{"id":"7cf2f7a3-6d6f-40cb-88fc-8d9406e2c2c4","arxiv_id":"2511.07015","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A physical theory is a 'model bundle': models lie in fibres over a moduli space, quasi-dualities are local transition functions, and dualities are global trivializations.","lead":"A philosophy-of-physics paper argues that a physical theory should be seen as a bundle of models over a space of parameters, with dualities and quasi-dualities playing the role of global and local gluing maps. The proposal aims to replace the 'collection of models' picture and to make inter-theoretic relations a structural part of what a theory is.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Model-bundle framework only applies to invertible, structure-preserving quasi-dualities; non-invertible cases acknowledged but not developed.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the model bundle construction relies on quasi-dualities being local isomorphisms (invertible and structure-preserving) so that they form a group acting vertically. The paper explicitly acknowledges in §3.1.3 that many physically important quasi-dualities are non-invertible or not structure-preserving, and only provides a sketch of how to generalize (semi-groups, sheaves), without developing a rigorous framework. This is indeed the most load-bearing issue: if these generalizations are not worked out, the central claim that quasi-dualities are transition functions (and that dualities are global transition functions) holds only for a special class of examples. The reader's CONDITIONAL verdict appropriately reflects that the central idea is promising but not yet established for the full intended scope. My analysis does not change that verdict; it reinforces the need for the stated generalization to be developed before the strongest reading of the thesis can be accepted. The concrete test proposed would directly check whether the semi-group/sheaf generalization can actually accommodate a concrete non-invertible quasi-duality, thereby settling whether the concern is fatal or merely an open technical issue.","tokens_in":33470,"tokens_out":8362,"duration_ms":85666,"concrete_test":"Take the non-invertible Kramers-Wannier quasi-duality cited in §3.1.3 (Seiberg & Shao 2023). Attempt to define a bundle with local trivializations whose transition functions take values in a semi-group containing this non-invertible map, and check whether the cocycle condition on triple overlaps can be satisfied. If no such semi-group bundle exists (or if it degenerates to a sheaf without a well-defined fibre), the central claim fails for this case. Alternatively, verify whether the non-invertible map can be factored into a local isomorphism and a projection, and whether the bundle constructed from the invertible part captures the duality's physical content.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that quasi-dualities are local transition functions between fibres and that the fibre is the local common core presupposes (in §3.1.1) that quasi-dualities are local isomorphisms: structure-preserving bijections above a neighbourhood, so they form a group G of fibre automorphisms. But the paper itself acknowledges in §3.1.3 that physically important quasi-dualities can be non-invertible (e.g., non-invertible Kramers-Wannier maps) or fail to preserve fibre structure (e.g., spin-structure-sensitive maps in Seiberg-Witten). For such cases, the map between models is not a bijection, so it cannot serve as a transition function of a fibre bundle (which by definition are isomorphisms of fibres). The paper only sketches a 'semi-group model bundle' and a 'sheaf' generalization, without showing that the bundle construction—local trivializations, cocycle conditions, vertical action—goes through. Thus the central claim, as stated, is only established for a restricted class of quasi-dualities; the general geometric view is not made precise beyond that class. The load-bearing assumption is the invertibility/local-isomorphism property; if it fails, the model-bundle construction does not apply, and the claim that quasi-dualities are transition functions collapses.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a 'geometric view' of physical theories, according to which a theory is not a bare set of models but a single structured geometric object: a set of models equipped with topological and geometric structure, and more specifically a 'model bundle' with a moduli space in the base and models in the fibres. Quasi-dualities are characterised as local transition functions between fibres, while dualities are recovered as globally defined transition functions, so that the bundle is trivial and the semantic conception of theories is recovered as a trivial product bundle. The proposal is illustrated with two examples: the Seiberg-Witten theory, which supplies the main finite-dimensional case study, and quantum cosmology, which illustrates an infinite-dimensional normed-space variant. The paper also