betaFeedback

WorksPhysics and Astronomy

A Generalized Gompertz Law for the Rise and Fall of Empires

Abstract

Human societies tend to evolve toward increasing scale and complexity, often followed by stagnation and decline. Historical empires provide a prominent example of this trajectory; however, existing quantitative models capture expansion and rarely account for decline within a unified framework. Here, we introduce a minimal dynamical model that describes both growth and collapse in a single formalism. The model yields a closed-form trajectory that generalizes the Gompertz law and admits equivalent maximum-entropy and optimal-control formulations. We show that this functional form captures the rise and fall of fourteen diverse empires across widely varying time scales. Despite substantial historical heterogeneity, all cases follow a common asymmetric rise-peak-decline pattern governed by a small number of interpretable parameters. We identify the mechanism driving this behavior as the chronophage: an exponentially accumulating burden of administrative and coordination costs that offsets the gains of expansion. These findings uncover a general quantitative principle, yielding new insights into how increasing complexity constrains large-scale social systems.

Introduction

Complex systems across the natural and social sciences often exhibit characteristic trajectories of growth, saturation, and eventual decline. Such dynamics are observed in domains as diverse as organismal ageing, tumour progression, and population growth, where evolution of different kinds proceeds through asymmetric rise–peak–decline patterns rather than indefinite expansion; a broad class of these processes is well described by Gompertz-type dynamics [1, 2, 3, 4, 5].

Historical empires provide a particularly rich and well-documented class of complex systems in which these dynamics unfold at large spatial and temporal scales. A longstanding question across history and the social sciences is: Why such systems, after sustained periods of growth, ultimately decline and collapse? Existing explanations have emphasized a wide range of mechanisms, including environmental stress, external shocks, and internal conflict [6, 7, 8]. A prominent theoretical perspective, articulated by Joseph Tainter, attributes collapse to diminishing marginal returns on increasing societal complexity, whereby the costs of maintaining large-scale organization eventually outweigh its benefits [9, 10]. While these perspectives provide important insights, they have not been integrated into a concise quantitative framework capable of jointly describing growth and decline in a low-dimensional setting.

Previous quantitative approaches have successfully captured the expansion phase of empires. In particular, models of logistic growth reproduce the territorial evolution of diverse historical cases using a small number of parameters, suggesting the presence of coarse-grained regularities despite substantial historical heterogeneity [11]. However, such models are intrinsically limited to monotonic growth and do not account for the subsequent loss of territory and eventual collapse. As a result, the full life cycle of large-scale social systems remains without a unified, minimal dynamical description.

Here, we introduce a general dynamical framework that captures both growth and decline within a single, parameter-sparse formalism. The model yields a closed-form trajectory describing rise, peak, and decline, which generalizes the Gompertz law and can be derived independently from maximum-entropy and optimal-control principles. Within this framework, decline emerges endogenously from a single mechanism that arrests and reverses expansion over time.

We identify this mechanism as the chronophage1: an exponentially accumulating administrative and coordination burden that offsets expansion and drives decline. As systems grow, this burden progressively offsets the gains of expansion, ultimately leading to stagnation and decline. In this sense, the chronophage provides a quantitative realization of the rising costs of complexity, embedding earlier qualitative theories within a predictive dynamical model.

We test this framework on a diverse set of historical empires spanning maritime, land-based, and wartime cases across widely varying time scales. Despite differences in geography, political organization, and historical context, all cases follow a common asymmetric rise–peak–decline trajectory governed by a small number of interpretable parameters. This empirical regularity suggests that the life cycles of large-scale social systems may be governed by universal dynamical constraints. More broadly, the proposed framework identifies the dynamics of empires within a wider class of Gompertz-type processes observed in biological and physical systems of varying size and complexity (Figure 1), including mortality dynamics, tumour growth, microbial populations, and epidemic spread.

Illustration of Gompertz-type growth and decline dynamic of living systems at multiple scales

Figure 1. Illustration of Gompertz-type growth and decline dynamic of living systems at multiple scales. Gompertz-like functions govern a variety of biological and social mechanisms at different length and time scales, ranging from multicellular life to epidemic outbreaks. The life cycle of various empires also follows a similar pattern, as shown by the fit of (2) to historical data.

The remainder of the paper proceeds as follows. We first introduce the dynamical framework, specify the proposed rise–peak–decline model, and describe the preprocessing and fitting procedure used for the historical territorial series. We then apply the model to maritime and land-based empires, showing that their territorial histories follow the predicted trajectory. Next, we examine wartime empires, whose expansion and contraction are captured by the same framework despite unfolding on much shorter time scales. Finally, we interpret the fitted parameters in historical terms and discuss the broader implications of the chronophage mechanism for understanding the rise and fall of large territorial systems. Two complementary theoretical derivations are provided in the appendices: Appendix A derives the rise–peak–decline envelope from a maximum-entropy construction, while Appendix B shows how the same form emerges from an equivalent optimal-control formulation and its reduced dynamics.

Methods and model

In this work, we adopt a dynamical-systems approach to model the life cycle of fourteen historical empires: four maritime empires, the British Empire [12], the French Empire [13], the Spanish Empire [14], and the Portuguese Empire [15]; six land empires, the Qing Empire and Republic of China [16], the Mongol Empire [17], the Islamic Caliphate [18], the Seljuk Empire [19], the Mughal Empire [20], and the Delhi Sultanate [21]. These ten long-run, Eurasian historical cases were modelled in their expansion phase using a logistic framework by Marchetti and Ausubel [11]; here, however, we broaden the comparison to include four wartime empires: Nazi Germany [22], Imperial Japan [23], Napoleonic France [24], and the Italian Empire [25].

The purpose of the model is to provide a low-dimensional description of the full territorial life cycle of these systems, including both expansion and decline. Although historical empires differ substantially in institutional structure, geography, duration, and external pressures, their territorial trajectories often display a common asymmetric rise–peak–decline pattern [26, 27, 28, 29]. The framework developed here treats these trajectories as instances of a generalized Gompertz-like process, in which growth slows exponentially over time and is ultimately offset by increasing constraints. In the context of empires, these constraints are interpreted as the accumulating administrative, fiscal, logistical, and coordination burdens associated with maintaining territorial scale.

Proposed model

The logistic growth equation constitutes a canonical model of bounded, density-dependent expansion and has been extensively employed across the social sciences and quantitative history to describe growth under finite constraints, see Figure 2. In cliodynamic and structural-demographic frameworks [30, 31], logistic-type dynamics often provide the baseline representation of population and aggregate expansion. [11] showed that the territorial expansion of many historical empires is well approximated by a logistic trajectory,

Figure 2. A common rise–peak–decline trajectory for empires. Faint curves show normalized empire trajectories with a shared rise–peak–decline pattern. The orange curve is the proposed model; the dashed brown curve is logistic growth without decline. The blue region shows post-peak divergence caused by the chronophage, Ω(t)\Omega(t): the growing administrative burden that offsets expansion and drives contraction. The green marker marks peak territory.

x˙=rx(1−x),\dot{x} = r x(1-x),

where x(t)x(t) denotes normalized extent of the area as a function of time tt, and rr is an intrinsic expansion rate.

