Trust Attractor Mathematics — The Trust Equation, Thermodynamic Games, and Computational Irreducibility

1. The Trust Equation

The preceding chapters discuss stability models, monitoring costs, scale-bounded coordination, and games. The compact accounting below proposes connections between them. Each dynamical or thermodynamic connection requires its own measured variables and model.

1.1 The Core Equation

Picture a team whose working hours split two ways: hours spent doing the shared work, and hours spent checking that everyone else is doing it. Every hour of checking is an hour the work loses. The Trust Equation makes that tradeoff exact.

Define the effective coordination capacity of a system:

$$\Omega = C - \kappa H$$

where:

• $\Omega$ = effective coordination capacity: the bandwidth actually available for productive coordination
• $C$ = total coordination channel capacity: what the system could achieve if all bandwidth served coordination
• $\kappa \in [0, 1]$ = control intensity: the fraction of system resources devoted to monitoring, enforcement, and compliance verification
• $H$ = enforcement overhead per unit control, calibrated in the same capacity units as $C$. A physical energy cost requires a separate conversion; the label compliance entropy does not supply one.

The equation has a simple physical meaning. Every system that coordinates has some channel capacity $C$: the total bandwidth available for exchanging signals, resources, commitments. When part of that bandwidth is consumed by enforcement (monitoring whether agents comply, punishing defectors, verifying reports), the effective capacity shrinks. $H$ measures enforcement overhead in this accounting model.

Consider the limiting cases:

Pure invitation ($\kappa = 0$): $\Omega = C$. No bandwidth consumed by enforcement. Full channel capacity available for coordination. The system coordinates through shared incentives, mutual benefit, and reputation: mechanisms that cost little beyond the coordination itself.

Pure coercion ($\kappa = 1$): $\Omega = C - H$. Maximum bandwidth consumed by enforcement. The system can still coordinate productively only if $C > H$: the total channel capacity must exceed the cost of total enforcement. For complex systems, this condition often fails.

Coordination collapse ($\kappa > C/H$): $\Omega < 0$. Enforcement costs exceed total channel capacity. The system is spending more on monitoring compliance than it produces through coordination. This is the point at which control becomes self-defeating. The inspectors need inspectors, the reports require reports on reports, and the bureaucratic overhead consumes the productive output it was meant to protect.

For constant $C,H>0$, this objective is linear and decreasing in $\kappa$, so its maximum on $[0,1]$ occurs at $\kappa=0$. Calling a maximizer an attractor additionally requires a dynamics for $\kappa$. Section 1.3 allows $C$ to depend on control, so even a boundary maximum at full control becomes possible.

Landauer’s principle¹ gives a reset cost for uncertain memory under specified thermal conditions. Measurement can be logically reversible: $(b,m)\mapsto(b,m\mathbin{\mathrm{XOR}}b)$ copies a bit into blank memory through a bijection. Cyclic apparatus may later erase information, but observation alone carries no universal per-bit-erasure floor. Converting physical reset energy to capacity overhead $H$ or game payoff $k$ requires a calibrated device and accounting model.

1.2 Reported Monitoring Comparison

The monitoring threshold experiments (Chapter 17b) provide direct measurements (Compliance Entropy Parameter Sweep 4.2, 252 parameter combinations, 2026-01-15; raw data in The Universal Algorithm/demos/experiments/results/compliance_entropy_sweep_20260115_040456.json):

The reported boost drops by about 42% between $\kappa=0$ and $0.5$, and reaches zero at $1$. These are cooperation differences, without a measurement of $C$ or $H$. A zero boost does not identify $\Omega=0$ or show that monitoring overhead exactly consumed coordination capacity. That causal accounting requires independent overhead and capacity measurements.

1.3 The Scale-Dependent Variational Form

The simple Trust Equation treats $C$ and $H$ as constants. In practice, both depend on the number of agents $N$ and the control regime $\kappa$. A richer formulation introduces this dependence explicitly:

$$\Omega^*(N) = \max_{\kappa} \left[ C(N, \kappa) - \kappa \, H(N) \right]$$

This is an optimization objective over $0\leq\kappa\leq1$, without a specified evolution law. At a smooth interior optimum, $\partial_\kappa C=H$; boundary optima remain possible. For example, $C=1+2\kappa$, $H=1$ gives $\Omega=1+\kappa$, maximized at full control. Positive overhead alone does not force $\kappa^*$ below one.

Small groups ($N < \sim 5\text{--}10$). In small groups, agents can track each other’s behavior through direct observation and reputation. Coordination capacity $C$ is largely independent of $\kappa$: adding formal enforcement does not increase what the group can achieve because informal mechanisms already suffice. Meanwhile, $H(N)$ is non-trivial; enforcement has real costs even in small groups. Result: $\kappa^* \to 0$. The Trust Attractor dominates. This matches the scale-boundedness data (Chapter 17b), which found trust-based coordination reliably superior for $N < \sim 5\text{--}10$ (revised downward from an earlier estimate of ~50 based on subsequent empirical work; see the empirical validation annex).

The monitoring-failure crossover ($N \approx 5$ up to the Dunbar number, $\sim 150$). As $N$ grows past roughly 5-10, informal monitoring fails. Agents cannot track everyone’s reputation through direct experience. (The Dunbar number, near 150, is the rough ceiling on how many stable relationships one person can track.) The function $C(N, \kappa)$ begins to increase with $\kappa$: some formal structure (norms, roles, protocols) genuinely increases coordination capacity. The optimal $\kappa^*$ rises above zero. This is the crossover observed in the scale-boundedness experiments, where the trust advantage begins to narrow.

Institutional scale ($N \gg 150$). For large populations, $C(N, 0)$ may be low: coordination without any structure is difficult among thousands or millions of strangers. Formal institutions (laws, contracts, and enforcement mechanisms) raise $C$ substantially. The optimal $\kappa^*$ may be moderate. The Trust Equation does not claim that all enforcement is wasteful at scale. Its optimum depends on the measured slope of $C(N,\kappa)$ relative to $H(N)$. An interior optimum and the decline beyond it need those comparative measurements.

A dynamical delay test needs its own gain and delay model. In $\dot y=-\alpha y(t-\tau)$, $1/e$ is a real-root collision and $\pi/2$ is the asymptotic-stability boundary. Neither $\kappa=\alpha$ nor a growing delay $\tau(\kappa)$ follows from the capacity accounting. Negative $\Omega$ and dynamical instability are separate hypotheses; oscillatory decay can remain stable.

1.4 Relationship to Existing Apparatus

The Trust Equation synthesizes several results already established in the main text:

Candidate Lyapunov function. $V_{trust}(x)=(1-a_1)^2+(1-a_2)^2+(1-trust)^2+control^2$ is squared distance from $(1,1,1,0)$. For dynamics $\dot x=f(x)$, its derivative is $\dot V=2(x-x^*)\cdot f(x)$. The capacity objective supplies no sign: $f=x-x^*$ gives $\dot V=2V>0$, while the negative flow gives $-2V$. A basin percentage requires the vector field, sampling measure, horizon, and convergence criterion; the reported 84/0/16 split cannot be derived from $\Omega$ or squared distance alone.