discusses a cautious realist interpretation of the geometric structures on the moduli space.","tokens_in":33745,"tokens_out":10599,"duration_ms":114124,"significance":"If the proposal holds, it offers a fresh answer to recent critiques of the semantic conception, connecting the structure-of-theories debate to moduli spaces, fibre bundles, and quasi-dualities from contemporary physics. The paper is commendably concrete: it engages with a substantial case study (Seiberg-Witten theory), draws on a developing body of work on dualities, and explicitly flags its schematic character and possible generalizations. The author is honest about limitations, including non-invertible quasi-dualities and non-bundle cases. The philosophical payoff would be significant if the geometric view can be shown to do independent explanatory work beyond re-describing known physics.","major_comments":[{"comment":"The central claim that quasi-dualities are local transition functions presupposes, as stated in §3.1.1, that quasi-dualities are local isomorphisms forming a group G. The paper itself acknowledges in §3.1.3 that physically important quasi-dualities can be non-invertible (e.g., non-invertible Kramers-Wannier maps) or fail to preserve fibre structure (spin-structure sensitivity in Seiberg-Witten), and it only sketches a 'semi-group model bundle' and a sheaf generalization without verifying the bundle axioms. Since the abstract and conclusion state the identification without this caveat, the thesis is currently established only for a restricted class. Please either restrict the claim to invertible, structure-preserving quasi-dualities or provide enough of the generalized construction (cocycle conditions, action of the semi-group) to show that the geometric view applies to the acknowledged c","section":"§3.1.1 and §3.1.3"},{"comment":"The 'recovery' of dualities as globally defined transition functions is partly definitional: §3.1.1 identifies the structure group G with the group of quasi-dualities, and §3.1.2 then defines a duality as a transition function covering the whole base. This is a redescription of the quasi-duality/duality distinction rather than an independent result. To make the proposal substantive, the paper should state more explicitly which aspects are stipulative and show that the local/global distinction tracks a physical criterion (e.g., extendability of the common core) not already built into the definition. The Seiberg-Witten example helps, but the general claim needs this clarification.","section":"§3.1.1–§3.1.2"},{"comment":"The identification of quasi-dualities with vertical fibre automorphisms assumes that the parameters on which they act are coordinates of the moduli space, not base points. This is argued for the Seiberg-Witten coupling τ, but not for other important cases. For instance, T-duality, discussed in §2.2 as an effective duality, acts on the radius R, which is normally a modulus (a base point). If so, T-duality relates fibres over different base points and is not a vertical transition function. The paper needs either a general criterion for deciding when a parameter is a coordinate rather than a base point, or a treatment of 'horizontal' quasi-dualities, if the framework is to cover the examples it cites.","section":"§3.1.1 (discussion of S-duality and footnotes 34–36)"}],"minor_comments":[{"comment":"The displayed central charge is misprinted as Z = n_e a_D + n_e a; the first term should be n_m a_D, as required by the subsequent transformation law (n_m, n_e) → (n_m, n_e) M^{-1} and by Eq. (1).","section":"Appendix A, Eq. (2)"},{"comment":"There are two distinct 2017 entries for De Haro in the references ('Spacetime and Physical Equivalence' and 'Dualities and emergent gravity'), but in-text citations do not distinguish them as 2017a/2017b. Please disambiguate.","section":"References"},{"comment":"The claim that the semantic conception is 'recovered' as a trivial model bundle is conceptually strained: the semantic conception does not include a moduli-space base or a structure group, so adding such structure is an expansion rather than a recovery. This should be phrased more cautiously.","section":"§3.2 and Conclusion"}],"recommendation":"major_revision","confidential_remarks":"The paper is an ambitious programmatic proposal and draws heavily on the author's prior joint work with Butterfield; this is legitimate but should not obscure the fact that several of the headline claims are stipulative. The framing as 'recovery' and 'subsumption' should be softened where the construction defines the distinction. The mathematical typo in Eq. (2) should be corrected. The paper is within scope for a general philosophy of science journal and is likely to generate useful discussion if revised."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The model-bundle proposal is worth reading and is genuinely new. Treating quasi-dualities as local transition functions and dualities as global trivializations gives philosophers of physics a concrete geometric picture that goes beyond the usual structured-view slogans. The paper is also honest: it says up front that the proposal is schematic, and it flags the main generalizations rather than burying them.