While this parsimonious form captures the rise and saturation (peak) of imperial growth, it does not capture the subsequent territorial decline or collapse. This limitation is structural: a one-dimensional autonomous ordinary differential equation (ODE) admits only fixed points as attractors; trajectories therefore equilibrate monotonically and cannot generate a single-peaked rise–fall cycle without additional structure. Hence, producing an asymmetric peak followed by contraction would require the simplest possible extension beyond a 1D autonomous model: either (i) a non-autonomous (explicitly time-dependent) modification or (ii) an equivalent two-dimensional autonomous system with an additional state variable. We adopt a particularly minimal extension of the logistic framework by introducing a single time-dependent term governed by a new parameter, α\alpha, yielding

x˙=rx(1−eαt).(1)\dot{x} = rx\left(1-e^{\alpha t}\right). \tag*{(1)}

This preserves the logistic-like early expansion but endogenizes a subsequent reversal as the effective growth factor turns negative, allowing the same closed-form law to capture the full historical rise–peak–decline trajectory. Equivalently, (1) describes a process in which the instantaneous expansion potential is progressively eroded by a constraint term that grows exponentially through time. We refer to this exponentially accumulating constraint as the chronophage: the time-dependent burden of administration, coordination, maintenance, and mobilization that increasingly offsets the gains of expansion.

The model we propose in (1) admits a closed-form solution that takes the form

x(t)=xmax⁡exp⁡{s+sα(t−tmax⁡)−seα(t−tmax⁡)},(2)x(t) = x_{\max}\exp\left\{s+s\alpha(t-t_{\max})-se^{\alpha(t-t_{\max})}\right\}, \tag*{(2)}

where xmax⁡x_{\max} denotes the peak territorial extent, tmax⁡t_{\max} the time at which this peak occurs, α\alpha the congestion-growth parameter, and s=r/αs=r/\alpha a dimensionless shape parameter controlling the relative steepness of ascent and decline.

The normalization of territorial series by their historical maxima prior to estimation implies that xmax⁡x_{\max} should lie close to unity across cases, as confirmed by the fitted values reported in Table 1. The parameters (r,α)(r,\alpha) determine the intrinsic expansion rate and the exponential accumulation of mismatch, respectively, while ss summarizes their ratio and thus the overall asymmetry of the trajectory.

PolitySpantmax⁡t_{\max}xmax⁡x_{\max}ssα\alphar=sαr = s\alphaRMSE
British Empire1600–198019231.021.87×10−11.87 \times10^{-1}7.05×10−27.05 \times10^{-2}1.31×10−21.31 \times10^{-2}3.69×10−23.69 \times10^{-2}
French Empire1815–198019331.135.92×10−15.92 \times10^{-1}5.24×10−25.24 \times10^{-2}3.10×10−23.10 \times10^{-2}1.09×10−11.09 \times10^{-1}
Spanish Empire1482–197517681.062.91×10−12.91 \times10^{-1}3.04×10−23.04 \times10^{-2}8.83×10−38.83 \times10^{-3}5.60×10−25.60 \times10^{-2}
Portuguese Empire1200–197518230.921.63×1001.63 \times10^{0}1.10×10−21.10 \times10^{-2}1.80×10−21.80 \times10^{-2}3.37×10−23.37 \times10^{-2}
Qing Dynasty–ROC1600–197918241.086.36×10−16.36 \times10^{-1}1.55×10−21.55 \times10^{-2}9.84×10−39.84 \times10^{-3}1.02×10−11.02 \times10^{-1}
Mongol-Yüan Empire1200–140012921.032.44×1012.44 \times10^{1}4.25×10−34.25 \times10^{-3}1.04×10−11.04 \times10^{-1}6.59×10−26.59 \times10^{-2}
Islamic Caliphate622–9457471.026.92×1026.92 \times10^{2}5.31×10−45.31 \times10^{-4}3.67×10−13.67 \times10^{-1}6.55×10−26.55 \times10^{-2}
Seljuk Empire970–119410921.106.48×1026.48 \times10^{2}9.98×10−49.98 \times10^{-4}6.47×10−16.47 \times10^{-1}1.09×10−11.09 \times10^{-1}
Mughal Empire1519–179816551.081.37×1011.37 \times10^{1}4.90×10−34.90 \times10^{-3}6.69×10−26.69 \times10^{-2}7.01×10−27.01 \times10^{-2}
Delhi Sultanate1040–139813150.871.28×1001.28 \times10^{0}1.58×10−21.58 \times10^{-2}2.03×10−22.03 \times10^{-2}7.21×10−27.21 \times10^{-2}
Nazi Germany1933–194519421.035.81×1025.81 \times10^{2}2.73×10−22.73 \times10^{-2}1.59×1011.59 \times10^{1}6.97×10−26.97 \times10^{-2}
Imperial Japan1895–194519441.021.68×10−21.68 \times10^{-2}8.00×1008.00 \times10^{0}1.34×10−11.34 \times10^{-1}6.00×10−26.00 \times10^{-2}
Napoleonic France1795–181518110.981.23×10−11.23 \times10^{-1}7.01×10−17.01 \times10^{-1}8.62×10−28.62 \times10^{-2}5.32×10−25.32 \times10^{-2}
Italian Empire1882–194719351.106.76×10−26.76 \times10^{-2}4.27×10−14.27 \times10^{-1}2.89×10−22.89 \times10^{-2}8.21×10−28.21 \times10^{-2}

Table 1. Fitted parameters for each polity under the proposed model. Span gives the data window used for fitting. tmax⁡t_{\max} and xmax⁡x_{\max} are in standard decimal; all other quantities use three significant digits in scientific notation. RMSE is the Root Mean Squared Error of the fit.

Eq. (2) is therefore a generalized Gompertz-like rise–peak–decline envelope. It differs from the standard logistic trajectory by allowing the effective growth rate to become negative after the accumulation of constraints exceeds the expansionary drive. In this sense, the model converts the qualitative idea of diminishing returns to increasing complexity into a simple quantitative mechanism: expansion initially dominates, but the chronophage grows over time, eventually suppressing growth and generating territorial contraction.

Data preprocessing

To facilitate the numerical fit procedure, we regularize each time series within its observed span using a shape-preserving cubic Hermite interpolator (PCHIP). This generates 20 evenly spaced points that pass exactly through the original observations, preserve local monotonicity, avoid spline overshoot, and never extrapolate beyond the data range. The interpolated points thus provide a smooth but faithful proxy trajectory on which the model can be estimated. For transparency, all figures show both the raw data (red) and the interpolated points used for fitting.

The recorded areas [28] show major discontinuities at specific times, e.g., the Portuguese Empire losing 90% of its area between 1820 and 1822 and the Mongolian Empire dropping to 11 million km2\mathrm{km}^{2} in 1310 from 24 million km2\mathrm{km}^{2} the year before (1309). The Delhi Sultanate series likewise contains an isolated point on the ascending branch that produces an artificial plateau immediately before the maximum. Such implausibly abrupt patterns likely reflect discrete reporting conventions rather than genuine one-step territorial shifts, and their removal yields smoother representations of both imperial expansion and retreat. For France, we add an observation for 1980 equal to metropolitan France plus its overseas departments and territories. This ensures consistency with Taagepera’s convention and makes the late-twentieth-century value directly comparable to the 1962 entry. For the Qing–Republic of China series, we follow Taagepera’s data up to 1920. As a continuation, we add two points: 1949 and 1979, using the area administered by the Republic of China on Taiwan after its relocation. For the wartime cases shown in Figure 5, the territorial series are compiled from the corresponding historical sources [32, 33, 34, 35].

Figure 5. Fit of the proposed model to wartime territorial expansions of empires. Red dots denote data on occupied territories for each wartime case. In (a) Nazi Germany, the blue dots represent the chronophage proxy data, scaled appropriately for visual clarity.

Across all historical cases, the same procedure is followed: raw observations are shown alongside their PCHIP-regularized counterparts, and the proposed model is fitted within the observed span. The resulting curves should be interpreted as stylized trajectories that capture the asymmetric rise–and–fall predicted by the model, while remaining transparent about the sparsity and conventions of the underlying historical data.

