Emergentism is a family of positions in philosophy concerning how organized wholes depend on their components while exhibiting properties, patterns, or causal powers that the components do not possess individually. Its central question is not simply whether wholes differ from their parts, but whether—and in what sense—the properties of wholes resist explanation, prediction, or reduction in terms of those parts and their interactions. Emergentism has been especially influential in debates about life, consciousness, and the relationship between the special sciences and fundamental physics.
The term covers importantly different commitments. Some emergentists maintain that higher-level phenomena are fully generated by lower-level processes but require autonomous models or can be predicted only by following those processes through. Others hold that genuinely new properties or causal powers appear at higher levels and cannot be derived even from complete lower-level knowledge. These positions are commonly distinguished as weak and strong emergence, although the terminology is not uniform.
Modern emergentism developed from nineteenth-century discussions of chemistry, biology, and mental activity, became a recognizable philosophical movement in Britain during the early twentieth century, and subsequently intersected with complexity science and contemporary philosophy of mind. Its continuing significance lies in the distinction between explaining how a system is constituted and explaining why its organized behavior has the particular character it does.
Dependence, novelty, and organization
Emergentist positions usually combine two claims: dependence on a lower-level basis and novelty relative to that basis. A liquid depends on molecular constituents and their interactions; an organism depends on organized physical and chemical processes; a conscious episode depends, on many accounts, on activity in a nervous system. Yet liquidity, biological regulation, and conscious experience are not properties ordinarily attributed to isolated molecules or individual nerve cells. The philosophical issue is what follows from this difference. Mere absence of a property in isolated components is insufficient to establish a substantial emergentist thesis: the mass of a composite object, for example, also differs from the mass of any one component.
A useful distinction separates novelty of properties from novelty of explanatory principles. A system can instantiate a property that none of its parts instantiates while remaining straightforwardly explainable through their arrangement. A triangle has geometrical features that no single line segment has, but this alone does not demonstrate irreducibility. Stronger emergentist claims concern whether a complete account of components, relations, and governing laws would suffice to derive the relevant system-level features. Still stronger claims concern whether the system has causal powers that cannot be identified with, or exhaustively accounted for by, lower-level powers.
Organization is therefore central. The relevant basis generally includes not just a list of constituents but their spatial arrangement, interactions, history, and environmental conditions. The same chemical constituents can participate in very different structures, while similar large-scale patterns can arise in systems made from different materials. An emergent property is consequently often attributed to an organized process rather than to a substance considered independently of its relations. This feature connects emergentism with ontology: disagreement concerns what kinds of entities and properties exist, not merely the vocabulary used to describe them.
Dependence is frequently expressed through supervenience. In a standard formulation, there cannot be a difference in higher-level properties without some difference in the relevant lower-level basis. Supervenience does not, by itself, establish reduction or emergence. Both a reductive identity theory and an emergentist theory can accept such covariance while disagreeing about explanation and causation. Nor does supervenience automatically specify whether the dependence holds necessarily, only under actual natural laws, or within a restricted class of systems.
Emergence must also be distinguished from the appearance of something new over time. A property may be emergent in a structural sense whenever the appropriate organization exists, even if that organization has existed for millions of years. Conversely, a historically unprecedented event need not involve an emergent property. Evolutionary novelty, developmental transitions, and the formation of new institutions raise questions about emergence, but temporal novelty alone does not settle them.
Historical formation of emergentism
Ancient discussions of wholes, form, and organization provide antecedents to emergentist thinking, but they should not be identified without qualification with modern emergentism. Aristotle treated living things as organized unities and explained them through form as well as material constituents. His hylomorphism, however, belonged to a different framework from the modern question of whether higher-level properties follow from lower-level scientific laws. Modern emergentism took shape against the development of mechanics, chemical theory, and evolutionary biology.
