (representing epistemic uncertainty about the exact post-intervention state) and let \( V \) be any non-empty open subset of \( \mathcal{M} \). Then there exists \( N \in \mathbb{N} \) such that \( f^n(U) \cap V \neq \varnothing \) for all \( n > N \).

The ensemble of possible futures compatible with our uncertain knowledge of the post-intervention state eventually overlaps with every region of \( \mathcal{M} \). The individual's influence is not merely preserved; it is distributed across the full topology of the system.

Proof Sketch:

Autonomous case. If no further impulses arrive after \( \tau_i \), the result follows directly from the Topological Mixing axiom (Axiom 6) applied to \( U \) and \( V \).

Piecewise-autonomous case. In reality, subsequent impulses \( I_{i+1}, I_{i+2}, \ldots \) interleave with the autonomous flow. The actual evolution of \( U \) is not \( f^n(U) \) but a composition \[ \Phi(U) = \bigl(\cdots \circ T_{I_{i+2}} \circ f^{n_2} \circ T_{I_{i+1}} \circ f^{n_1}\bigr)(U) \] where \( T_{I_j} \) denotes translation by impulse \( I_j \) and \( n_k \) counts the autonomous steps in the \( k \)-th inter-impulse interval.

We assume that the impulse sequence does not constitute a globally coordinated adversarial forcing that exactly cancels the expansive dynamics at every step—an assumption consistent with the finite, local nature of human agency (Axiom 8). Under this assumption, each autonomous segment partially stretches and folds open sets, and the interleaved translations further disperse them. The composite map \( \Phi \) retains asymptotic mixing, and \( \Phi^n(U) \cap V \neq \varnothing \) for sufficiently large \( n \).

The Two Results: Points vs. Ensembles

Proposition 10 is a statement about individual trajectories (points): the two histories never reconverge.
Proposition 12 is a statement about ensembles (open sets): the set of possible futures spreads across all of \( \mathcal{M} \).

Together: an individual's influence is both permanently real (the trajectory is different) and globally consequential (the range of its possible effects spans the entire state space).

The Two Theorems Together

Let us step back and see the full logical arc from Note 01 to this note:

  1. Axioms (Note 01): We assume a smooth state-space, a bounded attractor, topological mixing, a positive Lyapunov exponent, and macroscopic agency.
  2. Proposition 10: Any non-zero impulse creates a permanent separation. The two trajectories never reconverge. (Your action rewrites the future.)
  3. Truncation Fallacy: The universe has no mechanism to round this rewrite to zero.
  4. Proposition 12: The rewrite is not local. Via topological mixing, it eventually permeates the entire state space. (Your action rewrites all of the future.)

This is the mathematical core of Topoethics. Everything that follows in the paper—the refutation of counter-arguments, the layer separation between physical causality and institutional ethics, the analysis of death as structural mutation—builds on these two propositions.

As always, these are conditional results: they hold for any system satisfying the axioms. Whether any particular real-world system does so remains an empirical question outside the scope of the mathematics.