For each polity, the territorial area series is normalized by its empirical maximum before fitting. The model in (2) is then fitted to the PCHIP-regularized trajectory within the observed temporal span. The primary free parameters controlling the trajectory are tmax⁡t_{\max}, ss, and α\alpha, with the intrinsic expansion rate recovered as r=sαr = s\alpha. We also estimate xmax⁡x_{\max} as a scale parameter rather than fixing it to one, using it as a sanity check on the normalization; because the territorial series are normalized by their empirical maxima, fitted values of xmax⁡x_{\max} are expected to remain close to unity. Thus, the effective shape of the trajectory is governed by three parameters, while xmax⁡x_{\max} captures residual scale mismatch.

Goodness of fit is summarized using the root mean squared error (RMSE) between the fitted curve and the interpolated normalized series. The same fitting procedure is applied to all maritime, land-based, and wartime cases, allowing differences across empires to be expressed through parameter values rather than through changes in functional form.

Results

The territorial trajectories of the major maritime empires (British, French, Spanish and Portuguese) are well described by (2) under a three-parameter shape fit, with xmax⁡x_{\max} retained as a scale check. As shown in Figure 3, the fitted curve reproduces the characteristic empirical pattern: a sustained, approximately logistic rise, a single interior maximum at tmax⁡t_{\max}, and a prolonged, asymmetric decline. Within (2), early expansion is governed primarily by the linear growth component, whereas the nested exponential term progressively suppresses net growth, generating an endogenous turning point and a right-skewed retreat. Despite substantial heterogeneity in temporal features and scale, cross-case differences are captured parsimoniously through variation in (tmax⁡,s,α)(t_{\max},s,\alpha), while the functional form remains invariant. This suggests that imperial expansion conforms closely to a common rise–peak–decline envelope.

Figure 3. The fit of the proposed model to the territorial expansion of maritime empires. Red dots denote original data [26, 27, 28]; black curves show fits of the proposed model estimated on shape-preserving interpolated series. In (a) British Empire, the blue dots represent the chronophage proxy data, scaled appropriately for visual clarity.

A similar result holds for major land-based empires (Mongol, Qing–Republic of China (ROC), Caliphate, Seljuk, Mughal and Delhi). As shown in Figure 4, (2) provides an equally close fit to their territorial histories, capturing both rapid pre-peak expansion and extended post-peak contraction. Differences across cases are reflected primarily through parameter values (see Table 1), most notably the timing of the maximum and the relative strength of the congestion term, rather than through changes in functional structure. In all cases, the same mechanism encoded in (2) produces a single peak followed by asymmetric decline, indicating that maritime and land empires share a common quantitative growth profile despite distinct geopolitical contexts.

Figure 4. The fit of the proposed model to the territorial expansion of land-based empires. In (a) Qing Dynasty-ROC, the blue dots represent the chronophage proxy data, scaled appropriately for visual clarity.

We have seen so far that decline, rather than stability, is a common phenomenon that follows growth. To elucidate the mechanism underlying this process of decline, the model in (1) can be rewritten equivalently as an autonomous system of two coupled ordinary differential equations:

x˙(t)=x(t)[sα−Ω(t)],Ω˙(t)=αΩ(t).(3)\dot{x}(t) = x(t)\left[s\alpha- \Omega(t)\right], \qquad\dot{\Omega}(t) = \alpha\Omega(t). \tag*{(3)}

In this form the model resembles a predator–prey system in which territorial extent x(t)x(t) is suppressed by a “predator” Ω(t)\Omega(t) that grows autonomously and, unlike in the case of classical Lotka–Volterra dynamics, does not decay. This one-way coupling produces a single rise–peak–decline trajectory rather than oscillations, consistent with the historical patterns described above. The variable Ω(t)∼eαt\Omega(t) \sim e^{\alpha t} represents the chronophage, an exponentially growing stock of administrative congestion or coordination burden and substitutes the single-time dependent term from the prior formulation in (1).

The chronophage mechanism encoded in the parameter α\alpha can be reflected in concrete historical trends, relating to the multiplicative growth of coordination, monitoring and maintenance burdens as territorial systems grow. Two distinct historical proxies, administrative personnel in the British Empire (see fig:3a) and fiscal expansion in the Qing dynasty (see fig:4a) illustrate how such congestion can accumulate at rates commensurate with the estimated parameter values in (2).

In the late British Empire, the size of the Colonial Service expanded rapidly, rising from roughly 1,000 overseas officers around 1900 to approximately 8,000 by 1938, about 11,000 by 1947, and nearly 18,000 posts by 1954 [36]. Approximated as exponential growth, this corresponds to an annual increase on the order of 5–7%. This magnitude is comparable to the fitted congestion parameter for Britain (α≈7.05×10−2 yr−1\alpha\approx7.05 \times10^{-2}\ \mathrm{yr}^{-1}). Within the chronophage framework, the rapid expansion of overseas administrative personnel reflects escalating demands for governance, information processing and supervision across a geographically dispersed empire. As these demands compound, they absorb an increasing share of resources and progressively offset the gains from further territorial acquisition, contributing to the asymmetric post-peak decline generated by (2).

A parallel dynamic appears in the Qing dynasty, where the chronophage mechanism is reflected in rising fiscal strain. Reconstructed Qing fiscal series show that during the

nineteenth century state expenditures grew persistently relative to revenues, pushing the fiscal balance from surplus toward sustained deficit [37]. This deterioration is primarily attributed to the mounting costs of large-scale internal rebellions and associated military mobilization, which imposed prolonged upward pressure on state outlays. Expressed in growth terms, total expenditures increased at approximately 2% per year over extended late-imperial intervals, a rate consistent with the fitted chronophage parameter reported in Table 1. Within the framework of (2), this correspondence supports the interpretation of α\alpha as capturing the accumulation of systemic maintenance burdens required to preserve territorial integrity and internal order.

The functional form of (2) also describes a class of short-lived, highly expansionary polities whose territorial growth was closely tied to major military conflicts. Figure 5 shows four such cases with data available from associated sources: Nazi Germany [32], Imperial Japan [33], Napoleonic France [34] and the Italian Empire [35]. In contrast to the maritime and land empires discussed above, whose expansion unfolded over centuries, these polities expanded rapidly during periods of intense warfare and collapsed within a few years. Despite this difference in time scales, their territorial trajectories exhibit a structural pattern akin to the cases discussed above: a steep ascent, a peak located near the end of the expansion phase, and a rapid but asymmetric contraction. (2) captures these dynamics using the same minimal parameterization as in the historical cases. The main difference lies in the parameter values (see Table 1): wartime empires often exhibit substantially larger expansion rates rr, reflecting accelerated conquest dynamics, while the chronophage parameter α\alpha remains positive and produces the same endogenous turning point. The persistence of the same rise–peak–decline envelope across these very different historical contexts suggests that the mechanism encoded in the model is not limited to long administrative processes but can also govern rapid geopolitical expansions driven by military mobilization.

For Nazi Germany, the chronophage can be interpreted through the dynamics of the wartime economy. As a proxy for the scale of economic activity, Figure 5 plots the square root of German GDP during the late 1930s and early 1940s [38]. The use of the square root follows a simple mass-action intuition: if aggregate output arises from interactions within the productive system, total GDP scales approximately with the square of the economy’s effective degree of interaction. Taking the square root of GDP therefore provides an estimate of this underlying interaction. When expressed in this way, the resulting series grows at roughly three percent per year over the relevant interval, closely matching the chronophage growth rate reported for Nazi Germany in Table 1. Within this interpretation, the chronophage represents the increasing share of economic capacity absorbed by mobilisation, occupation administration and wartime logistics. Unlike the long-run imperial cases discussed above, where the chronophage reflects gradually accumulating coordination burdens, here it acts as an economic bottleneck to political and military ambitions, constraining further territorial expansion and contributing to the abrupt decline that follows the wartime peak.

