{"id":"7dbd9a6e-33a5-43e0-b5f0-d53c3ac1bd3a","arxiv_id":"2411.11816","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every ring can be recovered from a new topological object, its noncommutative spectrum, extending the classical duality between commutative rings and affine schemes.","lead":"This paper proves that every ring, even a noncommutative one, can be represented as a geometric space together with the functions on it, and can be recovered from that space. The result extends a classical duality in algebra and offers a starting point for studying quantum spacetime geometry.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The advertised anti-equivalence is either vacuous or unproven: Theorem 7.20 proves only faithfulness, which does not yield a full anti-equivalence onto a subcategory of PreRingSites; if 'subcategory' is allowed to be non-full, the statement is a tautology for any faithful functor.","rationale":"The reader's weakest_assumption concerns Proposition 4.4, the formal homotopy Zariski topology. That is a real technical risk, since the proof is a sketch relying on stability of homotopy epimorphisms under homotopy pushouts and on the homotopical pullback lemma from [40], and the functoriality of Spec^NC collapses if the coverage axioms fail. I do not dispute that concern. However, I read the most load-bearing issue as the gap between the statement of Theorem 1.1 and its proof: the paper proves faithfulness, not a full anti-equivalence, and the paper's own Section 7.2 explicitly says 'we do not achieve the goal of describing a perfect duality' and only constructs 'the less ambitious task of constructing a faithful functor.' A faithful functor is not an anti-equivalence onto a full subcategory; if 'subcategory' is taken to be non-full, then the anti-equivalence statement becomes a tautological reformulation of faithfulness. This is not an ad hominem or a disagreement with consensus: it is a mismatch internal to the paper's central claim, and it is testable by the concrete fullness check described above. Because the honest mathematical content, the existence of a faithful representation, may still be valuable if Proposition 4.4 is repaired, I keep the reader's CONDITIONAL verdict unchanged rather than moving to REJECT. The check on Example 7.16 is especially relevant because the non-local map Z_(p) -> Q is exactly the classical obstruction to fullness that PreRingSites, by dropping locality, admits.","tokens_in":59344,"tokens_out":16082,"duration_ms":177235,"concrete_test":"Fix definitions and take A = Z, B = Q, with V = Z_(p) an open in Zar_Z. Define a morphism of pre-ringed sites (Zar_Q,O_Q) -> (Zar_Z,O_Z) by sending the unique object of Zar_Q to V and taking the structure morphism on V to be the canonical inclusion Z_(p) -> Q, extended by pullback; check whether this satisfies the axioms of Definition 7.17 (continuity, cover preservation, naturality). If it does, the functor (7.5) is not full, so Theorem 1.1 cannot hold as a full anti-equivalence and must be weakened to 'faithful embedding.' If it does not, identify which axiom fails; either answer settles the status of the advertised anti-equivalence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem (Theorem 1.1, Corollary 7.21) is stated as \"Rings_Z is anti-equivalent to a subcategory of PreRingSites.\" The proof supplied is Theorem 7.20, which shows that the contravariant functor A -> (Zar_A,O_A) is faithful. Faithfulness alone does not establish an anti-equivalence in the standard sense: an equivalence is full, faithful, and essentially surjective. The paper gives no argument for fullness, and the commutative analogue already fails: Example 7.16's morphism Spec Q -> Spec Z_(p) is a morphism of pre-ringed spaces (locality is not imposed in PreRingSites) and is not induced by any ring map Z -> Q; the same construction should yield a morphism of the associated pre-ringed sites not in the image of (7.5). If one instead allows the target 'subcategory' to be non-full, then every faithful functor is trivially an anti-equivalence onto its image with inherited morphisms, and Theorem 1.1 becomes a restatement of faithfulness rather than a duality. Thus the precise content of the central claim is weaker than stated: a faithful embedding, not a Gelfand-type anti-equivalence. This concern is independent of, though complementary to, the technical gaps in Proposition 4.4; the faithfulness result