Agency formalism. The proposed $\mathcal A=|\partial J/\partial\pi|$³ measures an ambient sensitivity. Shrinking the feasible policy set need not shrink it: $J(\pi)=\pi$ has derivative one even on $[0,\epsilon]$. If practical agency means achievable outcome variation, the feasible directions and their scale must enter a separate measure. Full enforcement alone does not force this derivative to zero.

Causal entropic forces. Causal path entropy requires reachable trajectories, their probability measure, and a horizon.⁴ Capacity surplus $\Omega$ specifies none of these. More bandwidth can support a deterministic single path. Maximizing $\Omega$ has no demonstrated identity or monotonic relation with maximizing future path entropy.

What is new. The capacity equation and its scale-dependent optimum are proposed accounting tools. Their empirical calibration and connection to dynamics remain open; they supply no unifying variational theorem for all the models below.

2. Thermodynamic Game Theory — Three New Games

Classical game theory has been extraordinarily productive. Three extensions below ask how observation costs, changing payoffs, and participation rules alter particular coordination games. These idealizations describe specified models; their consequences require the dynamics and constraints of each extension.

Three modeling questions:

1. Monitoring costs. A standard Prisoner’s Dilemma specifies action-pair payoffs without a separate observation charge. A physical or economic implementation can add communication, measurement, and apparatus costs, which need their own accounting.

2. Static payoffs. The payoff matrix is fixed. Cooperate/defect yields the same rewards in round 1,000 as in round 1. In reality, systems change through interaction. Damage-and-repair cycles alter future payoffs. Relationships have memory in the physical, thermodynamic sense: they are path-dependent.

3. Population and participation. A two-player matrix specifies one encounter. Population dynamics additionally depend on matching, group size, and the right to leave. The scale-boundedness experiments test particular choices of those rules.

The extensions change particular payoff comparisons and equilibria. Their effects on cooperation depend on the parameters and population dynamics. In the Demon model below, positive monitoring cost can also stabilize the all-defector state; the fixed-population trust advantage supplies no universal favorable equilibrium direction.

2.1 The Demon Game

Relaxes: costless monitoring

The name honors Maxwell’s demon, the molecular gatekeeper of Chapter 17. The game stipulates a positive observation cost in payoff units; it does not derive that cost from Landauer’s reset bound.

Setup. Two players engage in a repeated interaction with standard Prisoner’s Dilemma payoffs: mutual cooperation yields the reward $R$ each, mutual defection yields the punishment $P$ each, unilateral defection yields the temptation $T$ for the defector and the sucker’s payoff $S$ for the cooperator, with $T > R > P > S$.

The modification: learning the other player’s action costs $k > 0$ units. Each round, a player first chooses a mode (either trust or verify).

• Trust mode: The player assumes the other cooperated and responds cooperatively. Cost: 0. Risk: if the other defected, the trusting player receives $S$.
• Verify mode: The player pays $k$, observes the other’s actual action, and follows a stipulated reciprocal response. Cost: $k$. Information: full.

The payoff matrix becomes:

$$\begin{aligned} \text{Trust-Trust:} \quad & (R, \; R) \\ \text{Trust-Verify:} \quad & (R, \; R - k) \quad \text{if both cooperate} \\ \text{Verify-Verify:} \quad & (R - k, \; R - k) \quad \text{if both cooperate} \\ \end{aligned}$$

When both players cooperate, verification is pure waste: the verifier pays $k$ to confirm what the trusting player got for free. The classical PD assumes $k = 0$, which makes verification costless and trust unnecessary. When $k > 0$, a verification arms race is self-defeating. Both players pay monitoring costs that produce no additional coordination.

Strategy comparison. The verifier is stipulated to cooperate with cooperators and defect against defectors. It is not a one-shot best response to a known cooperator: defection earns $T-k>R-k$. A repeated-game justification would need retaliation and discount assumptions. The comparison below is between the two stipulated strategies at a given population defection rate, without a universal threshold at $k=T-R$.

Worked example: Trust vs. Verify in a mixed population.

A player in trust mode against this population receives:

$$E[\text{trust}] = (1 - d) \cdot R + d \cdot S = (1 - d) \cdot 3 + d \cdot 0 = 3(1 - d)$$

The trusting player cooperates unconditionally. Against cooperators (fraction $1 - d$), they earn mutual cooperation $R = 3$. Against defectors (fraction $d$), they are exploited for $S = 0$.

A player in verify mode pays $k$, observes the opponent’s action, and follows a stipulated reciprocal response: cooperating with cooperators and defecting against defectors. Their expected payoff:

$$E[\text{verify}] = (1 - d) \cdot R + d \cdot P - k = (1 - d) \cdot 3 + d \cdot 1 - k = 3 - 2d - k$$

The verifier earns $R = 3$ against cooperators (same as trust mode) and $P = 1$ against detected defectors (the mutual-defection payoff, since the defector also faces a defecting response). The monitoring cost $k$ is subtracted regardless.

Trust mode dominates verify mode when:

$$3(1 - d) > 3 - 2d - k$$

Expanding: $3 - 3d > 3 - 2d - k$. Simplifying: $-3d > -2d - k$, which yields:

$$k > d$$

Trust dominates when monitoring cost exceeds the defection rate.

This result is notable for what it does not depend on. Carrying the general payoffs through the same derivation gives the threshold $k > d(P - S)$: trust dominates when the monitoring cost exceeds the defection rate times the gap between the mutual-defection and sucker payoffs, with no dependence on the cooperation payoff $R$. For the standard payoffs used here ($P - S = 1$), this reduces to the boxed result $k > d$. In a mostly-cooperative population ($d = 0.05$, i.e. 5% defectors), any monitoring cost above 5% of the payoff unit makes trust the rational strategy. Cheap monitoring is still wasteful if defection is rare enough.

Mixed population dynamics. The payoff advantage $k>d(P-S)$ compares trust with reciprocal verification, holding the opponent mix fixed. It does not show that defectors decline or cooperators assort. Defectors may exploit the additional trusting targets. Assortment requires a matching or exclusion rule, treated separately in the Invitation Game.

