{"id":"37a664b2-4817-4c62-8ba6-4cb016cb02a4","arxiv_id":"1908.04287","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Compactly generated spaces and quasi-spaces are defined and shown cartesian closed in any (T,V)-category, with new examples including an identification of Alexandroff approach spaces with metric approach spaces.","lead":"This paper builds a general framework in enriched category theory where compactly generated spaces and quasi-spaces are defined for (T,V)-categories, covering topology, approach spaces, and ordered or metric spaces. It proves the resulting categories are cartesian closed and shows Alexandroff approach spaces coincide with metric approach spaces.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.3's proof of (QS3) for the exponential quasi-structure relies on an unproved distributivity of (T,V)-Cat; unless the canonical pullback/coproduct comparison is an isomorphism, cartesian closedness of Qs(T,V)-Cat is not established.","rationale":"The reader's conditional verdict is appropriate. I read the paper's central claim as a genuine extension of [ELS04] and [Day68]: Theorem 2.3 and Section 5 are structurally plausible, and the negative NA-App example in §3.2 suggests that the author is attentive to hypotheses. However, Theorem 4.3 is the payoff of the quasi-space half, and its proof contains an explicit unproved appeal to distributivity. This is exactly the kind of hidden assumption that can invalidate a categorical construction even when all examples satisfy it. I am not arguing that the theorem is false; for compact Hausdorff (T,V)-spaces one expects C≅Set^T to make finite distributivity true. But a proof or citation, followed by a completed diagram chase, is required. Therefore the reader's CONDITIONAL verdict stands: accept only after the distributivity lemma is supplied. I agree with the reader that this is the weakest assumption.","tokens_in":31152,"tokens_out":27524,"duration_ms":278279,"concrete_test":"Prove the missing finite-distributivity lemma needed in §4.3: for C_i, B, C in the compact Hausdorff class and continuous maps η_i:C_i→C, show that the canonical map ∐_i(C_i×_C B)→(∐_i C_i)×_C B is an isomorphism in (T,V)-Cat, exploiting C≅Set^T and the fact that T preserves finite coproducts. Then insert this lemma into the (QS3) verification for the exponential quasi-structure and check that the diagram chase goes through; if no proof can be given, or if the chase still depends on an additional unnamed isomorphism, Theorem 4.3 is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing concern is in Section 4.3, in the verification of axiom (QS3) for the candidate exponential quasi-structure. A map β:C→Qs(X,Y) covered by admissible maps β_i:C_i→C is proved admissible by forming pullbacks D_i=C_i×_C B and then invoking 'distributivity of (T,V)-Cat' to obtain the comparison map µ:∐_i D_i→(∐_i C_i)×_C B. The argument that ev∘⟨β·h,α⟩ is covered by the family γ_i = ev∘⟨β_i∘η_i∘π_{C_i}, α∘π_i^B⟩ tacitly requires that the coproduct of the pullbacks be, at least canonically and compatibly, the pullback of the coproduct. Neither a proof nor a citation is supplied, and this is not a formal consequence of topologicalness over Set without checking that the relevant initial and final lift constructions commute. Since (QS3) is one of the defining axioms of a quasi-space, failure of this distributivity step would mean the function object is not a quasi-space and Theorem 4.3 is unsupported. The property is plausible for the compact Hausdorff class, where C≅Set^T with T preserving finite coproducts, so I do not claim a counterexample; I claim a genuine proof gap in a load-bearing step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a general framework of C-generated (T,V)-spaces and quasi-(T,V)-spaces, extending the classical notions of compactly generated topological spaces and Spanier quasi-topological spaces to categories of lax algebras for a monad T and a quantale V. Section 2 defines C-generated spaces, proves that (T,V)-Cat_C is coreflective in (T,V)-Cat, characterizes C-generated spaces as coequalizers of coproducts of generating spaces, and, under an explicit condition (EP), proves that both C-Map and (T,V)-Cat_C are cartesian closed. Section 3 applies the framework to the classes of compact Hausdorff spaces, the Sierpinski (T,V)-space (yielding Alexandroff spaces), exponentiable spaces, and injective spaces, with detailed