{"id":"a6b17b60-c661-4f16-9252-09eab2ecbda0","arxiv_id":"1908.02189","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Left noetherian rings are exactly those whose modules form a superstable class with embeddings, and left pure-semisimple rings are exactly those whose modules form a superstable class with pure embeddings.","lead":"An abstract elementary class analysis shows that left noetherian rings and left pure-semisimple rings can be characterized by whether their module categories are superstable under embeddings or pure embeddings. The paper gives several equivalent model-theoretic conditions, connecting two classical algebraic notions to the stability hierarchy.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Pure-semisimple characterization hinges on unproven [KuMa] Facts 4.3 and 4.15; without them Lemma 4.16 and Theorem 4.26 (2)⇒(9) collapse.","rationale":"The reader's weakest assumption identifies exactly the load-bearing dependency: Theorem 4.28's proof routes through Theorem 4.26, whose Lemma 4.16 depends on Facts 4.3 and 4.15 from [KuMa]. I agree that this is the most serious concern, because those facts are used in both directions of the pure-semisimple equivalence: to show that uniqueness at one cardinal forces all limit models to be Σ-pure-injective (Lemma 4.16), and to show that a Σ-pure-injective universal model yields uniqueness of limit models (Theorem 4.26, (8)⇒(2)). If Fact 4.3 were false, the transfer of Σ-pure-injectivity via elementary equivalence would fail and Theorem 4.28 would not follow. The reader's CONDITIONAL verdict is appropriate: the concern is real but not a demonstrated error, and it could be resolved by a revision that either proves these facts or gives precise, self-contained pointers to the corresponding results in [KuMa]. I also note the additional sketched proofs (Proposition 3.11, Lemma 4.6, Lemma 4.21, Theorem 4.36) and the questionable reliance on Fact 2.17.(3) in the (5)⇒(2) steps, but these are secondary to the dependence on Fact 4.3/4.15 for the central pure-semisimple characterization. Thus I recommend no change to the reader's verdict.","tokens_in":24561,"tokens_out":28273,"duration_ms":276208,"concrete_test":"Independently re-derive Fact 4.3 for KT = (R-Mod, ≤pp), the setting of Theorem 4.28, using only the lemmas proved in this paper (Fact 4.2, [KuMa, 3.16]) and standard pp-type technology, without citing [KuMa, 4.3]. Concretely, take an arbitrary (λ,α)-limit model M and (λ,β)-limit model N and verify Inv(M,φ,ψ) = Inv(N,φ,ψ) for all pp-formulas φ,ψ with ThR ⊢ ψ → φ, as in Lemma 4.12. If the derivation fails or silently requires closure under direct sums or an extra stability assumption not stated in Fact 4.3, then Lemma 4.16 has a hidden assumption and the pure-semisimple equivalence is not established as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 4.28 is derived from Theorem 4.26, and the critical step (2)⇒(9) inside Theorem 4.26 is Lemma 4.16. That lemma uses two facts imported from [KuMa] without proof: Fact 4.3 (any two limit models in KT are elementarily equivalent) and Fact 4.15 (in a class closed under direct sums, a (λ,ω)-limit model is isomorphic to N^(ℵ0) for N a (λ,|T|+)-limit model). From these, Lemma 4.16 concludes that every (λ,α)-limit model is Σ-pure-injective by transferring Σ-pure-injectivity from N* to an arbitrary limit model M via elementary equivalence (Fact 2.21). This transfer is not cosmetic: the direction (2)⇒(1) of Theorem 4.28 (superstability implies pure-semisimple) needs Lemma 4.16 to obtain condition (9), and the converse (1)⇒(2) also uses Fact 4.3 in the proof of (8)⇒(2) of Theorem 4.26. Fact 4.3 is highly nontrivial because in general AECs limit models of different lengths and bases need not be elementarily equivalent, yet the paper gives no proof and no explicit pointer to the exact statement in [KuMa] that covers the case T = ThR. The noetherian theorem has the same external dependency through Proposition 3.11, which is stated without proof as being analogous to [KuMa, 4.9]. The central claim is therefore only as secure as the correctness and applicability of those imported facts.