In A System of Logic (1843), John Stuart Mill distinguished cases in which causes combine according to a principle of composition from cases governed by what he called heteropathic laws. The combined effects of mechanical forces supplied a model of composition; chemical reactions appeared to produce effects qualitatively unlike those of the reactants considered separately. Mill's distinction helped establish the idea that knowledge of constituents might not suffice to deduce the characteristics of a compound without additional laws. His account concerned the structure of scientific explanation rather than a fully developed theory of emergent mental causation.
George Henry Lewes introduced the influential contrast between “resultants” and “emergents” in Problems of Life and Mind, published in multiple volumes beginning in 1874. Resultants could be understood through combinations of known factors, whereas emergents involved qualitatively different outcomes. During the early twentieth century, Samuel Alexander's Space, Time, and Deity (1920), C. Lloyd Morgan's Emergent Evolution (1923), and C. D. Broad's The Mind and Its Place in Nature (1925) developed distinct versions of a layered natural world. Matter, life, and mind were treated as connected stages or levels rather than wholly independent domains.
Broad provided a particularly clear statement of the issue. In his discussion of chemical compounds, an idealized investigator might know the constituents and their behavior in other combinations yet still be unable to deduce the properties of a new compound without additional information. The important claim was not that ordinary scientists lacked adequate calculating power, but that certain laws connecting composition and higher-level characteristics might be fundamental rather than derivable. The imagined possibility of such laws gave classical emergentism a more substantial content than the assertion that complex systems are difficult to understand.
The scientific examples supporting early emergentism changed as quantum mechanics and molecular biology expanded explanatory resources. Chemical bonding and many biological processes became intelligible through mechanisms unavailable to nineteenth-century investigators. These developments weakened particular arguments from unexplained chemistry or life without logically refuting every form of emergentism. From the later twentieth century onward, philosophical discussion increasingly separated ontological emergence from computational and explanatory autonomy. The resulting literature no longer constituted a single movement: debates about mind, collective behavior, and scientific reduction developed related but nonidentical concepts of emergence.
Weak, strong, epistemic, and ontological emergence
The distinction between weak and strong emergence is widely used, but its meaning depends on the author. In broad usage, weak emergence describes higher-level patterns that arise entirely through lower-level interactions, without requiring additional fundamental powers or laws. Strong emergence describes properties that are not fully derivable from the lower-level basis and may possess irreducible causal powers. This distinction is useful only when the relevant meaning of derivability and the proposed lower-level basis are made explicit.
Mark Bedau developed an influential account of weak emergence in which a macrostate can be derived from a system's microdynamics and external conditions, but only by simulation. The point is not that the macrostate lacks a lower-level explanation. Rather, obtaining it requires following the evolution of the underlying process instead of using an independently available shortcut. A cellular automaton provides a natural illustration: simple local update rules can produce persistent moving configurations whose behavior is much easier to describe at the pattern level than by enumerating individual cell updates.
John Conway's Game of Life, introduced in 1970, illustrates both the attraction and the limits of this approach. Its grid cells change state according to fixed rules, yet configurations such as gliders maintain recognizable patterns while moving across the grid. A glider is a real pattern in successive configurations, not an extra ingredient added to the cells. Some configurations have readily predictable behavior, whereas arbitrary configurations raise much harder computational questions. It would therefore be misleading to claim that every pattern in the game is unpredictable or that simulation difficulty demonstrates the existence of new causal forces.
A related distinction separates epistemic from ontological emergence. Epistemic emergence concerns limitations in knowledge, prediction, conceptual resources, or available explanations. Ontological emergence concerns the constitution of reality itself: whether there are genuinely irreducible properties, entities, laws, or powers. These distinctions do not coincide perfectly with weak and strong emergence. An author may call a property ontologically real while denying that it introduces any fundamental causal power; another may reserve ontological emergence for a much stronger break with reduction.
Strong emergence requires special care about the word “possible.” Failure of deduction might reflect limited information, finite computational resources, the absence of suitable concepts, or an in-principle insufficiency of lower-level laws. Only the last directly supports the strongest metaphysical conclusion. A phenomenon can also be surprising without being emergent: unexpected behavior may disappear once a missing interaction or boundary condition is supplied. Conversely, an emergent property need not remain surprising after its regularities become familiar. Successful classification therefore depends on specifying the contrast between what the lower-level account provides and what the higher-level account contributes.