Discussion

For the case of historical maritime and land empires (red points in Figure 3 and Figure 4), we use data from Taagepera’s compilations [28, 26]. A common feature of these datasets is the convention of recording the maximum territorial extent immediately prior to a major contraction. While consistent across empires and convenient for encyclopedic reference, this practice tends to accentuate abrupt breaks and understate the continuity of more gradual adjustments in influence, administration, and economic reach. As a result, declines may appear as sudden collapses rather than the drawn-out retreats they often were. The discrepancy between these step-like observations and the smoother dynamics implied by the model is visible, for example, in the Seljuk decline. To mitigate this effect, we smooth a small number of clear discontinuities, most notably in the Portuguese and Mongolian series, so that the fitted trajectories capture longer-run dynamics rather than isolated reporting steps; details are provided in the Methods section. In doing so, this can additionally introduce artificial plateaus or transient humps in some cases, such as the British, Qing, and Napoleonic empires, which should be interpreted as artifacts of the underlying data construction rather than substantive features of the dynamics.

Following the pre-processing, the functional form of (2) is fitted to the data. All resultant parameters, stated in Table 1, have direct interpretations in terms of observable quantities. The parameter tmax⁡t_{\max} identifies the timing of the maximum area and aligns closely with the historically recognized apex of each polity. The intrinsic rate rr governs the pace of territorial expansion during the early phase of growth, while α\alpha measures the rate at which the limiting mechanism represented by the chronophage accumulates over time. The dimensionless parameter s=r/αs = r/\alpha therefore captures the balance between expansionary momentum and the speed at which systemic constraints emerge. Together, these quantities map the empirical trajectories of territorial extent onto interpretable measures of timing, growth and decline.

Across the maritime and long-run land empires shown in Figure 3 and Figure 4, the fitted parameters reveal several distinct regimes. The maritime empires, together with the Qing and Delhi cases, exhibit intrinsic expansion rates of a few percent per unit time despite large differences in era and geography. By contrast, conquest-dominated land empires such as the Mongol, Caliphate, Seljuk, and Mughal cases occupy a markedly faster-growth regime, with substantially larger rr than the maritime cases. These differences in rr coexist with systematic variation in α\alpha: maritime empires typically exhibit relatively larger α\alpha, indicating that the limiting mechanism accumulates more rapidly in overseas systems than in most contiguous land empires, although Qing and Delhi lie closer to the maritime end of this spectrum. Because s=r/αs = r/\alpha, similar qualitative rise–fall shapes can therefore arise from very different decompositions into expansion intensity (rr) and constraint accumulation (α\alpha). In this sense, high rr is not unique to modern warfare, but also characterizes several historical episodes of rapid territorial consolidation across the land empires considered here.

However, wartime polities in Figure 5 do not form a unified parameter regime. Their fitted rr values are generally higher than those of the maritime colonial cases, but overlap with the above mentioned conquest-dominated land empires. Moreover, α\alpha varies widely across wartime cases: some exhibit very rapid constraint growth (large α\alpha), while others do not, precluding the classification of “wartime” empires as a uniform high-α\alpha category. Instead, the consistency is structural and temporal: these trajectories are compressed into short spans, with peaks tmax⁡t_{\max} occurring close to the end of the expansion window and contraction following quickly thereafter. In this setting the chronophage should be interpreted less as slowly accumulating coordination burdens and more as rapidly binding constraints associated with mobilisation, occupation and multi-front war, which can tighten on the same short horizon as territorial expansion. Although there are marked differences between the cases, the same functional form captures both long-run imperial expansion and short wartime territorial surges, with variation across cases expressed primarily through parameter values rather than through changes in the underlying dynamical mechanism.

At a fundamental level, the functional form can be motivated from an information-theoretic perspective. As shown in Appendix A, among unimodal asymmetric envelopes consistent with a fixed mean and an exponential moment reflecting the accumulation of multiplicative hazards or coordination burdens, the Gompertz-type curve arises as the maximum-entropy solution [39, 40]. In this interpretation the trajectory represents the minimum-bias curve compatible with the assumed structural constraints. The resulting envelope therefore introduces no additional structure beyond what is required by the constraints themselves, providing a neutral baseline against which the empirical rise and fall of large territorial systems can be analysed and interpreted.

This maximum-entropy envelope also possesses several practical advantages for empirical analysis. Imperial trajectories are typically asymmetric, with a relatively gradual phase of expansion followed by a more rapid and skewed decline. The envelope implied by (2) reproduces this pattern naturally, generating an approximately exponential rise followed by a faster contraction without requiring additional ad hoc parameters. At the same time, the formulation remains analytically simple and parsimonious: only a small number of parameters determine the scale, timing and asymmetry of the trajectory, and each maps directly onto interpretable quantities such as the expansion rate, the timing of the peak and the rate at which systemic burdens accumulate. Alternative unimodal curves, including Richards, skew–normal, generalized Gamma or Beta forms, can also reproduce asymmetric trajectories but typically either require additional shape parameters or lack clear interpretability, making cross-case comparison more difficult.

The same trajectory emerges naturally from simple dynamical formulations. In the reduced chronophage representation introduced above, territorial extent evolves under the influence of a burden that grows exponentially and gradually suppresses expansion, yielding a single rise–peak–decline cycle. From an economic perspective, the same envelope arises in an optimal-control framework in which the empire balances the gains from territorial expansion against the costs of sustaining support capacity. As detailed in Appendix B, territorial extent and support capacity evolve as coupled state variables, while imperial policy adjusts endogenously under expansion costs. Under this specification, the optimal policy generates a trajectory whose early transient reproduces the dynamics of (1). Expansion initially dominates but is later constrained as coordination costs accumulate and territorial reach outpaces support capacity.

To clarify the historical context in which the framework developed so far can reasonably apply, we identify a set of premises under which a single rise–peak–decline cycle can be meaningfully represented by a unified dynamical law. Over the fitted span we therefore require:

  • P1. A single coherent rule set. The polity operates under a broadly consistent strategic and institutional framework without major regime resets that would change the governing parameters within the fitted window.

  • P2. Aggressive and sustained expansion. Territorial growth is deliberate and persistent rather than opportunistic or episodic.

  • P3. Local dominance during ascent. The historical evolution is primarily determined by internal capacities rather than sustained subordination to stronger external rivals in the principal theatre.

  • P4. Early cheap-growth orientation. Initial territorial gains are pursued as if they can be incorporated with relatively limited increases in administrative, fiscal, or logistical capacity.

These premises together define a regime in which expansion initially proceeds rapidly but gradually generates internal coordination burdens that eventually reverse the trajectory.

The historical maritime and land empires examined here satisfy these conditions to a sufficient degree such that their rise and decline can be described by an autonomous dynamical system driven primarily by internal mechanisms. The British, French, Spanish, and Portuguese empires each pursued sustained overseas expansion under relatively stable administrative settings and enjoyed extended periods of naval or geopolitical advantage during their ascent. Similarly, the Mongol, Caliphate, Seljuk, Qing–ROC, and Mughal trajectories reflect coherent imperial projects characterized by purposeful territorial enlargement and early phases in which expansion occurred with comparatively low resistance. Under these conditions, the rise and subsequent decline can be interpreted as emerging endogenously from the interaction between expansionary drive and the gradual accumulation of coordination burdens, producing a single asymmetric expansion–contraction cycle. The Delhi Sultanate is more borderline: its trajectory can be treated within the same framework only if dynastic succession is understood not as a sequence of regime resets, but as preserving a sufficiently continuous outward imperial orientation over the fitted span, with later fragmentation interpreted as the decline of that same project; Timur’s invasion then acts less as a separate cycle than as an exogenous accelerator of an already weakening trajectory.

