Publications

This article engages with Mark Moffett’s proposed definition of society, which emphasizes shared group identification over social interaction, and argues for a different, problem-driven approach to definition in evolutionary social science. Rather than treating definitions as primary, I situate them within the broader research agenda aimed at explaining the Great Holocene Transformation—the dramatic expansion of human societies in scale and complexity over the past 10,000 years. My central analytical focus is on cooperation: the capacity of individuals to coordinate actions toward collective goals despite incentives to free-ride. Accordingly, society is defined here as a collective of individuals engaged in sustained cooperation, with interest group proposed as a more flexible and analytically useful term. Within this framework, polities represent a key subtype of interest groups, characterized by their role as independent political units whose persistence depends on maintaining cooperative cohesion, particularly among elites. Other commonly cited features of societies—such as group identification, territoriality, and intergenerational continuity—are treated as secondary mechanisms that support cooperation rather than defining properties. Ultimately, I argue that flexible, operational definitions are most valuable at intermediate stages of scientific inquiry, where they facilitate the construction and empirical testing of theoretical models.

How do large-scale human societies maintain functional integration? Over the last 12,000 years, polities grew over six orders of magnitude in population, far exceeding the scale at which the interpersonal mechanisms that promote cooperation in small-scale societies can operate. Here we test the hypothesis that cooperation and functional integration in large-scale societies is promoted by social (and, specifically, institutional) complexity. We model large-scale societies as territorial social networks subject to energetic, cognitive, and competitive constraints. We demonstrate that agricultural intensification generates increasing population density, raising the per capita rate of social interaction beyond what small-scale mechanisms can sustain and requiring institutional rules to substitute for interpersonal knowledge. The model predicts that territory and social complexity scale with polity population with exponents of 5/6 and 1/6. Tests against data describing hundreds of polities spanning the Holocene support these predictions, providing quantitative evidence that social complexity is a central integrative mechanism enabling large-scale human cooperation.

Cultural macroevolution (CME) is a subfield of cultural evolution that studies large-scale changes in the cultural traits of whole groups. A central question in CME is how and why characteristics of polities (such as chiefdoms, states, and empires) evolve over time. Multilevel selection, and especially its application to cultural evolution, provides a very useful theoretical framework for CME because many polity-level characteristics evolve under selection pressures that act in opposite directions at different levels of hierarchical organization. In this chapter I discuss the conceptual framework that multilevel selection provides for studying CME and summarize the main empirical results from my recent book (Turchin 2025). I conclude that major predictions of cultural multilevel selection theory enjoy substantial empirical support.

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve more than 90% accuracy on popular benchmarks such as Measuring Massive Multitask Language Understanding1, limiting informed measurement of state-of-the-art LLM capabilities. Here, in response, we introduce Humanity’s Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be an expert-level closed-ended academic benchmark with broad subject coverage. HLE consists of 2,500 questions across dozens of subjects, including mathematics, humanities and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable but cannot be quickly answered by internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a marked gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai.

Is it possible to forecast the dynamics of societal resilience and its obverse, sociopolitical unrest or even breakdown? This is the goal of Structural-Demographic Theory (SDT), which integrates mechanism-based models with data and focuses on the dynamics of structural drivers for instability over the long-term (thus, requiring a historical approach). Several recent studies utilizing the SDT framework have proven adept at predicting (or “retrodicting”) sociopolitical instability in c.20 past societies. It was also used in 2010 to successfully forecast the outbreak of US instability for 10 years in the future (in 2020). This study applies the SDT to explore the dynamics of another contemporary society. Our aim is to empirically test SDT in a most rigorous way, using it to forecast future dynamics of sociopolitical instability in Japan. Our research questions are: (1) how accurately (if at all) does the SDT framework predict future levels of sociopolitical instability; (2) What are the relative contributions of possible drivers of instability, including those proposed by SDT, as well as other theories, in explaining instability levels; and (3) are there key ‘leverage points’ that might help mitigate the negative consequences of instability? This article explains how we develop quantitative indices for SDT drivers of instability: mass immiseration, elite overproduction, and state fiscal distress. We then construct a Political Stress Index to track the evolution of these pressures from the 1990s to the present and implement a multipath forecasting model to project future trends through 2050, using a variety of intervention (or non-intervention) scenarios. Our plan is to revisit these predictions ten years in the future with the goal of assessing their accuracy.

