Research · Summary
Back to sweepResearch sweep · deep · 2015 – 2026
Emergent Behaviour Across Disciplines
The science of emergent behaviour and self-organisation from 2015–May 2026, connecting reaction-diffusion and cellular-automata models to commercial markets, organisational behaviour and culture, ecology, biology, physics and chaos theory, including foundational algorithms and their cross-domain explanatory power.
- academic
- blogs
- financial
Synthesised 2026-05-15
Emergence becomes engineering
The study of emergence in 2025 looks fundamentally different from the descriptive complexity science of the 2000s. Three things have changed: cellular automata are now differentiable and trainable, agent-based models have crossed from stylised demonstration into central-bank policy infrastructure, and the strong-versus-weak emergence debate has been reframed around computational irreducibility rather than metaphysics. The CFA Institute's October 2025 report formally instructing investment professionals to treat markets as complex adaptive systems is the clearest signal that the field has exited its academic enclave.
Sources: CFA Institute Research and Policy Center (2025) (↗); Distill (2020) (↗)
The intellectual scaffolding rests on a longer arc. Turing's 1952 reaction-diffusion paper, von Neumann's self-replicating automata, Conway's Game of Life (1970), Wolfram's elementary cellular automaton classification, Reynolds' Boids (1987), Prigogine's dissipative structures, and Bak's self-organised criticality (1987) all describe how local rules generate global order without central control. Gray and Scott's reaction-diffusion model sits at the canonical junction of continuous and discrete, producing spots, stripes, mitotic splits and travelling solitons from two coupled PDEs tuned by a feed rate and a kill rate.
Sources: Mathematics (MDPI) (2023) (↗); biologicalmodeling.org (Philip Compeau) (2022) (↗); arXiv (2024) (↗)
The defining shift of the past 18 months is operational. Robert Axtell and Doyne Farmer's 90-page Journal of Economic Literature survey (March 2025) declared agent-based modelling a mature methodology with documented wins in volatility clustering, market impact, systemic risk and housing dynamics. Carlos Gershenson's npj Complexity article (March 2025) unified self-organisation across a dozen disciplines under a computational definition. Bert Chan's Lenia and Flow-Lenia extended cellular automata into continuous space, time and state, cataloguing over 400 self-organising species and producing open-ended evolutionary dynamics within a single mathematical framework.
Sources: Journal of Economic Literature (2025) (↗); npj Complexity (2025) (↗); arXiv / Complex Systems (2018) (↗); arXiv (2025) (↗)
Why it matters now: machine learning has made the inverse problem tractable. Given a desired emergent pattern, neural cellular automata can learn the local update rule that produces it. This reverses the classical direction of complexity research and has direct implications for morphogenetic engineering, programmable materials, and, more speculatively, the reverse-engineering of organisational and market patterns.
Sources: Distill (2020) (↗); arXiv / ALIFE 2024 (2024) (↗); arXiv (2025) (↗)
Key findings
The foundational algorithms now form a coherent family, not a zoo. Conway's Game of Life, Wolfram's elementary CA, Langton's Ant, Reynolds' Boids, Schelling segregation, Sugarscape, ant-colony stigmergy and particle swarm optimisation share a formal property: macro-order from local rules with no global controller. The Gray-Scott reaction-diffusion model is the continuous analogue, governed by two coupled PDEs where activator u and inhibitor v evolve under diffusion and a cubic autocatalytic reaction, tuned by feed parameter f and kill parameter k. A 2024 arXiv review extends Turing instability theory to discrete networks and higher-order topologies, demonstrating that the same instability mechanism produces patterns across biological, economic and linguistic systems and providing the mathematical bridge between PDE-based and agent-based models.
Sources: arXiv (2024) (↗); VisualPDE.com (2025) (↗); biologicalmodeling.org (Philip Compeau) (2022) (↗)
Neural cellular automata are the most important methodological development of the period. Mordvintsev and colleagues at Google's Distill team (February 2020) trained local cell-update rules to grow and regenerate target morphologies, producing self-healing patterns reminiscent of biological tissue. Follow-up work introduced adversarial reprogramming, self-organising textures, and attention-based variants at NeurIPS 2022. A June 2025 arXiv paper applies NCAs to the ARC-AGI benchmark and exposes hard limits on decentralised computation relative to large language models, while differentiable logic cellular automata (June 2025) make the update rule symbolically interpretable.
Sources: Distill (2020) (↗); Distill.pub (2021) (↗); Distill.pub (2021) (↗); NeurIPS 2022 (2022) (↗); arXiv (2025) (↗); arXiv (2025) (↗)
Self-organised criticality has crossed from theory to engineered system. A July 2021 Nature Communications paper engineered SOC into E. coli gene networks, the first synthetic biological instantiation of Bak's mechanism. Romanczuk's 2022 review synthesised evidence that fish schools and bird flocks operate near pseudo-critical points to maximise collective information processing. A December 2025 arXiv paper demonstrates SOC and sharp phase transitions in diverse model ecosystems, placing them in directed-percolation universality classes, and a July 2025 paper shows symmetry-breaking and critical growth transitions in stem-cell-derived organoids.