verifications for App, NA-App, and (U,[0,1]⊙)-Cat. Section 4 defines quasi-(T,V)-spaces, proves that the forgetful functor Qs(T,V)-Cat -> Set is topological, and proves that Qs(T,V)-Cat is cartesian closed. Section 5 establishes that the category of compactly generated (T,V)-spaces is fully reflective in Qs(T,V)-Cat, generalizing Day's result for Top.","tokens_in":31496,"tokens_out":34833,"duration_ms":351489,"significance":"If the results hold, the paper provides a coherent and broad generalization of two classical remedies for the non-cartesian closedness of Top: the subcategory of compactly generated spaces and the supercategory of quasi-topological spaces. The main theorems are proven by standard initial and final lifting arguments, and the paper follows the strategy of Escardo-Lawson-Simpson and Day closely, which makes the claims concrete and falsifiable. The treatment of Alexandroff spaces and the detailed examples for approach spaces and (U,[0,1]⊙)-Cat are valuable. The explicit reliance on [CHR20] for exponentiability of injective spaces is acknowledged. The paper contains no machine-checked proofs, but the arguments are sufficiently explicit to be checked by hand; the statements are precise enough that a counterexample to any of the main theorems would be identifiable.","major_comments":[{"comment":"The proof of Theorem 4.3 contains the sentence \"We observe that we also use distributivity of (T,V)-Cat\" and describes the map µ: ∐_i(C_i ×_C B) → (∐_i C_i) ×_C B as a surjective (T,V)-continuous map. No proof or reference for this distributivity is supplied, and if the verification of (QS3) genuinely required it, Theorem 4.3 would be unsupported. On inspection, however, distributivity is not needed: the canonical map µ exists by the universal property of the coproduct, and the covering map can be taken to be π_B ∘ µ, whose surjectivity follows directly from the surjectivity of η and the pullback condition. The triangle (ev ∘ ⟨β·h, α⟩) ∘ (π_B ∘ µ) = ∐_i γ_i then holds by the definition of the pullback. I therefore recommend deleting or correcting the distributivity sentence and adding a one-line justification that π_B ∘ µ is a continuous surjection. This resolves the main proof gap without changing the statement of Theorem 4.3.","section":"Section 4.3, verification of (QS3) for the exponential quasi-structure"}],"minor_comments":[{"comment":"In the displayed definition of the exponential structure d_j, the projections are named π_X and π_Z and the evaluation is written as h(x); since the exponential is Z^{Y_j}, this should be π_{Y_j} and π_Z and h(y_j), following [CHT03].","section":"Section 2.3, formula for d_j"},{"comment":"The verification of (QS3) for the quotient quasi-structure is relegated to \"One can check\"; since (QS3) is a defining axiom of a quasi-space, please include the short argument that uses the finite coproduct of the covering maps together with closure of C under finite coproducts.","section":"Section 4.2, quotient quasi-structure"},{"comment":"The sentence \"ξ(v1)>ξ(v2)=0\" is confusing after the preceding claim ξ(v1)=ξ(v2)=v; please clarify that the value of the operation ξ(v1)>ξ(v2) is 0 and that this is what makes the relevant diagram non-commutative.","section":"Section 3.2, Example (3), NA-App"},{"comment":"The phrase \"for each X∈TTX\" should be typeset as an element of T^2X (or the intended iterated power of T) to avoid ambiguity.","section":"Section 1.3, before equation (1.x)"},{"comment":"In Theorem 1.3 the first hypothesis says diagram (1.ix) is commutative, but (1.ix) is displayed with a lax-commutativity symbol; please specify whether strict commutativity is intended and adjust the diagram accordingly.","section":"Section 1.3, Theorem statements and diagram (1.ix)"},{"comment":"The equalities QsApp = QsNA-App = QsTop and QsOrd = QsMet = QsUltMet = QsB1Met are asserted from the equivalences of the compact Hausdorff classes; a brief justification that the quasi-space definitions are preserved under these equivalences would be helpful for the reader.","section":"Section 4.3, Examples"}],"recommendation":"minor_revision","confidential_remarks":"The paper's central dependence on [CHR20], which is co-authored by the author, is legitimate because that paper is the cited source for exponentiability of injective (T,V)-spaces; still, the editor may want to confirm that the relevant results in [CHR20] are indeed published and not circular. The size illegitimacy of Qs(T,V)-Cat is acknowledged in the paper and is standard in this line of work. The paper fits the scope of a categorical topology journal such as TAC or Applied Categorical Structures."