The wartime cases operate under the same structural premises P1-P4, albeit in a markedly different geopolitical environment. The expansionary ambitions of regimes such as Nazi Germany, Imperial Japan, Napoleonic France, and the Italian Empire rapidly provoked coordinated responses from rival powers. As a result, the dynamics cannot be interpreted as purely autonomous: external resistance becomes progressively stronger as expansion proceeds. Nevertheless, these reactions feed back into the same internal mechanism represented in (1). Military commitments, administrative burdens, and economic mobilization increase the demands of maintaining territorial control, amplifying the internal constraints that limit expansion. The trajectory therefore reflects a reinforcing spiral in which expansion catalyzes international opposition, which in turn raises the internal costs of sustaining the enlarged system and accelerates the transition from rapid ascent to abrupt decline.

However, not all historical empires necessarily satisfy all the premises P1-P4 concomitantly. Table 2 illustrates how these premises help delimit the historical situations in which the proposed model provides a meaningful description. The empires listed share certain superficial similarities with the fitted cases, including large territorial scale, long duration, or participation in the same maritime or Eurasian geopolitical systems, yet their trajectories do not correspond to the single internally driven expansion–contraction cycle assumed by the model. In several cases, such as the Dutch and Danish empires [41], the difficulty arises because expansion was organised primarily through decentralized commercial networks rather than a unified territorial project.

PolityFailed premisesReason
Dutch EmpireP1, P2, P3, P4A dispersed commercial empire organised around chartered companies and trade networks rather than a unified territorial project. Expansion was episodic and profit-driven, local dominance was contested by rival maritime powers, and growth depended on commercial concessions rather than territorial integration.
Danish EmpireP1, P2, P3Colonial possessions were geographically scattered and strategically secondary to the metropolitan state. Expansion was limited and opportunistic rather than sustained, and Denmark rarely exercised clear regional dominance in the imperial theatre.
Roman EmpireP4Expansion relied on a highly integrated fiscal–military and administrative system that scaled with territorial reach. Frontier consolidation through taxation, infrastructure, and permanent forces implies expansion was not treated as low-friction growth, and the long plateau preceding decline reflects strong internal feedback mechanisms.
Holy Roman EmpireP1, P2, P3A polycentric and elective composite polity rather than a unified strategic project, lacking sustained aggressive expansion and centralised territorial dominance. Its structure prevents treatment as a single coherent expansionary trajectory.
Byzantine EmpireP1, P2, P3Repeated institutional and military reorganisations reset governing parameters, while the empire alternated between contraction and partial recovery under persistent pressure from peer powers. The resulting trajectory contains multiple cycles rather than a single rise–peak–decline pattern.
Ottoman EmpireP1, P3The empire underwent major administrative and institutional transformations across its long history, and expansion occurred in sustained competition with powerful regional rivals. These factors prevent treating the trajectory as a single internally driven cycle.
Safavid PersiaP1, P3, P4Repeated military and administrative reorganisations, persistent peer pressure, and high campaign costs prevent treatment as a single cheap–growth expansion.
Ming EmpireP2, P3Imperial strategy emphasised consolidation and internal governance rather than sustained outward territorial expansion. Regional rivals and frontier dynamics constrained dominance, producing a trajectory marked by plateaus and not just expansion or decline.

Table 2. Empires that do not satisfy one or more premises required for the single-cycle dynamics assumed by the proposed model.

In others, including the Byzantine [42] and Holy Roman [43] empires, repeated institutional transformations and persistent competition with comparable powers prevented the emergence of a single coherent expansionary trajectory. Large land empires such as Rome [44] or the Ottomans also illustrate additional limitations. Their long histories involved substantial administrative adaptation, multiple strategic phases, and extended plateaus that are historically meaningful, and are better interpreted as outcomes of complex feedback processes than of a single expansionary pulse. The Ming case highlights a different issue, where territorial dynamics reflect long phases of consolidation within a broader sequence of dynastic transitions [45]. The Safavid Empire likewise does not satisfy the premises of the model, as its territorial trajectory was shaped by sustained rivalry with the Ottoman Empire and ultimately ended through external conquest during the Afghan invasion of the early eighteenth century [46]. These examples thus go beyond highlighting discrepancies and instead contribute to clarifying the scope of the model. In summary, the framework developed in this work applies to cases where territorial growth and contraction form a relatively coherent imperial project

governed primarily by internal dynamics rather than repeated institutional resets or externally imposed discontinuities.

Conclusion

In this work, we propose a minimal extension of logistic growth that captures not only the expansion of empires but also their decline. Across cases of various historical maritime, land-based and wartime empires, the same closed-form law reproduces the characteristic observed pattern of a rapid rise, a subsequent peak and an asymmetric contraction. This is the central significance of the present work: despite the substantial variation in historical context, chronology, geography and political organization, the fitted trajectories reveal a common coarse-grained regularity in the rise and fall of territorial systems.

The robustness of this pattern is particularly striking given the heterogeneity of the examples considered. The maritime empires expanded through overseas control, the land empires through contiguous conquest and frontier incorporation, and the wartime cases through highly compressed episodes of military mobilization. Despite this, all these cases can be represented within the same dynamical framework, with differences expressed primarily through parameter values rather than through changes in functional form. This suggests that the model captures structural features of expansion and overextension that persist despite the many historical intricacies that distinguish individual empires.

A second contribution of the paper is interpretive. By introducing chronophage as the mechanism governing asymmetric decline, the model provides a concrete way of representing what has often been described more qualitatively as the rising burden of complexity, coordination and maintenance. In the long-run historical cases, this burden appears through administrative and fiscal expansion, whereas in the wartime cases it is amplified by mobilization, occupation and the feedback of international resistance. The same formalism therefore connects empirical fits to a substantive account of why territorial growth eventually becomes self-limiting.

Since the model is analytically solvable, parameter-sparse and can be expressed in terms of interpretable quantities, it allows for extension to additional historical and contemporary cases. Its tractability makes it useful not only for retrospective description but also for comparative analysis across systems that differ widely in size and duration. The broader implication of these results is that the rise and fall of empires can be understood simultaneously as a statistical regularity and as a structural consequence of increasing complexity. The generalized Gompertz form identified in this study is not just an empirical fit: it also arises as a maximum-entropy, minimally biased envelope for asymmetric rise–peak–decline trajectories. At the same time, the chronophage gives this abstract form a concrete historical interpretation, by representing the cumulative burdens of administration, coordination and maintenance that grow with scale and ultimately outweigh the gains of further expansion.

In this way, our work places Tainter’s theory of collapse on a quantitative footing: diminishing returns on complexity are expressed dynamically as the gradual accumulation of internal constraints that first suppress expansion and then drive contraction. The common trajectory shared by otherwise very different empires therefore points to a general principle whereby large-scale social systems decline not only because of contingent shocks, but because complexity itself can become self-limiting. Thus, the minimal nature of the framework developed in this work, along with its wide applicability, makes it a valuable tool in understanding the evolution of human systems at large.

Funding

This publication is supported by the European Union’s Horizon Europe research and innovation programme under the Marie Skłodowska–Curie Postdoctoral Fellowship Programme, SMASH co-funded under the grant agreement No. 101081355. The operation (SMASH project) is co-funded by the Republic of Slovenia and the European Union from the European Regional Development Fund.

Disclaimer

Co-funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.

Conflict of interest

The authors declare no competing interests.

