Until very recently, I was not aware that there exists a mathematical literature on revolutions and collective uprisings. That fact alone seemed worth pausing over. At first glance, the idea sounds slightly surprising: one does not instinctively associate revolutions with threshold functions, stochastic transitions, hazard rates, or metastable states. But on closer inspection this is not strange at all. Whenever a system consists of many interacting agents whose behavior depends on what they believe others are doing, the possibility arises that the macroscopic order may change abruptly even when the underlying parameters change only gradually.
The text below was generated by AI on the basis of prompts and editorial guidance. It should be read as an exploratory essay in mathematical and conceptual modeling, not as a contribution to political science in the strict disciplinary sense, and not as a predictive statement about any specific country.
Revolutions as Stochastic Phase Transitions
Threshold Models, Hidden Preferences, and Public Order
With Iran in Spring 2026 as a Case
Abstract
This article examines a class of mathematical approaches to revolutionary dynamics, with emphasis on threshold models, preference falsification, stochastic state transitions, mean-field approximations, and coupled protest-security systems. The central claim is limited but important: public political stability need not coincide with deep structural stability. Under certain conditions, regimes may occupy metastable states, appearing outwardly robust while underlying parameters drift toward a region in which abrupt macroscopic reorganization becomes possible. Iran in spring 2026 is treated here not as a predictive object but as a contemporary case that makes such questions vivid. The aim is not to reduce history to equations, but to clarify how mathematical structure may illuminate the relation between latent dissent, public conformity, and sudden collective activation.
Keywords. revolutionary dynamics; threshold models; preference falsification; stochastic processes; hazard functions; metastability; Markov jump processes; collective behavior; political instability; Iran
1. Introduction
The study of revolutions is usually situated within history, political sociology, comparative politics, or area studies. Yet several of its central questions are structurally intelligible in terms familiar from mathematical modeling. Why do societies sometimes remain publicly quiescent for long periods despite substantial private discontent? Why do regimes that appear stable occasionally undergo rapid collapse? Why do comparable levels of frustration produce protest in one case but not in another? And why do such transitions often appear less as smooth developments than as abrupt changes in the visible state of the system?
These questions do not require the assumption that history is reducible to mathematics. They require only the weaker claim that there may exist formal mechanisms capable of generating long periods of apparent equilibrium followed by sudden collective change. Among the most relevant such mechanisms are threshold-dependent activation, hidden preference structure, stochastic fluctuation, nonlinear feedback, and coupling between protest dynamics and coercive institutions.
Iran in spring 2026 offers a contemporary setting in which these issues become especially salient. A fragile ceasefire, continuing mediation, economic pressure, strategic maritime constraints, and uncertainty in state capacity together create a situation in which public stability may be a poor proxy for structural resilience. The present text does not attempt to predict whether regime collapse will occur. It asks instead what sort of mathematical architecture is appropriate when thinking about political systems that may be stable only in a superficial sense.
2. Threshold Models of Collective Activation
A natural point of departure is the threshold model of collective behavior. In its simplest form one may write
$$ a_i(t)= \begin{cases} 1, & \text{if } N_i(t)\ge \theta_i,\\ 0, & \text{if } N_i(t)<\theta_i. \end{cases} $$Here $a_i(t)\in\{0,1\}$ denotes whether individual $i$ is actively participating in protest at time $t$. The variable $N_i(t)$ is the level of participation visible to that individual. It need not represent the total number of protesters in an entire country. More plausibly, it denotes visible participation within the person’s local informational field: workplace, kinship network, neighborhood, online channels, or other socially proximate structures. The threshold $\theta_i$ is then the minimum visible participation required before individual $i$ enters the active state.
The mathematical interest of this model lies in the fact that simple local decision rules can generate sharply nonlinear collective behavior. Even when each individual obeys a straightforward activation rule, the aggregate system may display cascades, discontinuities, and strong sensitivity to the distribution of thresholds across the population.
The passage from local visibility $N_i(t)$ to a global or population-level description is not trivial. Network structure matters. A sparse or clustered network may prevent cascades that would occur in a well-mixed population, whereas a densely connected network may amplify even modest local activations. Later sections will therefore invoke mean-field dynamics only as an explicit approximation.