Replicator dynamics. Let $x_T$, $x_V$, and $x_D$ denote the population fractions of trust-mode, verify-mode, and defector players, with $x_T + x_V + x_D = 1$. Using the standard IPD payoffs already established ($T = 5$, $R = 3$, $P = 1$, $S = 0$), the fitness of each strategy in the well-mixed population is:

$$\begin{aligned} f_T &= R(x_T + x_V) + S \cdot x_D = 3(1 - x_D) \\ f_V &= (R - k)(x_T + x_V) + (P - k) \cdot x_D = 3 - 2x_D - k \\ f_D &= T \cdot x_T + P \cdot x_V + P \cdot x_D = 5 x_T + (1 - x_T) \end{aligned}$$

Trust-mode players cooperate unconditionally: they earn $R = 3$ from cooperators (both trust and verify) and $S = 0$ from defectors. Verify-mode players pay $k$ every round for observation (the cost is incurred before the opponent’s action is revealed), then follow the stipulated reciprocal response: they earn $R - k$ from cooperators and $P - k$ from detected defectors. Defectors earn $T = 5$ from exploitable trust-mode players and $P = 1$ from both verifiers (who detect and defect back) and other defectors.

The standard replicator equation governs each strategy’s growth: $\dot{x}_i = x_i(f_i - \bar{f})$, where $\bar{f} = x_T f_T + x_V f_V + x_D f_D$ is the population mean fitness. Strategies with above-average fitness grow; those below average shrink.

Interior fixed point. Setting $f_T = f_V = f_D$ yields two independent conditions. The first, $f_T = f_V$, gives $3(1 - x_D) = 3 - 2x_D - k$, which simplifies directly to $x_D = k$: the defection rate at equilibrium equals the monitoring cost. The equality $k=d$ is one condition for equal fitness; the strict inequality $k>d$ remains a pairwise fitness advantage at the current mixture. The second, $f_T = f_D$, gives $3(1 - x_D) = 1 + 4x_T$, which at $x_D = k$ yields $x_T = (2 - 3k)/4$. The candidate fixed point is interior exactly when $0<k<2/3$. At $k=2/3$ the trust fraction reaches the boundary; outside the interval these coordinates do not describe an interior population:

$$x_D^* = k, \quad x_T^* = \frac{2 - 3k}{4}, \quad x_V^* = \frac{2 - k}{4}$$

For $k = 0.10$: $x_T^* = 0.425$, $x_V^* = 0.475$, $x_D^* = 0.10$. For $k = 0.05$: $x_T^* = 0.463$, $x_V^* = 0.488$, $x_D^* = 0.05$. Trust and verify occupy nearly equal shares, with verify slightly dominant ($x_V^* - x_T^* = k/2$): the immune response slightly outweighs the coordination surplus at equilibrium, and the gap grows with monitoring cost.

Local instability. Eliminating $x_V=1-x_T-x_D$, the Jacobian at the interior point has

$$\operatorname{tr}J=\frac{k(k+2)}4>0,\qquad\det J=\frac{k(k-2)(3k-2)}4>0\quad(0<k<2/3).$$

Both eigenvalues have positive real part, so this is an unstable equilibrium, not a saddle. At $k=0.1$ they are approximately $0.02625\pm0.28295i$. It supplies no cooperation separatrix or critical mass $1-k$. On the invariant edge $x_V=0$, $f_D-f_T=1+x_T>0$: even 99% initial trusting cooperators lose to defectors. This countertrajectory refutes the proposed critical mass without establishing a global phase portrait. No link to the separate 84% basin follows.

Axelrod’s tournament results⁵ motivate asking how observation and computation costs change strategy rankings. In a score formed solely from action-pair payoffs, two implementations making the same moves earn the same score, however much one computes. An implicit computation-to-payoff penalty would require a specified penalty, deadline, or resource-to-action mechanism. The Demon Game supplies a separate costly-monitoring extension, without explaining the earlier tournament ranking through an unmeasured $k$.

The game-theoretic literature on costly monitoring is well-established. Ben-Porath and Kahneman (2003) formalized communication in repeated games with costly monitoring; Lehrer and Solan (2018) extended this to high-frequency interactions; Miyagawa, Miyahara, and Sekiguchi (2008) proved folk theorem variants for games with observation costs.^5a These works treat monitoring cost as an abstract economic parameter. The Demon Game uses an abstract payoff cost too. A physical implementation could add a measured reset cost, but no energy-to-payoff conversion is established here.

The Demon Game compares stipulated observation costs and exploitation losses. Landauer constrains erasure in a physical memory cycle; the payoff parameter $k$ and its dependence on scale require a separate model.

Expected payoff plotted against defection rate $d$. Each verification curve meets the trust line (no monitoring) at its own threshold $d^* = k$. Left of that crossing, where the defection rate falls below that curve’s monitoring cost, trust has the higher expected payoff among these two stipulated strategies: monitoring costs exceed the gains from catching defectors. Right of it, verification pays for itself.

2.2 The Invitation Game

Relaxes: mandatory participation

Setup. Players choose among three options: cooperate, defect, or exit. Exit yields a guaranteed fallback payoff $E$, where $P < E < R$. The fallback represents the outside option: what an agent can achieve alone or by finding other partners.

This apparently minor addition transforms the dynamics. In the standard PD, players are locked in: they must interact regardless of the other’s behavior. Coercive players can exploit cooperative ones indefinitely because the victim cannot leave. With exit available, the dynamics shift.

Exit can limit exploitation. Leaving ends one exploitive encounter. Whether a defector then receives $E$, finds another victim, or changes behavior depends on the matching and exclusion rules.

Assortment requires a mechanism. Exit ends a particular interaction. It does not supply identity memory, partner exclusion, preferential matching, reproduction, or removal of a defector from the population. Under anonymous random rematching, a defector may keep meeting new cooperators.

A conditional exclusion model. Suppose known defectors are permanently excluded from all resident cooperators, receive fallback $E$ thereafter, and payoffs are infinite-horizon time averages. A defector’s single exploitation is diluted away, leaving $E<R$; a resident earns $R$ per cooperative encounter. Under those premises it cannot invade that resident state. With discounting, its normalized payoff is $(1-\delta)T+\delta E$, and resistance requires $R>(1-\delta)T+\delta E$. At $T=5,R=3,E=2,\delta=0.1$, the mutant earns 4.7, so exclusion alone is insufficient for impatient players. This payoff comparison establishes no global convergence. With all-pairs cumulative payoffs a cooperator earns $R(n_c-1)$, rather than $Rn_c$; per-encounter and per-population payoffs must be kept distinct.

Optional participation is established territory in evolutionary games.^5b Loner strategies can generate cycling. The right to exit is one institutional component, with its consequences determined by matching, exclusion, and population dynamics.

Conditional sorting time. If a separate removal law stipulates $x_D(t+1)=(1-p_{exit})x_D(t)$ with $0<p_{exit}<1$, the first round reaching $x_D\leq\epsilon$ is

$$t_s=\left\lceil\frac{\log(x_D^0/\epsilon)}{-\log(1-p_{exit})}\right\rceil$$

for $x_D^0>\epsilon$. The coefficient $1/p_{exit}$ is only a small-$p_{exit}$ approximation. With the proposed $p_{exit}=(E-P)/(R-P)$, $E=2.7,R=3,P=1$ gives $p_{exit}=0.85$ and coefficient about 0.527; $E=1.5$ gives $p_{exit}=0.25$ and about 3.476. Exit from a pair is not the stipulated population removal law. A cooperator encounter factor $x_C(t)$ requires a different time-dependent calculation.