still depends on that proposition for the site structure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a noncommutative spectrum functor on the homotopy category of connective dg-algebras over Z, built from localizations that are homotopical epimorphisms. The central construction associates to each A the small site Zar_A of homotopical-epimorphism localizations of A, together with a structure presheaf O_A. The main result is stated as an anti-equivalence of Rings_Z with a subcategory of pre-ringed sites; the proof establishes that the contravariant functor A ↦ (Zar_A, O_A) is faithful. The paper also introduces a finer Zariski topology, a relative spectrum that agrees with the Grothendieck spectrum for finitely generated commutative C-algebras, and a notion of descendable presheaves intended as a noncommutative replacement for sheaves.","tokens_in":59494,"tokens_out":11112,"duration_ms":99443,"significance":"If the constructions are correct, the faithful embedding is a meaningful step: it shows that the site of homotopical localizations with its structure presheaf contains enough information to recover the ring, and the comparisons with the Grothendieck spectrum and with smashing subcategories give the construction independent interest. The paper is transparent about relying on a conjecture for the fine-topology results, and Theorem 7.20 itself is unconditional and proven by a short, convincing evaluation at the terminal object. However, the advertised Gelfand-type anti-equivalence is not obtained: fullness is neither proven nor claimed, and the paper explicitly defers the description of the essential image. The main value is thus the faithful embedding and the supporting machinery, not an anti-equivalence in the usual sense.","major_comments":[{"comment":"The statement that Rings_Z is 'anti-equivalent to a subcategory of PreRingSites' is not supported by the proof. Theorem 7.20 proves only faithfulness of the functor (7.5). No fullness or essential surjectivity is established; in fact, the text following Theorem 7.20 says the description of the essential image is deferred to a separate work, and Example 7.16 shows that the functor is not full on morphisms even for commutative rings: the map of pre-ringed spaces Spec Q → Spec Z_(p) is not induced by a ring map. If 'subcategory' is taken to be the non-full image subcategory, then Theorem 1.1 is a tautological restatement of faithfulness; if it is taken as a full subcategory, the claim is false. The theorem should be restated as a faithful contravariant embedding, and the phrase 'anti-equivalence' should be reserved for a future result or used only with an explicit warning about the non-full subcategory convention.","section":"Theorem 1.1 and Corollary 7.21; Section 7.2"},{"comment":"The proof that the formal homotopy Zariski topology satisfies the Grothendieck topology axioms is incomplete. The stability of homotopical epimorphisms under arbitrary homotopy pushouts is asserted without proof or reference, and the transitivity axiom is verified by a sketch that invokes the homotopy pullback lemma of [40] without spelling out the required hypotheses. Because this proposition is what guarantees that Zar_A is a site and that Spec^NC is a functor (Corollary 4.15), the constructions in Sections 4.3 and 7.2 are not fully justified until this proof is completed.","section":"Proposition 4.4"}],"minor_comments":[{"comment":"The sentence containing 'see Defition 4.3' has a typo: it should read 'see Definition 4.3'.","section":"Section 1"},{"comment":"The corollary states that the category of rings is 'equivalent' to a subcategory of PreRingSites; given the contravariance of the construction, the intended word appears to be 'anti-equivalent', or the statement should explicitly define the non-full subcategory convention.","section":"Corollary 7.21"},{"comment":"The notation for the pushout A ∗^L_{R[s]} R[s,s^{-1}] should be clarified: the map is given by R[x] → A, x ↦ s, so the base of the free product is R[s] viewed as a copy of the polynomial ring R[x].","section":"Example 4.8(vi)"},{"comment":"The term 'Amistur–Čech nerve' appears at least twice; the standard spelling is 'Amitsur–Čech'.","section":"Section 7.2"},{"comment":"In the paragraph discussing the ring k^n, the text says 'one computes Spec^NC(k)' but the computation is for the direct product k^n, so the symbol should be corrected accordingly.","section":"Section 