3. The Internal Structure of the Threshold Parameter
The symbol $\theta_i$ should not be treated as a primitive psychological constant. To do so would leave the model formally suggestive but conceptually underdeveloped. A more serious treatment interprets $\theta_i$ as a compressed representation of several latent variables. One may therefore write
$$ \theta_i = \Theta(c_i,r_i,b_i,m_i,\ell_i), $$where $c_i$ denotes practical cost, $r_i$ expected risk of repression, $b_i$ expected benefit if protest succeeds, $m_i$ moral or ideological disposition, and $\ell_i$ the individual’s estimate of others’ likely participation or continuing loyalty to the regime.
The threshold is thus not simply “inside” the individual. It is a reduced description of how much publicly visible coordination is required for private opposition to become overt action. The term $\ell_i$ remains implicit in the very use of $N_i$, since $N_i$ is not raw global participation but participation as perceived or inferred by the individual.
This may be made more explicit through a utility-difference formulation. Let
$$ \Delta U_i(N_i)=U_i(\text{participate}\mid N_i)-U_i(\text{abstain}\mid N_i). $$Participation becomes more likely as $\Delta U_i$ increases. A first-order ansatz is
$$ \Delta U_i(N_i)=\alpha_i N_i + m_i + b_i - c_i - r_i, $$where $\alpha_i>0$ measures responsiveness to visible participation by others. One then obtains the critical participation level
$$ \theta_i=\frac{c_i+r_i-m_i-b_i}{\alpha_i}. $$This should not be mistaken for a complete empirical law. It is a structural decomposition. It shows what an authoritarian regime must do in order to preserve public quiescence: increase repression risk $r_i$, increase practical cost $c_i$, reduce expected benefit $b_i$, and weaken the moral force $m_i$ of opposition. Opposition movements attempt the reverse: to lower effective thresholds by altering these parameters and by increasing visible participation itself.
4. Preference Falsification and Hidden Order
A threshold model remains incomplete unless one distinguishes between private opposition and public expression. In many authoritarian systems, individuals conceal their actual commitments. Public conformity may therefore coexist with extensive latent dissent.
This can be represented by two variables,
$$ x_i \in \{0,1\}, \qquad y_i \in \{0,1\}, $$where $x_i=1$ denotes private opposition and $y_i=1$ publicly expressed opposition. Preference falsification means that
$$ \Pr(y_i=1 \mid x_i=1) < 1. $$But this probability is itself endogenous. The willingness to express opposition depends on how much opposition is already publicly visible. Let
$$ Y=\sum_j y_j. $$Then one may write
$$ \Pr(y_i=1 \mid x_i=1, Y)=f(Y,\rho_i), $$where $\rho_i$ is an individual revelation threshold. The functional point is simple: people do not merely decide whether to oppose the regime; they also decide whether to reveal that opposition. Thus the same threshold logic may operate at two distinct levels. First, a threshold governs whether private opposition becomes public expression. Second, another threshold governs whether public expression becomes active participation.
The effective collective variable driving visible political dynamics is therefore not the hidden quantity $\sum_i x_i$, but the publicly observable quantity $\sum_i y_i$. A regime may appear stable not because private loyalty remains strong, but because visible opposition remains low. In formal terms, the observable macrostate is a projection of a richer and potentially very different microstate.
5. Stochastic Activation and a Logit Formulation
The deterministic threshold rule is conceptually useful but behaviorally rigid. A more realistic approach treats activation and deactivation as stochastic transitions between states. Let
$$ X_i(t)\in\{0,1\}, $$where $X_i=0$ denotes inactivity and $X_i=1$ active participation.
A natural stochastic extension of the utility-based formulation is obtained by assuming that the effective utility difference includes an idiosyncratic random term. If the latent utility difference is
$$ \Delta U_i(N_i)=\alpha_i N_i + m_i + b_i - c_i - r_i, $$and if choice is perturbed by a logistic-type noise term, then the probability of activation takes the familiar logit form
$$ P_i^{\text{on}}(N_i)=\frac{1}{1+\exp[-\beta\,\Delta U_i(N_i)]}. $$This immediately yields a bounded activation hazard in continuous time,
$$ \lambda_i(N_i)=\lambda_{\max}\,\frac{1}{1+\exp[-\beta(N_i-\theta_i)]}. $$Likewise, a corresponding deactivation hazard may be written as
$$ \mu_i(N_i)=\mu_{\max}\,\frac{1}{1+\exp[\gamma(N_i-\theta_i)]}. $$These forms have two advantages. First, they preserve the threshold intuition while remaining bounded above by $\lambda_{\max}$ and $\mu_{\max}$. Second, they connect the stochastic model directly to the utility structure introduced in the previous section. The probabilistic activation rule is not ad hoc; it is the stochastic counterpart of the deterministic threshold condition.