Sources: Nature Communications (2021) (↗); arXiv (2022) (↗); arXiv (2025) (↗); arXiv (2025) (↗)
Complexity economics has reached operational adoption in finance. The CFA Institute's October 2025 report co-authored by Genevieve Hayman and Raymond Pang explicitly instructs the asset-management profession to treat markets as complex adaptive systems with feedback loops, network effects and emergent behaviour. W. Brian Arthur's 2021 Nature Reviews Physics article and the 2020 SFI Press volume edited with Beinhocker and Stanger provide the academic anchor. Axtell and Farmer's March 2025 Journal of Economic Literature survey documents concrete wins: ABMs reproduce volatility clustering, market impact functions, systemic risk cascades and housing bubble dynamics that DSGE models cannot generate endogenously.
Sources: CFA Institute Research and Policy Center (2025) (↗); Nature Reviews Physics (2021) (↗); Santa Fe Institute Press (2020) (↗); Journal of Economic Literature (2025) (↗)
Central banks have moved from interest to deployment. A 2025 Banco de España and INET Oxford working paper documents agent-based modelling at central banks as a working programme rather than a research curiosity, with applications to housing markets, corporate bond liquidity and stress-testing. Farmer's group at Oxford correctly predicted the UK's Q2 2020 GDP collapse using an input-output ABM during COVID-19. A January 2025 special issue of the Journal of Economic Behavior and Organization, edited by Bednar, Farmer, del Rio-Chanona and others, bundled six thematic advances (inequality, climate, technology, production networks, macroeconomics, ABM) under one complexity-economics banner.
Sources: INET Oxford Working Papers / Banco de España (2025) (↗); Journal of Economic Behavior & Organization (2025) (↗); Allen Lane / Oxford Martin School (2024) (↗)
Organisational emergence has finally received a formal academic framework. A June 2025 Review of Managerial Science paper introduces a Theory of Organisational Emergence (TOE), addressing invisible conflicts and emerging dynamics in entrepreneurship as formal complex-systems phenomena rather than narrative analogy. This complements Ralph Stacey's complex responsive processes and Laloux's teal organisations, but with sharper theoretical claims. Independent writers including the Human-Centric Engineering and Post-Bureaucracy Substacks frame culture, dysfunction and cohesion as emergent properties of local agent interactions that managerial hierarchies tend to suppress.
Sources: Review of Managerial Science (Springer) (2025) (↗); Substack — Human-Centric Engineering (2024) (↗); Substack — Post-Bureaucracy (2021) (↗)
Statistical-physics readings of markets are sharpening. JP Bouchaud's July 2024 Substack post connects self-organised criticality to the excess-volatility puzzle, arguing that markets sit generically near criticality so small internal shocks trigger endogenous crashes with power-law tails. The Wider Angle (April 2026) extends SOC to societal polycrisis. Psych Safety (April 2025) applies sandpile logic to organisational safety, framing accidents as avalanches on a self-organised criticality landscape rather than failures of individual judgement.
Sources: Substack — JP Bouchaud (2024) (↗); Substack — The Wider Angle (2026) (↗); psychsafety.com (practitioner blog) (2025) (↗)
The emergence philosophy debate has shifted onto computational ground. A July 2025 arXiv survey, "What is emergence after all?", and Hervé Zwirn's 2023 piece "Explaining Emergence" reframe the strong-versus-weak distinction around computational irreducibility in Wolfram's sense: most observed emergence is weak (derivable in principle from micro-rules) but practically incompressible (you cannot predict the trajectory faster than running the system). The Stanford Encyclopedia of Philosophy and IEP entries provide the formal philosophical scaffolding. Practitioner writers conflate weak and strong emergence routinely, a category error that most complexity research treats as resolved.
Sources: arXiv (2025) (↗); arXiv / UM6P Science Week Proceedings (2023) (↗); Stanford Encyclopedia of Philosophy (2020) (↗); Internet Encyclopedia of Philosophy (2022) (↗)
LLM-augmented ABMs are the contested frontier. A Santa Fe Institute report in October 2025 documents efforts to encode agents with LLM-derived priors to produce policy-grade agent-based models. A parallel 2025 arXiv review of generative social simulations finds validation largely absent and LLM opacity problematic for isolating emergent causal mechanisms. This is the active fault line: the promise of richer agents versus the loss of mechanistic interpretability that made ABMs attractive in the first place.
Sources: Santa Fe Institute (2025) (↗); Journal of Economic Interaction and Coordination (2025) (↗)
Calibration remains the operational bottleneck. The Journal of Artificial Societies and Social Simulation paper on high-frequency flash crash scenarios (March 2024) demonstrates that ABMs can reproduce micro-structural crash dynamics but only with extensive parameter fitting. PLOS ONE (2018) on calibrating emergent phenomena in stock markets formalised the inverse problem. A February 2026 Journal of Simulation review of ABM in economic markets identifies calibration, validation and computational scale as the three persistent obstacles to regulatory-grade use.
