{"id":"3c721e27-c0b8-4a97-9039-9328c88279f0","arxiv_id":"2608.06120","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Quantalic linear lambda-calculus gains an additive disjunction operator with sound and approximately complete equational reasoning, plus probabilistic and quantum model constructions.","lead":"This paper adds case statements (if-then-else) to a quantitative lambda-calculus in which equality is replaced by distance in a quantale. It proves the extended rules sound and approximately complete, and uses Banach-space and quantum-channel models to reason about approximations in random walks.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Approximate completeness rests on an unproved enrichment of coproducts in Syn(T); the inequality shown in Theorem 3.13 Step 2 does not by itself establish the model condition.","rationale":"I read the paper in good faith and found no clear mathematical contradiction in the statements: the Banach-space and Karoubi-envelope models are coherent, and the continuity assumption on the quantale is explicit rather than hidden. However, the approximate-completeness proof is not self-contained: Theorem 3.13 Step 2 asserts that Syn(T) is V-Cat-sep-co-Cartesian but only sketches the key enrichment inequality, omitting quotient invariance, the enriched universal property, and the full soundness verification of the new case rule in general (possibly non-integral) quantales. The reader's conditional verdict already reflects this pattern of deferred proofs; my stress test sharpens it to the precise load-bearing point: the step where the syntactic category is shown to be a model. I therefore do not move the verdict, but I would ask the authors to supply the missing proof or a machine-checked development before the approximate-completeness claim is treated as fully established. The continuity assumption itself is not the problem; the problem is that the proof steps that lean on it are not written out in enough detail to certify them.","tokens_in":25278,"tokens_out":40386,"duration_ms":362179,"concrete_test":"Formalize Theorem 3.13 Step 2 in full: (a) prove with Q,R directed that (∨Q)∧(∨R)=∨_{q∈Q,r∈R}(q∧r); (b) show the co-pairing map is well-defined on Syn(T)/∼ and that the final rule of Figure 3 yields [t,s]≤_{q∧r}[t',s'] from t≤_q t' and s≤_r s'; (c) give the complete soundness chain for that rule using only Definitions 3.6–3.8, without assuming integrality of V. If the chain requires an additional hypothesis, either strengthen Definition 3.8 or restrict Theorem 3.13 accordingly.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The approximate-completeness theorem (Theorem 3.13) is the paper's central positive claim, and its proof depends on showing that the syntactic category Syn(T) is a V-Cat-sep-co-Cartesian autonomous model of T. Step 2 supplies one inequality, a(t,t')∧a(s,s') ≤ a([t,s],[t',s']), which is exactly the V-functor condition for the co-pairing map. But to conclude that the resulting coproducts are enriched in the sense of Definition 3.8, the paper must also show that this map is well-defined on the separated quotient, that the injections and mediating morphisms satisfy the enriched universal property, and that the last rule of Figure 3 used to derive the label q∧r is sound for arbitrary continuous quantales, not only integral ones. None of these steps is written out. In particular, the rule combines the distance p of the scrutinees with the product distance q∧r of the branches; verifying this in a general V-Cat-co-Cartesian category requires a chain of inequalities through composition, tensor, and the co-pairing enrichment, and that chain is not present. If any link fails, Theorem 3.13 collapses at this point. This is a verification gap rather than a demonstrated error, but it is load-bearing because the syntactic model is the only witness used to derive the approximate-completeness conclusion.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper extends the quantalic linear lambda-calculus of [DN22, DN23b] with additive disjunction and case statements. It adds a quantalic (in)equational system (Figure 3), defines V-Cat-co-Cartesian autonomous categories as the semantic setting, and claims soundness (Theorem 3.12) and approximate completeness over continuous quantales (Theorem 3.13). The paper also presents several classes of models, including generalized metric spaces, gluing, enriched presheaves, Banach spaces, and a Karoubi-envelope category of quantum channels, and illustrates the framework on quantitative reasoning about random walks.","tokens_in":25536,"tokens_out":22493,"duration_ms":200215,"significance":"If the central proofs are completed, this is a worthwhile contribution: it gives a finitary quantitative axiomatization of additive disjunction, an approximate-completeness result that avoids the Archimedean rule, and several nontrivial models with potential applications in probabilistic and quantum programming. The Lindenbaum-Tarski strategy is standard and non-circular, and the continuity assumption on the quantale is used in a principled way. The Banach-space and quantum-channel constructions are concrete and give useful evidence that the notion of model is not vacuous. However, the current manuscript leaves