Reduction, realization, and autonomous explanation
Emergentism is often presented as the opposite of reductionism, but reduction itself has several meanings. Ontological reduction identifies apparently distinct entities or properties with more basic ones. Explanatory reduction shows how a phenomenon follows from underlying mechanisms. Intertheoretic reduction concerns relations between theories, including whether the laws of one can be derived from another with suitable connecting assumptions. Methodological reduction is a research strategy that investigates components to understand a system. A position can reject one kind of reduction while accepting others.
An organism, for example, can be investigated through molecules, cells, tissues, and physiological systems without assuming that every biological explanation must ultimately be replaced by particle physics. Molecular mechanisms may explain how a process occurs, while evolutionary history explains why that mechanism became established in a lineage. These explanations answer different questions. Their coexistence establishes explanatory pluralism more readily than it establishes strong emergence. The existence of several useful explanatory levels does not by itself show that nature contains additional fundamental laws at each level.
Multiple realizability supplies an important argument for higher-level autonomy. A functional property may be instantiated by physically different systems. The same computational operation can be implemented in different hardware, and similar biological functions can be achieved through different anatomical arrangements. In functionalism, mental states are characterized partly through their causal roles rather than a single material composition. If a useful category groups together heterogeneous realizations, its explanatory generalizations may not map neatly onto one lower-level physical kind.
However, multiple realizability does not entail strong emergence. A higher-level property could be realized by several physical configurations while each instance remains fully explicable through its particular configuration. The question is then whether a general explanation at the higher level captures something that would be obscured by enumerating all implementations. Emergentist and nonreductive positions often emphasize this autonomy, whereas reductive approaches may permit domain-specific reductions or argue that the alleged diversity is smaller than it appears.
A formal way to express the relationship uses a microscopic state , a macrostate , and a coarse-graining map , such that . If microscopic dynamics are represented by , an autonomous macroscopic evolution would ideally satisfy
In actual scientific models, this relation often holds only approximately, statistically, or under restricted conditions. Its usefulness depends on timescale separation, stability, and the variables retained by the coarse-graining. Such autonomy can explain why higher-level laws remain informative even when no microscopic laws are violated. It does not automatically establish irreducible metaphysical novelty.
Downward causation and the causal exclusion problem
Downward causation is the idea that higher-level organization affects processes at lower levels. The expression became prominent in twentieth-century discussions of biology and systems theory, including work by Donald T. Campbell. Examples include cellular organization regulating molecular activity, an organism's behavior changing conditions within its tissues, and institutional rules influencing individual action. The philosophical difficulty lies in identifying what kind of influence is involved and whether it exceeds ordinary interactions among the system's constituents.
One interpretation treats downward causation as constraint. A membrane confines molecules; a tissue's geometry channels mechanical forces; an organism maintains conditions under which biochemical reactions occur. These are substantial explanatory facts, but they can often be described through spatial arrangements, forces, and interactions already recognized at the physical level. The higher-level account identifies a causally important organization rather than an additional force. On this interpretation, saying that the cell regulates a molecular process need not compete with a molecular account of how regulation happens.
A stronger interpretation attributes novel causal powers to the whole. This raises the causal exclusion problem, associated especially with Jaegwon Kim's work on mental causation. Suppose a higher-level property is realized by a physical state , and a later physical event already has a sufficient physical cause. If is distinct from that physical cause and independently sufficient for , the event appears systematically overdetermined. If it contributes nothing beyond the physical cause, the higher-level property risks becoming causally idle.
The argument draws on several assumptions: the causal closure of the physical domain, a distinction between higher-level and physical properties, and resistance to pervasive causal overdetermination. Emergentists respond in different ways. Some reject closure and propose additional fundamental powers. Others deny that a whole and its realizing parts should be counted as independent competing causes. Still others understand causation through stable difference-making relationships: a higher-level variable may identify a cause accurately even when its effects are physically implemented throughout.
