{"id":"d5f6fbb4-e63b-4ca7-a67f-1657d737896a","arxiv_id":"2602.12447","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For the 1D long-range ferromagnetic Ising model with J(r)=r^{-α} (1<α≤2), a convergent low-temperature cluster expansion is established and the two-point truncated correlation is shown to decay with the exact algebraic rate α.","lead":"A rigorous low-temperature cluster expansion is constructed for the one-dimensional long-range Ising model with interaction J(r)=r^{-α}, 1<α≤2. The paper proves the two-point correlation decays as a power of the distance with exponent α, answering a long-standing question in rigorous statistical mechanics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proof inherits Hypotheses 1–3 from [2] without proof; if any fails for a range of α, Theorems 1.1–1.3 collapse.","rationale":"After re-reading the main argument, I found no internal inconsistency or post-hoc fitting. The convergence proof is a careful tree-summing argument and the estimates appear to close provided Hypotheses 1–3 hold. However, these hypotheses are exactly the input from [2] and are not verified here. They are used at crucial steps: Hypothesis 1 in Lemma 3.8, Hypotheses 2–3 throughout Sections 5 and 7. Since the paper's strongest claims (Theorems 1.1–1.3) depend on the full interval α∈(1,2], any gap in [2] would invalidate them. The exact-α lower bound from [26] is a secondary external dependency. This is an external dependency rather than an internal flaw, but it is load-bearing. The reader's weakest_assumption correctly identifies this point. I therefore keep the CONDITIONAL verdict and propose a check that would give confidence (or reveal a counterexample) for the hypotheses.","tokens_in":47402,"tokens_out":13274,"duration_ms":105502,"concrete_test":"Run a numerical search over α∈{1.01,1.05,1.1,...,2} and over all contour pairs γ1,γ2 with H(γi)≤20 and distance up to 50, checking the quasi-additivity inequality H(γ1∪γ2) ≥ H(γ2)+7/8 H(γ1) (and symmetrically). If any violation occurs, Hypothesis 1 fails and Theorem 1.1 is invalid; if no violation within this range, the concern shifts to the full proof in [2] but is not refuted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2.4 states that Hypotheses 1–3 (quasi-additivity, contour counting, and exponential smallness) are load-bearing for the cluster expansion, but they are not proved here; they are attributed to [2]. These hypotheses are used in Lemma 3.8 to bound the polymer activity by an exponential of -βH, in Lemma 5.3 to control the one-contour leaf-pruning sum, in Corollary 5.8 for the contracting functions, and in Lemma 5.11 to close the polymer-tree sum. If Hypothesis 1 fails, the bound Σ_e Φ(e) ≤ (1/8)Σ_γ H(γ) in Lemma 3.8 fails, so the activity bound and hence the entire convergence proof in Theorem 1.1 break down. If Hypothesis 2 fails, the counting arguments in Corollaries 5.4, 5.8, and Lemmas 6.6, 7.3 fail. If Hypothesis 3 fails, the exponential sums over contour classes are not controlled. The theorems are therefore only as reliable as [2]. This is an external dependency, not an internal inconsistency, but it is the weakest point: if [2]'s proof has a hidden restriction (e.g., α>α* for some α*>1), the claimed all-α validity is not supported. The exact-α lower bound also relies on [26], a secondary dependency.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a convergent low-temperature cluster expansion for the one-dimensional ferromagnetic Ising model with interaction J(r)=r^{-α}, α∈(1,2]. The main results are: absolute convergence of the cluster expansion for log Z, analyticity in a magnetic-field polydisk (Theorem 1.1); an upper bound |<σ_x;σ_y>| ≤ 2 e^{-c β}|x−y|^{-α} (Theorem 1.2); and analogous tree-summation bounds for N-point truncated correlations (Theorem 1.3). Together with the known lower bound of Iagolnitzer–Souillard [26], Theorem 1.2 is claimed to give the exact algebraic decay rate α. The proof proceeds by constructing a polymer gas of positive Fröhlich–Spencer contours, bounding the polymer activity in Lemma 3.8, developing an abstract tree-summing formalism in Section 4, and then translating contour estimates into lattice-site estimates in Sections 6–7. The decisive input is the set of three 'Crucial estimates' (Hypotheses 1–3 in Section 2.4), which are not proved in this manuscript but are imported from the same-group preprint [2].","tokens_in":47717,"tokens_out":11139,"duration_ms":105476,"significance":"If