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves two algebraic-model-theoretic equivalences. Theorem 3.12 shows that a ring R is left noetherian if and only if the AEC (R-Mod, ⊆R) of left R-modules with ordinary embeddings is superstable, giving seven equivalent conditions involving uniqueness of limit models, superlimits, stability, and Σ-injectivity of limit models. Theorem 4.28 shows that R is left pure-semisimple if and only if the AEC K_{Th_R} = (R-Mod, ≤pp) with pure embeddings is superstable, giving nine equivalent conditions. The pure-embedding results are obtained through a general study of superstable classes K_T = (Mod(T), ≤pp) closed under direct sums (Theorems 4.23 and 4.26), and the paper also gives a partial solution to Conjecture 2 of [BoVan] for such classes (Theorem 4.34, Corollary 4.35) and a version without closure under direct sums (Theorem 4.36).","tokens_in":24791,"tokens_out":9699,"duration_ms":94341,"significance":"If the theorems stand, they establish a genuine bridge between stability-theoretic dividing lines and classical ring theory, with remarkably precise cardinal bounds: the good behavior starts exactly at |R|+ℵ0 or (|R|+ℵ0)+, improving the general eventual bounds of [GrVas17, 1.3]. The paper is carefully written for both model theorists and algebraists, and the main algebraic characterizations are derived from classical facts (Fact 2.23 and Fact 2.25) rather than by post-hoc selection. The most substantial weakness is that the deepest pure-embedding steps are quoted without proof from the companion paper [KuMa], and the noetherian argument likewise omits the proof of Proposition 3.11; the correctness of those imports is load-bearing for the main theorems.","major_comments":[{"comment":"The proof of Theorem 4.26(2)⇒(9), and hence the pure-semisimple characterization in Theorem 4.28, depends on two facts imported from [KuMa] without proof: Fact 4.3 (any two limit models in K_T are elementarily equivalent) and Fact 4.15 (in a class closed under direct sums, a (λ,ω)-limit model is isomorphic to N^(ℵ0) for N a (λ,|T|+)-limit model). These are not immediate consequences of the AEC axioms stated in Section 2, and they are applied here to K_{Th_R}, whose hypotheses the manuscript does not verify against the exact statements in [KuMa]. Since Lemma 4.16 uses these facts to conclude that every limit model is Σ-pure-injective, and Theorem 4.26 then uses that conclusion to derive pure-semisimplicity, the imported results are load-bearing. Please include full proofs of Facts 4.3 and 4.15 in the paper or an appendix, or at minimum state the exact theorems from [KuMa] with all hypotheses and verify those hypotheses for K_{Th_R}.","section":"§4.2, Facts 4.3 and 4.15; Lemma 4.16"},{"comment":"Proposition 3.11 is a load-bearing input in the proof of Theorem 3.12: it is used directly in the proofs of (3)⇒(1) and (1)⇒(7) to identify N^(ℵ0) as a (χ,ω)-limit model. The manuscript says the proof is 'basically the same as that of [KuMa, 4.9]' and omits it. This omission matters because the ambient class here is (R-Mod, ⊆R) with ordinary embeddings, whereas [KuMa, 4.9] is stated for K_T with pure embeddings, and the precise isomorphism statement is essential for the subsequent Σ-injectivity arguments. Please provide a proof of Proposition 3.11, or give a precise statement-and-proof in the cited reference together with a verification that its hypotheses apply to (R-Mod, ⊆R).","section":"§3, Proposition 3.11 and proof of Theorem 3.12"}],"minor_comments":[{"comment":"In the proof of (3)⇒(4), the text says 'By condition (5) K_T is λ-stable', but condition (5) has not yet been derived; it should say 'By (3) and Lemma 4.18'.","section":"§4.2, proof of Theorem 4.23, (3)⇒(4)"},{"comment":"The reference list contains two works labeled [Zim79] with different spellings (Zimmermann and Zimmermann-Huisgen); the in-text citations to [Zim79] are ambiguous and should be disambiguated, for example as [Zim79a] and [Zim79b].","section":"References"},{"comment":"The abstract's Theorem 0.2 lists only four equivalences, while the full Theorem 4.28 states nine; consider harmonizing the abstract with the full theorem or adding a note that the full statement appears in Section 4.","section":"Abstract and Theorem 4.28"},{"comment":"The proof sketch of the backward direction of Lemma 3.2 would be easier to check if it explained why the map f is well-defined on finite sums with coefficients in R and why the repeated use of amalgamation preserves the quantifier-free type; the phrase 'applying amalgamation a couple of times' is vague.","section":"Lemma 3.2"},{"comment":"There is a typo: 'the above theorem improves the bounds where the nice propertis show up' should read 'properties'.","section":"Remark 3.13"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans very heavily on [KuMa], a companion paper co-authored by the author. Before acceptance, I recommend verifying that [KuMa] is either published or available in a stable, accessible form, and that the exact statements of Facts 4.3 and 4.15 can be checked by a referee. The main theorems are plausible and the exposition is good, but as submitted the paper is not self-contained at precisely the points where the mathematics is hardest."