two load-bearing verification steps at the level of assertion, so the advertised soundness and approximate-completeness results are not yet fully established.","major_comments":[{"comment":"The proof that Syn(T) is a V-Cat^sep-co-Cartesian model of T is incomplete. The displayed chain in Step 2 verifies only the inequality a(t,t')∧a(s,s') ≤ a([t,s],[t',s']), i.e. the V-functor condition for the co-pairing map in Eq. (2). It does not verify that this map is well defined on the separated quotient, that the injections are V-functors, that the type A⊕B together with the injections and case elimination satisfies the enriched universal property, or that the final rule of Figure 3 is sound in an arbitrary V-Cat-co-Cartesian autonomous category, including non-integral quantales. Moreover, the displayed derivation needs an explicit use of the case rule with the unit label k to obtain [t,s]≤_{q∧r}[t',s'] and then the weak rule to pass to q; this step is absent. Since Syn(T) is the only model used to derive approximate completeness, this is a load-bearing gap.","section":"§3.3 (Theorem 3.13, Step 2)"},{"comment":"Soundness is asserted with the one-line proof 'follows straightforwardly by induction'. The new case rule in Figure 3, whose label is p⊗(q∧r), must be justified by a chain through enriched composition, enriched tensor, the V-functor property of co-pairing (Eq. (2)), and the distributivity isomorphism in a co-Cartesian autonomous category. No such chain is supplied. This matters not only for Theorem 3.12 itself but also for Theorem 3.13, because the same rule is used in the construction of the syntactic model. The reader is left without a verification of a central claim of the paper.","section":"§3.3 (Theorem 3.12)"}],"minor_comments":[{"comment":"The substitutions t[z/a] and s[w/b] are not consistent with the free-variable conventions used in Eq. (4), where a common fresh variable is substituted into two single-variable judgments; please rewrite this part with clear variables, for example t[z/x] and s[z/y], and avoid reusing a and b both as term names and as variable names.","section":"§3.3, Step 2 notation"},{"comment":"There is a typo: 'The latter is also a a basis' should read 'a basis'.","section":"§2.2, Example 2.3"},{"comment":"'A Vλ-theory ((G,Σ), Ax) is a triple' should be 'is a pair' or 'is data consisting of'; there are only two components displayed.","section":"§3.2, Definition 3.9"},{"comment":"The diagram of reflections and adjunctions involving V-Cat, V-Cat^sep, V-Cat^sym, and V-Cat^{sym,sep} is garbled in the text and should be redrawn with explicit functor directions and labels.","section":"§4.1"},{"comment":"The displayed limit line contains 'µ(-∞, q_n)' without the interval; it should read µ((-∞,q_n)). Also, 'linearcontractions' in §5.1 is missing a space.","section":"§5.1, Example 5.2"},{"comment":"The running title 'QUANT ALIC' appears to be a misspelling of 'QUANTALIC'.","section":"Title"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans heavily on the authors' prior work [DN22, DN23b] for the construction of Syn(T). I recommend that the editor require the authors to state precisely which parts of that construction are reused unchanged and which parts are new, and to supply the missing verification in Theorem 3.13 Step 2 and the missing soundness derivation for the case rule. I see no circularity problem: the Lindenbaum-Tarski construction is the standard technique. The concern is purely that central proof steps are currently asserted rather than demonstrated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real extension, not a repackaging. The earlier quantalic lambda-calculus line [DN22, DN23b] had no additives; here the authors add ⊕ with the right enrichment discipline—multiplicative structure enriched over ⊗, additive coproducts enriched over the Cartesian product structure of V-Cat. The approximate-completeness result over continuous quantales, with the Archimedean rule dropped, is the right theorem to aim for, and the Banach-space and quantum-channel models are concrete enough to show the framework does work.\n\nWhat's good: the worked random-walk example actually uses the equational system, not just the semantics; the Cauchy-sequence argument is coherent. The Karoubi envelope construction for CPTP is a genuine fix for the absence of coproducts, and the enrichment proof with the diamond norm is nontrivial. The gluing and enriched-presheaf sections are sketched but point to useful tools.\n\nSoft spots, in proportion: the two central theorems are not fully proved. Theorem 3.12 soundness is dismissed with 'straightforward induction,' but the new case rule needs a real verification in arbitrary V-Cat-sep-co-Cartesian autonomous categories—the composition/tensor chain for the label p⊗(q∧r) isn't shown. Theorem 3.13's Step 2 proves the key co-pairing inequality, which is the heart of the V-functor condition, but leaves well-definedness on the separated quotient, the injections, and the universal property implicit. I think those are routine and probably fine, but the paper should say so instead of leaving them as an exercise. The stress-test's worry that the last rule's soundness is load-bearing for Theorem 3.13 is off: that rule's soundness belongs to Theorem 3.12, not to the syntactic model construction. So the gap is real but not a demonstrated error.