6.1"},{"comment":"The notation table lists Spec^NC_fine as a functor on CRings_Z, whereas Definition 5.5 states it for A ∈ HRings_Z; these should be reconciled.","section":"Notation table and Definition 5.5"}],"recommendation":"major_revision","confidential_remarks":"The overstatement of Theorem 1.1 is likely to be the main point of controversy. The body of the paper is honest about the gap, and the core faithful-embedding theorem appears sound. However, the abstract and the statement of Theorem 1.1 should be corrected, and the proof of Proposition 4.4 must be brought to a fully detailed argument before the paper can be accepted. The paper's value lies in the construction and the faithful embedding; I would not support rejection, as the issues are addressable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Let me cut to the chase. The paper builds a genuinely new spectrum for noncommutative rings from homotopical epimorphisms and proves a faithful embedding of HRings_Z into pre-ringed sites. The faithful embedding is real and the proof is short and robust: it only inspects the structure presheaf on the terminal object. The handling of Reyes's no-go theorem is honest: they show finite-presentation variants collapse and that module-level conservativity fails (Example 5.1), so they sidestep the obstruction rather than pretending it away. The examples, like the kA_2 quiver, are worked out in enough detail to be persuasive. The formal homotopy Zariski topology and the descendable-presheaf framework are new, not a repackaging of prior spectra.\n\nBut the headline theorem is not what they prove. Theorem 1.1 and Corollary 7.21 claim 'anti-equivalent to a subcategory of pre-ringed sites.' Theorem 7.20 proves only faithfulness. Faithfulness alone does not give an anti-equivalence; if the subcategory is allowed to be non-full, then every faithful functor is trivially an equivalence onto its image. The paper's own introduction softens this to 'by this we mean that the functor is faithful,' but the abstract and Theorem 1.1 still overstate. Example 7.16 shows a morphism of pre-ringed spaces not induced by any ring map, so fullness is genuinely absent. This is a real gap between packaging and content.\n\nThe other soft spots are more minor. Proposition 4.4, proving the formal homotopy Zariski topology is a Grothendieck topology, is a sketch that leans on a cited homotopical pullback lemma; if that stability fails, the spectrum functor collapses. There are also unstated computations in several examples and a conjecture (5.18) assumed for the descendable-presheaf results. None of this undermines the faithfulness theorem, which is independent of the conjecture.\n\nWho gets value from this? Anyone working on noncommutative spectra, derived algebraic geometry, or the Reyes no-go program. The quantum-spacetime motivation in the abstract is not delivered and should be ignored.\n\nShould it be refereed? Yes. The construction is new, the central proof is solid, and the overstatement is fixable. Send it to a good algebra journal, but insist the authors either prove fullness or rewrite the main theorem as a faithful embedding into pre-ringed sites with the non-full image made explicit. The gap is repairable and the core content deserves publication.","headline":"The construction is genuinely new and the faithful-embedding result is solid, but the advertised anti-equivalence is really just faithfulness; the paper overstates its main theorem.","tokens_in":60225,"tokens_out":3507,"would_cite":true,"duration_ms":30472,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14A22","14A30","16E35","18F20"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that every ring, and more generally every connective dg-algebra over $\\mathbb{Z}$, is faithfully represented by the site of its homotopical-epimorphism localizations together with a structure presheaf, yielding a…","keywords":["noncommutative Gelfand duality","homotopical epimorphism","formal homotopy Zariski topology","noncommutative spectrum","derived algebraic geometry","pre-ringed sites","descendable presheaf","comonadic descent"],"falsifier":"Take a concrete noncommutative connective dg-algebra $A$ and a homotopy pushout square $A \\to B$, $A \\to C$ in $\\mathrm{HRings}_{\\mathbb{Z}}$; if the induced map $B \\to B \\ast^{\\mathrm{L}}_A C$ is not a homotopical epimorphism, or if two morphisms $C \\rightrightarrows D$ in $\\mathrm{HRings}_B$ become equivalent after every base change along a cover but are not equivalent directly, then the topology axioms of Proposition 4.4 fail and the spectrum functor $\\operatorname{Spec}^{\\mathrm{NC}}$ would not be functorial.","tokens_in":58925,"feed_emoji":"🔄","tokens_out":13510,"duration_ms":122571,"temperature":0.7,"pith_summary":"The paper tries to establish a noncommutative version of Gelfand duality for ordinary rings: every ring $R$ can be represented, faithfully, by a geometric object built from $R$'s own localizations. The representation is the pair $(\\mathrm{Zar}_R, \\mathcal{O}_R)$, where $\\mathrm{Zar}_R$ is the site whose 'open subsets' are the homotopical epimorphisms out of $R$ and whose covers are finite families whose base-change functors on algebras are jointly conservative, and $\\mathcal{O}_R$ is a structure presheaf of connective dg-algebras. The main theorem states that the category of rings is anti-equivalent to a subcategory of pre-ringed sites, meaning that distinct morphisms of rings induce distinct morphisms of the associated sites; the construction passes through the larger homotopy category $\\mathrm{HRings}_{\\mathbb{Z}}$ of connective dg-algebras, because even for ordinary noncommutative rings the natural structure presheaf can take values that are not concentrated in degree zero. This matters because it gives noncommutative rings a genuine geometric spectrum with points, opens, and functions, while classical attempts to extend the prime-ideal spectrum to all rings are blocked by a no-go theorem. If correct, it also supplies a framework in which the geometry of quantum-spacetime algebras could be studied rigorously.","feed_headline":"Rings are recoverable from their noncommutative spectra","feed_subtitle":"Using dg-algebra localizations, distinct rings yield distinct pre-ringed sites, a noncommutative Gelfand duality.","key_machinery":"The carrying object is a new Grothendieck topology on $\\mathrm{HRings}_{\\mathbb{Z}}^{\\mathrm{op}}$, the formal homotopy Zariski topology. Its open embeddings are homotopical epimorphisms: maps $A \\to B$ such that the fold/codiagonal map $B \\ast^{\\mathrm{L}}_A B \\to B$ is an equivalence (equivalently $B \\otimes^{\\mathrm{L}}_A B \\to B$ is an equivalence), the noncommutative replacement for flat epimorphisms and Zariski open immersions. Its covers are finite families $\\{A \\to B_i\\}$ for which the base-change functors $(-)\\ast^{\\mathrm{L}}_A B_i \\colon \\mathrm{HRings}_A \\to \\mathrm{HRings}_{B_i}$ are jointly conservative. From this topology one forms the small site $\\mathrm{Zar}_A$ of localizations of $A$; the site is coherent and has enough points, which yields the sober space $\\operatorname{Spec}^{\\mathrm{NC}}(A)$. The second load-bearing piece is the structure presheaf $\\mathcal{O}_A$ together with the notion of a descendable presheaf: reconstruction from a cover is performed with the Amitsur--Cech nerve and comonadic descent, rather than by the ordinary sheaf condition.","core_discovery":"The central claim is that the contravariant functor $\\mathrm{Spec}^{\\mathrm{NC}}\\colon \\mathrm{HRings}_{\\mathbb{Z}} \\to \\mathrm{PreRingSites}$, $A \\mapsto (\\mathrm{Zar}_A,\\mathcal{O}_A)$, is faithful (Theorem 7.20); composing with the fully faithful inclusion $\\mathrm{Rings}_{\\mathbb{Z}} \\hookrightarrow \\mathrm{HRings}_{\\mathbb{Z}}$ gives Theorem 1.1. Here $\\mathrm{HRings}_{\\mathbb{Z}}$ is the homotopy category of connective dg-algebras over $\\mathbb{Z}$, in which ordinary rings sit as discrete objects. The site $\\mathrm{Zar}_A$ is the opposite of the category of homotopical epimorphisms $A \\to B$, with finite covers given by families whose derived free-product functors $(-)\\ast^{\\mathrm{L}}_A B_i$ are jointly conservative on $\\mathrm{HRings}_A$; the structure presheaf sends such a localization to $B$ as a connective dg-algebra. The proof of faithfulness evaluates both sides on the terminal object: a morphism $A \\to B$ of homotopical rings is recovered as the map $\\mathcal{O}_A(X_A)=A \\to B$ induced on global sections, so two distinct maps produce two distinct maps of pre-ringed sites. The paper also shows the associated topos has enough points, so $\\mathrm{Zar}_A$ presents a sober topological space $\\operatorname{Spec}^{\\mathrm{NC}}(A)$, and that $\\mathcal{O}_A$ is a descendable presheaf satisfying comonadic descent, which is what replaces the sheaf