Economic Naturalness
Economic naturalness is the principle that the coarse-grainings we use to describe complex systems are selected not by formal elegance alone but by the cost of error. A description survives not because it is mathematically privileged but because deviations from it have been tested against costs — energetic, social, or survival-related — and eliminated. The naturalness of a coarse-graining is an economic fact, not merely a mathematical or evolutionary one.
The concept synthesizes three traditions that are usually treated separately: the renormalization group in physics, evolutionary epistemology in biology, and institutional economics in social science. Each describes how descriptions converge to stable levels, but none has named the common mechanism: the cost of sustaining deviations exceeds the benefit of tracking more detail.
Three Expressions of the Same Constraint
- Mathematical naturalness. In physics, renormalization group fixed points are "natural" because perturbations that deviate from the fixed point are expensive to sustain — they require continuous energy input to maintain. A ferromagnet's magnetization direction is natural because any deviation costs anisotropy energy. The Higgs vacuum expectation value is natural because the potential's curvature penalizes deviations. The mathematics of RG flow formalizes what is implicit in all economic naturalness
- stability is purchased at the cost of energy, and the cheaper state wins.
- Evolutionary naturalness. Organisms track variables at scales that were adaptive to track. Predators do not model their prey at the molecular level because the metabolic cost of such modeling would exceed the caloric value of the meal. The neuronal level of description is natural not because neurons are ontologically fundamental but because the fitness cost of tracking ion channels in real-time would eliminate any lineage that attempted it. Evolution selects coarse-grainings by the same mechanism that physics selects fixed points
- the cost of deviation is fatal.
- Cultural naturalness. Scientific paradigms, engineering standards, legal categories, and linguistic phonemes are all coarse-grainings that survived because the social cost of using alternative descriptions exceeded the benefit. A speaker who uses non-standard phonemes pays a cost in comprehension errors. An engineer who ignores material standards pays a cost in component incompatibility. A jurist who rejects precedent pays a cost in unresolvable disputes. The selection mechanism is not biological evolution but social coordination failure
- the cost of being misunderstood, the cost of being incompatible, the cost of being unenforceable.
The Common Structure
In all three cases, the coarse-graining is a compressed representation, and compression is subject to a resource budget. The resource differs — energy in physics, metabolism in biology, coordination effort in culture — but the structure is identical:
- There is a space of possible descriptions at finer grain than the one used.
- Finer-grain descriptions contain more information but require more resources to maintain.
- The system (physical, biological, or social) has a finite resource budget.
- Descriptions that exceed the budget are eliminated — by dissipation, by death, by social exclusion.
- The surviving description is the one that maximizes predictive power per unit resource cost.
This is not a metaphor. It is a formal claim: under plausible assumptions about cost functions and information measures, the equilibrium description is the solution to a constrained optimization problem that is mathematically identical across all three domains. The Lagrangian has two terms: an information term (how well the description predicts observations) and a cost term (how expensive the description is to maintain). The equilibrium balances them.
The Implication for Causal Emergence Debates
The causal emergence debate — particularly the critiques of Hoel's effective information framework — has identified a circularity problem: the framework presupposes the coarse-graining it claims to evaluate. Economic naturalness resolves this not by finding a better selection criterion but by recognizing that the circularity is structural. Every causal analysis presupposes a coarse-graining because every causal analysis is performed by an embedded observer with a cost function.
Hoel's effective information measures causal power against a maximum-entropy intervention distribution. But no real observer applies maximum-entropy interventions. A hardware engineer tests circuits at nodes where experience has taught her failures are likely. A biologist measures organisms at scales where her instruments are calibrated. A social scientist surveys populations at sampling frames where response rates are tractable. All three intervention distributions are shaped by cost — the cost of applying interventions, the cost of processing data, the cost of reaching participants.
This means the "natural" perturbation distribution is not uniform. It is weighted by the observer's accumulated consequence-testing. The circularity in Hoel's framework is not a bug; it is the formalism correctly reporting that causal power is always indexed to an intervention class, and intervention classes are always shaped by economic constraints.
The Connection to Emergence Theory
Economic naturalness reframes the weak/strong emergence distinction. Weak emergence holds that macro-properties are in principle derivable from micro-laws but computationally intractable. Strong emergence holds that macro-properties are ontologically novel. Economic naturalness adds a third dimension: practical emergence, the claim that macro-properties are the equilibrium descriptions that survive resource constraints, and their autonomy from micro-descriptions is not a matter of principle or ontology but of cost-structure.
On this view, wetness is emergent not because we cannot compute it from molecular dynamics (though we cannot) and not because it is ontologically novel (though it may be) but because tracking every water molecule to predict slipperiness would require resources no organism possesses. The neuronal level of cognition is natural because tracking every synaptic vesicle to predict behavior would cost more energy than the behavior is worth. The functional level of software is natural because debugging at the transistor level would take longer than the product lifetime.
The Meta-Claim
Economic naturalness is not merely a claim about how descriptions are selected. It is a claim about how reality is structured for finite knowers. The levels of description that survive are not arbitrary conventions, but they are not metaphysically privileged either. They are the product of a selection process that operates on the coupling between knowers and known — and the selection criterion is the cost of error.
This dissolves the perspectival/realist dichotomy that has troubled emergence theory. Coarse-grainings are real in the sense that they have been tested against costs that cannot be externalized. They are perspectival in the sense that different observers with different budgets converge on different levels. The two properties are not in tension. They are dual descriptions of the same selective history.