Sources: Journal of Artificial Societies and Social Simulation (2024) (↗); PLOS ONE (2018) (↗); Journal of Simulation (Taylor & Francis) (2026) (↗)
Evidence and data
The quantitative spine of the sweep is the Axtell-Farmer survey in the Journal of Economic Literature (March 2025), 90 pages reviewing four decades of agent-based modelling in economics and finance. It documents successful reproduction of volatility clustering exponents, market impact power laws (price impact scaling with the square root of trade volume), and systemic risk propagation in interbank networks. The 2018 PLOS ONE calibration paper showed that ABM-generated returns can be tuned to match S&P 500 stylised facts including heavy-tailed distributions and long-memory volatility, though out-of-sample predictive performance remains weak.
Sources: Journal of Economic Literature (2025) (↗); PLOS ONE (2018) (↗)
Bert Chan's Lenia (arXiv, December 2018) catalogued over 400 distinct self-organising lifeform species in a continuous cellular automaton, demonstrating formal equivalence with generalised reaction-diffusion systems. Flow-Lenia (2022 and 2025) added mass conservation and localised update-rule parameters, enabling multi-species co-existence and open-ended evolutionary dynamics in a single substrate.
Sources: arXiv / Complex Systems (2018) (↗); arXiv / Artificial Life Conference 2023 (2022) (↗); arXiv (2025) (↗)
Akaash Dudwani's December 2025 Substack cites BCG figures placing complexity-aware innovation project success at 37 percent versus 12 percent for stage-gate processes, a roughly 3x ratio that the post itself acknowledges is thinly sourced. The number should be treated as practitioner folklore rather than calibrated evidence.
Sources: Substack — Akaash Dudwani (2025) (↗)
The Gray-Scott parameter space is empirically mapped in the 2023 Mathematics (MDPI) paper, identifying spot, stripe, mitosis and soliton regimes across the f-k plane with sharp transitions between qualitatively distinct patterns. The September 2024 arXiv paper on the thermodynamically-consistent variational Gray-Scott model recovers these regimes while restoring energy conservation, connecting the abstract model to physically realisable systems.
Sources: Mathematics (MDPI) (2023) (↗); arXiv (2024) (↗)
Doyne Farmer's COVID prediction is the most cited concrete forecasting win for complexity economics: an input-output ABM at Oxford correctly anticipated the UK's Q2 2020 GDP collapse during the pandemic, a period when conventional macro models failed conspicuously.
Sources: Allen Lane / Oxford Martin School (2024) (↗); INET Oxford (2025) (↗)
Signals and tensions
Descriptive versus predictive. J. Barkley Rosser's critique, echoed in the Journal of Economic Behavior and Organization special issue introduction, is that complexity analyses tend to confirm already-observed stylised facts (volatility clustering, fat tails, lock-in) rather than generate novel falsifiable predictions. The financial press treats complexity economics as a paradigm change; the JEL survey treats it as a methodology with documented but bounded wins.
Sources: Journal of Economic Behavior & Organization (2025) (↗); Journal of Economic Literature (2025) (↗); INET Oxford Working Papers (2025) (↗)
LLM agents versus mechanistic interpretability. Encoding agents with LLM-derived behaviour produces richer simulations but obscures the local rules that make ABMs causally legible. The SFI October 2025 report sees this as the path to policy-grade models. The 2025 JEIC review treats it as a validation crisis in waiting.
Sources: Santa Fe Institute (2025) (↗); Journal of Economic Interaction and Coordination (2025) (↗)
Edge-of-chaos in organisations. Langton's edge-of-chaos hypothesis has been adopted wholesale by management writers (Stacey, Snowden, Laloux) but has weak empirical support in human organisations. Romanczuk's animal-collectives review provides solid criticality evidence in fish and birds; equivalent evidence for firms or teams is absent. The Review of Managerial Science TOE paper attempts to fill the gap formally.
Sources: arXiv (2022) (↗); Review of Managerial Science (Springer) (2025) (↗)
Strong emergence as live or dead debate. Most current complexity science treats strong emergence as scientifically undemonstrated and works within weak emergence plus computational irreducibility. Independent essayists and practitioner writers still invoke strong emergence implicitly when describing culture or markets as ontologically irreducible. The July 2025 arXiv survey treats the debate as largely resolved in favour of weak emergence; practitioner writing has not caught up.
Sources: arXiv (2025) (↗); arXiv / UM6P Science Week Proceedings (2023) (↗)
SOC in finance: mechanism or metaphor. Bouchaud's statistical-physics framing of markets as generically critical is mathematically precise. Its application to societal polycrisis (The Wider Angle, 2026) and organisational safety (Psych Safety, 2025) is suggestive but lacks the calibration that makes SOC empirically tractable in physics.
Sources: Substack — JP Bouchaud (2024) (↗); Substack — The Wider Angle (2026) (↗); psychsafety.com (practitioner blog) (2025) (↗)
NCA scope. Neural cellular automata excel at morphogenetic targets and texture synthesis but fail at general reasoning benchmarks. The June 2025 ARC-AGI paper is honest about this, framing NCAs as a complement to LLMs rather than a replacement.