The dispute cannot be resolved simply by observing feedback. A thermostat's output changes the temperature subsequently measured by its sensor, but this circular organization is compatible with an entirely ordinary physical mechanism. Nor does a higher-level variable's predictive usefulness automatically establish a new causal power. The central issues are whether causal explanations at different levels compete, what makes two descriptions descriptions of the same causal process, and whether causal completeness at one level excludes legitimate causal attribution at another.
Collective phenomena in physics and complexity science
Physical science supplies some of the clearest examples of organized collective behavior, although their interpretation as strong emergence is contested. Temperature, pressure, fluid flow, and magnetic order characterize large collections through variables that are not simply properties of isolated constituents. Statistical mechanics explains how such variables relate to microscopic states and their distributions. In a classical monatomic ideal gas at equilibrium, for example, the mean translational kinetic energy per particle is
where is the Boltzmann constant and is absolute temperature. This relation has a specified domain; temperature is not universally identical to mean translational kinetic energy.
Phase transitions are especially important because qualitative changes can accompany gradual variation in control parameters. Ferromagnetic ordering, for instance, involves collective alignment below a critical temperature. In idealized infinite systems, some thermodynamic quantities become nonanalytic at a transition. Finite systems do not exhibit exactly the same mathematical singularities, but sufficiently large systems can closely approximate the associated behavior. The distinction between finite reality and infinite-system idealization is central to philosophical arguments about whether phase transitions can be reduced to microscopic mechanics.
Renormalization-group methods explain why systems with different microscopic details can exhibit the same critical behavior. Repeated changes of descriptive scale identify features that remain important and features whose influence diminishes. Systems belonging to the same universality class can share critical exponents despite differences in their materials. Philip W. Anderson's 1972 essay “More Is Different” emphasized that knowing fundamental laws does not eliminate the need to discover new principles governing organized many-body systems. This supports the autonomy of higher-level physics without, by itself, settling whether fundamental causal powers emerge.
Complexity science also studies pattern formation far from equilibrium. Convection cells arise in suitably heated fluids; oscillating chemical reactions produce organized temporal behavior; interacting oscillators can synchronize. Such examples often involve nonlinear feedback and continuing exchanges of energy or matter with the environment. Their macroscopic organization depends on precise physical conditions, not on an unexplained tendency of complexity to generate order. Equilibrium collective phenomena and nonequilibrium self-organization therefore belong within a broad discussion of emergence but should not be treated as one mechanism.
Chaos theory adds another distinction. Deterministic systems can display sensitive dependence on initial conditions, making long-term prediction practically unreliable even when their governing equations are known. This is not automatically strong emergence, and emergence does not require chaos. Stable macroscopic order can arise in systems whose microscopic trajectories are complicated, while chaotic behavior can occur in comparatively low-dimensional models. Prediction failure, collective order, and irreducibility are separate claims requiring separate arguments.
Organisms, ecosystems, and social systems
Biological emergentism concerns the relation between organized living activity and its physical and chemical basis. Historically, the appearance of life as a distinct level encouraged claims that living systems possess properties unavailable in nonliving matter. Emergentism nevertheless differs from vitalism, especially where vitalism postulates a special life force. Many emergentist approaches accept that organisms consist entirely of ordinary physical materials while emphasizing the organization that enables self-maintenance, reproduction, and adaptive activity. Whether those capacities require irreducible laws remains a further question.
A living cell illustrates how interdependence complicates simple component-based explanation. Its membrane helps preserve chemical gradients; metabolic reactions supply usable energy and molecular precursors; genetic and regulatory processes contribute to maintaining the machinery that carries out those reactions. Feedback in gene expression can generate switches, oscillations, and differentiated cell states. Such behaviors are investigated through explicit molecular networks and experimentally testable models. Calling them emergent is informative when it identifies how network organization produces the behavior, rather than substituting a label for an undiscovered mechanism.