\n\nWhat it does well: it connects the structured-view literature (Halvorson, Halvorson–Tsementzis) with the physics of moduli spaces and gives a worked-out main example from Seiberg-Witten theory. The move of putting strong/weak coupling duality into a vertical automorphism of the same fibre, rather than a horizontal move in the base, is a clarifying contribution to how philosophers should think about S-duality. The discussion of the quantum cosmology example is brief but shows the framework is not limited to finite-dimensional cases.\n\nThe soft spots are real but mostly localized. The load-bearing assumption is in §3.1.1: quasi-dualities are taken to be local isomorphisms, forming a group, so that the bundle machinery applies. §3.1.3 acknowledges non-invertible quasi-dualities, like non-invertible Kramers-Wannier maps, but only sketches a semi-group model bundle. So the central claim that quasi-dualities are transition functions is established only for a restricted class, and the general geometric view is not made precise beyond that class. The paper admits this, so it is not a hidden flaw, but it does mean the framework is less general than the headline suggests. There is also a mild circularity: since the structure group is identified with the group of quasi-dualities, recovering dualities as globally defined transition functions is partly a redescription of the definitions. Given that the paper is proposing a way to organize examples rather than proving a theorem, this is acceptable, but readers should not expect a deep mathematical payoff.\n\nOne concrete error needs fixing: in Appendix A the monodromy group is called Γ0(4), but later in §A.4 it is called Γ0(2). That is a clear inconsistency in the main case study. The heavy self-citation to De Haro and Butterfield is noticeable but not inappropriate, since the book develops the schema being extended.\n\nWho benefits: philosophers of physics working on dualities, theoretical equivalence, or the semantic view will get real value from the model-bundle picture. It deserves a serious referee. A good referee should push on the invertibility assumption and the non-invertible generalization, and ask the author to fix the monodromy group slip.","headline":"A genuinely new geometric template for dualities and quasi-dualities, honest about being schematic; the main caveat is that it only works for invertible, structure-preserving quasi-dualities, plus a localized math slip in the Seiberg-Witten example.","tokens_in":34228,"tokens_out":2211,"would_cite":true,"duration_ms":24761,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A physical theory is a single geometric object: a bundle of models over a moduli space.","keywords":["philosophy of physics","semantic conception of theories","dualities","quasi-dualities","model bundle","moduli space","Seiberg-Witten theory","geometric view of theories"],"falsifier":"Examine a non-invertible quasi-duality, such as the duality of the quantum Ising model, and ask whether it can be realized as a transition function of a locally trivial fibre bundle. If it cannot be written as a family of local bijections preserving the fibre structure, then the model-bundle construction as stated fails for a class of quasi-dualities the paper itself cites, and the claim that theories are model bundles holds only for invertible quasi-dualities.","tokens_in":33351,"feed_emoji":"📐","tokens_out":9321,"duration_ms":85361,"temperature":0.7,"pith_summary":"The paper claims that a physical theory should be treated not as a bare set of models but as one geometric object: a 'model bundle' whose fibres are the models and whose base is a moduli space of parameters. On this picture, quasi-dualities—maps between models that fall short of being full isomorphisms—are local transition functions that glue the fibres together, while genuine dualities are transition functions defined over the whole base, making the bundle trivial. If the proposal is right, the semantic conception of theories is only the featureless special case of a product bundle, and the structure that physics actually displays (moduli spaces, metrics, monodromies, dualities) becomes internal to the theory rather than an external add-on. The author supports the view with a detailed case study of the low-energy Seiberg-Witten supersymmetric gauge theory and a quantum-cosmology example in which the space of wave-functions is a normed space of conformal-field-theory partition functions.","feed_headline":"A theory is a bundle, not a set of models","feed_subtitle":"Quasi-dualities glue fibres, dualities span the whole base, and the old 'set of models' view is only the trivial case.","key_machinery":"The central object is the model bundle, a fibre bundle π:E→M. The base M is a moduli space—a parameter space for coupling