Defensive Pessimism: Growth, Culture, and the Survival of Evidence

Technical Summary: Defensive Pessimism: Growth, Culture, and the Survival of Evidence

Problem Statement
The paper addresses a paradox of social learning: a society that is highly successful in adopting a productive technology may simultaneously become incapable of recognizing or correcting that technology's eventual failure. Success drives the expansion of a common social technology, which increases aggregate output but also raises the hazard of an "appropriative regime" (a state of systemic failure, congestion, or capture). Crucially, the very process of successful integration erodes the "outside practice"—the culturally sustained, skeptical alternative that provides the necessary counterfactual comparison and operational capacity to diagnose and reverse the failure. The central problem is determining when a growing society should preserve skeptical jurisdictions or populations, and under what conditions growth leads to "endogenous self-blinding," where the capacity for correction is destroyed faster than the failure arrives.

Methodology
The author employs a dynamic stochastic model of a growing society characterized by three core mechanisms:

1. Endogenous Scale and Hazard: Total social scale ($m_t$) grows exponentially. The hazard rate of transitioning from a productive regime ($P$) to an adverse regime ($Z$) is an increasing function of scale ($\lambda_t = \eta m_t^a$).
2. Practice Erosion: During the productive spell, the normalized stock of outside-practice capital ($s_t$) erodes at a rate dependent on scale ($\dot{s}_t = -b m_t^c s_t$). This capital represents the skills, records, and organizational vocabulary required to interpret failure and implement correction.
3. Sequential Correction: Correction is only possible after the adverse regime is recognized. Recognition depends on the absolute mass of cross-action comparisons (skeptics vs. adopters), while organizational correction depends on the remaining stock of practice capital ($s_T$). The expected delay in correction is inversely proportional to $s_T$.

The analysis focuses on the "hazard-clock" realization of the breakdown time and compares the elasticities of the breakdown hazard ($a$) and the erosion process ($c$). The paper also extends the model to political economy, distinguishing between the absolute mass of skeptics (which generates evidence) and their relative political share (which determines the ability to enact change). Finally, it introduces a mechanism for "collateralized cancellation," treating culture as a capitalized liability that must be funded if repealed.

Key Contributions and Results

1. The Scale–Attrition Trichotomy (Theorem 3.2)
The paper establishes a critical boundary condition based on the relative elasticities of breakdown and erosion:

• Hazard-Dominant ($c < a$): Growth accelerates the arrival of breakdown faster than it erodes corrective capacity. The expected time to correction is finite.
• Proportional ($c = a$): Growth accelerates both clocks equally. The distribution of capacity at breakdown is invariant to the growth rate, provided the erosion coefficient is sufficiently low ($b < \eta$).
• Erosion-Dominant ($c > a$): Growth erodes the outside practice faster than the breakdown risk accumulates. While breakdown occurs almost surely, the unconditional mean of the correction delay becomes infinite. This represents "endogenous self-blinding": the society manufactures the danger while destroying the institution needed to understand it.

2. The Mass–Share Separation (Proposition 3.3 & Theorem 4.1)
The paper demonstrates a divergence between epistemic viability and political mortality.

• Evidence vs. Power: Statistical evidence depends on the absolute mass of skeptics ($m_t \chi_t$), while political correction depends on their relative share ($\chi_t$).
• The Paradox: In a growing society where the skeptical share declines ($\chi_t \to 0$) but the total population grows faster ($g > \beta$), the absolute number of skeptics and the volume of diagnostic evidence can grow indefinitely. However, their political power vanishes.
• Majoritarian Unraveling: Under any fixed non-unanimous voting threshold, the skeptical bloc will eventually be outvoted and repeal the protective practice, even if the absolute mass of evidence is growing. The time to repeal is finite, leading to a loss of the counterfactual before the failure is recognized.

3. Asymmetric Reversal (Proposition 3.4)
The model formalizes the asymmetry between adoption and reversal. While a society can switch to a new worldview or technology immediately, rebuilding an abandoned outside practice takes time proportional to the duration of the successful spell that destroyed it. A society may recognize a failure overnight but remain operationally incapable of acting on that recognition for a prolonged period.

4. Capitalized Replication and Collateralized Cancellation (Theorems 5.1 & 5.3)
The paper challenges the view of culture as an irreplaceable ideology, reframing it as a "stock of epistemic capital" with a capitalized service path.

• Replication Boundary: A fully funded, externally enforced "pilot" can exactly replicate the action, evidence, and correction services of a skeptical lineage, provided the discount factor and growth of the practice satisfy $\delta M < 1$.
• Funded Cancellation: Because culture is a stock, cancelling it creates a liability. The paper proposes a "rolling performance bond" mechanism: a coalition wishing to repeal a protective practice must forfeit a bond to an independent trust. These funds finance the rebuilding of the practice (or a substitute pilot).
• Efficient Exit: This mechanism blocks opportunistic repeal (where the coalition cannot afford the replacement cost) while permitting reforms that are valuable enough to finance the replacement experiment. It ensures that the "external continuation value" of future evidence is internalized in the cancellation decision.

Significance and Claims
The paper claims to identify a specific margin of failure in social learning that is distinct from standard models of strategic experimentation or organizational dissent.

• Defensive Pessimism: The value of skepticism is not that pessimists are "wiser," but that their different behavior creates a causal counterfactual. Their practice is a "recovery technology" that becomes essential only after the dominant model fails.
• Institutional Design: The results suggest that optimal constitutional design must account for the depreciation of corrective capacity caused by success. Merely preserving a skeptical share is insufficient if the absolute mass of evidence is not protected or if the political share is allowed to vanish.
• Political Economy of Evidence: The paper highlights a fundamental tension: growth can make diagnostic evidence abundant while simultaneously making the political organization required to act on it negligible.
• Modest Scope: The author explicitly notes that the model does not claim to explain the Neolithic transition or rely on literary evidence (e.g., Golding's The Inheritors) as historical proof. Instead, it offers a formal analogue for "asymmetric reversal." The model assumes that statistical diagnosis does not automatically trigger correction and that outside practices do not rebuild autonomously before diagnosis; relaxing these assumptions would alter the results.

In summary, the paper argues that successful expansion can endogenously depreciate the very institutions required to diagnose and reverse that expansion's failures. The preservation of "defensive pessimism" is not a matter of ideology but of maintaining a capitalized stock of operational capacity and a counterfactual practice that can survive the political erosion of its share.