Data availability

No new data were collected for this study. The code and processed data used to preprocess the historical series, fit the model, generate the figures, and reproduce the optimal-control simulations are available on Zenodo: https://doi.org/10.5281/zenodo.22160644

The code repository contains two main notebooks. One notebook reads in the pre-processed historical territorial time series, fits the closed-form rise–peak–decline model to the normalized data for each case, and reproduces the empirical figures reported in the paper. The second notebook implements the full optimal-control formulation, numerically integrates the coupled state–costate system, and generates the simulated trajectories and diagnostic outputs shown in the manuscript. Together, the code reproduces the model fitting, parameter estimation, figure generation, and optimal-control simulation underlying the results presented here.

Maximum-entropy characterization of the rise–fall envelope

The functional form of (2) can also be obtained from a simple maximum-entropy construction [47]. Let x(t)≥0x(t) \ge0 denote a non-negative temporal envelope, and define its normalized profile p(t)p(t) as:

p(t)=x(t)∫−∞∞x(u) du,∫−∞∞p(t) dt=1.(4)p(t) = \frac{x(t)}{\int_{-\infty}^{\infty} x(u)\,\mathrm{d}u}, \qquad\int_{-\infty}^{\infty} p(t)\,\mathrm{d}t = 1. \tag*{(4)}

We seek the least-biased profile consistent with two weak structural constraints: a fixed mean time and a fixed exponential moment. The mean fixes a characteristic temporal location, while the exponential moment captures the presence of an exponentially growing burden, hazard, or coordination cost. We therefore maximize the Shannon entropy, H[p]H[p]:

H[p]=−∫−∞∞p(t)ln⁡p(t) dt(5)H[p] = -\int_{-\infty}^{\infty} p(t)\ln p(t)\,\mathrm{d}t \tag*{(5)}

subject to

∫p(t) dt=1,∫t p(t) dt=μ,∫eαtp(t) dt=mα,(6)\int p(t)\,\mathrm{d}t = 1, \qquad\int t\,p(t)\,\mathrm{d}t = \mu, \qquad\int e^{\alpha t}p(t)\,\mathrm{d}t = m_{\alpha}, \tag*{(6)}

where α>0\alpha> 0 is treated as a fixed parameter, μ\mu is the mean, and mαm_{\alpha} is the prescribed exponential moment. Introducing Lagrange multipliers λ0,λ1,λ2\lambda_0,\lambda_1,\lambda_2, we extremize the resulting Lagrangian L[p]\mathcal{L}[p]:

L[p]=−∫pln⁡p dt−λ0(∫p dt−1)−λ1(∫t p dt−μ)−λ2(∫eαtp(t) dt−mα).\mathcal{L}[p] = -\int p\ln p\,\mathrm{d}t - \lambda_0\left(\int p\,\mathrm{d}t - 1\right) - \lambda_1\left(\int t\,p\,\mathrm{d}t - \mu\right) - \lambda_2\left(\int e^{\alpha t}p(t)\,\mathrm{d}t - m_{\alpha}\right).

Taking the first variation with respect to pp gives

δLδp=−ln⁡p(t)−1−λ0−λ1t−λ2eαt=0.(7)\frac{\delta\mathcal{L}}{\delta p} = -\ln p(t) - 1 - \lambda_0 - \lambda_1t - \lambda_2e^{\alpha t} = 0. \tag*{(7)}

Hence

p(t)=exp⁡[−1−λ0−λ1t−λ2eαt].(8)p(t) = \exp\left[-1-\lambda_0-\lambda_1t-\lambda_2e^{\alpha t}\right]. \tag*{(8)}

For the empirically relevant unimodal case on t∈Rt \in\mathbb{R}, we require λ2>0\lambda_2 > 0 and λ1<0\lambda_1 < 0. Writing

β=−λ1>0,γ=λ2>0,(9)\beta= -\lambda_1 > 0, \qquad\gamma= \lambda_2 > 0, \tag*{(9)}

p(t)p(t) can now be rewritten as

p(t)=1Zexp⁡[βt−γeαt],(10)p(t) = \frac{1}{Z}\exp\left[\beta t-\gamma e^{\alpha t}\right], \tag*{(10)}

where ZZ is the normalization constant. Thus, the maximum-entropy profile is an exponential-of-exponential envelope. To connect this expression to the form used in the main text, note that the mode tmax⁡t_{\max} satisfies

ddtln⁡p(t)=β−αγeαt=0,(11)\frac{\mathrm{d}}{\mathrm{d}t}\ln p(t) = \beta-\alpha\gamma e^{\alpha t} = 0, \tag*{(11)}

so that

eαtmax⁡=βαγ.(12)e^{\alpha t_{\max}} = \frac{\beta}{\alpha\gamma}. \tag*{(12)}

Now, define

s=βα>0,γ=se−αtmax⁡;(13)s = \frac{\beta}{\alpha} > 0,\qquad\gamma= s e^{-\alpha t_{\max}}; \tag*{(13)}

this gives

p(t)∝exp⁡[sα(t−tmax⁡)−seα(t−tmax⁡)].(14)p(t) \propto\exp\left[s\alpha(t-t_{\max})-s e^{\alpha(t-t_{\max})}\right]. \tag*{(14)}

Finally, the empirical trajectory x(t)x(t) is an unnormalized envelope rather than a probability density, so we restore an arbitrary vertical scale and impose the peak condition x(tmax⁡)=xmax⁡x(t_{\max})=x_{\max}. This yields

x(t)=xmax⁡exp⁡[s+sα(t−tmax⁡)−seα(t−tmax⁡)],(15)x(t)=x_{\max}\exp\left[s+s\alpha(t-t_{\max})-s e^{\alpha(t-t_{\max})}\right], \tag*{(15)}

which is exactly the functional form used in (2).

In this sense, (2) can be interpreted as the minimum-bias, or maximum-entropy, unimodal asymmetric envelope compatible with a fixed temporal location and a fixed exponential moment. The resulting form does not impose any additional structure beyond these constraints, and therefore provides an information-theoretically neutral baseline for the observed rise–peak–decline trajectories.

Optimal control formulation and reduced dynamics

Following standard textbook treatments of infinite-horizon optimal control, we formulate territorial expansion as an optimal control problem in which territorial extent x(t)>0x(t)>0 is supported by territorial support capacity y(t)>0y(t)>0 [48, 49, 50]. Both x(t)x(t) and y(t)y(t) are measured in units of area. Territory grows at a saturating rate relative to available capacity, while capacity itself can be adjusted through policy effort. The state equations are

x˙(t)=rx(t)(1−x(t)y(t)),y˙(t)=v(t)y(t),(16)\dot{x}(t)=r x(t)\left(1-\frac{x(t)}{y(t)}\right),\qquad\dot{y}(t)=v(t)y(t), \tag*{(16)}

where r>0r>0 is the intrinsic expansion rate and v(t)v(t) is the proportional adjustment rate of support capacity.

The empire chooses the control v(t)v(t) to maximize the discounted lifetime objective [48]:

J=∫0∞e−ρtL(x(t),y(t),v(t)) dt.(17)J=\int_{0}^{\infty}e^{-\rho t}L(x(t),y(t),v(t))\,dt. \tag*{(17)}

Here JJ denotes the total discounted net payoff from managing territorial extent and support capacity over time, ρ>0\rho>0 is the discount rate, and L(x(t),y(t),v(t))L(x(t),y(t),v(t)) is the instantaneous or running payoff. Maximizing JJ therefore means choosing a policy that evolves over time, so as to balance the gains from territorial expansion against the costs of sustaining and adjusting the capacity required to support that expansion.