the imported Hypotheses 1–3 are valid for all α∈(1,2], the paper gives a substantial extension of Imbrie's α=2 cluster expansion to the full phase-transition region of the one-dimensional long-range Ising model, without the artificial large nearest-neighbor interaction used in earlier contour/cluster works. The abstract weighted-tree lemmas (Section 4) are stated with explicit constants and are potentially reusable. The correlation bounds are falsifiable and quantitative. The manuscript is not machine-checked and contains no code, but it does provide a coherent chain of explicit estimates. Its significance, however, is conditional: the geometric energy-entropy core is outsourced to an unpublished preprint from the same group, so the paper cannot stand alone as a proof of the stated all-α theorem.","major_comments":[{"comment":"The three crucial estimates — quasi-additivity of contour energies, the contour-counting bound, and exponential smallness of the contour activity — are load-bearing for every main theorem, but they are not proved here. They are used in Lemma 3.8 (activity bound), Lemma 5.3 and Corollaries 5.4/5.8 (tree bounds), and Lemma 5.11 (closing the polymer-tree sum). If any of them fails for some α∈(1,2], the convergence proof of Theorem 1.1 and the correlation bounds in Theorems 1.2–1.3 break down. Since [2] is a same-group arXiv preprint, the manuscript's central claim 'α∈(1,2]' is not self-contained. Please either include proofs of Hypotheses 1–3 in an appendix or explicitly state Theorems 1.1–1.3 as conditional on these hypotheses.","section":"§2.4, Hypotheses 1–3; used in §3.3, §5.2, §5.3"},{"comment":"The passage from contour sums to site sums relies on Lemma 5.7, whose proof is terse at the point where the summation over B is replaced by factors f_k(B). Because Corollaries 5.8–5.10 and Lemmas 7.1/7.5 all depend on this inequality, the proof should spell out exactly how |B| and diam(B) compensate for the dropped b-sums and how the summation over the last A_i is counted. If this step is not correct, the site-translation estimates in §7 would not follow. I am not asserting a definite error, but this is a central step that needs fuller justification.","section":"§6–§7, Lemma 5.7 and its descendants"}],"minor_comments":[{"comment":"The claim 'decay rate exactly α' depends on the external lower bound [26]. Please state the precise form of that bound, including its temperature/constant assumptions, so the combination of Theorem 1.2 and [26] is unambiguous.","section":"§1.2 / statement of Theorem 1.2"},{"comment":"The phrase 'αPp1,3´3 log 2s' appears to contain a typo; the text elsewhere uses 3−log_2 3. Please correct the formula.","section":"§8, Concluding Remarks"},{"comment":"The notation 'E_φ' and the ordering condition are not fully explained; in particular, the displayed inequality after 'Finally, since dist(...)' skips the bookkeeping of the b-summation. Expanding this two-line argument would greatly improve readability.","section":"Lemma 5.7 proof, around Eq. (58)"},{"comment":"A few notational inconsistencies (e.g. 'CΛ' vs 'C_Λ', 'H˚h,γ' subscripts, and the use of 'Sym(k)' vs 'Sympkq') should be cleaned up. They do not affect the mathematics.","section":"Notation, §2.3–2.4"}],"recommendation":"major_revision","confidential_remarks":"The core mathematical architecture appears coherent, and the abstract tree-summing framework is a real strength. However, the all-α claim rests on Hypotheses 1–3 from a preprint by the same group. If the journal accepts reliance on that preprint as standard, the paper could be close to acceptance after local clarifications; otherwise the authors must either prove the hypotheses or reword the theorems as conditional. I recommend major revision rather than rejection because the dependence is explicit and could in principle be repaired within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper gives the first convergent low-temperature cluster expansion for the one-dimensional long-range ferromagnetic Ising model over the whole α∈(1,2] range, and uses it to prove the two-point truncated correlation decays as |x−y|^{−α}, matching the old Iagolnitzer–Souillard lower bound. Imbrie had α=2; Cassandro–Merola–Picco–Rozikov had a restricted interval. The abstract tree machinery in Sections 4–5 is the real content: they introduce trees of contours with global compatibility, prove many-vertex contraction estimates with explicit constants (Theorem 4.21), and translate the polymer sums cleanly into lattice-site sums. The constants are explicit and the logic is coherent. This is careful, honest mathematical physics.