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a competent, workmanlike generalization of two classical constructions to (T,V)-Cat. The genuinely new payoff is in Section 3.2: the Alexandroff (T,V)-space characterization via A^\\circ and the identification of Alexandroff approach spaces with metric approach spaces. The rest is largely a port of [ELS04] and [Day68] into the (T,V) setting, but the port is done carefully and the examples are checked in detail.\n\nThe Section 2 theorem — C-generated (T,V)-Cat is cartesian closed under condition (EP) — follows the standard final/initial lifting pattern and looks correct. The quasi-space construction in Section 4 also follows the Day/Spanier template, and the topologicalness and reflection results in Sections 4.2 and 5 are what one would expect.\n\nThe real soft spot is Theorem 4.3, cartesian closedness of Qs(T,V)-Cat. In verifying (QS3) for the exponential quasi-structure, the author invokes \"distributivity of (T,V)-Cat\" without proof or citation. I think the stress-test note overstates the damage: the proof only needs a surjective continuous map from \\coprod_i (C_i \\times_C B) to B, and the projection to B supplies one directly, without any distributivity. The comparison map \\mu the author introduces is continuous by the universal properties of coproduct and pullback, and set-level surjectivity is automatic. So the gap is repairable and does not sink the theorem. But as written, the unproved distributivity assertion is a genuine omission; a referee should ask for it to be removed or justified.\n\nOne other thing worth noting: the paper leans on [CHR20] for exponentiability of injective (T,V)-spaces, and the author is a coauthor of that paper. This is not circular — it is prior work — but for the Alexandroff examples the dependence should be made explicit. The size illegitimacy of Qs(T,V)-Cat is acknowledged, but the main theorem is phrased as if it were an ordinary category; a sentence on universe enlargement would help.\n\nWho is this for? People in monoidal topology who want the unified (T,V) statement and the approach-space examples. It does not introduce a new mechanism, but it is a useful reference. I would send it to a competent referee rather than desk-reject; the Section 4.3 gap is minor and fixable. My own verdict is conditional acceptance after a small revision.","headline":"Solid (T,V)-generalization of compactly generated and quasi-spaces; the Alexandroff examples are the payoff, and the flagged gap in 4.3 is real but repairable.","tokens_in":32006,"tokens_out":16604,"would_cite":false,"duration_ms":173251,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18B30","18D15","54A05","54B30","54C35","54D50"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that compactly generated spaces and quasi-spaces, two classical constructions around the category Top, both exist and are cartesian closed in the general setting of (T,V)-categories.","keywords":["(T,V)-categories","compact and Hausdorff spaces","compactly generated spaces","Alexandroff spaces","cartesian closedness","quasi-spaces","quasi-topological spaces","topological functors"],"falsifier":"Within one of the paper's own examples, say App = $(\\mathsf{U},P_+)$-Cat, take finite families of compact Hausdorff objects $C_i$ and $B$ for which the canonical map $\\coprod(C_i\\times_C B)\\to(\\coprod C_i)\\times_C B$ can be computed; if any such map is not an isomorphism in (T,V)-Cat, the unstated distributive premise behind Section 4.3 fails, and the proof of cartesian closedness of Qs(T,V)-Cat collapses. Conversely, checking these maps in each row of the table would confirm the premise the argument silently uses.","tokens_in":2145,"feed_emoji":"📐","tokens_out":2768,"duration_ms":115241,"temperature":0.7,"pith_summary":"The paper sets out to prove that two classical strategies for repairing the failure of cartesian closedness in the category of topological spaces—restricting to compactly generated spaces and enlarging to quasi-spaces—work at the much higher level of generality of (T,V)-categories, a framework that includes ordered, metric, topological, and approach spaces in a single formalism. It shows that for any class C of 