NOTE-695: Esfeld & Lam — Moderate structural realism about space-time

Contribution

The paper names and defends a third structural realism. It is ontic, unlike Worrall’s, and it keeps objects, unlike French and Ladyman’s (§1). It works the view out for classical general relativity in the tensor formulation (§2) and the fibre-bundle formulation (§3). It places the view in the substantivalism–relationalism debate (§4). It says what survives if points go (§5).

What it adds is a reply to the question “if there are objects, don’t they require intrinsic properties as identity condition?” (p. 32; fn. 7 credits a referee and Wüthrich for pressing it). The reply has two parts. Concrete relations can do whatever identity work intrinsic properties could. And where nothing qualitative distinguishes the objects, a primitive numerical distinction is enough.

Key insight

Read the “master argument” for intrinsic properties (p. 29) as two premises: (1) relations require relata; (2) “these objects have to be something in themselves”, with intrinsic properties beyond their relations. Epistemic structural realism accepts both. Radical ontic structural realism rejects both. Moderate structural realism accepts (1) and rejects (2) (p. 31). “An object as such is nothing but that what bears the relations” (p. 32). The remainder, what makes two indistinguishable bearers two, is supplied by primitive numerical diversity. This is “not a primitive thisness, for it does not establish an identity in time—or any other sort of an identity” (p. 33). The paper says this is what the mutual dependence of objects and relations rests on: “we get the relata and the relations at once as the internal structure of a whole” (p. 34).

Assumptions

• What “intrinsic” means. Langton and Lewis: “having or lacking an intrinsic property is independent of accompaniment or loneliness” (p. 28).
• “Properties” exclude identity properties by stipulation. “Properties such as being that individual (e.g. being John Smith …) will hence not be considered in this paper” (p. 28). So the denial of intrinsic properties does not, by itself, deny haecceities. Haecceity is rejected separately, by the hole argument for points and by the “spirit of structural realism” (p. 33).
• The scope is fundamental intrinsic properties of basic objects. “It goes without saying that there are intrinsic properties of complex systems as a whole” (p. 28). The thesis concerns properties that “cannot be traced back to more fundamental properties” and that would ground relations.
• The argument that knowledge is relational is assumed, not argued. Jackson’s (1998) argument that science reaches only relations is taken “without further discussion” (p. 29).
• Scientific realism is presupposed. Structural realism is not used to defend it. The paper argues that Worrall’s first two theses are in tension, so structure is not immune to pessimistic induction (p. 29).
• Structure means concrete, determinate, first-order relations, particulars rather than universals. “These relations being as determinate as intrinsic properties are supposed to be” (p. 32). That is how it claims to escape the Newman objection (fn. 8, citing Chakravartty 2004 §3).
• The physics is classical GR, mostly vacuum. The paper limits itself mainly to “empty” solutions without non-gravitational fields. It asserts that this “does not alter the argumentation” (pp. 34–35), and §4 shows that matter does change point labelling (p. 43). It treats the metric only, not the connection (p. 35). It sets aside the algebraic formulation (p. 34) and quantum gravity (fn. 9).
• Leibniz equivalence is taken from the physics. Diffeomorphic models represent the same physical situation, “in agreement with the common practice among working relativists” (p. 36), after Earman & Norton and Stachel.

Key results

This is a philosophy paper with no theorem. Its results are arguments, of uneven strength.

§1: the general position (pp. 27–34)