\n\nThe soft spot is the one the reader flags: Hypotheses 1–3—quasi-additivity, the counting bound on contours, and exponential smallness of contour activity—are imported from [2], a preprint by the same group, and are not proved here. Section 2.4 states this plainly. So the convergence theorem is conditional: if [2] has a hidden restriction on α, Theorems 1.1–1.3 collapse. That is a real external dependency, and the ideal version of this paper would either reproduce those estimates or give theorem-numbered citations in the companion paper. It is not circularity—the hypotheses are prior independent statements, not the target result. The lower bound from [26] is a secondary external input, but that is classical and not a concern.\n\nOtherwise the paper reads straight. The bounds for the activity (Lemma 3.8), the leaf-pruning and total-edge estimates, and the contour-to-site lemmas are all presented with explicit constants. The many-point correlation bound in Theorem 1.3 is a nice bonus. I do not see post-hoc fitting or unsupported claims beyond the imported estimates.\n\nWho gets value: anyone working on contour methods or cluster expansions for long-range interactions, and especially readers using the multidimensional companion [6]. It should be refereed, not desk-rejected; the referee should verify the tree estimates and check exactly which statements of [2] are being used. Recommend: send out.","headline":"Full-α cluster expansion and exact α decay rate for the 1D long-range Ising model, built on the authors' own contour estimates—serious work, referee it.","tokens_in":48224,"tokens_out":2702,"would_cite":true,"duration_ms":25214,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B20","82B05","82B26"],"pacs":["05.50.+q"],"model":"deepseek-v4-flash","headline":"At low temperature, the 1D long-range Ising model's two-point correlation decays at the exact rate |x-y|^{-α} for all α∈(1,2].","keywords":["long-range Ising model","cluster expansion","M-contours","polymer gas","power-law decay","low-temperature correlations","one-dimensional lattice models"],"falsifier":"Find, for some α∈(1,2], a contour or collection of contours violating Hypothesis 1, e.g., H(Γ∖γ)+(7/8)H(γ)>H(Γ), or violating Hypothesis 3, e.g., a contour with anomalously low energy relative to its size; such an example would falsify Lemma 3.8's activity bound and, with it, Theorems 1.1–1.3. A direct numerical check of these hypotheses at α close to 1 for large β would be a concrete way to look for it.","tokens_in":47259,"feed_emoji":"🧲","tokens_out":7975,"duration_ms":72168,"temperature":0.7,"pith_summary":"This paper proves that the one-dimensional long-range ferromagnetic Ising model with interaction J(r)=r^{-α}, for every α in (1,2], has a convergent low-temperature cluster expansion in terms of a hard-core gas of polymers built from the model's M-contours. The convergence is absolute and the expansion is analytic in a polydisk of external fields. As an application, the paper shows that the two-point truncated correlation decays as |x-y|^{-α}; combined with a lower bound known since the 1970s, the decay rate is exactly α. It also gives explicit tree-graph upper bounds for N-point truncated correlations. The whole construction is carried out without imposing an artificially large nearest-neighbor coupling, covering the full region where the 1D model is known to order.","feed_headline":"1D long-range Ising correlations decay exactly like 1/r^α","feed_subtitle":"Low-temperature cluster expansion pins two-point decay at the interaction exponent for α∈(1,2].","key_machinery":"The central objects are M-contours: finite collections of spin flips that cannot be split into two pieces spaced farther apart than M times the smaller diameter to the 3/2 power; they play the role of local excitations with controlled energy and count. The paper packages compatible M-contours into polymers via a positive-compatibility relation and rewrites the partition function as a hard-core polymer gas. Each polymer's activity is bounded by z̄_β(Γ)=∏_{γ∈Γ} e^{-βH(γ)/2} Σ_{T∈T(Γ)} ∏_{edges} Φ(edge), with Φ the interaction between minus-interiors. The convergence argument is carried by a hierarchy of tree-sum estimates—one-vertex, two-vertex, and many-vertex—that reduce sums over