'generating' objects satisfying a natural product condition, the C-generated (T,V)-spaces form a cartesian closed category, and that taking C to be the compact Hausdorff spaces yields the expected compactly generated case. It then defines quasi-(T,V)-spaces by axiomatizing which maps from compact Hausdorff objects are admissible, proves the resulting category is topological and cartesian closed, and shows that compactly generated spaces embed into quasi-spaces as a full reflective subcategory. The upshot is that the compactly generated / quasi-space dichotomy is not a peculiarity of Top but a structural feature of the entire (T,V)-landscape.","feed_headline":"Compactly generated and quasi-space categories extend beyond Top","feed_subtitle":"A single setting covers ordered, metric, topological, and approach spaces, and both categories stay cartesian closed.","key_machinery":"The machinery has two load-bearing pieces. First is the notion of a generating class $\\mathcal{C}$ together with 'probes': continuous maps $C\\to(X,a)$ with $C\\in\\mathcal{C}$. The $\\mathcal{C}$-generated structure $a_c$ is the final (T,V)-structure with respect to all probes, so that a space is $\\mathcal{C}$-generated exactly when it is a coequalizer of a coproduct of generating spaces; the exponential structure on $\\mathcal{C}$-Map$(Y,Z)$ is built from the initial structure induced by composing with probes. Second is the quasi-space structure: a set $X$ together with, for each compact Hausdorff $C$, a set $Q(C,X)$ of admissible maps closed under constant maps, precomposition, and a covering axiom (QS3) that says a map is admissible precisely when a finite family of admissible maps covers it through a surjective continuous map from a coproduct. The cartesian closedness of Qs(T,V)-Cat is carried by the exponential quasi-structure on Qs$(X,Y)$, whose verification of (QS3) uses pullbacks of covering families and the distributivity of (T,V)-Cat.","core_discovery":"On the paper's own terms, the discovery is that both classical constructions survive the passage from Top to (T,V)-Cat. Theorem 2.3 establishes that C-Map and (T,V)-Cat$_\\mathcal{C}$ are cartesian closed whenever each generating object is exponentiable and products of generating objects are C-generated; Theorem 4.3 establishes that Qs(T,V)-Cat, whose objects are sets equipped with admissible-map data $Q(C,X)$ satisfying axioms (QS1)–(QS3), is also cartesian closed; and Section 5 establishes that the compactly generated (T,V)-spaces form a full reflective subcategory of Qs(T,V)-Cat, with the reflector sending a quasi-space to the final structure induced by its admissible maps. Along the way the paper recovers, in the table of examples, the classical facts that compactly generated topological spaces are quotients of disjoint sums of compact Hausdorff spaces, that Alexandroff topological spaces are exactly the C-generated spaces for the Sierpiński generator, and that quasi-spaces in the approach and $(\\mathsf{U},[0,1]_\\odot)$ settings coincide with quasi-topological spaces because the compact Hausdorff objects reduce to $\\mathsf{Set}^{\\mathsf{U}}$.","pith_inferences":["The paper leaves the distributive law in Section 4.3 unproved; a natural next step is to verify it for each row of the table, and the case-by-case diagram checks already used in Section 3.2 suggest the law may be forced by the algebraic extension $\\xi$ whenever the quantale is integral and totally ordered.","Because quasi-spaces only consult the compact Hausdorff objects, the equality QsApp = QsTop suggests a general principle: the quasi-space category over a (T,V)-category depends only on the 'compact Hausdorff core' $\\mathsf{Set}^{\\mathsf{T}}$, so two doctrines with the same T-algebras will have identical quasi-space categories.","A question the paper does not address is whether the reflector from quasi-spaces to compactly generated spaces preserves finite products; if it does, the quasi-space exponential would automatically restrict to the compactly generated subcategory, giving an alternative route to Theorem 2.3.","For Alexandroff spaces, the two diagram conditions, commutativity of (3.ii) and inequality (3.iii), are shown to be sufficient for $A_\\circ$ to give all Alexandroff (T,V)-spaces from Alexandroff V-spaces; an extension the paper does not pursue is whether these conditions are also necessary in general."],"forward_implications":["Every row of the paper's