\n\nThe circularity burden is low: Lindenbaum-Tarski is the standard tool, and the same-author prior results are peer-reviewed. I did not find a fatal flaw.\n\nThis paper deserves a serious referee. I'd send it to review and ask for the omitted proofs, especially in Sections 3.2–3.3, before acceptance.","headline":"Genuinely new additive disjunction for quantalic lambda-calculus with solid model constructions, but the central soundness and approximate-completeness proofs are sketches that need to be written out before I would trust them.","tokens_in":26055,"tokens_out":7224,"would_cite":true,"duration_ms":62146,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["03B70","03F52","18D20","68Q55"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper extends quantalic linear lambda-calculus with additive disjunction and proves the resulting (in)equational system sound for enriched co-Cartesian autonomous categories and approximately complete over continuous quantales.","keywords":["quantalic lambda-calculus","additive disjunction","case statements","quantale","approximate completeness","enriched category theory","generalised metric spaces","probabilistic programming"],"falsifier":"Check the proof of Theorem 3.13 at the step where coproduct enrichment in Syn(T) is proved: compute whether sup{q∧r : q∈Q, r∈R} equals (sup Q)∧(sup R) for directed sets Q,R of labels. This identity is exactly what continuity supplies via q∧(−), so a non-continuous quantale where it fails would show approximate completeness cannot extend beyond the stated continuity assumption.","tokens_in":24999,"feed_emoji":"🧮","tokens_out":9588,"duration_ms":89207,"temperature":0.7,"pith_summary":"The paper extends quantalic linear $\\lambda$-calculus — an equational system in which every (in)equation $\\Gamma \\vdash t \\leq_q s$ is labelled by an element $q$ of a quantale, so that the label behaves like a distance between terms — with additive disjunction $\\oplus$, the type constructor that gives case statements. Its main results are soundness and approximate completeness: the extended system is sound for $\\mathcal{V}$-Cat-separated co-Cartesian autonomous categories (categories whose morphism sets are separated $\\mathcal{V}$-categories and whose coproducts respect the product enrichment), and over continuous quantales any (in)equation true in all models is provable at every sufficiently close approximation $r \\ll q$ of its label $q$. The new congruence rule for case statements compounds labels as $p \\otimes (q \\wedge r)$, a worst-case combination of the test's distance and the two branch distances. The paper exhibits models based on generalised metric spaces, gluing, enriched presheaves, Banach spaces, and a Karoubi envelope of quantum channels, and uses the Banach-space model to justify reasoning about Cauchy sequences of random walks.","feed_headline":"Quantalic lambda calculus adds case statements, stays sound","feed_subtitle":"Proof system for quantitative conditionals is sound and approximately complete over continuous quantales.","key_machinery":"The central object is the quantale $\\mathcal{V}$ with its way-below relation $\\ll$, together with the new congruence rule for case statements at the bottom of Figure 3. That rule states that from $t \\leq_p s$, $u \\leq_q v$, and $w \\leq_r o$ one may infer $\\mathrm{case}\\, t\\{ \\ldots \\} \\leq_{p \\otimes (q \\wedge r)} \\mathrm{case}\\, s\\{ \\ldots \\}$; the label $p \\otimes (q \\wedge r)$ encodes a worst-case assumption, since the branch whose terms differ more is controlled by the smaller of $q$ and $r$, compounded with the distance $p$ between the tested terms. The other load-bearing piece is the basis $B$ of the continuous lattice: hom-objects of the syntactic category $\\mathrm{Syn}(T)$ are formed by taking suprema over labels $q$ for which $t \\leq_q s$ is a theorem, and rule (join) keeps those sets directed, so an approximation $r \\ll q$ can be extracted as a finite join of theorems. Continuity of $\\mathcal{V}$ enters through the preservation of directed suprema by $q \\wedge (-)$, which makes the coproducts of $\\mathrm{Syn}(T)$ enriched over products.","core_discovery":"The central claim is that the quantalic (in)equational system obtained by adding additive disjunction to quantalic linear $\\lambda$-calculus is sound and approximately complete: Theorem 3.12 states that every provable inequation is satisfied in every model, and Theorem 3.13 states that over a continuous quantale, if $\\Gamma \\vdash t \\leq_q s$ holds in all models of a $\\mathcal{V}\\lambda$-theory $T$, then for every basis element $r \\ll q$ the inequation $\\Gamma \\vdash t \\leq_r s$ is a theorem of $T$. This recovers ordinary completeness when $q$ is compact, and it deliberately avoids the Archimedean rule, which involves infinitely many premisses. The proof builds a syntactic category $\\mathrm{Syn}(T)$ whose hom-objects are separated $\\mathcal{V}$-categories of terms, shows its coproducts are enriched over products using the distributivity of $q \\wedge (-)$ over directed suprema, and then uses the way-below relation together with the (join) rule to extract a finite