• Against Worrall. If science describes only relations, then no line separates structure from content within a theory. Theory change is then change of structure, and the pessimistic induction applies to structure too (p. 29).
• Three arguments for ontic structural realism (pp. 30–31): from coherence of metaphysics with epistemology, from parsimony, and from quantum physics. The authors accept the first two. They reject the quantum argument as French and Ladyman give it, on two grounds. Primitive thisness is “a purely metaphysical position that one can take anyway”, so there is no empirical underdetermination (p. 30). And quantum field theory favours fields over particles, so the particle/field choice is not underdetermined either (Cao 2003a; p. 31).
• Against radical OSR: relations presuppose relata. Concrete relations are instantiated, and need something that instantiates them. Without relata, the position “is not able to distinguish between structure as an abstract entity and the structure of the physical world” (p. 31). Fn. 6 cites Cao, Chakravartty (2003), Busch and Psillos for this objection.
• Mutual dependence. Objects “can neither exist nor be conceived without relations in which they stand”. Relations “can neither exist in the physical world nor be conceived as the structure of the physical world without objects” (p. 32).
• Relations as identity conditions. “Insofar as intrinsic properties account for identity conditions, relations can perform that task as well. For instance, if A is bigger than B, heavier than C, etc., these relations individuate A” (p. 32). The analogy is the web of belief: beliefs individuated by inferential role (Quine 1951; pp. 32–33).
• The cases where neither suffices. “Quantum systems of the same kind whose states are entangled are indistinguishable … These systems do not have an identity in time.” And “space-time points can stand in exactly the same spatio-temporal relations and, yet, be numerically distinct”, “on certain symmetry assumptions about space-time” (p. 33). These cases tell against bundle theories of both kinds, built of intrinsic properties or of relations.
• Primitive numerical distinction. Haecceity is rejected (the hole argument; it gives objects priority). “One can simply accept a numerical distinction (diversity)—among quantum systems or space-time points—as primitive (a similar view is held by Pooley, 2005, Sect. 4)” (p. 33). It tells us “that there is a number of objects that is greater than one … and that is all that it tells us.” The motivating situation “is independent of structural realism. Any position in the metaphysics of science has to come to terms with” it.
• Why no priority either way (p. 34). Objects cannot be dispensed with, “on pain of running into absurdity”. Relations cannot be prior, because bundles fail to distinguish entangled systems and symmetric points. Objects cannot be prior, because that “would commit us to primitive thisness”.
• Quantum physics. The formalism “always refers to cases in which there is a definite number of quantum systems taken for granted (although these systems are not individuals)”. “Unless one goes for hidden variables, there are no intrinsic properties underlying the entanglement” (p. 34). It refers to Esfeld 2004 for the full case.

§2: GR in the tensor formulation (pp. 34–39)

• Epistemic SR implies intrinsic individuation of points, independent of the metric (p. 35). That is what the hole argument punishes. Take a diffeomorphism that is the identity outside a hole. If points are individuated independently of the metric, the two models are distinct solutions, and Cauchy data fail to fix the solution in the hole. Two space-times “may agree till a time t and then disagree for any time t′ > t” (pp. 35–36).
• Radical SR faces intelligibility. “What does a space-time relation (of the space-time structure) relate? How can the space-time structure be represented by the metric alone … given that the metric tensor field itself is defined as an assignment of a metric tensor at each manifold point?” (p. 36).
• The bare manifold point is ambiguous. It has no physical meaning, by active diffeomorphism invariance, yet it is “conceptually indispensable” (pp. 36–37). “There is no bijective correspondence between physical space-time points and mathematical bare manifold points; indeed, an equivalence class of diffeomorphic manifold points corresponds to a space-time point” (p. 36).
• The moderate reading. The relata are “space-time points (not bare manifold points!)” (p. 37). “The metric tensor field completely determines the structural identity of the space-time points” (p. 37). Coordinatization is a “purely mathematical” individuation with “no physical meaning” (pp. 37–38). If p ∈ M were the genuine representative of a space-time point, g′(p) ≠ g(p) for a drag-along g′ would break Leibniz equivalence (p. 38).
• A concrete construction: Bergmann–Komar, as developed by Lusanna and Pauri (2005). Four functions of the four independent eigenvalues of the Weyl tensor serve as a non-coordinate “pseudo-coordinatization” of points (pp. 38–39). Imposed as a gauge fixing in the constrained Hamiltonian framework, this labelling “constitutes the structural identity of space-time points” (p. 39). The authors concede three limits. It works “for space-times with no symmetries”. It needs “explicit breaking of active general covariance”. It holds for “a rather specific class of solutions” (pp. 39). They add that moderate SR “is however not tied to this gauge-fixing approach”.