The resulting system may be treated as a continuous-time Markov jump process. The total activation propensity is
$$ a_{\text{on}}(t)=\sum_{i:X_i=0}\lambda_i(N_i), $$and the total deactivation propensity is
$$ a_{\text{off}}(t)=\sum_{i:X_i=1}\mu_i(N_i). $$Hence the total event rate is
$$ a_0(t)=a_{\text{on}}(t)+a_{\text{off}}(t). $$The waiting time to the next event is then
$$ \tau \sim \mathrm{Exp}(a_0). $$This is formally analogous to the structure familiar from stochastic reaction systems. The analogy is not sociologically exact, but it is mathematically exact at the level of event-driven continuous-time dynamics. One thereby obtains not a single deterministic trajectory but an ensemble of possible paths: long quiet intervals, failed local activations, transient cascades, and occasionally large-scale transitions.
6. Population-Level Dynamics and the Master Equation
To pass from individual transitions to a collective description, let
$$ K(t)=\sum_i X_i(t) $$be the number of active participants in a population of size $M$. If one now adopts a well-mixed approximation, so that individuals respond only to the active fraction
$$ x=\frac{K}{M}, $$then the population-level transitions become
$$ K \to K+1 \quad \text{with rate} \quad (M-K)\lambda(x), $$ $$ K \to K-1 \quad \text{with rate} \quad K\mu(x). $$The probability $P(K,t)$ of observing $K$ active participants then satisfies the master equation
$$ \frac{dP(K,t)}{dt} = (M-(K-1))\lambda(x_{K-1})P(K-1,t) + (K+1)\mu(x_{K+1})P(K+1,t) - \Big[(M-K)\lambda(x_K)+K\mu(x_K)\Big]P(K,t), $$with $x_K=K/M$.
In the large-$M$ limit, one obtains the mean-field equation
$$ \frac{dx}{dt}=(1-x)\lambda(x)-x\mu(x). $$This equation is simple enough for qualitative analysis and rich enough to exhibit multiple fixed points. For example, if one takes
$$ \lambda(x)=\lambda_0 + a x^2, \qquad \mu(x)=\mu_0, $$then
$$ \frac{dx}{dt}=(1-x)(\lambda_0 + a x^2)-\mu_0 x. $$Depending on the parameters $\lambda_0$, $a$, and $\mu_0$, this cubic nonlinearity may admit three real fixed points, of which two are stable and one unstable. If $\lambda_0>0$, the low-activity fixed point is not exactly $x=0$ but a small positive value $x^\ast$, which may be interpreted sociologically as a background level of sporadic opposition rather than total silence. In that case the system is bistable. A low-activity regime and a high-activity regime may both be dynamically stable, separated by an unstable threshold. The low-activity state is then not necessarily deeply stable; it may be only metastable.
This is the formal basis for describing revolutionary episodes as stochastic phase transitions. The term should not be understood in a strict thermodynamic sense, but as an analogy to systems that admit multiple macroscopic regimes and abrupt noise-assisted transitions between them.
7. Metastability and Kramers-Type Escape
The language of metastability should be taken seriously rather than merely rhetorically. In a system with two competing macroscopic regimes, the low-activity state may be represented as a local basin rather than a globally stable equilibrium. Transition to a high-activity regime is then not impossible under the “stable” dynamics; it is merely suppressed.
In the simplest analogy, the escape rate from a metastable basin has Kramers-type form
$$ k \sim \exp\!\left(-\frac{\Delta V}{D}\right), $$where $\Delta V$ is an effective barrier height and $D$ the noise strength. The political interpretation is straightforward. As long as repression, uncertainty, and coordination failure keep the barrier high, large-scale protest remains unlikely. But if economic stress rises, legitimacy weakens, or visible dissent lowers effective thresholds, then the barrier $\Delta V$ may shrink. Under such conditions, events that would previously have died out may instead trigger macroscopic reorganization.
This gives mathematical substance to the claim that a regime may look stable while becoming progressively easier to dislodge. What changes first is not necessarily the visible macrostate, but the barrier structure governing escape from that macrostate.