Convergence remains a hypothesis. No compliance-entropy functional and cost-modified vector field have been defined to show $\dot V\leq0$ or rule out cycles. Fixed asymmetric costs alone establish no spiral toward cooperation. Faster learned exit requires an additional adaptive state. Section 3.4 proposes a model to test patterns of coordination, without deriving their convergence here.

Real-world mapping. The Invitation Game makes testable predictions about systems with varying exit costs. The key variable is $E$, the value of the outside option. The conditional exclusion model motivates investigating whether better outside options reduce exploitation under actual matching and mobility rules. Five domains illustrate this with particular clarity:

1. Labor markets. Visa-tied employment, employer-controlled housing, and debt burdens can raise exit costs. Portable skills and alternative employers can improve an outside option. A real comparison must account for access to those alternatives and labor-market rules.

2. Open source software. Forking permits departure with the code, while maintainers, contributor networks, infrastructure, and migration costs still affect the outside option. Whether a fork forms a durable cooperative community is an observed governance outcome.

3. Platform lock-in. Network effects, data portability, and accumulated social graphs affect switching costs. Raising portability motivates a test of whether users can credibly leave and whether the platform responds.

4. International relations. Alternative trading partners can make departure from a particular arrangement more credible. Their value depends on actual terms, geography, and resources; the payoff model alone predicts no universal improvement in bilateral conduct.

5. The Berlin Wall. Restrictions on departure illustrate how preventing exit can help sustain a coercive arrangement. The historical collapse supplies no measurement of an unbounded $E$ or a mathematically identified Trust Attractor phase transition.

Top: four snapshots of the spatial simulation at t = 0, 10, 50, and 200 (payoffs T=5, R=3, P=1, S=0, exit E=2), showing agents self-sorting into cooperator clusters (green) as defectors (red) fragment and loners who exit (blue) buffer the boundaries. Bottom: time series showing the cooperator fraction rising as defectors decline; the curves are a nine-step moving average of the raw counts.

2.3 The Hormetic Game

Relaxes: static payoffs

Hormesis is biology’s name for the pattern in which a low dose of a stressor, followed by recovery, leaves the organism stronger than before: exercise damages muscle, and the rebuilt muscle is stronger. The Hormetic Game writes that pattern into the payoff matrix.

Setup. A repeated game where the payoff matrix evolves based on the history of play. Specifically, a damage-and-repair cycle (adversarial round followed by cooperative round) increases future cooperation payoffs.

Define the cooperation payoff at round $t$ as:

$$R_t = R_0 + \beta \cdot D_t$$

where $R_0$ is the baseline cooperation payoff, $\beta > 0$ is the hormesis coefficient, and $D_t$ counts the number of completed damage-repair cycles up to round $t$. A damage-repair cycle is defined as an adversarial round (at least one player defects) followed within $w$ rounds by a cooperative round (both players cooperate).

The hormesis coefficient $\beta$ captures the anti-fragility finding from the manuscript’s experimental data: systems that experience adversarial pressure followed by cooperative repair develop stronger coordination than systems that never faced adversity. The virgin-vs-post-repair immunity result (Chapter 17b) showed exactly this pattern: post-repair systems took zero damage from subsequent adversarial pressure, while naive systems were vulnerable.¹

Implications for equilibrium analysis:

In the standard repeated PD, the Folk Theorem establishes that cooperation can be sustained as an equilibrium through threat of punishment. The Hormetic Game adds a new incentive: cooperation is more valuable after conflict than before it. This means:

1. Cooperative repair can raise joint payoffs. A jointly realized repair path raises $R_D$ under the stipulated rule. It does not strictly dominate permanent defection against arbitrary opponents. A permanent defector earns $T$ repeatedly against an unconditional cooperator; against an unconditional defector, cooperating earns $S<P$ and mutual repair cannot occur.

2. Forgiveness has conditional value. Repair can increase future cooperation payoffs. Its value depends on repair probability, the cost and timing of damage, and discounting, as the Bellman model below makes explicit.

3. Strategic vulnerability is a hypothesis. Deliberately accepting damage requires a comparison with available undamaged paths under the same opponent and shock law. A rising cooperation payoff alone establishes no optimal vulnerability policy.

Standard game theory largely lacks mechanisms for this, though work has begun to explore the territory. Su, McAvoy, Wang, and Nowak (2019) formalized evolutionary dynamics with game transitions: mutual cooperation can shift the game to a more rewarding variant, while defection shifts it to a less rewarding one.^5c Their framework captures endogenous payoff evolution driven by interaction history. The Hormetic Game extends this by adding the specific prediction that damage-repair cycles produce stronger cooperation than undamaged cooperation, the anti-fragility signature observed in the manuscript’s experimental data. Su et al. model transitions; the Hormetic Game models strengthening through adversity.

The Hormetic Game makes cooperation payoffs endogenous and path-dependent. Dynamic programming handles this by including repair history in the state. Nash and subgame-perfect equilibrium remain applicable to the resulting specified dynamic game; an expanding payoff frontier changes the incentive calculation and the appropriate state description.

2.3.1 The Bellman Equation for Hormetic Cooperation

Define the state variable $D$ as the number of completed damage-repair cycles. The per-round cooperation payoff is $R_D = R_0 + \beta D$.

The Bellman equation is dynamic programming’s central recursion: the value of a state is today’s payoff plus the discounted value of whatever state comes next. The value of being in a cooperative relationship at hormesis level $D$ satisfies:

$$V_C(D) = (R_0 + \beta D) + \delta \left[ (1 - p_d) \cdot V_C(D) + p_d \cdot V_{\text{damage}}(D) \right]$$

where $\delta \in (0, 1)$ is the discount factor, and $p_d$ is the per-round probability of an adversarial shock (exogenous conflict). After damage:

$$V_{\text{damage}}(D) = S + \delta \left[ p_r \cdot V_C(D+1) + (1 - p_r) \cdot V_{\text{defect}} \right]$$

where $p_r$ is the probability of successful repair, and $V_{\text{defect}} = P / (1 - \delta)$ is the value of permanent mutual defection.

For the expected discounted return with the stated linear payoff growth, $V_C(D+1)>V_C(D)$ when $\beta>0$, $0<\delta<1$, and $0\leq p_d,p_r\leq1$. The bounded-growth solution of the recursion has slope $\beta/[1-\delta+\delta p_d(1-\delta p_r)]>0$. This orders the stipulated state values; deliberately taking damage still requires the separate cost comparison below.