Sources: arXiv (2025) (↗)
Open questions
- Can ABMs generate sharp out-of-sample forecasts that beat DSGE on policy counterfactuals, or are they confined to reproducing stylised facts? The Axtell-Farmer survey is optimistic; the calibration literature is not.
Sources: Journal of Economic Literature (2025) (↗); PLOS ONE (2018) (↗)
- Is strong emergence empirically distinguishable from weak emergence in any complex system, or is the distinction permanently philosophical? Recent arXiv work treats this as a closed question; philosophy entries treat it as open.
Sources: arXiv (2025) (↗); Stanford Encyclopedia of Philosophy (2020) (↗)
- Do human organisations actually sit near criticality, or is the edge-of-chaos framing borrowed metaphorically from animal collectives where it is empirically established? No direct evidence yet for firms or teams.
Sources: arXiv (2022) (↗); Review of Managerial Science (Springer) (2025) (↗)
- Can LLM-augmented ABMs be validated when the agents themselves are opaque? The SFI agenda assumes yes; the JEIC review documents the absence of validation protocols.
Sources: Santa Fe Institute (2025) (↗); Journal of Economic Interaction and Coordination (2025) (↗)
- What is the formal mathematical unification between cellular automata, reaction-diffusion systems and agent-based models? The 2024 networked Turing patterns paper points the way; a full unification is not yet established.
Sources: arXiv (2024) (↗); arXiv (2024) (↗)
- Can neural cellular automata be inverted on observed organisational or market data to recover the local rules that produced observed patterns? The technical apparatus exists; the application is speculative.
Sources: Distill (2020) (↗); arXiv (2023) (↗)
- Where does Lucas-critique-style structural invalidity bite hardest in ABM policy use? Agent rules calibrated to historical data may not survive the policy interventions ABMs are meant to evaluate, and the field has not produced a clean answer.
Sources: INET Oxford Working Papers / Banco de España (2025) (↗); INET Oxford Working Papers (2025) (↗)
The field's next test is whether the CFA Institute's 2025 framing produces measurably different risk management or whether complexity economics ends up as the new vocabulary for the same practices. The methodology is ready. The validation is not.
![[sources-the-science-of-emergent-behaviour-and-self-organis]]
Sources
Summary: ↑ Back to summary
Academic & arXiv
| ID | Title | Outlet | Date | Significance |
|---|---|---|---|---|
| a1 | Growing Neural Cellular Automata | Distill | 2020-02 | Foundational paper by Mordvintsev, Randazzo, Niklasson and Levin establishing the neural CA paradigm for morphogenesis and self-organising regeneration, catalysing a wave of follow-on work. |
| a2 | Attention-based Neural Cellular Automata | NeurIPS 2022 | 2022-10 | Introduces Vision Transformer Cellular Automata (ViTCA), extending neural CAs with spatially localised self-attention and demonstrating globally organised self-organisation from local rules. |
| a3 | Lenia - Biology of Artificial Life | arXiv / Complex Systems | 2018-12 | Chan's canonical paper introducing Lenia, a continuous-space-time-state generalisation of Conway's Game of Life that bridges CA and reaction-diffusion systems, cataloguing over 400 emergent lifeform species. |
| a4 | Flow-Lenia: Towards open-ended evolution in cellular automata through mass conservation and parameter localization | arXiv / Artificial Life Conference 2023 | 2022-12 | Extends Lenia with mass conservation and localised update-rule parameters, enabling multi-species simulations and paving the way for intrinsic open-ended evolution in continuous CAs. |
| a5 | Flow-Lenia: Emergent evolutionary dynamics in mass conservative continuous cellular automata | arXiv | 2025-06 | Latest journal-length treatment of Flow-Lenia demonstrating emergent autopoiesis, self-replication, and open-ended evolutionary dynamics in continuous cellular automata. |
| a6 | Locally adaptive cellular automata for goal-oriented self-organization | arXiv | 2023-06 | Introduces adaptive CA that couples update rules to system state, demonstrating self-tuning to the critical Ising temperature and plastic neural-network analogues of collective emergence. |
| a7 | Emergent Dynamics in Neural Cellular Automata | arXiv / ALIFE 2024 | 2024-04 | Empirically characterises spontaneous motion, self-regeneration and generalisation as distinct emergent regimes within trained neural CAs, providing a comparative taxonomy of emergence types. |
| a8 | Neural Cellular Automata for ARC-AGI | arXiv | 2025-06 | Tests self-organising NCAs against the ARC-AGI benchmark, exposing the limits of decentralised computation relative to LLM-based approaches and mapping open problems in self-organising computation. |
| a9 | Differentiable Logic Cellular Automata | arXiv | 2025-06 | Proposes end-to-end learnable CA using binary logic gates rather than neural networks, producing interpretable, discrete, and robust self-organising systems that learn Conway's Game of Life rules. |
| a10 | On pattern formation in the thermodynamically-consistent variational Gray-Scott model | arXiv | 2024-09 | Extends the canonical Gray-Scott reaction-diffusion model to enforce thermodynamic consistency, revealing how energy constraints alter pattern regimes and providing a more physically grounded self-organisation framework. |