condition in the noncommutative setting.","pith_inferences":["Read in light of the previously established no-go theorem for spectrum functors valued in sets, the present construction suggests that the obstruction is not to geometric duality itself but to requiring points to be prime ideals: replacing topological spaces by sites plus structure presheaves restores a faithful duality.","The faithfulness argument only uses the terminal object of the site, so the same strategy could be tried in analytic settings: any contravariant representation that records the identity section at the terminal open will automatically be faithful, independently of whether the structure presheaf is a sheaf.","For physical applications, the spectrum gives a concrete way to ask what a point of a quantum spacetime is: for algebras such as noncommutative tori or symplectic twisted group algebras one could compute the lattice of homotopical epimorphisms and read off the resulting sober space, and the paper's examples show such point spaces are computable in nontrivial cases.","A natural test of the framework is the paper's Conjecture 7.14: finding a homotopical Zariski cover whose Amitsur--Cech complex for the structure presheaf is not exact would keep the faithful duality intact but would block the upgrade to ordinary ringed spaces."],"forward_implications":["Every ordinary ring, and every connective dg-algebra over $\\mathbb{Z}$, is distinguished by its localization site: distinct ring homomorphisms induce distinct morphisms of pre-ringed sites.","The spectrum $\\operatorname{Spec}^{\\mathrm{NC}}(A)$ is functorial, nonempty, and often large; its points are completely prime ultrafilters on the lattice of localizations, and its open subsets correspond to smashing localizations of the category of modules over $A$.","For commutative input the construction is compatible with classical geometry: fields and discrete valuation rings have the same spectra as the Grothendieck spectrum, the fine spectrum of a finitely generated commutative $\\mathbb{C}$-algebra is homeomorphic to the Grothendieck spectrum, and in general there are canonical continuous maps relating the noncommutative, fine, and Grothendieck spectra.","The structure presheaf of a discrete noncommutative ring can take values in nonzero cohomological degrees, so a faithful geometric representation of all rings forces homotopical and dg rings into the picture.","Since the structure presheaf satisfies comonadic descent but not necessarily the sheaf condition, the classical notion of sheaf is inadequate as the glueing rule for noncommutative geometry; the paper proposes descendable pre-sheaves as the correct replacement, with ordinary sheaf theory recovered in the commutative case."],"supporting_citations":[{"why":"Proves homological and homotopical epimorphisms coincide, the equivalence that legitimises homotopical epimorphisms as the open immersions of the topology.","marker":"[23]"},{"why":"Shows that flat epimorphisms of finite presentation between commutative rings are homotopical epimorphisms, giving the compatibility with the classical Zariski topology.","marker":"[60]"},{"why":"Supplies the homotopical pullback lemma used in Proposition 4.4 to verify that the proposed classes form a Grothendieck topology.","marker":"[40]"},{"why":"Formulates the no-go theorem that no spectrum functor to sets can simultaneously be nonempty, functorial, and extend the prime-ideal spectrum, the obstruction this framework is designed to bypass.","marker":"[50]"},{"why":"Classifies homotopical and flat epimorphisms of commutative Noetherian rings, used to compute spectra and to compare them with the Grothendieck spectrum.","marker":"[2]"},{"why":"Identifies smashing localizations of the module category of a commutative ring with sets of primes, underpinning the commutative comparison of spectra.","marker":"[46]"},{"why":"Provides the commutative notion of descendability that the paper generalises; its terminology and descent arguments are adapted for the noncommutative setting.","marker":"[44]"},{"why":"Gives the categorical Barr--Beck type conditions (Proposition 10.5) used to decide when local-to-global reconstruction is comonadic in Section 7.","marker":"[53]"},{"why":"Supplies the classical descent theorem for faithful monads on triangulated categories that underlies the comonadic descent step.","marker":"[5]"},{"why":"Provides the coherence and