We propose the following ansatz for the running payoff, L(x(t),y(t),v(t))L(x(t), y(t), v(t)), in (17):

L(x,y,v)=Cx2y+κ[(r−γ)(1−xy)−α]v−κ2v2,(18)L(x,y,v) = C\frac{x^2}{y} + \kappa\left[(r-\gamma)\left(1-\frac{x}{y}\right)-\alpha\right]v-\frac{\kappa}{2}v^2, \tag*{(18)}

with C,κ>0C,\kappa> 0, γ≥0\gamma\ge0, and α≥0\alpha\ge0. The parameter CC scales the direct return to effective territorial control, which is increasing in territorial extent but moderated by the support capacity available to sustain it. The parameter κ\kappa sets the overall strength of capacity adjustment, including both the incentive to change support capacity and the quadratic cost that makes rapid adjustment increasingly expensive. The parameter γ\gamma governs the strength of the balancing force that tends to restore proportionality between territorial extent and support capacity, while α\alpha represents the increasing burden associated with sustaining territorial overextension.

To characterize the optimal policy, it is useful to introduce the Hamiltonian, which combines the current payoff with the effect of present choices on the future evolution of the state variables [48, 50]. Using current-value costates μx\mu_x and μy\mu_y, the Hamiltonian is

H(x,y,v,μx,μy)=L(x,y,v)+μxrx(1−xy)+μyvy.(19)H(x,y,v,\mu_x,\mu_y) = L(x,y,v) + \mu_xrx\left(1-\frac{x}{y}\right) + \mu_yvy. \tag*{(19)}

The costates (also referred to as shadow values in the literature) μx\mu_x and μy\mu_y measure the marginal value of an additional unit of territorial extent and support capacity, respectively [48, 49].

Pontryagin’s maximum principle [50] provides the necessary conditions that any optimal path must satisfy: the state equations, the costate equations, and the stationarity condition for the control. These are

x˙=∂H∂μx,y˙=∂H∂μy,μ˙x=ρμx−∂H∂x,μ˙y=ρμy−∂H∂y,∂H∂v=0.(20)\dot{x} = \frac{\partial H}{\partial\mu_x}, \qquad\dot{y} = \frac{\partial H}{\partial\mu_y}, \qquad\dot{\mu}_x = \rho\mu_x-\frac{\partial H}{\partial x}, \qquad\dot{\mu}_y = \rho\mu_y-\frac{\partial H}{\partial y}, \qquad\frac{\partial H}{\partial v}=0. \tag*{(20)}

The stationarity condition in (20) gives the following optimal feedback rule:

v∗=(r−γ)(1−xy)−α+μyyκ.(21)v^* = (r-\gamma)\left(1-\frac{x}{y}\right)-\alpha+\frac{\mu_y y}{\kappa}. \tag*{(21)}

The last term, μyy/κ\mu_y y/\kappa, reflects the forward-looking value of expanding support capacity: when additional capacity is valuable in future periods, the optimal policy raises v∗v^* accordingly. Since μy\mu_y is the costate associated with yy, it measures the marginal contribution of an additional unit of support capacity to the discounted objective. Thus, μyy/κ\mu_y y/\kappa captures the extent to which the empire increases current adjustment effort because extra capacity is valuable for future expansion. To isolate the core rise–fall mechanism, it is useful to work with the congestion ratio

z(t)=x(t)y(t).(22)z(t) = \frac{x(t)}{y(t)}. \tag*{(22)}

This variable measures the degree to which territorial extent outpaces support capacity. Rewriting the dynamics in terms of z(t)z(t) clarifies that the central mechanism is the accumulation of mismatch between expansion and the capacity required to sustain it. Using (22) in (16), we obtain

z˙=z[r(1−z)−v∗].(23)\dot{z}=z\left[r(1-z)-v^{*}\right]. \tag*{(23)}

Substituting (21) into (23) gives

z˙=γz(1−z)+αz−μyyκz.(24)\dot{z}=\gamma z(1-z)+\alpha z-\frac{\mu_y y}{\kappa}z. \tag*{(24)}

To obtain a transparent reduced form, we consider the regime in which the term μyy/κ\mu_y y/\kappa is small relative to the direct state-dependent terms during the initial transient. This approximation isolates the basic mechanism by which mismatch between territorial extent and support capacity evolves. (24) then reduces to

z˙=γz(1−z)+αz.(25)\dot{z}=\gamma z(1-z)+\alpha z. \tag*{(25)}

(25) contains two distinct mechanisms. The term γz(1−z)\gamma z(1-z) represents a logistic contribution, which tends to restore balance between territorial extent and support capacity. The term αz\alpha z captures the self-reinforcing growth of congestion, that is, the accumulation of mismatch between territorial integration and administrative capability.

Logistic case (α=0\alpha=0). When the mismatch term vanishes (α=0\alpha=0), (25) reduces to

z˙=γz(1−z);(26)\dot{z}=\gamma z(1-z); \tag*{(26)}

this is the standard logistic equation. Its solution is

z(t)=11+Ae−γ(t−t0),A=1−z0z0.(27)z(t)=\frac{1}{1+Ae^{-\gamma(t-t_0)}},\qquad A=\frac{1-z_0}{z_0}. \tag*{(27)}

Using

x˙x=r(1−z),(28)\frac{\dot{x}}{x}=r(1-z), \tag*{(28)}

integration yields

x(t)=x0[1+A1+Ae−γ(t−t0)]r/γ.(29)x(t)=x_0\left[\frac{1+A}{1+Ae^{-\gamma(t-t_0)}}\right]^{r/\gamma}. \tag*{(29)}

Thus, when α=0\alpha=0, the model produces a symmetric logistic-type saturation trajectory for territorial expansion.

Rise–fall case (α>0\alpha>0). When mismatch accumulates (α>0\alpha>0), the additional term in (25) contributes to the growth of the congestion ratio, z(t)z(t). In the limiting regime where this term dominates the restoring mechanism, (25) reduces to

z˙=αz,(30)\dot{z}=\alpha z, \tag*{(30)}

Figure B1. Simulated trajectories of territorial extent and support capacity. The figure displays the full optimal-control dynamics for territorial extent x(t)x(t) (dashed) and support capacity y(t)y(t) (dotted). Both x(t)x(t) and y(t)y(t) are normalized, whereas time is expressed in number of years, starting at zero. Territorial extent grows at a saturating rate relative to available support, whereas support capacity evolves through the control variable. Both trajectories are normalized to a maximum of one to emphasize their relative temporal profiles.

with solution

z(t)=z0eα(t−t0).(31)z(t) = z_0 e^{\alpha(t-t_0)}. \tag*{(31)}

Substituting (31) into (28) gives

x˙=rx(1−z0eα(t−t0)).(32)\dot{x} = r x \left(1-z_0 e^{\alpha(t-t_0)}\right). \tag*{(32)}

Integrating this gives us

x(t)=x0exp⁡[r(t−t0)−rz0α(eα(t−t0)−1)].(33)x(t) = x_0 \exp\left[r(t-t_0)-\frac{r z_0}{\alpha}\left(e^{\alpha(t-t_0)}-1\right)\right]. \tag*{(33)}

Let r=sαr = s\alpha; recentering time at the peak tmax⁡t_{\max} defined by z(tmax⁡)=1z(t_{\max}) = 1 yields

x(t)=xmax⁡exp⁡[s+sα(t−tmax⁡)−seα(t−tmax⁡)],(34)x(t) = x_{\max}\exp\left[s+s\alpha(t-t_{\max})-s e^{\alpha(t-t_{\max})}\right], \tag*{(34)}

which is the functional form used in (2).

The two behaviors therefore arise within the same dynamical framework. When α=0\alpha= 0, territorial expansion approaches a symmetric logistic saturation. When α>0\alpha> 0, the accumulation of mismatch between territorial reach and support capacity drives an asymmetric rise–peak–decline trajectory.

