{"id":"8ce11156-3294-405c-a40e-753f66fca37e","arxiv_id":"2504.19690","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An algebraic curve construction proves the periodic TASEP Bethe ansatz equations have binomial(L,N) solutions with multiplicity, and yields explicit genus, degeneracy, and zeta(3) free energy results.","lead":"This paper develops an algebraic-geometric method to solve the Bethe ansatz equations for the periodic totally asymmetric exclusion process. It proves the expected number of solutions and derives explicit formulas for the curve's genus and component multiplicities, plus a free energy given by the Riemann zeta function.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unproved c*-monodromy assertion (Section 3.1) is the main gap: it drives the component/genus classification. A short Puiseux expansion at t=-N/(L-N) would settle it.","rationale":"I read the paper in good faith as an algebro-geometric proof of Bethe-ansatz completeness for periodic TASEP, with secondary results on the irreducible decomposition of the resulting curve and the genus of its components. The central counting argument in Theorem 2.7 is sound: the poles of h = v/w are correctly localized at w = 0, the local expansion near each G1-orbit gives the stated pole orders, and the combinatorics of N1(o) and N2(o) yields exactly C(L,N) poles. This establishes the completeness claim independently of the c*-monodromy, so the reader's strongest claim is not threatened. The genuine weakness is in Section 3.1: the assertion that the c*-monodromy is sigma^{-1} tau (the transposition (N L)) is stated without proof, yet it underlies the classification of connected components (Proposition 3.4), the orbit-size formula, and the genus formula of Eq. (33). I checked the internal consistency of the claim: for several small cases (L,N) = (4,2), (5,2), (6,3), the permutation sigma^{-1} tau is indeed a transposition, and it equals (N L). This is the only permutation consistent with the known monodromies around 0 and infinity and the product relation around the three branch points. Thus the concern is not a fatal error but an omitted proof. The paper should supply the short Puiseux derivation at the simple critical point t = -N/(L-N). I also note a minor typographical issue in the displayed definition of phi (Eq. (5)), which appears without the negative exponent in the OCR text; the surrounding computations clearly use phi(t) = t^N/(1-t)^L. This does not affect the mathematical argument. Overall, the reader's CONDITIONAL verdict is appropriate: the paper is likely correct, but the omitted c*-monodromy proof should be supplied before unconditional acceptance.","tokens_in":44389,"tokens_out":33005,"duration_ms":319447,"concrete_test":"For (L,N)=(5,2), numerically continue the five branches of t^2(1-t)^{-5}=w along a small loop around w=c* and record the permutation of the labels defined near w=0; verify that it is (2 5) = sigma^{-1} tau. More generally, derive the transposition analytically from the local expansion w - c* ~ A (t + N/(L-N))^2 and match the two colliding branches to the labels t_N and t_L via analytic continuation; if confirmed, the G2 construction and genus formulas stand.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1 asserts without proof that the monodromy of w: X -> P around the critical value c* is generated by sigma^{-1} circ tau (equivalently, by the transposition (N L)). This assertion is load-bearing for the advertised irreducible decomposition and genus formulas: it defines G2 = , which labels connected components (Proposition 3.4) and enters the Riemann-Hurwitz computation in Eq. (33). If the c*-monodromy were a different transposition, the orbit classification and the multiplicities of the v-values would change. The proof of Theorem 2.7 itself only needs the monodromy around 0, so the completeness count is unaffected. The claim is very likely correct: the local degree-2 ramification at t = -N/(L-N) forces a transposition, and the product relation with the known monodromies around 0 (sigma) and infinity (tau) determines that transposition to be (N L) = sigma^{-1} tau. However, the paper never supplies this derivation; the examples in Appendix B check only small (L,N), and the reader is left to take the assertion on faith. A rigorous proof should be added.