2.3.2 The Phase Transition: When the Dilemma Dissolves

The Prisoner’s Dilemma is a dilemma because $T > R$: the temptation to defect exceeds the reward for cooperation. In the Hormetic Game, the cooperation payoff rises with $D$:

$$R_D = R_0 + \beta D$$

At the critical cycle count:

$$D^* = \frac{T - R_0}{\beta}$$

the cooperation payoff equals the temptation payoff: $R_{D^*}=T$. Above it, cooperation is the strict best response to a cooperating opponent. Against a defector, $S<P$ still makes defection optimal. The resulting game is a coordination game with both mutual cooperation and mutual defection as strict Nash equilibria. For integer $D\geq0$, the first strict crossing is $\max(0,\lfloor D^*\rfloor+1)$ when $\beta>0$.

Worked example. With standard IPD payoffs ($T = 5, R_0 = 3, P = 1, S = 0$) and $\beta = 0.5$:

$$D^* = \frac{5 - 3}{0.5} = 4 \text{ cycles}$$

After four cycles, $R_4=5=T$: cooperation and defection tie against cooperation. After five cycles, $R_5=5.5>T$: cooperation becomes the strict best response to cooperation, while defection remains optimal against defection.

The Hormetic Game has a phase transition at $D^*$. Below $D^*$, the game is a genuine dilemma where defection tempts. Above $D^*$, cooperation is the best response to cooperation; mutual defection remains an equilibrium. Damage-repair cycles do not merely sustain cooperation through reputation or punishment. They transform the game itself.

2.3.3 Conditions for the Phase Transition

The phase transition requires two conditions:

Condition 1: Sufficient hormesis coefficient. $\beta$ must be large enough that $D^*$ is reachable within the expected lifetime of the relationship. If $\beta$ is tiny, $D^*$ is astronomically large. An immunity result in another model supplies no numerical payoff coefficient here. Calibrating $\beta$ requires measuring how this payoff changes after a completed repair.

Condition 2: Sufficient patience. Consider the favorable case of one damage round at time zero followed by certain repair, with a permanent bonus $\beta$ beginning at time one and no further shocks. Relative to continuous cooperation at level $D$, the net value is

$$-(R_D-S)+\frac{\delta\beta}{1-\delta}.$$

The cycle pays only when $\delta>(R_D-S)/(R_D-S+\beta)$, assuming $R_D>S$. At $D=0$, $R_0=3,S=0,\beta=0.5$, this is $\delta>6/7\approx0.8571$. The future bonus is discounted because it starts after damage. Repair failure, additional delay, and later higher opportunity cost require their own Bellman comparisons; this favorable threshold alone supplies no universal forgiveness rule.

Top: the payoff frontier expands. The per-round cooperation payoff $R_D = R_0 + \beta D$ rises with each damage-repair cycle and crosses the flat temptation line $T = 5$ at $D^* = 4$, where cooperation and defection tie against a cooperator. Bottom: this plotted Bellman calculation caps strengthening after eight cycles, using $V_C(9)=V_C(8)$. It is a finite capped variant of the uncapped recursion displayed above. The curves show cooperative value for those stipulated shocks, repair probabilities, and discounts; they provide no unilateral dominance proof.

2.4 Synthesis

The three games test different mechanisms. Costly reciprocal verification yields a fixed-population payoff comparison. Exit plus permanent exclusion can resist defector invasion under sufficient patience. Repair raises cooperation payoffs under a stipulated history rule, producing a coordination game above threshold. Their distinct premises supply no universal convergence theorem or proof that every thermodynamic correction favors trust.

3. The Wolfram-Trust Connection

Stephen Wolfram’s classification of cellular automata into four behavioral classes has become one of the organizing frameworks of complexity science.⁷ The classification is empirical (derived from exhaustive computational surveys) and robust across different automaton rules, dimensionalities, and initial conditions.

The four classes offer an illustrative vocabulary for coordination patterns: uniformity, periodic organization, disorder, and localized complexity. No map from CA rules and trajectories to $C,H,\kappa$ has been specified. Uniformity does not force $C=H$ or $\Omega=0$. A finite $n$-cell binary system has Shannon entropy at most $n\ln2$; chaos does not make enforcement cost $H$ infinite.

*Each dot represents one of Wolfram’s 256 elementary cellular automaton rules, positioned by how well its space-time pattern compresses (horizontal: the gzip size ratio, low for repetitive patterns, high for random ones) and by persistence (vertical). The four coordination regimes are visible as tendencies, not as clean territories. Chaotic defection is concentrated on the low-persistence floor at the right.

Frozen coercion, periodic institutions and the complex Trust Attractor all crowd into the highly compressible column at the left, and the latter two also spread across the whole persistence range. The four space-time diagrams below show one representative rule per class: Rule 32 frozen, Rule 108 periodic, Rule 30 chaotic, Rule 110 complex. Class labels come from a threshold heuristic over these same two metrics rather than Wolfram’s canonical classification, and the marked Trust Attractor region is drawn by hand around the labeled complex rules: illustrative, not a measured boundary.*

This mapping is analogical; the figure supplies descriptive metrics without identifying capacity or enforcement cost.

3.1 Coercion as Computational Reduction

Computational universality enables encodings of arbitrary computations. It does not require full equal-energy simulation for every prediction or control task. A universal machine padded with idle steps can admit a faster simulator, and particular Rule 110 trajectories are readily predictable. Control of one observable or invariant can require far less information than prediction of the full trajectory.

A monitoring-cost theorem would need a control objective, admissible inputs, complexity measure, reduction, and physical energy model. None is supplied here. The halting problem limits general total prediction; it gives no hardware-independent per-agent energy lower bound or impossibility of every control scheme.^8,9

3.2 Trust as Limited Prediction

Trust can mean coordinating without a complete model of another agent, using shared incentives, prior interaction, and outcome monitoring. That is a practical proposal with its own costs and failure modes. A compressed trust score does not guarantee alignment or a minimum-energy implementation.

Even if overhead $H(n)$ grows linearly with complexity, capacity $C(n)$ may grow too. With $C(n)=2n,H(n)=n$, the surplus is $n(2-\kappa)>0$ for every $\kappa\in[0,1]$. A collapse at large $n$ needs relative growth such as $C/H\to0$ and a fixed positive lower bound on $\kappa$. Unbounded growth in a size limit does not make cost infinite at any finite size. The claimed universal control impossibility and trust-only necessity are withdrawn.

3.3 The Edge-of-Chaos Connection

Class 4 dynamics occupy the boundary between order (Classes 1 and 2) and chaos (Class 3). This is Langton’s “edge of chaos,” the phase transition where computational capacity is maximized.¹⁰

A coordination model could test whether patterns combine persistent local structure with adaptability. Calling pure coercion Class 1/2 or pure defection Class 3 remains an analogy; control intensity alone specifies no CA rule. Power laws or long correlations would require measured scaling diagnostics before an edge-of-chaos or universality-class interpretation.

3.4 Proposed Experiment: Two-Layer Trust Automaton

The arguments above are theoretical. They can be tested computationally.