§3: GR in the fibre-bundle formulation (pp. 40–42)

• The bundle. The orthonormal frame bundle over the manifold. The metric is derived from the frame field, and the Einstein equations can be written in it (p. 40).
• Sections define the points. Stachel: “the map σ [a section] defines the points of space-time … space-time points are not defined before a cross-section is given” (Stachel 1986, quoted p. 40).
• The three claims (p. 41):
  1. base-to-fibre mappings encode the structural identity of points;
  2. base and fibres are “soldered”, so the bare manifold has no independent weight;
  3. active general covariance is invariance under fibre-preserving automorphisms of the bundle. These determine base diffeomorphisms uniquely, and a base diffeomorphism fixes them up to vertical ones. On this formulation it is “impossible to even formulate the hole argument” (Stachel 2002, quoted p. 41).
• Against both rivals at once (p. 42). Nothing in the formalism gives points identity independent of the fibres, which tells against epistemic SR. Points cannot be eliminated “on pain of running into mathematical, physical and conceptual difficulties”, which tells against radical SR. That the base space can be defined as a quotient does not eliminate points (Stachel & Iftime 2005).

§4: substantivalism and relationalism (pp. 42–43)

• Three conceptions (p. 42): (1) Newtonian substantivalism, with space-time and matter distinct; (2) Leibnizian relationalism, with space-time reduced to non-spatiotemporal relations or properties of matter; (3) Cartesian–Spinozan substantivalism, with space-time and matter identical.
• The position fits (1) or (3), not (2). Its “purely relational” claim is non-reductive. It is a form of “sophisticated substantivalism” (p. 43). It leaves open whether matter is distinct from space-time.
• Matter contributes to point identity. In Lusanna and Pauri’s framework, matter fields modify the labelling of points and so “also provide structural (but non-spatio-temporal) identity to space-time points” (p. 43). The text cites this as Lusanna and Pauri (2003). The reference list has only a 2005 paper.

§5: conclusion (pp. 43–45)

• The paper’s own summary. Coherence, parsimony and quantum theory support the position, which accounts for active general covariance better than its rivals (pp. 43–44).
• Not tied to points. The view is not tied to a set-theoretic representation by points. In algebraic GR, points correspond to maximal ideals, and their identity can be understood structurally at the algebraic level (p. 44). “Spelling out the details would obviously require a separate paper”.
• The broader claim. In a broader sense, moderate SR “is not necessarily committed to the existence of space-time points”. The basic constituents “could also stem from a generalization of the notion of a space-time point”. Quantum gravity “would require a rather radical modification of the space-time ontology presented here” while fitting “the main metaphysical claim” (pp. 44–45).

Claims

id	claim	strength	support
C1	Worrall’s epistemic SR is internally strained: if science describes only relations, structure is exposed to pessimistic induction too	moderate	informal argument, p. 29, conditional on the Jackson-style argument the paper assumes
C2	Radical OSR is unintelligible as a theory of the physical world, because concrete relations need relata	moderate (standard objection)	informal argument, p. 31; endorsed, not developed; fn. 6 cites four sources including Chakravartty 2003
C3	French and Ladyman’s quantum underdetermination arguments fail	weak–moderate	two short replies, pp. 30–31: haecceity is metaphysics anyway; QFT favours fields (Cao)
C4	Concrete relations can provide identity conditions wherever intrinsic properties can	moderate	argument from determinacy, p. 32; “A is bigger than B” example; web-of-belief analogy, pp. 32–33. Applies only where relations discern
C5	Where neither relations nor intrinsic properties discern (entangled same-kind systems; symmetric space-times), numerical diversity can be primitive without being primitive thisness	weak (stipulation with a motivation)	pp. 33–34; the “not a thisness” point is that it gives no identity in time or “any other sort of an identity”; credited as similar to Pooley 2005
C6	Objects and relations are mutually ontologically and conceptually dependent, with no priority either way	moderate as a statement of the view; weak as an argument	p. 32; the argument by elimination on p. 34 assumes the three options (bundles, thisness, primitive diversity) are exhaustive
C7	Epistemic SR about space-time individuates points intrinsically and so falls to the hole argument	moderate	pp. 35–36; relies on the standard Earman–Norton/Stachel reading; that ESR must individuate points intrinsically is asserted, and fn. 10 notes Slowik’s agnostic ESR does not
C8	The metric “completely determines the structural identity” of space-time points	weak as stated generally	p. 37; it holds generically. The paper’s own §1 symmetric case (p. 33) and the Bergmann–Komar restriction to space-times without symmetries (p. 39) are exceptions it does not reconcile with this sentence
C9	Bergmann–Komar/Lusanna–Pauri labelling realizes structural identity of points concretely	moderate within its scope	pp. 38–39; reported from Lusanna & Pauri 2005; vacuum, no symmetries, a specific class of solutions, gauge-fixed
C10	The fibre-bundle formulation “naturally receives” the moderate reading and makes the hole argument unformulable	moderate (interpretive fit)	pp. 40–42, resting on Stachel 1986, 2002 and Stachel & Iftime 2005; claims (1)–(3) are correspondences, not derivations
C11	Moderate SR about space-time is a non-reductive, “sophisticated” substantivalism, compatible with (1) or (3) and not with (2)	moderate	definitional placement, pp. 42–43
C12	The view survives the loss of points in algebraic GR or quantum gravity	weak (programmatic)	p. 44; details deferred to “a separate paper”