References

  1. [1]Thomas B. L. Kirkwood. Deciphering death: a commentary on Gompertz (1825) “on the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of life contingencies”. Philosophical Transactions of the Royal Society B: Biological Sciences, 370(1666):20140379, 2015.DOI
  2. [2]L. L. Eberhardt and J. M. Brewick. Models for population growth curves. ISRN Ecology, 2012:815016, 2012.
  3. [3]Jeffrey West and Paul K. Newton. Cellular interactions constrain tumor growth. Proceedings of the National Academy of Sciences, 116(6):1918–1923, 2019.
  4. [4]Ana Cascon and William F. Shadwick. Predicting the course of COVID-19 and other epidemic and endemic disease. medRxiv, 2021.DOI
  5. [5]Lara Veylit, Bernt-Erik Sæther, Jean-Michel Gaillard, Eric Baubet, and Marlène Gamelon. Many lifetime growth trajectories for a single mammal. Ecology and Evolution, 11(21):14789–14804, 2021.DOI
  6. [6]Karl W. Butzer. Collapse, environment, and society. Proceedings of the National Academy of Sciences, 109(10):3632–3639, 2012.DOI
  7. [7]Graeme S. Cumming and Garry D. Peterson. Unifying research on social–ecological resilience and collapse. Trends in Ecology & Evolution, 32(9):695–713, 2017.DOI
  8. [8]Sabin Roman. Collapse. In Dictionary of Ecological Economics, pages 72–73. Edward Elgar Publishing, 2023.
  9. [9]Joseph A. Tainter. The Collapse of Complex Societies. Cambridge University Press, Cambridge, 1988.
  10. [10]Sabin Roman. Long-term feedback mechanisms underlying societal growth and collapse. Journal of Big History, 8(3):76–97, 2025.
  11. [11]Cesare Marchetti and Jesse H. Ausubel. Quantitative dynamics of human empires. International Journal of Anthropology, 27(1-2):1–62, 2012.
  12. [12]John Darwin. The Empire Project: The Rise and Fall of the British World-System, 1830–1970. Cambridge University Press, Cambridge, 2009.
  13. [13]Martin Thomas. Fight or Flight: Britain, France, and Their Roads from Empire. Oxford University Press, Oxford, 2014.
  14. [14]J. H. Elliott. Empires of the Atlantic World: Britain and Spain in America, 1492–1830. Yale University Press, New Haven, 2006.
  15. [15]A. R. Disney. A History of Portugal and the Portuguese Empire, volume 2. Cambridge University Press, Cambridge, 2009.
  16. [16]Peter C. Perdue. China Marches West: The Qing Conquest of Central Eurasia. Harvard University Press, Cambridge, MA, 2005.
  17. [17]David Morgan. The Mongols. Wiley-Blackwell, Oxford, 2 edition, 2007.
  18. [18]Hugh Kennedy. The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth to the Eleventh Century. Routledge, London, 3 edition, 2016.
  19. [19]A. C. S. Peacock. The Great Seljuk Empire. Edinburgh University Press, Edinburgh, 2015.DOI
  20. [20]John F. Richards. The Mughal Empire. The New Cambridge History of India. Cambridge University Press, Cambridge, 1993.
  21. [21]Peter Jackson. The Delhi Sultanate: A Political and Military History. Cambridge University Press, Cambridge, 1999.
  22. [22]Richard J. Evans. The Third Reich at war. Allen Lane, London, 2008.
  23. [23]Louise Young. Japan’s Total Empire: Manchuria and the Culture of Wartime Imperialism. University of California Press, Berkeley, 1998.
  24. [24]Charles Esdaile. Napoleon’s Wars: An International History, 1803–1815. Allen Lane, London, 2007.
  25. [25]Angelo Del Boca. Gli italiani in Africa orientale. Vol. 2: La conquista dell’impero. Laterza, Rome-Bari, 1979.
  26. [26]Rein Taagepera. Size and duration of empires: systematics of size. Social Science Research, 7(2):108–127, 1978.
  27. [27]Rein Taagepera. Size and duration of empires: Growth-decline curves, 600 B.C. to 600 A.D. Social Science History, 3(3–4):115–138, 1979.
  28. [28]Rein Taagepera. Expansion and contraction patterns of large polities: context for Russia. International Studies Quarterly, 41(3):475–504, 1997.DOI
  29. [29]Samuel Arbesman. The life-spans of empires. Historical Methods: A Journal of Quantitative and Interdisciplinary History, 44(3):127–129, 2011.
  30. [30]Peter Turchin. Historical dynamics: Why states rise and fall. Princeton University Press, 2018.
  31. [31]Agata Angelika Sojecka and Aleksandra Drozd-Rzoska. Global population: from Super-Malthus behavior to doomsday criticality. Scientific Reports, 14(1):9853, 2024.DOI
  32. [32]Richard Overy. The Penguin Historical Atlas of the Third Reich. Penguin Books, London, 1996.
  33. [33]Martin Gilbert. The Routledge Atlas of the Second World War. Routledge, 2 edition, 2010.
  34. [34]Angus Konstam. Historical Atlas of the Napoleonic Era. Lyons Press, 2003.
  35. [35]Ruth Ben-Ghiat and Mia Fuller, editors. Italian Colonialism. Palgrave Macmillan, New York, 2005.
  36. [36]Anthony Hamilton Millard Kirk-Greene. The thin white line: the size of the British colonial service in Africa. African Affairs, 79(314):25–44, 1980.
  37. [37]Georg Orlandi, Daniel Hoyer, Hongjun Zhao, James S. Bennett, Majid Benam, Kathryn Kohn, and Peter Turchin. Structural-demographic analysis of the Qing Dynasty (1644–1912) collapse in China. PLoS ONE, 18(8):e0289748, 2023.
  38. [38]Mark Harrison, editor. The Economics of World War II: Six Great Powers in International Comparison. Cambridge University Press, Cambridge, 1998.
  39. [39]Jayanth R Banavar, Amos Maritan, and Igor Volkov. Applications of the principle of maximum entropy: from physics to ecology. Journal of Physics: Condensed Matter, 22(6):063101, 2010.DOI
  40. [40]Sabin Roman. Maximum entropy models for unimodal time series: case studies of Universe 25 and St. Matthew Island. In Discovery Science, volume 16090 of Lecture Notes in Computer Science, pages 32–44. Springer, 2025.
  41. [41]Femme S. Gaastra. The Dutch East India Company: expansion and decline. Walburg Pers, Zutphen, 2003.
  42. [42]Jonathan Shepard, editor. The Cambridge History of the Byzantine Empire c.500–1492. Cambridge University Press, Cambridge, 2008.
  43. [43]Barbara Stollberg-Rilinger. The Holy Roman Empire: A Short History. Princeton University Press, Princeton, 2018.
  44. [44]Sabin Roman and Erika Palmer. The growth and decline of the Western Roman Empire: quantifying the dynamics of army size, territory, and coinage. Cliodynamics, 10(2):76–98, 2019.
  45. [45]Sabin Roman. Historical dynamics of the Chinese dynasties. Heliyon, 7(6):e07293, 2021.
  46. [46]Andrew J. Newman. Safavid Iran: Rebirth of a Persian Empire. I.B. Tauris, London, 2006.
  47. [47]Jagat Narain Kapur and Hiremaglur K. Kesavan. Entropy Optimization Principles with Applications. Academic Press, 1992.
  48. [48]Morton I. Kamien and Nancy L. Schwartz. Dynamic Optimization: The Calculus of Variations and Optimal Control in Economics and Management. North-Holland, Amsterdam and New York, second edition, 1991.
  49. [49]Atle Seierstad and Knut Sydsæter. Optimal Control Theory with Economic Applications. North-Holland, 1987.
  50. [50]Suresh P. Sethi and Gerald L. Thompson. Optimal Control Theory: Applications to Management Science and Economics. Springer, New York, 2 edition, 2000.

Paper details

vlnc:W_g3vr4pyk
Version
v1 · latest
Published
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents