Dysmemic Pressure: Selection Dynamics in Organizational Information Environments

Executive Summary

Organizations staffed by intelligent people converge on collective delusion with predictable regularity. This paper identifies a structural mechanism, dysmemic pressure: the compound selection force that emerges when organizations compress information to coordinate at scale.

The analysis reframes organizational dysfunction from moral failure to selection outcome, which explains why standard interventions fail and what structural countermeasures might succeed.

The Mechanism

Three dynamics drive the pressure: strategic communication degradation (Crawford and Sobel, 1982), adverse selection in idea markets (Akerlof, 1970), and transmission bias (Boyd and Richerson, 1985). The paper formalizes the first through a hierarchical Crawford-Sobel model. In the single-channel case, as preference divergence b approaches 1/4 the equilibrium partition count falls to one and mutual information goes to zero: the channel degenerates to babbling.

A parameter sweep over 126 hierarchical configurations (7 layer counts, 6 bias levels, 3 noise levels, 50 Monte Carlo iterations each) shows that without noise the partition structure is preserved through the hierarchy, but with noise mutual information degrades monotonically, as the data processing inequality predicts. Near the babbling boundary (b = 0.24) even the first layer transmits only about 0.14 bits. The paper operationalizes dysmemic pressure as DP(k) = 1 − I(report_k, θ) / I(report₀, θ), a measurable index of information destruction at each layer.

Testable Propositions

The mechanism generates specific predictions:

• Preference divergence: communication accuracy correlates with incentive alignment between layers
• Evaluation coupling: information quality falls when the recipient of information also evaluates the sender
• Process capture: processes whose outputs evaluate participants come to optimize for evaluation rather than purpose
• Intervention decay: changing expressed norms without changing payoffs yields initial improvement, then regression
• External correction: under strong pressure, only information or consequences from outside the selection environment correct the organization

Three case studies (Boeing 737 MAX, Theranos, and Wirecard) illustrate the mechanism across industries. The paper notes they are illustrative, not controlled tests.

Relationship to the Cage

When compression and selection interact long enough, constructed reality, the population of signals, and the selection criteria become mutually consistent. That equilibrium is the Cage. No one designs it; it resists displacement because careers have adapted to it, the frame filters evidence about the frame, and unilateral deviation is punished. Countermeasures must change the selection environment: evaluation decoupling, external validation, prediction tracking with proper scoring rules, and protected channels for dissent.

Key References