NOTE-189: Ontic Structural Realism and Economics

Contribution

The paper extends ontic structural realism from fundamental physics to economics. It argues that economic science, as distinct from economic engineering, has progressed cumulatively by learning about abstract optimisation structures, not about people. It also argues that OSR explains why the Kuhnian “anti-neoclassical” histories of economics misdescribe that progress. This gives OSR a non-physics, consilience-based motivation.

Key insight

The objects that economic theory quantifies over (consumers, agents, games) are not people. They are structural positions in mathematical models: a “consumer” is a triple of a constant, a commodity set and a budget, and an agent is “the gravitational centre of a consistent preference field”. People stand to economic theory as tables and rocks stand to physics. They are real, and the theory explains their behaviour, but they are not what the theory is fundamentally about.

Assumptions

• OSR is the right philosophy of physics (§1). It is not argued here but taken from Ladyman 1998 and 2002, French & Ladyman 2003 and Ladyman & Ross (then forthcoming).
• The unity of science: the consilience argument holds “on the assumption that the world is unified”. Dupré’s and Cartwright’s disunity is set aside, not refuted (§1).
• OSR needs mature formal theory: structures have definite content only via mathematical models (on a non-Platonist reading) or as mathematical structures (on a Platonist one) (§1).
• Science and engineering are separable: claims about the ontology of economic science should ignore economic engineering, which uses folk, individualist ontology (§§2–3).
• Earlier work of Ross’s is assumed for the account of persons: persons are not identical to H. sapiens organisms (Ross 2005; Ross 2006, Cognitive Systems Research).

Key results

The results are argued theses, not proofs.

• Two conceptions of economics (§2). On the “Debreuvian” conception, economics is any theory that “generalizes over maximizing, optimizing, or meliorating relationships among (i) utility functions, (ii) scarce production inputs, and (iii) reallocations of (ii) with reference to (i)”, with no reference to people. On the “nostalgic” conception (Sen, Rosenberg), economics generalises over the behaviour of specified classes of people.
• History of progress (§2): from Jevons onward the discipline moves steadily “in the direction of minimizing reliance on psychological hypotheses”. Jevons black-boxes marginal utility. Edgeworth and Fisher eliminate cardinal utility. Hicks & Allen (1934) show that convexity of demand does not need diminishing marginal utility. In Samuelson (1947) “the individual agent disappeared from the core of microeconomics”. Debreu (1959) formalises general equilibrium, Arrow (1965) extends it to risk, Stiglitz takes up information asymmetry, and Stigum (1990) axiomatises the whole in first-order logic and set theory.
• The consumer in theory (§2): “a triple consisting of a constant, a set of scarce states that constitute improvements on some metric (‘commodities’), and a set of achieved states that can be exchanged … (‘a budget’)”. What is maximised are “values on preference fields – that is, variables in simultaneous field equations”.
• Game-theoretic ontology (§3): “Games are mathematical structures, networks of relationships whose relata are distinguished as such by the mathematical representations of the structures in question – just like quarks and bosons. OSR thus predicts the right sort of ontology for economic science.” Solutions to non-cooperative games are strategy vectors, and outcome distributions are “structural properties par excellence”.
• Individuals as bookkeeping (§3): in physics, individual objects are “book-keeping devices”. People are “normatively regulated virtual constructs” and “stabilization devices in social-evolutionary dynamics”. Economic agents “may be as transient as a modeler likes”, so a person can be modelled as a succession of agents.
• Methodological individualism in economics is traced to pedagogy and to the Chicago School’s policy prominence, not to the theory’s mathematics (§2).

Claims

id	claim	strength	support
C1	OSR resolves the Kuhn vs referential-realism impasse by direct appeal to physics	assertion	§1, citing Ladyman 1998 and 2002 and French & Ladyman 2003
C2	Physics has abandoned the idea that its fundamental relata are kinds of entities	assertion	citations to French & Krause 2006, d’Espagnat 2006, L&R forthcoming, §1
C3	Neither eliminativism nor reductionism about the special sciences is viable, because the special sciences are not trending that way	assertion	§1, one sentence
C4	The progress of economic theory has consisted in deepening knowledge of abstract structures, not of people	informal argument	a history-of-thought sketch, §2. Ross calls this his “more modest claim”
C5	Nothing in postwar economic theory’s mathematics assumes that people maximise	informal argument	the reading of consumer theory and game theory, §2
C6	Games are structures whose relata are individuated only by the structure, so OSR is the right ontology for economic science	informal argument	§3
C7	Given the unity of science, OSR gains consilient support from economics	informal argument	§1, conditional on the unity assumption
C8	People are not identical to H. sapiens organisms. They are socially constructed stabilisation devices	assertion here	summary of Ross 2005 and 2006, §3
C9	The belief that economics is individualist comes from pedagogy and Chicago-School ideology	informal argument	the author’s experience as a policy consultant (“I have never estimated a personal utility function on the job”), §2

Concepts

• Economic science vs economic engineering (after Mankiw 2006): theory of abstract optimisation structures, versus application to real persons’ problems in folk ontology.
• Debreuvian vs nostalgic economics: see Key results.
• Agent (economic): the gravitational centre of a consistent preference field. No thesis about which entities implement it or for how long.
• Individuals as bookkeeping devices: OSR’s deflationary reading of objects in partial models.