Development provides a particularly concrete example. Alan Turing's 1952 paper “The Chemical Basis of Morphogenesis” showed how interacting and diffusing substances could, under appropriate conditions, destabilize an initially uniform state and generate spatial patterns. The proposal demonstrated a mechanism by which local interactions can produce large-scale organization without a central patterning agent. Reaction–diffusion mechanisms are relevant to some biological patterning, but development also involves gene regulation, tissue mechanics, cell movement, and environmental interaction. No single emergence model accounts for all organismal form.
In an ecosystem, population cycles, food-web structure, and material flows arise through interactions among organisms and their surroundings. Higher-level regularities may be useful even when detailed individual histories remain inaccessible. Nevertheless, ecosystem boundaries are often open and scale-dependent, so an alleged emergent property must be specified carefully: stability can mean resistance to disturbance, speed of recovery, or persistence of particular populations, and these measures need not agree. General claims that greater complexity necessarily produces greater stability are therefore inadequate.
Social emergence introduces agents who interpret situations, follow rules, and sometimes respond to theories about their own behavior. Thomas Schelling's neighborhood models showed how individual location preferences can generate large-scale segregation without every individual preferring the resulting aggregate pattern. The result depends on the model's assumptions and does not establish a complete explanation of actual residential segregation. More generally, markets, organizations, and institutions depend on individual actions while also shaping subsequent choices. Philosophical disputes concern whether such reciprocal explanations require irreducible social properties or can be reconstructed through individuals, their relations, and historically established rules.
Mind, consciousness, and the limits of emergent explanation
The mind–body problem has remained a central motivation for emergentism. Mental activity appears systematically dependent on the brain, yet conscious experience seems unlike the electrical and chemical processes used to explain neural function. Emergentist theories propose that sufficiently organized physical activity gives rise to mental properties without requiring an independently existing mental substance. They therefore differ from substance versions of mind–body dualism, although strong versions can count as forms of property dualism.
The relation to physicalism depends on what physicalism requires. If it requires only that everything be physically constituted and that mental events have physical realizations, some emergentist theories may qualify. If it requires every property and causal power to be reducible to fundamental physical properties and powers, strong emergentism is excluded. The terminology can obscure this substantive difference: agreement that brains are necessary for ordinary human consciousness does not settle whether conscious properties are reducible, causally efficacious, or derivable from physical theory.
John Searle describes consciousness within biological naturalism as a higher-level feature caused by lower-level neuronal processes. His use of emergence is intended to preserve consciousness as a real biological phenomenon without introducing a separate mental substance. Other philosophers distinguish ordinary emergence, such as the appearance of collective physical properties, from the stronger emergence that might be required for subjective experience. The question is whether explaining information processing, discrimination, and behavioral control also explains why any of those processes has an experienced character.
This issue is expressed through the hard problem of consciousness and debates about qualia. A strong emergentist may posit basic laws connecting certain physical organizations with conscious qualities. Such laws would not follow from existing physical descriptions but would supplement them. The proposal has a clear logical structure, yet an explanatory burden remains: it must specify the relevant physical conditions, the resulting conscious properties, and the principles connecting them. Merely stating that consciousness emerges from complexity leaves these matters undecided.
Research on artificial intelligence creates a related but distinct setting. A system may display an ability that becomes noticeable only after changes in scale, training, or evaluation. That observation can support a claim about system-level performance without supporting a claim about consciousness or strong emergence. Apparent abruptness may also depend on the scoring metric: a thresholded success measure can change sharply while underlying performance changes gradually.
Across these debates, a substantive emergentist explanation must identify its lower-level basis, the conditions of organization, the higher-level property, and the precise respect in which ordinary derivation or reduction fails. Empirical discovery can establish dependencies, mechanisms, and robust regularities. The further claim that these regularities require fundamentally new properties or powers remains a distinct philosophical thesis, assessed through its explanatory commitments and its compatibility with the relevant sciences.