table—ordered, metric, ultrametric, bounded metric, topological, approach, non-Archimedean approach, and $(\\mathsf{U},[0,1]_\\odot)$-categories—has a cartesian closed category of compactly generated objects.","The Alexandroff construction generalizes: for an integral totally ordered quantale and $T=I$, the Sierpiński $(V,\\mathrm{hom})$ object generates a cartesian closed category of Alexandroff V-spaces, and in Top this recovers classical Alexandroff spaces, equivalent to preordered sets.","Quasi-(T,V)-spaces form a topological category over Set, hence are complete and cocomplete, and they are cartesian closed with the expected exponential Qs$(X,Y)$.","A (T,V)-space is compactly generated exactly when, for every (T,V)-space $Y$, (T,V)-continuous maps $X\\to Y$ coincide with quasi-continuous maps between the associated quasi-spaces; consequently compactly generated spaces form a full reflective subcategory of all quasi-spaces.","In several important cases the quasi-space category is independent of the metric choice: QsApp, QsNA-App, Qs$(\\mathsf{U},[0,1]_\\odot)$-Cat, and QsTop all coincide because their compact Hausdorff objects are all $\\mathsf{Set}^{\\mathsf{U}}$."],"supporting_citations":[{"why":"supplies the generating-class approach, the productivity condition (EP), and the comparison with Top that Section 2 follows","marker":"[ELS04]"},{"why":"introduces quasi-topological spaces and the admissible-map presentation that Section 4 generalizes, including an example of a quasi-space not associated with any space","marker":"[Spa63]"},{"why":"provides the template for the relationship between quasi-spaces and compactly generated spaces that Section 5 extends","marker":"[Day68]"},{"why":"supplies the definitions and theory of compact and Hausdorff (T,V)-spaces, including the equivalence with T-algebras used throughout","marker":"[HST14]"},{"why":"establishes exponentiability of injective (T,V)-spaces under the stated hypotheses, used to verify condition (EP) for the examples","marker":"[CHR20]"},{"why":"shows the size illegitimacy of the quasi-space category, which the paper acknowledges while retaining the terminology","marker":"[HR83]"},{"why":"records the coreflectivity of compactly generated spaces in Top, the classical statement whose generalization motivates Theorem 2.1","marker":"[Mac71]"}],"fun_headline_variants":["Compactly generated and quasi-spaces extend to enriched categories","Cartesian closed: compactly generated and quasi-space categories","Beyond Top: compactly generated and quasi-spaces in any (T,V)-Cat","Generalized enriched categories keep classical space constructions"],"cache_read_input_tokens":34048,"weakest_assumption_plain":"The proof that function quasi-spaces are closed under the covering axiom assumes, without proof, that (T,V)-Cat is distributive in the specific sense that the canonical comparison $\\coprod_i(C_i\\times_C B)\\to(\\coprod_i C_i)\\times_C B$ is an isomorphism; if that fails, the verification of axiom (QS3) for the exponential quasi-structure collapses, and Theorem 4.3 is not established.","fun_headline_variants_meta":{"raw":{"variants":["Compactly generated and quasi-spaces extend to enriched categories","Cartesian closed: compactly generated and quasi-space categories","Beyond Top: compactly generated and quasi-spaces in any (T,V)-Cat","Generalized enriched categories keep classical space constructions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00023,"raw_usage":{"total_tokens":1461,"prompt_tokens":905,"completion_tokens":556,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":521,"completion_tokens_details":{"reasoning_tokens":486}},"tokens_in":521,"tokens_out":556,"duration_ms":6470,"temperature":1.0,"reasoning_tokens":486,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:47:04.177471+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Within one of the paper's own examples, say App = $(\\mathsf{U},P_+)$-Cat, take finite families of compact Hausdorff objects $C_i$ and $B$ for which the canonical map $\\coprod(C_i\\times_C B)\\to(\\coprod C_i)\\times_C B$ can be computed; if any such map is not an isomorphism in (T,V)-Cat, the unstated distributive premise behind Section 4.3 fails, and the proof of cartesian closedness of Qs(T,V)-Cat collapses. Conversely, checking these maps in each row of the table would confirm the premise the argument silently uses.","supporting_citations":[],"review_version":1}