Design. A two-layer cellular automaton:

• Layer 1 (Action): Each cell occupies a state in $\{cooperate, defect\}$. The update rule depends on neighbors’ actions and on the cell’s trust score from Layer 2.
• Layer 2 (Trust): Each cell maintains a continuous trust score $\tau_i \in [0, 1]$, updated based on the outcomes of interaction with neighbors. Cooperation from a neighbor increases $\tau_i$; defection decreases it. The update rule is asymmetric: trust builds slowly and erodes quickly (matching empirical findings on trust dynamics).

Update rules:

$$\tau_i^{t+1} = \Pi_{[0,1]}\!\left(\tau_i^t + \alpha \sum_{j \in \mathcal N(i)}[\gamma_+\mathbf1[a_j^t=C]-\gamma_-\mathbf1[a_j^t=D]]\right)$$

where $\Pi_{[0,1]}(z)=\min(1,\max(0,z))$ explicitly preserves the declared trust domain, $\gamma_+<\gamma_-$, and $\alpha$ is the learning rate. The earlier unprojected rule can exceed both bounds.

Action update: cell $i$ cooperates with probability $s(\tau_i^t-\theta)$, with $s(z)=1/(1+e^{-z})$ and threshold $\theta$. For $\tau\in[0,1]$ and $\theta\in[0.3,0.7]$, probabilities stay between about 0.3318 and 0.6682, so neither all-cooperation nor all-defection is absorbing. A steeper sigmoid would be a distinct parameterized model.

Finite-horizon hypotheses:

1. The system self-organizes into Class 4-like dynamics: localized cooperation clusters with complex boundary behavior, neither frozen order nor uniform chaos.
2. At a prespecified observation horizon, a majority of sampled runs have a cooperation fraction above 0.5 over the final 100 steps. This is a hypothesis about the measured finite-horizon success fraction below, without an established Lyapunov basin or literal convergence of the nonabsorbing action process.
3. The system exhibits hormesis: cooperation clusters that survive adversarial perturbation (random injection of defection) develop higher average trust scores and greater resilience than clusters that never faced adversity.
4. The dynamics exhibit scale-dependent behavior: small clusters ($N < \sim 5\text{--}10$ cells) maintain cooperation without formal structure; large clusters require emergent institutional patterns (stable boundary cells, internal trust gradients) to sustain coordination.

If these hypotheses hold, the automaton would supply finite-horizon evidence for cooperation patterns under these local rules. Establishing an attractor, a stationary law, or a Wolfram class requires the relevant dynamical and scaling diagnostics; the finite-horizon success fraction alone establishes none of them.

Implementation notes and sensitivity. The following parameter ranges and comparison metrics are recommended for the two-layer trust automaton experiment.

Proposed search ranges. Learning rate $\alpha\in[0.01,0.1]$, trust-build rate $\gamma_+\in[0.05,0.2]$, trust-erode rate $\gamma_-\in[0.2,0.5]$ with $\gamma_->\gamma_+$, and threshold $\theta\in[0.3,0.7]$ are proposed settings. No derivation supplies universal slow, oscillatory, or frozen cutoffs at their endpoints. Degree, clipping, runtime, and initial conditions all matter.

Grid sizes. A geometric sequence such as $50^2,100^2,200^2,400^2$ can test finite-size sensitivity. Fifty cells per side is a proposed starting size, without a proven minimum for cluster formation or phase-transition-like behavior. Finite simulations require a stated horizon and mixing diagnostic. The stochastic action rule above precludes literal frozen action states.

Comparison metrics. Prespecify a horizon of at least 100 steps, grid, update schedule, and sampling law before evaluating four finite-horizon observables. (1) Cooperation fraction: the fraction of cells cooperating at each step, averaged over the final 100 steps of the stated horizon. (2) Mean cluster size: the average number of contiguous cooperating cells, measured by connected-component analysis on the binary cooperation layer. (3) Spatial autocorrelation length $\xi$: the decay length of the two-point correlation function $C(r) = \langle a_i a_j \rangle - \langle a_i \rangle^2$ for cells separated by distance $r$, fitted to $C(r) \sim e^{-r/\xi}$. Large $\xi$ indicates long-range coordination; $\xi \sim 1$ indicates fragmented dynamics. (4) Finite-horizon success fraction: the fraction of independent runs, sampling initial trust scores uniformly on $[0,1]$ and initial actions as independent fair cooperate/defect choices, whose final-100-step mean cooperation fraction exceeds 0.5. Include update randomness in each run and report run count, uncertainty, and horizon sensitivity. This operational success fraction is distinct from a dynamical basin fraction.

Connection to the stability question. Basin comparisons require the continuous and discrete models’ own flows or transition kernels and the same operational convergence definition. The squared-distance function supplies no 84% target. A finite-horizon cooperation fraction is a measured outcome, without a convergence theorem from the capacity objective.

4. Geometric Questions About Coordination

4.1 Path-Law Costs and Redundant Descriptions

For a suitable stochastic-control model, a state cost and a path-law KL penalty can be combined as¹¹

$$\mathcal J[u]=\mathbb E_u\!\left[\int_0^T V(x_t)dt\right]+\lambda D_{KL}(p_u\|p_0).$$

The full path-law KL is charged once, outside the time integral. Its existence requires absolute continuity and the control/noise assumptions of the chosen model. A likelihood ratio creates no gauge redundancy. For $dX=u\,dt+\sigma\,dW$ with constant drift, $\mathbb E_uX_T=uT$ and $D_{KL}(p_u\|p_0)=u^2T/(2\sigma^2)$. Distinct drifts have distinct path laws; $u$ and $-u$ share a KL value while their distributions differ.

Distinct methods achieving one endpoint objective need not be physically identical path laws. To compare Mission Command with Detailed Command, specify the same passive measure and task, and nested feasible policy families. The broader family’s optimal cost can be no greater. No gauge orbit or Faddeev-Popov operator follows from this optimization comparison.

4.2 Candidate Symmetries

Participant permutation symmetry is discrete, so Noether’s continuous-symmetry theorem¹² supplies no fairness current. It also allows stable asymmetric equilibria: the symmetric potential $(x_1+x_2-1)^2+x_1^2x_2^2$ has strict minima at (1,0) and (0,1), with Hessian eigenvalues $3\pm\sqrt5>0$.

Time-translation symmetry yields the associated energy/Hamiltonian charge only in a suitable specified variational dynamics. An autonomous trust-decay law $\dot t=-t$ conserves no trust stock. Rotations of a specified variational action yield angular momentum, without a conserved option volume or compensatory reopening of alternatives. The proposed fairness, trust, and optionality currents are research descriptions, with no derived conservation laws here.