Historians and social scientists have long been preoccupied with understanding and documenting periods of crisis. Such emphasis is only growing, and becoming more pressing, as the world continues to face a number of interrelated stressors in the form of irreversible climate change, major ecological shocks and disease outbreaks, eruptions of military violence, economic disruptions and deepening inequalities, political polarization and unrest, the rise of authoritarian and nationalist regimes. Crises in these domains are not new, but have been recurrent features of past societies. Although these periods have typically led to massive loss of life, the failure of critical institutions, and even complete societal collapse, there are instances in the historical record of societies managing to turn the tide of crisis even as violence and social turmoil grow. Here, we focus on four such cases of crises mitigated with structural reforms revealed from our previous historical analyses: early Republican Rome, mid-19th century England and Russia, USA during the the late 20th to early 21st centuries. Utilizing structural demographic theory as a lens to explore these cases, we seek to expose the pressures that built up leading to crisis and the early signs of violent confrontation revealed by these societies, as well as to examine the conditions and key decisions made in the midst of this unrest that allowed these societies to turn the tide and enact significant structural adaptations. Our findings have clear relevance to understanding and navigating similar crises in contemporary societies.

The scientific understanding of the complex dynamics of global history – from the rise and spread of states to their declines and falls, from their peaceful interactions with economic or diplomatic exchanges to violent confrontations – requires, at its core, a consistent and explicit encoding of historical political entities, their locations, extents and durations. Numerous attempts have been made to produce digital geographical compendia of polities with different time depths and resolutions. Most have been limited in scope and many of the more comprehensive geospatial datasets must either be licensed or are stored in proprietary formats, making access for scholarly analysis difficult. To address these issues we have developed Cliopatria, a comprehensive open-source geospatial dataset of worldwide states from 3400BCE to 2024CE. Presently it comprises over 1600 political entities sampled at varying timesteps and spatial scales. Here, we discuss its construction, its scope, and its current limitations.

Soil fertility depletion presents a negative feedback mechanism that could have impacted early adopters of agriculture. We consider whether such feedback can lead to population cycles among early agriculturalists, such as the boom-and-bust patterns suggested by an increasing amount of evidence for Neolithic Europe. Using general mathematical arguments, we show that this is unlikely, due to the interplay of two factors. First, there is an important mathematical difference between biotic (i.e., logistic) and abiotic resource replenishment; soil nutrients are better modeled by the abiotic case, which leads to more stable dynamics. Second, under realistic conditions, the resource replenishment process operates on fast time scales compared to attainable population growth rates, reinforcing the tendency towards stable dynamics. Both these factors are relevant for early agricultural societies and imply that nutrient depletion is likely not the main contributing factor to boom-and-bust cycles observed in the archaeological record.

The impact of inter-group conflict on population dynamics has long been debated, especially for prehistoric and non-state societies. In this work, we consider that beyond direct battle casualties, conflicts can also create a ‘landscape of fear’ in which many non-combatants near theatres of conflict abandon their homes and migrate away. This process causes population decline in the abandoned regions and increased stress on local resources in better-protected areas that are targeted by refugees. By applying analytical and computational modelling, we demonstrate that these indirect effects of conflict are sufficient to produce substantial, long-term population boom-and-bust patterns in non-state societies, such as the case of Mid-Holocene Europe. We also demonstrate that greater availability of defensible locations act to protect and maintain the supply of combatants, increasing the permanence of the landscape of fear and the likelihood of endemic warfare.

The scientific understanding of the complex dynamics of global history – from the rise and spread of states to their declines and falls, from their peaceful interactions with economic or diplomatic exchanges to violent confrontations – requires, at its core, a consistent and explicit encoding of historical political entities, their locations, extents and durations. Numerous attempts have been made to produce digital geographical compendia of polities with different time depths and resolutions. Most have been limited in scope and many of the more comprehensive geospatial datasets must either be licensed or are stored in proprietary formats, making access for scholarly analysis difficult. To address these issues we have developed Cliopatria, a comprehensive open-source geospatial dataset of worldwide states, political groups, events, and rulers from 3400BCE to 2024CE. Presently it comprises over 1800 political entities sampled at varying timesteps and spatial scales. Here, we discuss its construction, its scope, and its current limitations.