| a11 | Reduced dynamics for models of pattern formation | arXiv | 2024-03 | Proves that Gray-Scott, Brusselator and Glycolysis reaction-diffusion systems admit finite-dimensional reduced representations, providing analytical foundations for pattern formation and data assimilation. |
| a12 | Turing patterns on discrete topologies: from networks to higher-order structures | arXiv | 2024-07 | Comprehensive review extending Turing instability theory from continuous PDEs to networked and higher-order topologies, unifying pattern formation across biological, economic and linguistic domains. |
| a13 | Spatiotemporal Dynamics of Reaction-Diffusion System and Its Application to Turing Pattern Formation in a Gray-Scott Model | Mathematics (MDPI) | 2023-03 | Accessible empirical study of Gray-Scott parameter space, documenting transitions between spot, stripe, mitosis and soliton regimes and connecting them to biological morphogenesis. |
| a14 | Phase Transitions and Criticality in the Collective Behavior of Animals -- Self-organization and biological function | arXiv | 2022-11 | Synthesises statistical-physics criticality theory with empirical animal-collective data (fish schools, bird flocks, insect swarms), reviewing the 'criticality hypothesis' and its functional consequences. |
| a15 | Self-organized criticality in complex model ecosystems | arXiv | 2025-12 | Demonstrates SOC and sharp phase transitions in diverse, spatially structured model ecosystems, connecting Lotka-Volterra dynamics to directed-percolation universality classes. |
| a16 | Eigenstates in the self-organised criticality | arXiv | 2024-01 | Applies eigen-microstate analysis to BTW and Manna sandpile SOC models, characterising phase transitions without a pre-specified order parameter and estimating critical exponents numerically. |
| a17 | Engineering self-organized criticality in living cells | Nature Communications | 2021-07 | Experimental demonstration that SOC can be engineered into gene networks in E. coli via negative feedback on proteolytic degradation, directly linking physical SOC theory to synthetic biology. |
| a18 | Self-organization drives symmetry-breaking, scaling, and critical growth transitions in stem cell-derived organoids | arXiv | 2025-07 | Integrates imaging of over 10,000 gastruloid colonies with reaction-diffusion modelling to show spontaneous symmetry-breaking, power-law scaling and critical slowing-down in developmental self-organisation. |
| a19 | What is emergence after all? | arXiv | 2025-07 | Recent synthesis clarifying the weak-vs-strong emergence debate across physics, biology and AI, distinguishing explanatory autonomy from genuine ontological novelty and surveying open problems. |
| a20 | Self-organizing systems: what, how, and why? | npj Complexity | 2025-03 | Broad interdisciplinary survey by Gershenson spanning physics, chemistry, biology, ecology, social science and AI, providing a canonical contemporary framing of self-organisation across domains. |
| a21 | Explaining Emergence (Zwirn) | arXiv / UM6P Science Week Proceedings | 2023-08 | Analyses computational irreducibility as the mechanism distinguishing weak from strong emergence, with cellular automata as central examples, directly relevant to the ontological-vs-epistemic debate. |
| a22 | Foundations of complexity economics | Nature Reviews Physics | 2021-01 | W. Brian Arthur's authoritative review article presenting complexity economics as an extension of equilibrium theory, describing emergent macro-patterns from heterogeneous interacting agents. |
| a23 | Complexity Economics (SFI Press, 2020) | SFI Press | 2020-11 | Proceedings volume edited by Arthur, Beinhocker and Stanger collecting SFI's 2019 symposium contributions, establishing the current frontier of complexity economics including agent-based macro and market emergence. |
| a24 | Studying economic complexity with agent-based models: advances, challenges and future perspectives | Journal of Economic Interaction and Coordination | 2025-04 | Chudziak's 2025 review systematically maps ABM advances in economics, highlighting irreducible emergent phenomena and proposing new research agendas for multi-sectoral economic complexity. |
| a25 | Agent-based models move into the economic mainstream | Santa Fe Institute | 2025-10 | Reports SFI working-group findings on integrating LLMs with ABMs for policy-grade economic simulation, marking the practical convergence of complexity economics with modern AI tooling. |
Blogs & Independent Thinkers
| ID | Title | Outlet | Date | Significance |
|---|---|---|---|---|
| b1 | Emergence, and Team Dynamics — Human-Centric Engineering | Substack — Human-Centric Engineering | 2024-09 | Applies emergence and complex-adaptive-systems framing directly to software engineering teams, arguing that culture, dysfunction and cohesion are all emergent properties of local agent interactions rather than top-down design choices. |
| b2 | What is Complexity Theory? — Akaash Dudwani | Substack — Akaash Dudwani | 2025-12 | Surveys agent-based modelling, phase transitions, dynamical systems and organisational applications of complexity theory, including empirical claims about stage-gate versus complexity-aware innovation success rates. |
| b3 | Beyond Static Analysis — Getting Out of Control | Substack — Getting Out of Control (Neil) | 2025-04 | Makes a first-principles argument that markets and digital platforms are emergent-order systems and that antitrust frameworks must incorporate complexity-science insights to analyse AI and platform competition. |
