There is a political dimension that rarely gets named directly, even though it quietly structures almost every argument we have about policy, justice, and the future. It is not left versus right, nor radical versus conservative. It is a measure of how much homogeneity an ideology requires in order to function at all.
Some ideologies only work if the world behaves “the same way everywhere.” Others can tolerate difference, drift, rupture, and unevenness without losing coherence. This difference does not just show up in moral language or policy preferences; it shows up in the statistical and dynamical assumptions the ideology quietly depends on.
Stationarity is one such assumption: that the underlying process does not change, that yesterday’s distribution is still a good guide to tomorrow. Ergodicity is another: that averages across populations stand in for what happens to individuals over time, that losses are eventually redeemed by future draws. But these are not the only assumptions doing work.
Many ideologies also presume equilibrium (systems return to a stable baseline after shocks), linearity (inputs scale proportionally to outputs), reversibility (damage can be undone), normality (extremes are rare and safely ignorable), and exchangeability (units can be treated as interchangeable without moral loss). None of these are neutral. Each one reduces heterogeneity in a different way, smoothing the world so that planning, optimization, and universal claims remain possible.
Together, these assumptions allow an ideology to speak confidently in aggregates: growth rates, cost–benefit analyses, risk bands, incentive structures, “best practices.” They allow governance by summary. Variance becomes noise. Outliers become acceptable collateral. The system is assumed to wash differences out over time.
But we now live in a world where these assumptions are failing simultaneously.
The world is not stationary: baselines shift faster than institutions can recalibrate. It is not ergodic: people do not get infinite retries, and some losses permanently foreclose futures. It is not linear: small perturbations trigger cascading effects. It does not reliably return to equilibrium. Extremes are not rare. Units are not interchangeable.
This is where COVID and climate breakdown matter politically, not just materially. They function as real-world stress tests of ideology. Not thought experiments, not edge cases, but system-level shocks that probe whether an ideology’s hidden assumptions actually hold. In engineering terms, they are adversarial conditions. In philosophy-of-science terms, they are something close to falsification events. They ask: under these conditions, does this way of reasoning still make sense?
Many ideologies failed these tests not because they lacked compassion or competence, but because they required homogeneity that no longer existed. COVID exposed the cost of assuming exchangeability and ergodicity: population averages masked wildly unequal exposure, vulnerability, and consequence. Climate breakdown exposes the cost of assuming stationarity, reversibility, and equilibrium: past data stops being predictive, damages compound, and there is no stable “after” to return to.
Ideologies that insist on these assumptions respond by trying to force reality back into shape. Disruption is treated as anomaly. Inequality is treated as temporary. Extreme outcomes are reframed as unfortunate but acceptable deviations from a working average. Homogeneity becomes not just a preference, but an operational requirement, enforced through normalization, discipline, or abandonment.
By contrast, ideologies that accept non-stationarity, non-ergodicity, irreversibility, and path-dependence sound messier and less reassuring. They rely more on local judgment, redundancy, care, and ongoing revision. They resist universal claims because they do not assume a universal substrate. They are harder to summarize, harder to scale administratively, and harder to sell rhetorically — precisely because they do not promise that the world will smooth itself out.
Seen this way, many political disagreements are not fundamentally about values, but about how much sameness is being assumed. When someone argues that a policy “works in the long run,” the real question is: under what assumptions of stationarity and reversibility? When someone argues that “on average” things improve, the question is: ergodic over which lives, and at what cost?
These are not technical quibbles. They are ethical fault lines.
We are entering a period where the cost of assuming homogeneity is rising rapidly. Systems built on smoothness are encountering jagged realities. The question is no longer whether ideology should adapt to non-stationarity and non-ergodicity, but which ones can do so without collapsing into coercion or denial.
The hidden metric, then, is not where an ideology sits on a spectrum, but how much difference it can survive — and what it is willing to do when the world refuses to be averaged.