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a real result, not a repackaging. Theorems 3.12 and 4.28 give new characterizations of left noetherian and left pure-semisimple rings in terms of superstability and limit models of module classes, with bounds at |R| + ℵ0 instead of the much larger bounds from [GrVas17]. The paper is transparent that Vasey independently had part of the noetherian equivalence and Shelah had related remarks; the full equivalence lists, the pure-semisimple characterization, and the general theory of superstable classes KT closed under direct sums are the genuinely new content.\n\nWhat the paper does well: the noetherian proof is mostly self-contained and fairly elegant—identifying long limit models with injective modules and then using Σ-injectivity to force uniqueness is clean. The pure-semisimple section builds a useful general framework around the theory T̃, connecting stability of KT with stability of the complete theory T̃, and then Theorem 4.28 drops out as a corollary. The background material is well chosen for the intended mixed audience.\n\nWhere the soft spots are: the stress-test note is right that the pure-semisimple direction leans heavily on two imported facts from [KuMa], namely Fact 4.3 (limit models in KT are elementarily equivalent) and Fact 4.15 (in classes closed under direct sums, a (λ,ω)-limit model is N^(ℵ0) for N a (λ,|T|+)-limit model). Lemma 4.16 and direction (8)→(2) of Theorem 4.26 really need those. The noetherian theorem has a similar dependency at Proposition 3.11, which is stated without proof as being analogous to [KuMa, 4.9]. Since [KuMa] is accepted, this is not fatal, but it is a real soft spot: a referee should verify that those statements hold in exactly the needed generality, especially Fact 4.3, which is not true for arbitrary AECs. Lemma 4.6 and Theorem 4.36 are also sketched; the latter is peripheral to the ring-theoretic main theorems, but still worth tightening.\n\nI do not see a load-bearing flaw in the central argument. The characterizations are coherent, the proofs that are present are careful, and the citations to Shelah and Vasey are honest rather than buried. The main theorems are conditional on cited results, but that is normal mathematical practice when the cited results are accepted.\n\nBottom line: this deserves serious peer review. I would send it to a referee who knows both AEC limit-model theory and module theory, ask them to check the [KuMa] dependencies carefully, and accept after that check. I would bring it to a reading group and would cite the main theorems if I worked in this area.","headline":"Genuine new equivalences for noetherian and pure-semisimple rings via superstability of module classes; the main theorems are solid, but the referee needs to check the load-bearing facts imported from [KuMa].","tokens_in":25448,"tokens_out":3245,"would_cite":true,"duration_ms":34869,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03C48","16B70","03C45","03C60","13L05","16P40","16D10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Superstability of module classes characterizes noetherian and pure-semisimple rings.","keywords":["Superstability","Noetherian ring","Pure-semisimple ring","Limit models","Abstract elementary classes","Pure embeddings","Sigma-pure-injective modules"],"falsifier":"Construct a ring $R$ and a class $K_T$ of left $R$-modules with pure embeddings, closed under direct sums, such that $K_T$ has uniqueness of limit models at some $\\lambda \\geq (|R| + \\aleph_0)^+$ but some module in $K_T$ is not pure-injective; that would separate condition (3) from condition (1) of Theorem 4.28. Alternatively, exhibit any class of modules in which the $(\\mu, \\omega)$-limit model is not isomorphic to the countable direct sum of the $(\\mu, |T|^+)$-limit model, giving a concrete counterexample to Fact 4.15.","tokens_in":24245,"feed_emoji":"🧮","tokens_out":13210,"duration_ms":112080,"temperature":0.7,"pith_summary":"This paper proves that a model-theoretic dividing line, superstability, is not just a formal property of abstract elementary classes: it exactly detects two classical ring-theoretic properties. For any ring $R$, the class of left $R$-modules with embeddings is superstable precisely when $R$ is left noetherian (Theorem 3.12). The class of left $R$-modules with pure embeddings is superstable precisely when $R$ is left pure-semisimple (Theorem 4.28). The proof works by identifying the model-theoretic limit models in these classes with algebraic objects—injective modules in the first case and pure-injective modules in the