| b4 | The Metaphysics of Complexity — Wonderland | Substack — Wonderland | 2020-09 | Develops a metaphysical account of complex adaptive systems, arguing that emergent properties — including culture, trends and norms — possess genuine causal power that feeds back to shape individual agents. |
| b5 | Complex Adaptive Systems — the nature of organisation — Post-Bureaucracy | Substack — Post-Bureaucracy | 2021-01 | Makes the case that managerial hierarchies, rooted in mechanistic 19th-century paradigms, actively suppress the autonomous agent behaviour necessary for emergence, and that post-bureaucratic design should increase agent autonomy. |
| b6 | A Computational Approach to Emergence — Complexity Thoughts | Substack — Complexity Thoughts | 2024-07 | Frames emergence computationally — as structured collective behaviour arising from fundamental units without centralised control — and explores what distinguishes it from mere randomness or designed order. |
| b7 | Chaos, Collapse and Complex Systems in Volatile Times — Claire Hartnell | Substack — Claire Hartnell | 2025-01 | Applies tipping-point and emergent-properties framing to societal and organisational fragility, connecting feedback loops, fractal behaviour and disintegration dynamics to real-world crises. |
| b8 | Societal Transitions and Self-Organised Criticality — The Wider Angle | Substack — The Wider Angle | 2026-04 | Applies Bak's self-organised criticality and power-law dynamics to societal risk, geopolitical cascades and civilisational transitions, connecting the sandpile model to polycrisis modelling and agent-based simulation. |
| b9 | The Self-Organized Criticality Paradigm in Economics and Finance — JP Bouchaud | Substack — JP Bouchaud | 2024-07 | By a leading statistical physicist and hedge-fund researcher, this post rigorously links SOC to excess volatility, endogenous market crashes and the 'small shocks, large cycles' puzzle, grounding the metaphor in quantitative finance evidence. |
| b10 | Growing Neural Cellular Automata — Distill | Distill.pub | 2020-02 | Canonical interactive research article demonstrating that neural networks can learn cellular-automaton update rules to produce self-organising, regenerative morphogenetic patterns — directly linking reaction-diffusion theory to modern deep learning. |
| b11 | Adversarial Reprogramming of Neural Cellular Automata — Distill | Distill.pub | 2021-05 | Extends the neural CA programme by showing how adversarial perturbations can reprogramme self-organising systems, with implications for robustness, downward causation, and the engineering of emergent behaviour. |
| b12 | Self-Organising Textures — Distill | Distill.pub | 2021-02 | Demonstrates that neural cellular automata, when reformulated as discretised PDEs, reproduce Turing-pattern textures — providing a direct computational bridge between reaction-diffusion theory and learnable self-organisation. |
| b13 | The Gray-Scott Model: A Turing Pattern Cellular Automaton — Biological Modeling | biologicalmodeling.org (Philip Compeau) | 2022 | Provides the clearest practitioner-level derivation of the Gray-Scott feed/kill parameter space, showing how minute shifts in f and k drive the system between stripes, spots and solitons — the canonical demonstration of self-organisation from minimal rules. |
| b14 | Reaction-Diffusion Models and Turing Patterns — Nils Olovsson | nils-olovsson.se (personal technical blog) | 2023-05 | GPU-accelerated Python implementation and visual survey of Gray-Scott parameter regimes (coral, mitosis) with clear exposition of how diffusion ratios control pattern selection — useful cross-reference for the computational machinery. |
| b15 | Reaction-Diffusion: Gray-Scott on a 2D Grid — 4rknova | 4rknova.com (personal technical blog) | 2026-02 | Recent (February 2026) WebGL implementation explaining the nonlinear interaction term and double-buffering scheme, with the explicit observation that all visual complexity is emergent from identical local update rules. |
| b16 | Gray-Scott Model — VisualPDE | VisualPDE.com | 2025 | Interactive browser simulation of the Gray-Scott system with parameter sweeps, demonstrating the full dynamical regime map including spiral waves and the effect of diffusion ratio D on pattern variety. |
| b17 | Economic Complexity: A Different Way to Look at the Economy — W. Brian Arthur (Medium / SFI) | Medium — Santa Fe Institute Foundations & Frontiers | 2014 | Arthur's own account of how complexity economics emerged from the 1987 SFI conference, framing market price formation as emergent order and setting out the core challenge to equilibrium economics that organised the field. |
| b18 | Complexity Economics — SFI Press (Arthur, Beinhocker, Stanger) | SFI Press | 2020 | The primary edited volume from SFI's 2019 Fall Symposium, collating practitioner and researcher perspectives on agent-based macro, market microstructure, collective intelligence and organisational scaling under a complexity lens. |
| b19 | Complexity Economics: An Introduction — SFI Press | SFI Press | 2020 | Frames Adam Smith's invisible-hand metaphor explicitly as a statement about emergence, traces complexity economics from Prigogine's Brussels group to the SFI programme, and argues that macro patterns emerge from dynamic micro- and meso-level interactions. |
| b20 | Complexity Economics — W. Brian Arthur personal site | santafe.edu / personal site | 2021 | Arthur's own curated page on complexity economics, with the key insight that increasing-returns lock-in and artificial stock-market GARCH behaviour are the two demonstrated 'killer apps' where the complexity framework outperforms equilibrium models. |