Concepts

• Master argument for intrinsic properties (p. 29). (1) Relations require relata. (2) Relata must be something in themselves, with intrinsic properties beyond their relations. After Langton 1998.
• Moderate structural realism. An ontic structural realism accepting (1) and denying (2). There are objects, “but instead of being characterized by intrinsic properties, all there is to the basic physical objects are the relations in which they stand” (p. 31).
• Radical structural realism. French and Ladyman (2003): “there is only structure in the sense of relations, and there are no objects standing in the relations” (p. 30).
• Numerical distinction (diversity), primitive. The bare fact that there are more than one, “and that is all that it tells us” (p. 33). It establishes no identity, in time or otherwise. That is the stated difference from primitive thisness.
• Structural identity. The identity of a space-time point as fixed by its position in the network of chronogeometrical, inertio-gravitational and causal relations given by the metric (p. 38), or by a section of the frame bundle (p. 41). The paper also calls it “physical identity”.
• Space-time point vs bare manifold point. A bare manifold point has only mathematical identity, through coordinatization. A space-time point is what the metric makes of manifold points. It corresponds to an equivalence class under diffeomorphism, not bijectively to a manifold point (pp. 36–38).
• Leibniz equivalence. Diffeomorphic models represent the same physical situation (p. 36).
• Soldering. The mutual determination of base-space and fibre geometry in the frame bundle (p. 40).

Connections

• Chakravartty (LIT-196, read in NOTE-149). He is the paper’s ally, not its critic. Fn. 6 cites Chakravartty 2003 for the intelligibility objection to radical OSR, and fn. 8 cites Chakravartty 2004 for concrete relations escaping Newman. His 2012 dilemma is aimed at this paper (his fn. 9 and §10.6, per NOTE-149). On his map, the paper takes the horn with no intrinsic identity-fixing features. It answers in advance the “what is the what?” question that §10.6 asks of its p. 32 sentence. The “what” is a bearer numerically distinct from others, with no further identity. NOTE-149 records no engagement by Chakravartty with the primitive-diversity reply. I have not reread his chapter to check.
• Chakravartty’s mass-and-charge point (NOTE-149 C6) is not met here. The quantum discussion concerns entanglement and indistinguishability only. State-independent properties such as mass and charge are never discussed, so the paper does not say whether they are intrinsic. The p. 28 restriction to “fundamental intrinsic properties of basic physical objects” leaves the question where Chakravartty presses it.
• Leitgeb & Ladyman (LIT-154, NOTE-133). The move is the same: plurality that no relation discerns, accepted as primitive. Leitgeb and Ladyman make it for mathematical places and withhold it for physics. This paper makes it for physics: entangled quantum systems and space-time points under symmetry (p. 33), and credits the view to Pooley (2005 §4).
• The Stanford survey (LIT-045, NOTE-038). Its thesis 4, “Esfeld’s ‘moderate structural realism’”, is glossed as “there are objects, but all their properties are relational”. That is faithful, with one scope point. The paper denies only fundamental intrinsic properties of basic objects, and grants them to composites (p. 28). NOTE-038’s thesis 5, identity grounded in structure through weak discernibility, is a different view. The paper never uses weak discernibility. It goes straight from discerning relations to primitive diversity.
• Morganti (LIT-173, NOTE-095). His §3.2 says moderate OSR posits property-less “bare occupants”. The paper rejects “the view of objects as bare particulars” alongside bundle theories (p. 33). Its objects are characterized by their relations and are numerically, not haecceitistically, distinct. Whether that difference survives his objection is for a reading of Morganti to say. The “drift toward substance ontology” he cites is in Esfeld & Lam 2011, not here.
• Lam & Wüthrich and Bain (LIT-222, LIT-223; THEORY-038). Bain names this paper among those pressing the incoherence charge on radical OSR (NOTE-193), and Lam & Wüthrich build on it (NOTE-194). The p. 44 remark that manifold points correspond to maximal ideals in algebraic GR is the point Lam & Wüthrich later turn against Bain: Einstein algebras drop points “in name only”.
• Wallace (LIT-151, NOTE-142). This paper is a clear case of what Wallace’s §13 calls the second structural-realist project: priority and anti-individualism theses inside the object/relation framework.
• Ladyman 2001 (LIT-217, NOTE-192). Both are non-eliminative. Ladyman grants entangled particles “charge, mass and so on”. This paper denies basic objects any fundamental intrinsic property. So, unlike Ladyman 2001, it is on the no-intrinsic side of THEORY-034’s table.