4.3 Suppressed Preferences and Gauge-Fixing

Faddeev-Popov gauge-fixing removes redundant descriptions and preserves gauge-invariant physics.¹³ Its determinant requires a gauge condition, group action, and operator $\delta F/\delta\theta$. Rescaling $F$ changes the determinant’s normalization. It cannot thereby measure physically suppressed preferences, or prove a coercive equilibrium becomes a saddle. Those preferences can affect an organization, but the effect needs actual behavioral dynamics and a stability calculation.

No ghost determinant or critical ghost fraction has been derived for coordination. An absorbing state is a transition-law property. Its presence alone does not assign directed-percolation universality: symmetries, interaction range, conserved fields, disorder, and dimension also matter. The two-dimensional Ising exponent 0.125 and directed-percolation exponent about 0.583 refer to different specified model classes, without a proved institutional class transition.

4.4 Holonomy and Alignment Drift

Parallel transport around a curved surface can rotate a carried vector on return. Some loops have identity holonomy even when curvature is nonzero. A geometric alignment model¹⁴ would need a context manifold, fibers, and a measured connection before it could compare contextual drift in RLHF and bilateral methods.

A local gauge relabeling transforms curvature and loop holonomy by conjugation; their gauge-invariant content is unchanged. It cannot flatten physical curvature or remove drift. Ongoing interactions might change a connection or the actual update dynamics. Whether that reduces alignment drift is an empirical and model-dependent question, without the flatter-bundle conclusion asserted by the earlier gauge analogy.

4.5 Dimension, Symmetry, and Coordination

Accessibility of cooperate and defect is not $Z_2$ invariance. A two-state chain with rates 1 and 2 allows both states while breaking label-swap symmetry. Universality classification requires its actual symmetry, interaction range, dynamics, conservation laws, disorder, and dimension. Effective dimension and a pair of available actions alone do not identify the Ising class or the Trust Attractor.²

For the finite-temperature short-range equilibrium Ising model, one dimension is the lower-critical boundary: the necessary dimensional comparison is $d>1$, rather than $d\geq1$. This is neither a sufficient condition for a coordination phase nor a theorem for absorbing-state dynamics. A driven one-dimensional contact process can have an active/absorbing transition. An autoregressive token sequence supplies no finite-range equilibrium Ising mapping, so token ordering alone gives no impossibility theorem for language-model correction.

The reported C6p logit-intervention and two-pass-correction results remain empirical comparisons (Chapter 22). Probe AUROC around 0.85 is an operating point in the tested setup, without an Ising coupling identification. The reported topology simulations and connectome fits likewise test particular models; they do not establish a universal social mapping merely from network dimension. Assignment to a universality class requires the full model and finite-size diagnostics.

Notes

¹ Landauer, Rolf, “Irreversibility and Heat Generation in the Computing Process,” IBM Journal of Research and Development 5(3) (1961): 183–191. Bennett later showed that computation itself can be thermodynamically reversible, but erasure (discarding information) cannot. A cyclic monitoring apparatus can incur reset costs for erased uncertainty. Logical observation itself can be reversible; physical accounting and an energy-to-payoff conversion remain separate.

² Wallace, Rodrick, “Cognitive Command and Control,” in Computational Psychiatry (Springer, 2017). See also Chapter 17 of the main text for the full development of Wallace’s stability analysis applied to coordination systems.

³ The agency formalism $\mathcal{A} = |\partial J / \partial \pi|$ provides a scalar measure of how responsive a system’s outcomes are to its own policy choices. [The original source for this formalism could not be verified; the formulation presented here follows the author’s reconstruction.]

⁴ Wissner-Gross, Alexander D., and Cameron E. Freer, “Causal Entropic Forces,” Physical Review Letters 110(16) (2013): 168702.

⁵ Axelrod, Robert, The Evolution of Cooperation (Basic Books, 1984). Axelrod’s tournaments demonstrated that simple, transparent, forgiving strategies outperformed complex, opaque, punitive ones, a result consistent with the Demon Game’s prediction that low-monitoring strategies dominate when monitoring costs are real.

⁶ Taleb, Nassim Nicholas, Antifragile: Things That Gain from Disorder (Random House, 2012). Taleb’s framework identifies systems that benefit from volatility. The Hormetic Game formalizes this in game-theoretic terms, connecting anti-fragility to the endogenous evolution of payoff matrices.

⁷ Wolfram, Stephen, A New Kind of Science (Wolfram Media, 2002). The four-class taxonomy was first published in Wolfram, “Statistical Mechanics of Cellular Automata,” Reviews of Modern Physics 55(3) (1983): 601–644.

⁸ Wolfram, Stephen, “Undecidability and Intractability in Theoretical Physics,” Physical Review Letters 54(8) (1985): 735–738. Computational irreducibility is related to, but distinct from, Turing undecidability. A system can be computationally irreducible (no shortcut exists) without being undecidable (the computation halts and produces a definite answer; you just can’t skip ahead to it).

⁹ This is a consequence of the computational universality of Class 4 systems. Wolfram conjectured, and Cook (2004) proved for Rule 110, that Class 4 cellular automata are Turing-complete. Universality supports arbitrary computation encodings and general undecidability results. It supplies no uniform finite-horizon simulation-speed or hardware-energy lower bound, and no requirement to predict full trajectories for every control objective.

¹⁰ Langton, Christopher G., “Computation at the Edge of Chaos: Phase Transitions and Emergent Computation,” Physica D 42 (1990): 12–37. Langton showed that a parameter $\lambda$ (the fraction of non-quiescent transitions in the rule table) controls the transition from Class 1/2 to Class 3, with Class 4 dynamics concentrated at the critical $\lambda$ value: the edge of chaos.

^5a Ben-Porath, Elchanan, and Michael Kahneman, “Communication in repeated games with costly monitoring,” Games and Economic Behavior 44 (2003): 227–250. Lehrer, Ehud, and Eilon Solan, “High frequency repeated games with costly monitoring,” Theoretical Economics 13(1) (2018): 87–113. Miyagawa, Eiichi, Yasuyuki Miyahara, and Tadashi Sekiguchi, “The folk theorem for repeated games with observation costs,” Kyoto University Working Paper 597 (2008).

^5b Batali, John, and Philip Kitcher, “Evolution of altruism in optional and compulsory games,” Journal of Theoretical Biology 175 (1995): 161–171. Hauert, Christoph, Silvia De Monte, Josef Hofbauer, and Karl Sigmund, “Volunteering as Red Queen mechanism for cooperation in public goods games,” Science 296 (2002): 1129–1132. Szabó, György, and Christoph Hauert, “Phase transitions and volunteering in spatial public goods games,” Physical Review Letters 89 (2002): 118101.

^5c Su, Qi, Alex McAvoy, Long Wang, and Martin A. Nowak, “Evolutionary dynamics with game transitions,” Proceedings of the National Academy of Sciences 116(51) (2019): 25398–25404.

¹¹ Kappen, Hilbert J., “Path integrals and symmetry breaking for optimal control theory,” Journal of Statistical Mechanics: Theory and Experiment 2005(11) (2005): P11011. Kappen showed that the Hamilton-Jacobi-Bellman equation for stochastic optimal control can be transformed into a path integral by exponentiating the cost-to-go, yielding a partition function whose free energy is the optimal cost. The KL penalty depends on the control and noise assumptions. A path probability ratio supplies no gauge redundancy by itself.

¹² Noether, Emmy, “Invariante Variationsprobleme,” Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-Physikalische Klasse (1918): 235–257. The theorem establishes a one-to-one correspondence between continuous symmetries of the action and conserved currents. The proposed coordination interpretations have no specified symmetry action or derived conserved current here.

¹³ Faddeev, Ludwig D., and Victor N. Popov, “Feynman diagrams for the Yang-Mills field,” Physics Letters B 25(1) (1967): 29–30. The Faddeev-Popov procedure resolves the overcounting problem in gauge-theory path integrals by introducing ghost fields that cancel unphysical degrees of freedom. Suppressed preferences are a behavioral modeling question; no Faddeev-Popov determinant or coordination stability theorem is established here. See also Yang, Chen-Ning, and Robert L. Mills, “Conservation of isotopic spin and isotopic gauge invariance,” Physical Review 96(1) (1954): 191–195, for the original non-Abelian gauge theory that the Faddeev-Popov procedure was developed to quantize.

¹⁴ The fiber bundle formulation of gauge theory is standard in mathematical physics; see Nakahara, Mikio, Geometry, Topology and Physics, 2nd ed. (CRC Press, 2003), chapters 9–11. The application to alignment is analogical: alignment is treated as a section of a principal bundle over context space, with the connection encoding how alignment transforms under context changes. The holonomy measures the net rotation of the alignment vector after parallel transport around a closed loop, geometrically identical to the Berry phase in quantum mechanics.

The formal relationship between Landauer’s principle and game-theoretic monitoring costs remains an open question for future work, as does the empirical measurement of monitoring costs in organizational settings and the integration of endogenous payoff evolution into repeated-game frameworks.

5. Delay Kernels and Their Distinct Boundaries

The lattice comparison and scalar delayed-feedback model describe different dynamics. Their numerical coincidences require a derived coarse-graining map before they can be identified.

5.1 Fixed Delay

For the stated equation¹⁵

$$\dot X(t)=\beta-\alpha X(t-\tau),$$

center at $X^*=\beta/\alpha$ and write $y=X-X^*$. With $\alpha,\tau>0$, $\lambda\tau=W_j(-\alpha\tau)$. At $\alpha\tau=1/e$, the leading real roots meet; beyond it they become complex. Oscillatory decay remains stable for $1/e<\alpha\tau<\pi/2$. The first imaginary-axis crossing has $\cos(\omega\tau)=0$ and $\omega=\alpha$, giving $\alpha\tau=\pi/2$. At equality asymptotic stability is lost. Neither the real-root collision nor the linear stability crossing supplies a catastrophic-collapse theorem for organizations.

5.2 Exponential Distributed Delay

An exponential feedback kernel is a convolution of past state, not simply a distribution of event waiting times:

$$\dot y(t)=-\alpha\int_0^\infty m e^{-ms}y(t-s)ds.$$

Its characteristic roots solve $\lambda^2+m\lambda+\alpha m=0$:

$$\lambda=-m/2\pm\sqrt{m^2-4\alpha m}/2.$$

For all $\alpha,m>0$, both roots have negative real part. The value $\alpha/m=1/4$ is the real-to-complex root boundary, not loss of stability. Initial-condition coefficients also matter when interpreting monotonicity of a particular observable; real roots alone guarantee no universal nonoscillatory organizational response.

5.3 Erlang Kernels

For $k$ sequential exponential stages of rate $m$, the mean delay is $\bar\tau=k/m$ and the kernel transform is $(m/(m+\lambda))^k$. The characteristic equation is $\lambda+\alpha(m/(m+\lambda))^k=0$. For $k>1$, the first imaginary-axis crossing satisfies

$$\arctan(\omega/m)=\frac{\pi}{2k},\qquad\alpha\bar\tau=k\tan\!\left(\frac{\pi}{2k}\right)\sec\!\left(\frac{\pi}{2k}\right)^k.$$

For $k=2$ this boundary is 4, for $k=3$ it is $8/3$, and as $k\to\infty$ at fixed mean delay it tends to $\pi/2$. The $k=1$ exponential model has no finite stability boundary.

The separate negative-real-root collision occurs at $\lambda=-m/(k+1)$ and $\alpha\bar\tau=(k/(k+1))^{k+1}$. Its values are:

These are collision values, without a theorem that every root is real or every response is monotone up to the table entry. For $k=3$, the characteristic quartic has complex roots even at $m=1$, $\alpha=0.01<27/256$: approximately $-1.11265748\pm0.17375243i$, with real roots $-0.76436907$ and $-0.01031598$. The 47% change between endpoint products supplies no organizational stability percentage. At fixed stage rate, the collision gain is $\alpha=m k^k/(k+1)^{k+1}$, which decreases from $m/4$ to $4m/27$ to $27m/256$ as stages are added. Fixed-stage-rate and fixed-total-mean comparisons must be kept distinct.

5.4 The Lattice Mapping Remains Open

Poisson spin-update waiting times are event holding times in a Markov generator. They do not yield an exponential convolution of past state. Discrete random-site updates have geometric waiting times, approaching exponential under suitable scaling. Identifying coercion fraction $p$ with dimensional gain $\alpha$, and making mean feedback delay one through a correlation-time unit, requires a derivation. Holding time and correlation time are different quantities.

The reported A15v2 finite-size crossover near $p=0.25$ and AS12 finite-size drift toward zero are lattice results, without a common threshold supplied by the exponential-delay model. The value $1/4$ is its root-character boundary, without a stability or channel-capacity identification.

5.5 Organizational and Token Dynamics

Erlang order could model a measured sequence of exponential processing stages. Organization depth alone gives neither exponential stage times nor a fixed total mean delay. Extra stages can increase latency, so the collision-product increase does not imply deeper hierarchies tolerate greater gain.

Autoregressive tokens condition on the prefix and are generally not memoryless as a token process. Enlarging the state to the full prefix can make it Markov, without an exponential feedback kernel or an assigned Erlang order. Chain-of-thought, transformer layers, and planning depth likewise supply no measured $k$. Mapping these architectures to a delay model remains a research programme, rather than a $1/4$ control bound.

¹⁵ Wallace, Rodrick, “Cognitive Command and Control,” in Computational Psychiatry (Springer, 2017), Section 5. The delay-differential equation and Lambert W analysis.

¹⁶ Wallace, Rodrick, “Cognitive Command and Control,” in Computational Psychiatry (Springer, 2017). The scalar fixed-delay and distributed-delay calculations above distinguish root collisions from stability boundaries. The Erlang calculation is the author’s extension; no lattice or organizational mapping is derived here.