join of theorems from semantic validity.","pith_inferences":["A natural testable extension the paper leaves implicit is additive conjunction $\\&$: the same enriched-coproduct machinery should dualize to products, with the worst-case label presumably flipping from $q \\wedge r$ to $q \\vee r$, though the paper does not prove this.","The $p \\otimes (q \\wedge r)$ rule is a template for quantitative case analysis beyond lambda-calculus: any deductive system whose labels form a continuous quantale could adopt the same rule and inherit approximate completeness, provided its syntactic category supports a directed join of premises.","The quantum model suggests a metric treatment of measurement as a conditional: once idempotents are split via the Karoubi envelope, a qubit measurement type $2 \\to 1 + 1$ becomes representable, and the paper's random-walk approximation technique could be repeated for quantum walks with distance bounds given by the diamond norm."],"forward_implications":["Quantitative extensionality for disjunctive types becomes a theorem: if $t$ restricted to the left injection is within $q$ of $s$, and $t$ restricted to the right injection is within $r$ of $s$, then the unrestricted terms are within $q \\wedge r$.","Over the metric quantale, semantic verification of an approximation can be converted into syntactic proofs up to arbitrary rational precision, because the extended non-negative rationals form a basis of the quantale.","The gluing construction extends to the enriched quantalic setting, enabling categorical study of meta-theoretic properties such as conservativity and concrete completeness for the extended calculus.","In the Banach-space model, perturbations of predicates or actions propagate to $k$-step random walks with explicit metric bounds, so the calculus supports certified approximations of probabilistic programs.","In the quantum setting, the Karoubi envelope of quantum channels supplies the missing coproducts, so conditionals and higher-order structure coexist in a metric-enriched model."],"supporting_citations":[{"why":"Introduces quantalic linear lambda-calculus and its enriched categorical semantics, which this paper extends with additive disjunction.","marker":"[DN22]"},{"why":"Establishes the syntactic category Syn(T) and the V-Cat-autonomous structure of the original calculus, reused in the approximate-completeness proof.","marker":"[DN23b]"},{"why":"Supplies the way-below relation, continuity of lattices, and the fact that q∧(−) preserves directed suprema in continuous lattices.","marker":"[GHK+03]"},{"why":"Provides V-categories and generalised metric spaces as the basis of enrichment used throughout the semantics.","marker":"[HST14]"},{"why":"Gives the classical case-statement rules and co-Cartesian autonomous category axiomatics that are quantalized here.","marker":"[Cro93]"},{"why":"Shows the gluing construction for linear logic models, which the paper adapts to the enriched quantalic setting.","marker":"[HS03]"},{"why":"Provides the reflection theorem that endows finite-product-preserving enriched presheaves with autonomous structure.","marker":"[Day72]"},{"why":"Gives the multi-sorted universal algebra framework used to define the left adjoint in the presheaf reflection.","marker":"[AR09]"},{"why":"Provides Kraus representation and the diamond norm used to enrich the quantum-channel model.","marker":"[Wat18]"}],"fun_headline_variants":["Quantalic lambda calculus quantifies case statements soundly","Sound extension: quantalic lambda calculus with additive disjunction","Quantitative case analysis via additive disjunction in lambda calculus","Approximate completeness for quantalic lambda calculus with additive disjunction","From equivalence to quantity: lambda calculus gains additive disjunction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the quantale's underlying lattice is continuous, so that intersecting with any fixed label preserves directed unions.","fun_headline_variants_meta":{"raw":{"variants":["Quantalic lambda calculus quantifies case statements soundly","Sound extension: quantalic lambda calculus with additive disjunction","Quantitative case analysis via additive disjunction in lambda calculus","Approximate completeness for quantalic lambda calculus with additive disjunction","From equivalence to quantity: lambda calculus gains additive disjunction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000827,"raw_usage":{"total_tokens":3590,"prompt_tokens":898,"completion_tokens":2692,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":514,"completion_tokens_details":{"reasoning_tokens":2611}},"tokens_in":514,"tokens_out":2692,"duration_ms":16987,"temperature":1.0,"reasoning_tokens":2611,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:31:11.648735+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Check the proof of Theorem 3.13 at the step where coproduct enrichment in Syn(T) is proved: compute whether sup{q∧r : q∈Q, r∈R} equals (sup Q)∧(sup R) for directed sets Q,R of labels. This identity is exactly what continuity supplies via q∧(−), so a non-continuous quantale where it fails would show approximate completeness cannot extend beyond the stated continuity assumption.","supporting_citations":[],"review_version":1}