Bearing on the record

• THEORY-034. Its title says that among the record’s readings “only Floridi’s informational structural realism” keeps relata while denying them every intrinsic identity-fixing feature. With this reading that is no longer true. Esfeld & Lam is a second, and the founding statement of the no-priority variant Chakravartty names. Its promote_when says the account “is refuted if a held non-eliminative OSR text denies relata every intrinsic identity-fixing feature and meets the dilemma”. This text meets the first condition. Whether primitive numerical diversity “meets” the dilemma is a judgement THEORY-034 should make, not this NOTE. Chakravartty would say a bearer with no intrinsic features “is not an object at all” (p. 197). The paper says the bearer is something, and that being numerically one of several is not an intrinsic feature or an identity. I have not edited THEORY-034.
• CLAIM-136 (the owner’s individuation thesis). The paper holds the claim’s weak version, not its strict one. It says relations individuate where they can, and accepts primitive numerical plurality where they cannot (pp. 33–34). The claim says “the weak version has a held precedent, for mathematics only”, and that the assistant “placed it ‘close to’ Esfeld and Lam (A194), then weakened it”. The record now holds a physics precedent for the weak version. And the “weakening” moved the thesis toward Esfeld & Lam, not away from them. The strict version, on which nothing exists as an individual beyond relational differentiation, is stronger than anything this paper asserts. A dated note is added to the claim.
• CLAIM-134 (the objection). It is confirmed that this is the position Chakravartty targets. Two points correct it. First, the paper extends the weak option to physical objects, so the claim’s “Leitgeb and Ladyman take the weak option for mathematical places only” is no longer the record’s only precedent. Its “Nothing in the record extends it to persons or utterances” stays true. Second, the paper has a reply to the step “A primitive fact of numerical distinctness is a non-relational ground of identity”. It denies that numerical distinction is an identity of any sort (p. 33), and it treats primitive diversity as a constraint every metaphysics of physics faces, structuralist or not. The objection may still stand, but it now has a held reply to meet. A dated note is added to the claim.
• CLAIM-138 (the Newman objection). The paper’s answer to Newman is the one that claim says concedes non-structural input. Relations are “specific, concrete” and “as determinate as intrinsic properties are supposed to be” (p. 32, fn. 8). So the paper is a held instance of that remedy. It does not say what makes a relation concrete beyond its determinacy. I have not edited the claim.
• CLAIM-137 says the record holds Esfeld’s moderate structural realism “through the Stanford survey”. It now holds it directly. Its description is accurate.
• For filing. metaphysics first: the subject is the ontology of objects, relations and intrinsic properties. philosophy-of-science: structural realism and the interpretation of GR. individuation: the identity and numerical diversity of space-time points and quantum systems is the paper’s hinge. quantum-foundations is not proposed. Entanglement is a supporting case in two paragraphs (pp. 33–34), and the quantum argument is referred to Esfeld 2004. No topic names general relativity. philosophy-of-science covers “the interpretation of physical theories”. The paper carries nothing for ML practice, and no anthology topic holds it.