Societal ‘crises’ are periods of turmoil and destabilization in socio-cultural, political, economic, and other systems, often accompanied by varying amounts of violence and sometimes significant changes in social structure. The extensive literature analyzing societal crises has concentrated on relatively few historical examples (large-scale events such the fall of the Roman Empire or the French and Russian Revolutions) emphasizing different aspects of these events as potential causes or consistent effects. To investigate crises and prior approaches to explaining them, and to avoid a potential small-sample size bias present in several previous studies, we sought to uniformly characterize a substantial collection of historical crises, spanning millennia, from the prehistoric to post-industrial, and afflicting a wide range of polities in diverse global regions; the Crisis Database (CrisisDB). Here, we describe this dataset which comprises 168 crises suggested by historians and characterized by a number of significant 'consequences' (such as civil war, epidemics, or loss of population) including also institutional and cultural reforms (for example improved sufferance or constitutional changes) that might occur during and immediately following the crisis period. Our analyses show that the consequences experienced by each crisis is highly variable. The outcomes themselves are uncorrelated with one another and, overall, the set of consequences is largely unpredictable when compared to other large-scale properties of society suggested by previous scholars such as its territorial size, religion, administrative size, or historical recency. We conclude that there is no ‘typical’ societal crisis of the past, but crisis situations can take a variety of different directions. We offer some suggestions on the forces that might drive these varying consequences for exploration in future work.

The goal of this study is to empirically test hypotheses about wars of attrition by evaluating their predictions for the conflict in Ukraine. Evaluation will occur after the war is over and authoritative data sources become available for analysis. This pre-registration document presents two quantitative hypotheses that make opposite predictions about the course of the War in Ukraine: (1) the Economic Power hypothesis, which predicts a win for Ukraine and (2) the Casualties Rates hypothesis, which predicts a win for Russia. Additionally, I consider an alternative hypothesis, according to which the outcome will be determined by random unforeseen events. The document includes four main parts: 1. An introduction providing the conceptual background and the rationale for this study. 2. The mathematical framework and a computational model that incorporates both Economic Power and Casualties Rates hypotheses as special cases. 3. An analysis plan that defines model outputs (what is predicted) and model inputs (parameter values and initial conditions), which need to be estimated from data. 4. An interim assessment (as of Summer 2023) using non-authoritative sources illustrating how, after the end of the war, input parameters will be estimated and the accuracy of predictions assessed. At the time of pre-registration (November 2023) the conflict is still unresolved. Neither side has made significant territorial gains for over a year (since the late Fall of 2022). Furthermore, no authoritative source for data, needed to accurately estimate inputs, is currently available. Estimates published in the press differ wildly depending on the source. As a consequence, the alternative predictions discussed in the interim assessment should not be taken as predicting the future course of the conflict. They instead are meant to demonstrate how these specific scientific hypotheses about war dynamics will be assessed after the war concludes.

The impact of inter-group conflict on population dynamics has long been debated, especially for prehistoric and non-state societies. In this work, we consider that beyond direct battle casualties, conflicts can also create a “landscape of fear” in which many non-combatants near theaters of conflict abandon their homes and migrate away. This process causes population decline in the abandoned regions and increased stress on local resources in better protected areas that are targeted by refugees. By applying analytical and computational modeling, we demonstrate that these indirect effects of conflict are sufficient to produce substantial, long-term population boom-and-bust patterns in non-state societies, such as the case of Mid-Holocene Europe. We also demonstrate that greater availability of defensible locations, by acting to protect and maintain the supply of combatants, increases the permanence of the landscape of fear and the likelihood of endemic warfare.

Climate variability and natural hazards like floods and earthquakes can act as environmental shocks or socioecological stressors leading to instability and suffering throughout human history. Yet, societies experience a wide range of outcomes when facing such challenges: some suffer from social unrest, civil violence or complete collapse; others prove more resilient and maintain key social functions. We currently lack a clear, generally agreed-upon conceptual framework and evidentiary base to explore what causes these divergent outcomes. Here, we discuss efforts to develop such a framework through the Crisis Database (CrisisDB) programme. We illustrate that the impact of environmental stressors is mediated through extant cultural, political and economic structures that evolve over extended timescales (decades to centuries). These structures can generate high resilience to major shocks, facilitate positive adaptation, or, alternatively, undermine collective action and lead to unrest, violence and even societal collapse. By exposing the ways that different societies have reacted to crises over their lifetime, this framework can help identify the factors and complex social–ecological interactions that either bolster or undermine resilience to contemporary climate shocks.