Connections

It applies the OSR of r1 and Ladyman 1998 to economics. It is the special-science companion of ETMG (then forthcoming). It answers Mäki 1992 on Friedman’s “instrumentalism” and Sen 1977 and 1987 on “rational fools”. Its games-as-structures claim reappears in r2 (“Strategies and equilibria among them are aptly thought of as emergent real patterns”, p. 2). Ross’s later real-patterns work in economics is the MIT volume chapter by Alexseev, Harrison, Lau & Ross (2026), which was not read. It contrasts with LIT-196, as noted under corrections.

Bearing on the record

It is the record’s only reading by Don Ross and its only argument for OSR from outside physics. It fixes Ross’s reading of OSR’s anti-individualism: individuals are heuristics, and ordinary objects are not denied. For ML practice there is nothing to carry over. The formal-structures-not-persons reading of agents is a philosophical point about economic modelling, not a practice recommendation.

Limitations

• The history of economics is a two-page sketch. Ross himself disclaims “a simple fact of the matter about the true narrative”. The progress claim C4 is stated as conditional (“insofar as economics has made progress”).
• The consilience argument rests on an unargued unity-of-science assumption (§1).
• Claim C2, the premise from physics, is a citation, not an argument. Compare r1’s more cautious underdetermination framing.
• The claims about persons (C8) are summaries of other work.
• It is a preprint. The 2008 journal text may differ.

Open questions

• Does the argument extend to less formalised special sciences once they acquire canonical models, as Ross implies sociology might?
• If agents are individuated only by the structure, what distinguishes two games with isomorphic payoff structures but different implementing systems? This is the economics analogue of LIT-154’s structuralist identity question.
• Does behavioural economics’ reintroduction of psychology count against the structuralist history, as the Kuhnian reading claims, or is it engineering?

Corrections to the seeded skim

• There is no dossier or skim, so these are corrections to the record’s existing accounts.
• OSR’s scope is broader than the record’s accounts give it. Every OSR reading in the record (LIT-045, LIT-173, LIT-196, LIT-170, LIT-151) treats OSR as a thesis about fundamental physics. LIT-184 and r2 extend it through real patterns and scale relativity. This paper is the other extension, by the co-author of Every Thing Must Go. Its argument is a consilience argument: “individualism fails generally across the sciences (Kincaid 1997), OSR can be motivated by inquiries far removed from physics, and on the assumption that the world is unified OSR is thus supported by a consilience argument appealing to all of the sciences together” (§1). It limits the argument to sciences with “canonical formal theory”, such as economics and not (yet) sociology, because OSR’s structures “have definite content” only through mathematical models (§1).
• On eliminativism: individuals are demoted, not eliminated. LIT-196 and NOTE-149 treat OSR’s options as eliminative or priority-based non-eliminative. Ross states a third, deflationary line. “According to OSR, individual objects in physics are basically heuristics: book-keeping devices that help investigators manipulate partial models of reality”, and “OSR as applied to physics doesn’t deny that there are tables or rocks … (I.e., it denies that fundamental reality is a collection of objects like rocks, but smaller.)” (§3). On Ross’s reading, OSR is eliminativist about individuals as the constituents of fundamental reality. It is not eliminativist about objects as such, which are real at their own scales. That is the rainforest-realist reading r2 makes explicit. It is neither of the positions LIT-196’s dilemma is aimed at.
• The quantum premise is stated as an empirical fact, stronger than r1. Ross says “physics itself has abandoned the idea that its fundamental relata (quarks, bosons, etc.) are kinds of entities at all (French and Krause 2006; D’Espagnat 2006; Ladyman and Ross forthcoming)” (§1). He also says “the advocate of OSR makes the case that the world is not fundamentally ‘made of individuals’ strictly by appeal to physics”, with entanglement as the crucial premise. Compare r1 (Ladyman 2001), where the same physics yields only an underdetermination between individuals and non-individuals. Ross’s gloss is an assertion with citations, not an argument. LIT-173’s objection that particle kinds have irreducible intrinsic features applies to this formulation directly.
• It does not mention real patterns or rainforest realism by name. The special-science ontology is argued from the formalism of economics, not from Dennett’s criterion. The link to rainforest realism is only the tables-and-rocks passage and the claim that people are “stabilization devices”. So this paper does not state the rainforest-realism thesis. Use r2 for that.
• LIT-151 (Wallace), LIT-170 (French) and LIT-152 (Floridi): no bearing beyond the above. Ross’s point that OSR’s content comes from mathematical models (§1) is a prose form of LIT-151’s math-first thesis.