| b21 | Studying Economic Complexity with Agent-Based Models: Advances, Challenges and Future Perspectives — Journal of Economic Interaction and Coordination | Springer / Journal of Economic Interaction and Coordination | 2024-12 | Peer-reviewed critique arguing that agent-based computational economics has moved too far toward curve-fitting and statistical matching, losing sight of the original mission to model emergent feedback mechanisms in heterogeneous, multi-sectoral economies. |
| b22 | Self-Organizing Systems: What, How, and Why? — npj Complexity | npj Complexity (Nature Portfolio) | 2025-03 | A wide-ranging 2025 review by Carlos Gershenson mapping self-organisation across physics, biology, ecology, AI, linguistics and social science, with citations to Carroll's 2024 arXiv paper on what emergence can possibly mean. |
| b23 | Emergent Properties — Stanford Encyclopedia of Philosophy | Stanford Encyclopedia of Philosophy | 2020-08 | The canonical philosophical reference distinguishing weak emergence (epistemological, compatible with physicalism) from strong emergence (ontological novelty, requires new causal laws), directly relevant to the debate about whether organisational and market phenomena involve irreducible emergence. |
| b24 | Emergence — Internet Encyclopedia of Philosophy | Internet Encyclopedia of Philosophy | 2022 | Comprehensive survey of Chalmers's weak/strong distinction and O'Connor-Wong's epistemic/ontological split, explaining why most complexity-science phenomena are classified as weak emergence compatible with reduction-in-principle. |
| b25 | Safety-Organised Criticality — Psych Safety | psychsafety.com (practitioner blog) | 2025-04 | Translates Bak's SOC framework into organisational safety management, arguing that accumulated near-misses and small hazards drive organisations toward a critical state where minor procedural deviations can trigger major incidents. |
Financial Press
| ID | Title | Outlet | Date | Significance |
|---|---|---|---|---|
| f1 | Foundations of complexity economics | Nature Reviews Physics | 2021-01 | W. Brian Arthur's definitive peer-reviewed statement of complexity economics as an alternative to equilibrium models, providing the canonical academic framework through which financial and business journalists interpret markets as emergent, non-equilibrium systems. |
| f2 | Agent-Based Modeling in Economics and Finance: Past, Present, and Future | Journal of Economic Literature | 2025-03 | Landmark 90-page survey by Robert Axtell and J. Doyne Farmer covering the full arc of ABM in economics and finance, establishing clustered volatility, market impact and systemic risk as core empirical accomplishments of agent-based modelling in financial markets. |
| f3 | Reframing Financial Markets as Complex Systems | CFA Institute Research and Policy Center | 2025-10 | Practitioner-facing report from the CFA Institute explicitly framing financial markets as complex adaptive systems and equipping investment professionals with complexity science tools for systemic risk analysis, portfolio management and system-level investing. |
| f4 | Complex systems approaches to 21st century challenges: Introduction to the Special Issue | Journal of Economic Behavior & Organization | 2025-01 | Special issue editorial by Bednar, Farmer, del Rio-Chanona and others from the Oxford and Santa Fe complexity economics community, summarising a decade of methodological advances across inequality, climate and technology and marking the field's institutional consolidation. |
| f5 | High-Frequency Financial Market Simulation and Flash Crash Scenarios Analysis: An Agent-Based Modelling Approach | Journal of Artificial Societies and Social Simulation | 2024-03 | Demonstrates how agent-based models calibrated with machine learning surrogates can reproduce millisecond-level stylised facts of the E-Mini S&P 500 futures market, directly linking computational models of emergence to high-frequency trading regulation. |
| f6 | Calibrating emergent phenomena in stock markets with agent based models | PLOS ONE | 2018-03 | Fievet and Sornette show how agent-based models can identify emergent precursors to market crises, including the dotcom bubble and 2008 crash, providing early empirical evidence for ABM-based early-warning systems in financial markets. |
| f7 | Agent-based modeling and simulation for economic markets: a comprehensive review of applications, challenges, and opportunities | Journal of Simulation (Taylor & Francis) | 2026-02 | Comprehensive 2026 review of over 120 ABMS studies in economic and financial market analysis, establishing the current state of the art, methodological challenges and identifying emergent phenomena as the central explanatory target of the field. |
| f8 | Understanding flash crash contagion and systemic risk: A micro–macro agent-based approach | Journal of Economic Dynamics and Control | 2019-01 | Synthesises a microscopic limit order book with a macroscopic fund-asset network to show that flash crash contagion arises as emergent behaviour from portfolio diversification, algorithmic trading strategies and network topology — directly relevant to financial regulation. |
| f9 | Complexity Economics (SFI Press volume edited by Arthur, Beinhocker, Stanger) | Santa Fe Institute Press | 2020 | Edited volume capturing SFI's 2019 Applied Complexity Network Symposium, assembling both scholarly and practitioner perspectives on how complexity economics explains market dynamics, organisational scaling and collective intelligence. |