second—and then showing that uniqueness of limit models at one large cardinal forces the strong $\\Sigma$ versions that imply the ring-theoretic condition. A reader should care because the equivalences are exact and the cardinals at which stability begins are computed: $|R| + \\aleph_0$ for noetherian, $(|R| + \\aleph_0)^+$ for pure-semisimple.","feed_headline":"Superstability characterizes noetherian and pure-semisimple rings","feed_subtitle":"For a ring R, noetherianity and pure-semisimplicity are equivalent to superstability of its modules, at exact cardinals","key_machinery":"The machinery is the limit model: a module of size $\\lambda$ obtained as the union of a chain of $\\lambda$-sized modules in which each successor is universal over its predecessor. In the pure-embedding setting the paper attaches to the class $K_T$ the complete first-order theory $\\tilde{T} = \\mathrm{Th}(\\tilde{M}_T)$, where $\\tilde{M}_T$ is the $(2^{|T|}, \\omega)$-limit model; this lets arguments pass between the abstract class and ordinary module theory. The load-bearing identity is Fact 4.15, quoted from [KuMa]: if $K_T$ is closed under direct sums, a $(\\lambda, \\omega)$-limit model is isomorphic to $N^{(\\aleph_0)}$ for any $(\\lambda, |T|^+)$-limit model $N$. Combined with purity facts, this identity converts uniqueness of limit models at a single cardinal into $\\Sigma$-pure-injectivity of all limit models, from which the ring-theoretic equivalences follow. For the noetherian theorem the key algebraic input is the existence criterion for injective universal modules quoted from [Ekl71, Proposition 3].","core_discovery":"The paper's central discovery is a pair of algebraic characterizations of superstability. Theorem 3.12 says that for a ring $R$ the following are equivalent: $R$ is left noetherian; the class $(R\\text{-Mod}, \\subseteq_R)$ of left $R$-modules with embeddings is superstable; for every $\\lambda \\geq |R| + \\aleph_0$ there is $\\chi \\geq \\lambda$ with uniqueness of limit models of size $\\chi$; the class is $\\lambda$-stable for every $\\lambda \\geq |R| + \\aleph_0$; and every limit model is $\\Sigma$-injective. Theorem 4.28 gives the analogous chain for pure-semisimplicity: $R$ is left pure-semisimple iff the class $K_{\\mathrm{Th}_R}$ of left $R$-modules with pure embeddings is superstable, iff there is a single $\\lambda \\geq (|R| + \\aleph_0)^+$ with uniqueness of limit models, iff every limit model is $\\Sigma$-pure-injective, and iff a $\\Sigma$-pure-injective universal model exists at some $\\lambda \\geq (|R| + \\aleph_0)^+$. The engine behind Theorem 4.28 is a general analysis of classes of modules with pure embeddings that are closed under direct sums: in such classes uniqueness of limit models at one cardinal propagates to all cardinals and forces every module to be pure-injective.","pith_inferences":["Editorial extension: the one-cardinal-to-everywhere propagation of Lemma 4.20 depends only on closure under direct sums and on long limit models being pure-injective; if those hold in other categories of modules, superstability of the class should again be equivalent to a $\\Sigma$-pure-injectivity statement, giving a template for new algebraic characterizations.","Editorial extension: condition (8) of Theorem 4.28 makes pure-semisimplicity testable by the existence of a single universal module; for finite-dimensional algebras this suggests a possible route toward the long-standing pure-semisimple conjecture, though the paper does not address representation type.","Editorial extension: Theorem 4.34 suggests a general principle that the limit-model spectrum of an algebraic abstract elementary class is controlled entirely by the complete theory $\\tilde{T}$; one could test this by computing $\\kappa(\\tilde{T})$ for concrete module classes and predicting exactly where non-uniqueness appears."],"forward_implications":["If $R$ is left pure-semisimple, then its pure-embedding class is $\\lambda$-stable for every $\\lambda \\geq |R| + \\aleph_0$ and has uniqueness of limit models and superlimits at all those cardinals, so the nice model-theoretic behavior starts exactly at the optimal cardinal.","If the pure-embedding class has uniqueness of limit models at even one cardinal $\\lambda \\geq (|R| + \\aleph_0)^+$, then every left $R$-module is pure-injective, so a single model-theoretic uniqueness statement carries the full algebraic content of pure-semisimplicity.","For a left noetherian ring, every limit model in the embedding class is $\\Sigma$-injective, which connects chain conditions on ideals to the structure of large modules built from universal chains.","For first-order axiomatizable