8. Coupled Protest and Security Dynamics
A regime rarely collapses merely because dissatisfaction is widespread. Collapse more often occurs when coercive institutions lose coherence, loyalty, or operational effectiveness. It is therefore mathematically insufficient to model protest behavior alone.
The next step is best understood explicitly as a mean-field approximation of a coupled two-population system. Let $x(t)$ denote the fraction of active protesters and $s(t)$ the fraction of security actors remaining loyal to the regime. One may then write, schematically,
$$ \frac{dx}{dt}=(1-x)\lambda(x,s)-x\mu(x,s), $$ $$ \frac{ds}{dt}= -\,\phi(x,E)\,s + \psi(R)\,(1-s). $$Here $E$ denotes economic stress and $R$ stabilizing resources such as patronage, repression capacity, ideological cohesion, or external support. The function $\phi(x,E)$ is a defection rate that increases with protest scale and stress, whereas $\psi(R)$ captures mechanisms of re-stabilization.
The point of this coupled system is not precise prediction. It is to show that the stability structure of the protest dynamics may itself depend on the loyalty variable $s$. If $s$ falls below a critical level, the low-activity equilibrium for $x$ may weaken or disappear altogether. Protest and coercion are therefore not separate domains but coupled dynamical fields.
This level shift from stochastic microdynamics to deterministic mean-field equations should be read as an approximation, not as a change of subject. One could in principle model security actors as a second population of stochastic Markov-type agents. The present form simply compresses that richer structure into a tractable macroscopic picture.
9. Public Order as a Metastable Projection
The mathematical picture that emerges is not one of simple accumulation but of structured instability. A regime may occupy a metastable state. Small protests fail. Local activations decay. Public conformity reproduces itself. Yet the deeper system may already be drifting through parameter space: repression becomes costlier, legitimacy weaker, visibility of dissent greater, economic stress more acute, and loyalty within coercive institutions less secure.
When this occurs, public order may still appear intact while its basin of stability has narrowed considerably. The event that triggers macroscopic reorganization may then be relatively small in itself: a strike, an arrest, a payment crisis, a military failure, or a symbolic act of defiance. What matters is not the triggering event in isolation, but the system’s altered susceptibility to amplification.
This helps explain why revolutionary episodes are so often experienced as surprising. The surprise does not primarily concern the existence of discontent. It concerns the discrepancy between observable stillness and latent instability. Public order, in this sense, is not identical to the underlying social microstate. It is a visible projection sustained by hidden dispositions, selective revelation, local coordination, and mutual uncertainty.
10. Iran in Spring 2026 as a Case
Iran in spring 2026 should be approached in exactly this framework. A fragile ceasefire, continuing mediation, economic stress, strategic pressure around maritime chokepoints, and uncertainty in state capacity do not imply imminent regime collapse. No mathematically serious model would justify such confidence.
What these conditions do imply is that relevant threshold and hazard parameters may be unusually labile. Repression costs may rise. Visible dissent may become more consequential. Confidence in state durability may weaken. Loyalty within the security apparatus may become more sensitive to economic and strategic strain. Under such conditions, public order may remain observable while structural resilience deteriorates.
Iran therefore functions here not as a verdict but as a case: a historically concrete instance in which the distinction between public stability and structural stability becomes analytically salient.
11. Conclusion
Mathematical models of revolution do not render history computable in any strong sense. They do, however, clarify a class of mechanisms by which public stability may persist despite latent opposition, and by which abrupt macroscopic transitions may occur without long visible preparation.
The central lesson is methodological. Public political order is not identical to the sum of private mental states. It is a relationally sustained structure of visible expectations, hidden dispositions, local coordination, and mutually conditioned restraint. A revolution, in this perspective, is not merely a change of rulers. It is a reorganization of the publicly accessible social world.
For that reason, the mathematical study of revolutionary dynamics remains conceptually valuable even where prediction remains weak. It helps distinguish deep stability from metastability, visible conformity from latent dissent, and gradual parameter drift from sudden collective transition.
References
Granovetter, M. (1978). Threshold models of collective behavior.
Kuran, T. (1991). Now out of never: The element of surprise in the East European revolution of 1989.
Gillespie, D. T. (1977). Exact stochastic simulation of coupled chemical reactions.
Klein, G. R., & Regan, P. M. (2018). Dynamics of political protests.
Kommentarer