points theorems that turn the small site into a sober topological space with enough points.","marker":"[43]"}],"fun_headline_variants":["Faithful noncommutative spectra recover rings","Noncommutative Gelfand duality: rings from spectra","A faithful spectrum for all rings via derived sites","Rings recoverable from noncommutative spectrum functor"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction stands or falls on one premise: the chosen classes of 'open subsets' and 'covers' truly satisfy the axioms of a Grothendieck topology; the proof relies on the stability of homotopical epimorphisms under homotopy pushouts and on the homotopical pullback lemma, and any morphism where those fail would destroy the spectrum functor.","fun_headline_variants_meta":{"raw":{"variants":["Faithful noncommutative spectra recover rings","Noncommutative Gelfand duality: rings from spectra","A faithful spectrum for all rings via derived sites","Rings recoverable from noncommutative spectrum functor"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00039,"raw_usage":{"total_tokens":2085,"prompt_tokens":1010,"completion_tokens":1075,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":626,"completion_tokens_details":{"reasoning_tokens":1010}},"tokens_in":626,"tokens_out":1075,"duration_ms":9053,"temperature":1.0,"reasoning_tokens":1010,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T18:09:21.908972+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a concrete noncommutative connective dg-algebra $A$ and a homotopy pushout square $A \\to B$, $A \\to C$ in $\\mathrm{HRings}_{\\mathbb{Z}}$; if the induced map $B \\to B \\ast^{\\mathrm{L}}_A C$ is not a homotopical epimorphism, or if two morphisms $C \\rightrightarrows D$ in $\\mathrm{HRings}_B$ become equivalent after every base change along a cover but are not equivalent directly, then the topology axioms of Proposition 4.4 fail and the spectrum functor $\\operatorname{Spec}^{\\mathrm{NC}}$ would not be functorial.","supporting_citations":[{"cited_title":"Chuang, A","cited_arxiv_id":null,"evidence_quote":"Proves homological and homotopical epimorphisms coincide, the equivalence that legitimises homotopical epimorphisms as the open immersions of the topology."},{"cited_title":"Brave New","cited_arxiv_id":null,"evidence_quote":"Shows that flat epimorphisms of finite presentation between commutative rings are homotopical epimorphisms, giving the compatibility with the classical Zariski topology."},{"cited_title":"Lurie, Higher topos theory , Princeton University Press, (2009)","cited_arxiv_id":null,"evidence_quote":"Supplies the homotopical pullback lemma used in Proposition 4.4 to verify that the proposed classes form a Grothendieck topology."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Formulates the no-go theorem that no spectrum functor to sets can simultaneously be nonempty, functorial, and extend the prime-ideal spectrum, the obstruction this framework is designed to bypass."},{"cited_title":"Angeleri H¨ ugel, F","cited_arxiv_id":null,"evidence_quote":"Classifies homotopical and flat epimorphisms of commutative Noetherian rings, used to compute spectra and to compare them with the Grothendieck spectrum."},{"cited_title":"Neeman, M","cited_arxiv_id":null,"evidence_quote":"Identifies smashing localizations of the module category of a commutative ring with sets of primes, underpinning the commutative comparison of spectra."},{"cited_title":"Mathew, The Galois group of a stable homotopy theory , Adv","cited_arxiv_id":null,"evidence_quote":"Provides the commutative notion of descendability that the paper generalises; its terminology and descent arguments are adapted for the noncommutative setting."},{"cited_title":"Scholze, Condensed mathematics , Lecture notes based on joint work with D","cited_arxiv_id":null,"evidence_quote":"Gives the categorical Barr--Beck type conditions (Proposition 10.5) used to decide when local-to-global reconstruction is comonadic in Section 7."},{"cited_title":"Balmer, Descent in triangulated categories , Math","cited_arxiv_id":null,"evidence_quote":"Supplies the classical descent theorem for faithful monads on triangulated categories that underlies the comonadic descent step."},{"cited_title":"Mac Lane, I","cited_arxiv_id":null,"evidence_quote":"Provides the coherence and points theorems that turn the small site into a sober topological space with enough points."}],"review_version":1}