Limitations

• The GR sections argue fit, not entailment. The tensor and bundle formulations “naturally receive” the moderate reading (p. 41), and that is as far as they go. Leibniz equivalence and Stachel’s bundle reading are equally the basis of “sophisticated substantivalism” and of other structuralisms. The paper says itself that the endorsement of Leibniz equivalence is called sophisticated substantivalism (p. 43).
• Identity is equivocal across §§1–2. §1 says symmetric space-times have points the relations do not discern, and assigns them primitive diversity (p. 33). §2 says the metric “completely determines the structural identity of the space-time points” (p. 37). §2 never returns to the symmetric case, and the concrete construction it cites works only without symmetries (p. 39). In those cases, which are the physically familiar exact solutions, structural identity does not fix which point is which. The work is done by the primitive diversity of §1.
• The construction it cites breaks the symmetry it relies on. The Bergmann–Komar labelling gives structural identity “through the explicit breaking of active general covariance” (p. 39). The paper treats active general covariance as what “physically grounds” its view (p. 38). The tension is acknowledged in a clause, not discussed.
• The no-priority argument assumes an exhaustive list. It runs by elimination over bundles, thisness and primitive diversity (p. 34). Other options, such as weak discernibility, Morganti-style partial dependence and grounding accounts, are not considered.
• “Space-time point” has no fixed referent. It is an equivalence class of manifold points (p. 36), what a section defines (p. 40), and the values of the gravitational degrees of freedom (Dorato & Pauri, p. 39). The paper endorses all three without saying which it means.
• Quantum physics is asserted, not argued. The paper refers to Esfeld 2004. It does not discuss mass, charge or spin. “Unless one goes for hidden variables” (p. 34) sets aside Bohmian views, which Esfeld later adopted.
• The vacuum restriction matters. It is said not to alter the argument (p. 35). But §4 reports that matter fields change the labelling of points, and matter supplies part of their identity (p. 43).
• Slips.
  • The author line prints “V. Lan” for Lam (p. 27).
  • The text cites “Lusanna and Pauri (2003)”, but the references have only 2005 (p. 43).
  • “Explanatory less successful” (p. 31).
  • The Psillos URL begins “ttp://” (p. 45).
  • “Dordercht” (p. 46).

Open questions

• Is primitive numerical diversity coherent without thisness? It counts without identifying, and it is “not empirically accessible” in the way thisness is said to be (p. 33). A reply from Chakravartty, or a held defence of the distinction (Pooley 2005 §4 is the paper’s own source), would settle whether it escapes the dilemma or is the dilemma’s “thin” concession, as CLAIM-134 has it.
• What makes a relation concrete? The paper’s answer to Newman is that relations are concrete and determinate, but it gives no criterion. Without one, the escape from Newman is a stipulation.
• What do points become in symmetric space-times? Do points in Minkowski or de Sitter space have a structural identity at all, or only primitive diversity? The paper’s two sections give different answers.
• Does it survive quantum gravity? The paper promises that moderate SR survives the loss of points in algebraic GR or quantum gravity, but defers the details to “a separate paper” (p. 44). Whether Esfeld and Lam wrote it (their 2011 and Lam & Esfeld 2012, cited in NOTE-194) is unchecked.

Corrections to the registration

• The registration was accurate on the thesis. The summary and the three key takeaways quote the abstract and match the body.
• The keywords were the preprint’s. The published keywords are “Fibre bundles · Hole argument · Intrinsic properties · Quantum entanglement · Relations · Space-time points · Structural realism” (p. 27). The registration lacked “fibre bundles” and “quantum entanglement”.
• The quoted phrase. It is “an object as such is nothing but that what bears the relations” (p. 32). NOTE-149 and the registration give it without “that”, presumably as Chakravartty quotes it.
• The question the registration posed. “Which horn of Chakravartty’s dilemma the paper takes, and whether its spacetime points have any intrinsic identity-fixing feature”. The answer: the no-intrinsic horn. Points have no intrinsic identity-fixing feature, and haecceity is explicitly rejected through the hole argument. Their plurality, where relations fail to discern, is primitive numerical diversity, which the authors deny is an identity.