classes of modules with joint embedding and amalgamation, the spectrum of limit models is eventually constant above $\\mathrm{LS}(K)^+$ (Corollary 4.35), giving a positive solution above $\\mathrm{LS}(K)^+$ to Conjecture 2 of [BoVan].","A ring is left pure-semisimple iff increasing chains of $\\lambda$-saturated models remain $\\lambda$-saturated for every large $\\lambda$, so a purely structural union property is equivalent to an algebraic finiteness condition."],"supporting_citations":[{"why":"Supplies Fact 4.15, the identity $N^{(\\aleph_0)} \\cong$ (the $(\\lambda,\\omega)$-limit model), which turns one-cardinal uniqueness into $\\Sigma$-pure-injectivity of every limit model.","marker":"[KuMa, 4.9]"},{"why":"Supplies Fact 4.14: long limit models in pure-embedding classes are pure-injective, the algebraic identification behind Theorem 4.28.","marker":"[KuMa, 4.5]"},{"why":"Supplies Fact 4.3: all limit models in $K_T$ are elementarily equivalent, letting the paper attach a single complete theory $\\tilde{T}$ to the class.","marker":"[KuMa, 4.3]"},{"why":"Stated as Fact 2.19: $M$ is $\\Sigma$-pure-injective iff $M^{(\\aleph_0)}$ is pure-injective, used to promote countable-direct-sum information to $\\Sigma$-pure-injectivity.","marker":"[Zim79, 3.4]"},{"why":"Stated as Fact 2.22: $\\Sigma$-pure-injective modules have $\\lambda$-stable theories, used in Lemma 4.18 to pass from one-cardinal uniqueness to global $\\lambda$-stability.","marker":"[Pre88, 3.2]"},{"why":"Stated as Fact 3.5, the existence criterion for an injective universal module in the embedding class, which is central to the noetherian theorem.","marker":"[Ekl71, Proposition 3]"},{"why":"Provides the general framework showing that several candidate definitions of superstability agree in tame abstract elementary classes; it motivates conditions (4) through (7) in Theorems 3.12 and 4.28.","marker":"[GrVas17, 1.3]"},{"why":"Stated as Fact 2.23: noetherianity is equivalent to every injective module being $\\Sigma$-injective, a key step in Theorem 3.12.","marker":"[Pre09, 4.4.17]"},{"why":"Stated as Fact 2.25: pure-semisimplicity is equivalent to every module being $\\Sigma$-pure-injective, which connects Theorem 4.26 to the ring-theoretic theorem.","marker":"[Pre88, 11.3]"},{"why":"Provides the Schr\\\"oder-Bernstein fact that injective, and analogously pure-injective, modules with mutual embeddings are isomorphic; this is used to prove uniqueness of limit models from injectivity or pure-injectivity.","marker":"[Bum65]"}],"fun_headline_variants":["Superstability captures noetherian and pure-semisimple rings","Superstability equals noetherian and pure-semisimple rings","Superstability unlocks noetherian and pure-semisimple rings","Superstability pins down noetherian and pure-semisimple rings","Superstability identifies noetherian and pure-semisimple rings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on a quoted technical fact, not proved in this paper: in any class of modules with pure embeddings that is closed under direct sums, a limit model built with countable steps is isomorphic to the countable direct sum of a limit model built with long steps; if that fact is wrong, the proof that one-cardinal uniqueness forces pure-semisimplicity collapses.","fun_headline_variants_meta":{"raw":{"variants":["Superstability captures noetherian and pure-semisimple rings","Superstability equals noetherian and pure-semisimple rings","Superstability unlocks noetherian and pure-semisimple rings","Superstability pins down noetherian and pure-semisimple rings","Superstability identifies noetherian and pure-semisimple rings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000989,"raw_usage":{"total_tokens":4310,"prompt_tokens":1182,"completion_tokens":3128,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":798,"completion_tokens_details":{"reasoning_tokens":3036}},"tokens_in":798,"tokens_out":3128,"duration_ms":50316,"temperature":1.0,"reasoning_tokens":3036,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:51:21.165785+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a ring $R$ and a class $K_T$ of left $R$-modules with pure embeddings, closed under direct sums, such that $K_T$ has uniqueness of limit models at some $\\lambda \\geq (|R| + \\aleph_0)^+$ but some module in $K_T$ is not pure-injective; that would separate condition (3) from condition (1) of Theorem 4.28. Alternatively, exhibit any class of modules in which the $(\\mu, \\omega)$-limit model is not isomorphic to the countable direct sum of the $(\\mu, |T|^+)$-limit model, giving a concrete counterexample to Fact 4.15.","supporting_citations":[],"review_version":1}