constants, expectation values, and similar data—and the fibre over t is a model M(t), a set of quantities and states with a dynamics. The structure group G is the group of quasi-dualities, defined as local isomorphisms of the fibre; the transition functions between local trivializations of the bundle, subject to cocycle conditions on triple overlaps, are the quasi-dualities. This machinery converts inter-theoretic relations into geometry: dualities are globally defined transition functions (a trivial bundle); monodromies around singularities in the base encode non-perturb","core_discovery":"The central claim is that a physical theory is a model bundle: a fibre bundle whose base is the moduli space of parameters, whose fibre over a point t is a model M(t)=Q(t)×S(t) (quantities and states, with dynamics), and whose structure group is the group of quasi-dualities. Quasi-dualities are defined locally, over overlapping regions of the moduli space, and act vertically on the fibres; they are the transition functions that specify how the local pieces of the bundle are glued. A duality is recovered when the transition functions extend over the whole moduli space, so the bundle is globally trivializable and the models are globally defined; the semantic conception is recovered only as the","pith_inferences":["If the bundle picture is adopted, a natural criterion of theoretical equivalence suggests itself: two theories are equivalent when their model bundles are isomorphic, with dualities as global trivializations; the paper gestures at this but does not develop it.","The acknowledged limitation to invertible quasi-dualities points to a likely next step: non-invertible dualities such as those in the quantum Ising model may require a semigroup or category-theoretic generalization of transition functions, turning the model bundle into a stack-like or sheaf-like object.","One testable extension is to treat any known family of dual models as a putative bundle and compute its monodromy group; the monodromy around a singularity would then predict the existence of new massless states, as in the worked supersymmetric gauge theory example.","The geometric view implies that parameter values are part of the identity of a theory: coupling constants and radii are organized in a space whose geometry is physically meaningful, so 'same equations, different constants' is not automatically the same theory."],"forward_implications":["A theory is one structured object, so comparing theories means comparing bundles, not just comparing model sets.","Quasi-dualities acquire a positive structural role: they are the local gluing maps that build the theory, not merely approximations to dualities.","Dualities become globally defined transition functions, so dual models are coordinate descriptions of one global fibre; their shared common core is literally the fibre.","The semantic conception is subsumed as the trivial product-bundle case, which lacks the structure needed for moduli-space metrics and monodromies.","Geometric features of the moduli space—its metric, its singularities, the monodromies around them—carry physical information, such as the emergence of massless non-perturbative states."],"fun_headline_variants":["Theory is a bundle, not a set of models","Model bundles: where theories live in fibre","Quasi-dualities glue the fibres of theory","Dualities are just global quasi-dualities","From set of models to geometric bundle"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The construction assumes quasi-dualities are reversible, structure-preserving maps so that they form a group acting vertically on the fibres; if important physical quasi-dualities are irreversible—as the paper acknowledges—the model bundle as developed does not cover them.","fun_headline_variants_meta":{"raw":{"variants":["Theory is a bundle, not a set of models","Model bundles: where theories live in fibre","Quasi-dualities glue the fibres of theory","Dualities are just global quasi-dualities","From set of models to geometric bundle"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000282,"raw_usage":{"total_tokens":1525,"prompt_tokens":787,"completion_tokens":738,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":531,"completion_tokens_details":{"reasoning_tokens":683}},"tokens_in":531,"tokens_out":738,"duration_ms":7895,"temperature":1.0,"reasoning_tokens":683,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T23:08:20.387878+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Examine a non-invertible quasi-duality, such as the duality of the quantum Ising model, and ask whether it can be realized as a transition function of a locally trivial fibre bundle. If it cannot be written as a family of local bijections preserving the fibre structure, then the model-bundle construction as stated fails for a class of quasi-dualities the paper itself cites, and the claim that theories are model bundles holds only for invertible quasi-dualities.","supporting_citations":[],"review_version":1}