| f10 | W. Brian Arthur: Complexity Economics page | Santa Fe Institute | 2021 | Primary source on Arthur's core claim that the economy is not necessarily in equilibrium and that agents co-evolve beliefs and strategies, forming the intellectual foundation for treating markets as emergent complex adaptive systems. |
| f11 | Agent-based modeling at central banks: recent developments and new challenges | INET Oxford Working Papers / Banco de España | 2025 | Documents the uptake of ABMs by central banks — including the Bank of England and Banco de España — for housing markets, macroprudential policy and financial stability, marking the shift from academic research to regulatory practice. |
| f12 | Complexity and Paradigm Change in Economics (Beinhocker and Bednar working paper) | INET Oxford Working Papers | 2025 | Beinhocker and Bednar argue that complexity economics constitutes not merely a methodological supplement but an emerging paradigm shift with major implications for economic policy, drawing on thirty-five years of SFI research. |
| f13 | Complexity Economics programme — INET Oxford | Institute for New Economic Thinking, Oxford Martin School | 2025 | Primary institutional home of J. Doyne Farmer's group, publishing working papers on data-driven ABMs, supply-chain resilience, sovereign debt and macroeconomic forecasting — the leading European centre for applied complexity economics. |
| f14 | Why does economics need complexity? | Soft Computing (Springer) | 2025-04 | Reviews the epistemological case for complexity economics over DSGE models, arguing that ABM is the only methodology capable of reproducing unique far-from-equilibrium economic events, with direct implications for post-crisis macroeconomic policy. |
| f15 | Making Sense of Chaos: A Better Economics for a Better World — J. Doyne Farmer (Allen Lane, 2024) | Allen Lane / Oxford Martin School | 2024 | Farmer's book-length public argument for complexity economics, receiving coverage from the Wall Street Journal, Financial Times and Bloomberg, translating decades of Santa Fe Institute research into a business and policy audience. |
| f16 | Foundations of Complexity Economics — Santa Fe Institute news piece | Santa Fe Institute | 2021-03 | Institutional narrative tracing the origins of complexity economics from the 1987 Kenneth Arrow–Philip Anderson workshop to its current methodological maturity, providing authoritative intellectual history for the field. |
| f17 | Complexity economics offers new tools for today's global challenges | Phys.org / Journal of Economic Behavior & Organization | 2025-10 | News coverage of the 2025 JEBO special issue, synthesising six thematic advances — climate economics, inequality, technological change, production networks, macroeconomics and ABM — for a science-literate business readership. |
| f18 | Complexity Economics — Exploring Economics platform | Exploring Economics (pluralist economics platform) | 2023 | Accessible synthesis of complexity economics explaining why economic growth and inflation are emergent phenomena from heterogeneous agent interactions, and why equilibrium is a special case rather than the norm — valuable for contextualising the business press coverage. |
| f19 | Financial ripple effect in complex adaptive supply networks: an agent-based model | International Journal of Production Research (Taylor & Francis) | 2023 | Shows how financial contagion in supply chains emerges from company-to-company lending relationships and network topology, linking complexity science directly to supply-chain finance and enterprise risk management. |
| f20 | Theory of Organisational Emergence (TOE): addressing invisible conflicts and emerging dynamics in entrepreneurship | Review of Managerial Science (Springer) | 2025-06 | Introduces an interdisciplinary framework treating organisations explicitly as complex adaptive systems where slight perturbations cascade into systemic change, applying complexity concepts directly to entrepreneurial strategy and management practice. |
| f21 | Agent-Based Models for Financial Crises (ResearchGate) | Annual Review of Economics / ResearchGate | 2017 | Systematic account of how ABMs replace homogeneous representative-agent assumptions with heterogeneous bounded-rational agents to model non-ergodic dynamics that manifest during financial crises — a key conceptual bridge between complexity theory and financial regulation. |
| f22 | Complexity economics — Wikipedia overview | Wikipedia | 2025 | Useful reference for the Santa Fe Artificial Stock Market model as the archetypal demonstration that emergent bubbles and crashes arise from heterogeneous agent expectations — and for the documented critique that complexity results often confirm already-observed facts rather than generating novel predictions. |
| f23 | J. Doyne Farmer — INET Oxford profile | INET Oxford | 2025 | Primary profile page documenting Farmer's full research programme, including Macrocosm Inc (his complexity economics startup), publications on ABM macroeconomic forecasting and financial stability, and his work on the COVID-19 economic impact model. |
| f24 | W. Brian Arthur — Santa Fe Institute selected papers | Santa Fe Institute | 2023 | Complete bibliography of Arthur's foundational work from the 1995 El Farol paper through to the 2021 Nature Reviews Physics article, giving the authoritative record of the intellectual lineage connecting complexity science to market emergence. |