trees of p","core_discovery":"The claim is that at sufficiently low temperature the partition function of the 1D long-range Ising model with J(r)=r^{-α}, 1<α≤2, admits a genuine convergent cluster expansion: log Z_{Λ,β} is an absolutely convergent series over polymers built from M-contours, and the same series is analytic in a polydisk of external fields. From this expansion the paper derives concrete correlation bounds: the two-point truncated correlation is at most 2e^{-c₃β}|x-y|^{-α}, which, together with a lower bound known since the 1970s, makes the decay rate exactly α. For N points, the correlation is bounded by a sum over trees of products of |a-b|^{-α}.","pith_inferences":["The theorems inherit their validity from the three contour estimates stated in §2.4 and proved in a companion paper; the present text does not reproduce those proofs, so the reader should check them for the full interval (1,2] before relying on Theorem 1.1.","The site-estimate lemmas, such as the product-to-tree inequality for lattice points, are stated quite generally and look reusable for nearby problems—random-field perturbations, decaying fields, or phase separation in 1D long-range models—without repeating the contour analysis.","At α=2 the upper bound matches the critical decay; a known intermediate phase with a slower algebraic exponent exists elsewhere in the phase diagram, so the exact-α statement should be read as a low-temperature statement, not a claim about all β.","The argument uses α>1 through summability of 1/r^α and cannot extend to α≤1, where no long-range order occurs; the behavior right at α=1 is therefore untouched by this method."],"forward_implications":["For every α∈(1,2] and sufficiently large β, the two-point truncated correlation has the exact low-temperature decay |x-y|^{-α}; the upper bound supplied here complements a lower bound known since the 1970s.","The pressure series converges absolutely and is analytic in an external-field polydisk, so thermodynamic derivatives such as magnetization, susceptibilities, and n-point correlations may be computed term by term.","N-point truncated correlations obey a tree-graph bound with weights |a-b|^{-α} on each tree edge—the long-range analogue of the usual exponential decay bound for short-range models.","The cluster expansion covers the entire phase-transition interval 1<α≤2 without imposing an artificially large nearest-neighbor coupling, removing an assumption used in earlier contour proofs.","The same expansion provides a rigorous starting point for further low-temperature quantities, such as phase-separation points and surface tensions in the long-range 1D model."],"fun_headline_variants":["Cluster expansion proves exact 1/r^α decay in 1D Ising","Low-T cluster expansion yields exact α decay for 1D Ising","Ising decay exponent pinned to α by cluster expansion","At low T, cluster expansion fixes Ising decay at 1/r^α"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the three contour estimates of §2.4 (quasi-additivity, a counting bound, and exponential smallness of activity) hold for every α∈(1,2]; the paper states them as hypotheses and cites a companion paper for their proof, and if any of them fails the cluster expansion and the correlation theorems collapse.","fun_headline_variants_meta":{"raw":{"variants":["Cluster expansion proves exact 1/r^α decay in 1D Ising","Low-T cluster expansion yields exact α decay for 1D Ising","Ising decay exponent pinned to α by cluster expansion","At low T, cluster expansion fixes Ising decay at 1/r^α"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002053,"raw_usage":{"total_tokens":7757,"prompt_tokens":602,"completion_tokens":7155,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":346,"completion_tokens_details":{"reasoning_tokens":7076}},"tokens_in":346,"tokens_out":7155,"duration_ms":41630,"temperature":1.0,"reasoning_tokens":7076,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T23:49:01.945667+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find, for some α∈(1,2], a contour or collection of contours violating Hypothesis 1, e.g., H(Γ∖γ)+(7/8)H(γ)>H(Γ), or violating Hypothesis 3, e.g., a contour with anomalously low energy relative to its size; such an example would falsify Lemma 3.8's activity bound and, with it, Theorems 1.1–1.3. A direct numerical check of these hypotheses at α close to 1 for large β would be a concrete way to look for it.","supporting_citations":[],"review_version":1}