This paper analyzes the collapse of the Qing dynasty (1644–1912) through the lens of the Structural Demographic Theory (SDT), a general framework for understanding the drivers of socio-political instability in state-level societies. Although a number of competing ideas for the collapse have been proposed, none provide a comprehensive explanation that incorporates the interaction of all the multiple drivers involved. We argue that the four-fold population explosion peaking in the 19th century, the growing competition for a stagnant number of elite positions, and increasing state fiscal stress combined to produce an increasingly disgruntled populace and elite, leading to significant internal rebellions. We find that while neither the ecological disasters nor the foreign incursions during the 19th century were sufficient on their own to bring down the Qing, when coupled with the rising internal socio-political stresses, they produced a rapid succession of triggering events that culminated in the Qing collapse.

Are human societies dynamical systems? Can they be studied—and, perhaps, to a degree predicted—with the methods of complexity science, such as agent-based models and big data analytics? If yes, what are the limits to prediction? A particularly challenging question is, can we forecast the dynamics of societal resilience and its obverse, sociopolitical unrest or even breakdown? So far efforts to predict onset of rebellions and civil wars using theory-free big data approaches have proved unsuccessful. An alternative approach, based on Structural-Demographic Theory (SDT), which integrates mechanism-based models with data and focuses on the dynamics of structural drivers for instability over the long-term (thus, requiring a historical approach), has shown better promise. Specifically, several recent studies utilizing the SDT framework have proven adept at predicting (or “retrodicting”) sociopolitical instability in c.20 past societies. It was also used in 2010 to successfully forecast outbreak of US instability 10 years in the future (in 2020). Collectively, this work is producing a growing body of evidence showcasing the ability of SDT-based approaches to uncover critical societal dynamics in the deep past as well as more contemporary cases. The next step in these efforts is to employ these insights towards the future. Here, we outline the SDT approach and document how it can be employed to explore the dynamics of any number of past and contemporary societies, appealing to researchers to pursue this line of research in as many cases as possible. The overall goal of this research is to empirically test SDT in a most rigorous way, using it to forecast coming periods of unrest. Because the theory is likely to fail in many ways, the second goal is to learn from these errors so that we can further refine the theory (or develop better-working alternatives). Specifically, we will learn: (1) how accurately (if at all) does the SDT framework predict future levels of sociopolitical instability (integrating incidence of antigovernment demonstrations, violent riots, and armed conflict); (2) What are the relative contributions of possible drivers of instability, including those proposed by SDT, as well as other theories, in explaining instability levels; and (3) are there key ‘leverage points’ that might help mitigate the negative consequences of instability?

Archaeological evidence suggests that the population dynamics of Mid-Holocene (Late Mesolithic to Initial Bronze Age, ca. 7000-3000 BCE) Europe are characterized by recurrent booms and busts of regional settlement and occupation density. These boom-bust patterns are documented in the temporal distribution of 14C dates and in archaeological settlement data from regional studies. We test two competing hypotheses attempting to explain these dynamics: climate forcing and social dynamics leading to inter-group conflict. Using the framework of spatially-explicit agent-based models, we translated these hypotheses into a suite of explicit computational models, derived quantitative predictions for population fluctuations, and compared these predictions to data. We demonstrate that climate variation during the European Mid-Holocene is unable to explain the quantitative features (average periodicities and amplitudes) of observed boom-bust dynamics. In contrast, scenarios with social dynamics encompassing density-dependent conflict produce population patterns with time scales and amplitudes similar to those observed in the data. These results suggest that social processes, including violent conflict, played a crucial role in the shaping of population dynamics of European Mid-Holocene societies.

Why did complex societies, characterized by densely-populated walled cities, first arise in northern China, jump-starting early Chinese civilization? We explore this question in three steps. First, the North, especially the alluvial plains along the Yellow and the Yangtze Rivers, had generally flatter terrains than the South. Second, by dividing China’s landmass into 100 km ×100 km grid-cells and using our archaeological database, we demonstrate that cells with flatter terrains faced higher war threats in prehistoric and early historic times, where war threats are respectively proxied by each cell’s number of excavated military grave goods for the Neolithic period (8000−1700 BCE) and by its number of recorded conflicts for the Eastern Zhou (770−221 BCE, the earliest period for which war data are available). Third, we establish that during both the Neolithic and the Eastern Zhou, higher war threats led to the construction of more settlements with defensive walls and moats, resulting in more walled cities (i.e., early cradles of civilization). Thus, warfare was a key driver of the evolution of complex societies. This finding is robust after controlling for irrigation potential, agricultural productivity, and threats from the steppe, as well as under alternative specifications.