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops an algebro-geometric method for the Bethe ansatz equations of the periodic totally asymmetric simple exclusion process (TASEP) with L sites and N particles. The authors embed the Bethe equations into a plane curve C0 whose coordinates are the product v of Bethe roots and the auxiliary variable w, via the rational map t^N/(1-t)^L. They prove (Theorem 2.7) that the Bethe equations have binom(L,N) physical roots counted with multiplicity, by counting poles of the meromorphic function h=v/w on the associated Riemann surface X. They then study the connected-component decomposition of X in terms of the monodromy group G2 = <σ,τ> acting on the set of N-element subsets of [L], and derive explicit formulas for the number of connected components, the genus of each component, and the total genus (Lemmas 3.12, Eq. (33), Eq. (39)), recovering Prolhac's table. They also give a Golinelli-Mallick-type degeneracy analysis (Propositions 5.1, 5.2, 5.5) and apply special Bethe roots with v=1 to evaluate the norm (partition function) of the five-vertex model, obtaining a free energy F = 7ζ(3)/(16π²) in the thermodynamic limit.","tokens_in":44667,"tokens_out":19818,"duration_ms":177718,"significance":"The pole-counting proof of completeness is a clear advance: it is rigorous, self-contained in the algebraic formulation, and does not rely on the numerical-ansatz approach of earlier works. The connected-component and genus formulas, together with the degeneracy classification, provide a systematic framework that explains and extends the examples of Prolhac and Golinelli-Mallick. The free-energy evaluation is a new exact result. I also note the paper's positive features: the algebraic curve realization is explicitly computable for small cases (Appendix B), and the counting argument in Theorem 2.7 is machine-checkable. However, the component/genus/degeneracy results depend on an unproved assertion about the monodromy around the third critical value c*, which is the main obstacle to accepting the paper in its current form.","major_comments":[{"comment":"The statement \"We also show that the monodromy action around c∗ is generated by ˆσ−1 ◦ˆτ\" is not followed by a proof. This is a load-bearing assertion: it defines the group G2, whose orbit decomposition gives the connected components (Prop. 3.4) and which enters the Riemann-Hurwitz computation in Eq. (33) and the factorization in Prop. 5.8. Without a proof, the explicit component/genus formulas are not rigorously established. The claim is plausible and can be verified by a short local analysis at the double critical point t = -N/(L-N) (where the map has ramification index 2, forcing the monodromy to be a transposition, and the product with the known monodromies around 0 and ∞ determines the transposition to be (N L) = σ^{-1}τ). The authors should supply this derivation in the manuscript.","section":"Section 3.1"},{"comment":"The proof contains the step \"By the identity theorem in complex analysis, we can replace I′,J′ with g(I′),g(J′), where g is an arbitrary element of the monodromy group G2.\" As written, this step is not justified because the identity theorem applies to single-valued functions on a connected Riemann surface, whereas the branches t_i(w) are multivalued functions on the base; the proposed replacement changes the branch and may move the equality to a different sheet. The intended argument likely uses the Galois action on the field of algebraic functions generated by the t_i, and it should be written out carefully. This is load-bearing because Proposition 5.2 is the converse half of the degeneracy criterion (eq. (41) and the classification in Section 5.1).","section":"Section 5.1, Proposition 5.2"}],"minor_comments":[{"comment":"These equations are missing the division sign: the correct map is ϕ(t) = t^N/(1-t)^L, and similarly w = t_i(w)^N/(1-t_i(w))^L and ω_e^{k-1} w^{1/e} = t^{N'}/(1-t)^{L'}.","section":"Eqs. (5), (6), (31)"},{"comment":"The condition for injectivity of the permutation action on the N-subsets is misprinted: \"if 01$, the surface splits into connected components whose orbit structure under a monodromy group explains how often a given product value occurs, reproducing spectral degeneracies of the Markov matrix, and the same geometric input yields explicit formulas for the number of components, ramification points, and total genus, as well as partition functions of the five-vertex model whose thermodynamic free energy is $7\\zeta(3)/(16\\pi^2)$.","feed_headline":"Counting poles proves the TASEP Bethe ansatz is complete","feed_subtitle":"The expected binomial number of solutions is proven with multiplicity, and curve components explain eigenvalue degeneracies.","key_machinery":"The central object is the Riemann surface $X$, obtained by analytically continuing the $L$ inverse branches $t_1(w),\\dots,t_L(w)$ of $\\varphi(t)=t^N(1-t)^L$ and realized algebraically as the (possibly singular, non-reduced) plane curve $C_0\\subset\\mathbb{C}^2$ with coordinates $v=\\prod_{i\\in I}t_i(w)$ and $w$, defined by the equation $\\prod_{I\\in\\Omega}(v-\\prod_{i\\in I}t_i(w))=0$ whose coefficients are elementary symmetric polynomials in the branches. The counting argument is carried by the meromorphic function $h=v/w$: its zero set on $X$ outside the origin matches the Bethe line, and its poles are concentrated on the fiber $w=0$, where the local expansion of $v$ is governed by the cyclic monodromy $\\hat\\sigma$ of order $\\mathrm{lcm}(N,L-N)$. The component classification is carried by the larger monodromy group $G_2=\\langle\\hat\\sigma,\\hat\\tau\\rangle$, where $\\hat\\tau$ shifts all $L$ indices cyclically; orbits of $G_2$ on $N$-subsets are in bijection with cyclic orbits of the package-counting tuples $(A_1,\\dots,A_e)$, and these orbits determine the connected components, their ramification data, and, through the Riemann-Hurwitz formula, the genus.","core_discovery":"The central claim is Theorem 2.7: the Bethe equations $z_i^N(1-z_i)^L = (-1)^{N+1}e^{L\\gamma}\\prod_j z_j$ for $i=1,\\dots,N$ have $\\binom{L}{N}$ physical roots counted with multiplicity. The argument realizes the equations on a compact Riemann surface $X$ formed from the $L$ branches of the equation $w=t^N(1-t)^L$. Each sheet is labelled by an $N$-subset $I\\subset\\{1,\\dots,L\\}$, and the function $v=\\prod_{i\\in I}t_i(w)$ is single-valued on $X$; the Bethe line $w=(-1)^{N+1}e^{L\\gamma}v$ cuts the corresponding plane curve $C_0$ exactly at the physical roots. The proof counts poles of $h=v/w$ at the fiber $w=0$, using the cyclic monodromy $\\hat\\sigma$ that permutes the first $N$ branches among themselves and the last $L-N$ branches among themselves; summing the pole orders over all sheets gives $\\binom{L}{N}$, so for generic $\\gamma$ every root is simple and the Bethe ansatz is complete (Corollary 2.8). For $\\gcd(L,N)=e>1$, the paper classifies the connected components of $X$ by orbits of the full monodromy group $G_2=\\langle\\hat\\sigma,\\hat\\tau\\rangle$ on $N$-subsets, equivalently by cyclic orbits of the $e$-tuple $(A_1,\\dots,A_e)$ counting how many selected indices lie in each residue class modulo $e$. Components that differ by replacing one full residue class with an empty one map to the same irreducible component of $C_0$, which is the algebro-geometric mechanism behind Golinelli-Mallick-type eigenvalue degeneracies. Explicit formulas for the number of components, ramification indices, and the total genus follow from Riemann-Hurwitz plus character computations, and for half-filling the special roots $t_k+t_{N+k}=2+\\omega_N^{1/2-k}$, $t_kt_{N+k}=1$ evaluate five-vertex partition functions in terms of roots of unity.","pith_inferences":["The completeness theorem depends only on monodromy around $w=0$, so the unproved statement about monodromy around $c^*$ has no bearing on Theorem 2.7; the classification of components, degeneracies, and genus formulas are the parts that would need revisiting if that monodromy statement failed.","The same pole-counting scheme should transfer to the ASEP with $q\\neq 1$, where the rational map $w=t^N/(1-t)^L$ is replaced by its two-parameter ASEP analogue; a natural test is whether the component classification reproduces the known degeneracies of the ASEP Markov matrix.","The appearance of $\\zeta(3)$ in the half-filling free energy suggests a possible numerical probe: computing the same free energy through independent transfer-matrix methods for larger $N$ should approach $0.05328\\ldots$, a check not performed in the paper.","The explicit factorizations in Appendix B identify the roots missed by the earlier Cassini-oval ansatz at $(L,N)=(10,5)$; one could test the completeness machinery by verifying numerically that those roots (coming from the degree-15 factor in $f_1$) satisfy the Bethe equations and give independent eigenvectors."],"forward_implications":["For generic fugacity $\\gamma$, the Bethe ansatz is complete at the level of eigenvectors: the Bethe vectors built from the roots span the full $\\binom{L}{N}$-dimensional space, with the zero-root steady state handled separately at $\\gamma=0$.","When $L$ and $N$ share a common factor, the Markov matrix exhibits spectral degeneracies exactly when two connected components of the curve are related by replacing a full package (residue class) by an empty one; in the most degenerate case the component is defined by a power of $v-1$ and the common eigenvalue is explicitly $h(-L'+N')$.","The closed formulas for the number of connected components, the ramification structure, and the total genus give complete algebro-geometric data for every $(L,N)$, reproducing all previously computed examples.","At half-filling $L=2N$, the special Golinelli-Mallick-type roots $t_k+t_{N+k}=2+\\omega_N^{1/2-k}$, $t_kt_{N+k}=1$ make the on-shell Bethe-vector norm an explicit product over roots of unity; its thermodynamic-limit free energy per site is $7\\zeta(3)/(16\\pi^2)$.","Particle-hole duality holds at the level of the curve: the Bethe equations for $N$ and $L-N$ particles give the same values of $v$ and the same eigenvalues $E=\\sum_i z_i/(1-z_i)$, and overlaps with alternating initial states vanish for Golinelli-Mallick type tuples at half- and one-third filling."],"supporting_citations":[{"why":"Introduces the Riemann surface for TASEP that the paper realizes as a plane curve.","marker":"[39]"},{"why":"Supplies the Bethe ansatz equations for periodic TASEP that are the object of study.","marker":"[13]"},{"why":"Provides the package construction and spectral-degeneracy argument that the paper reinterprets in algebro-geometric terms.","marker":"[21]"},{"why":"Gives the standard construction of plane curves from elementary symmetric polynomials in branches, used to write the defining equation of $C_0$.","marker":"[63]"},{"why":"Supplies the determinant representation and conventions used for Bethe vectors and for checking the ansatz in examples.","marker":"[41]"},{"why":"Gives the determinant representation of the on-shell Bethe-vector norm used in the five-vertex model partition function.","marker":"[49]"},{"why":"Provides the half-filling norm and long-time asymptotic results that the special-root evaluations extend.","marker":"[50]"},{"why":"Supplies the factorized norm and overlap formulas, including the Grothendieck-polynomial setting, used for the free energy and vanishing results.","marker":"[51]"}],"fun_headline_variants":["Riemann surface counts Bethe roots in TASEP","Algebraic curve proves Bethe ansatz complete","Curve splitting yields TASEP degeneracies","Counting on a curve completes TASEP Bethe"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the monodromy of the covering $w:X\\to\\mathbb{P}^1$ around the critical value $c^*$ is generated by the permutation $\\hat\\sigma^{-1}\\circ\\hat\\tau$ (stated in Section 3.1 without proof); if that monodromy were different, the classification of connected components, the degeneracy multiplicities, and the genus formulas would change, while the pole count of Theorem 2.7 would remain valid.","fun_headline_variants_meta":{"raw":{"variants":["Riemann surface counts Bethe roots in TASEP","Algebraic curve proves Bethe ansatz complete","Curve splitting yields TASEP degeneracies","Counting on a curve completes TASEP Bethe"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000269,"raw_usage":{"total_tokens":1794,"prompt_tokens":1291,"completion_tokens":503,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":907,"completion_tokens_details":{"reasoning_tokens":441}},"tokens_in":907,"tokens_out":503,"duration_ms":5248,"temperature":1.0,"reasoning_tokens":441,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:47:03.809571+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct way to settle the claim is to compute the actual permutation induced on the sheets of $w:X\\to\\mathbb{P}^1$ by a small loop around $w=c^*$ for a case like $(L,N)=(6,3)$ or $(8,4)$ and compare it with $\\hat\\sigma^{-1}\\circ\\hat\\tau$; a mismatch would falsify the component/genus classification, whereas the pole-counting proof of completeness can be checked independently by computing the total pole order of $h=v/w$ on the fiber $w=0$.","supporting_citations":[{"cited_title":"Prolhac, Riemann surface for TASEP with periodic bou ndaries, J","cited_arxiv_id":null,"evidence_quote":"Introduces the Riemann surface for TASEP that the paper realizes as a plane curve."},{"cited_title":"Gwa and H","cited_arxiv_id":null,"evidence_quote":"Supplies the Bethe ansatz equations for periodic TASEP that are the object of study."},{"cited_title":"Golinelli and K","cited_arxiv_id":null,"evidence_quote":"Provides the package construction and spectral-degeneracy argument that the paper reinterprets in algebro-geometric terms."},{"cited_title":"W eyl, and G","cited_arxiv_id":null,"evidence_quote":"Gives the standard construction of plane curves from elementary symmetric polynomials in branches, used to write the defining equation of $C_0$."},{"cited_title":"Prolhac, KPZ fluctuations in finite volume, SciPost Ph ys","cited_arxiv_id":null,"evidence_quote":"Supplies the determinant representation and conventions used for Bethe vectors and for checking the ansatz in examples."},{"cited_title":"Bogoliubov, Determinantal Representation of the Time-Dependent Stationary Correlation Function for the Totally Asymmetric Simple Exclusion Model, SIGMA 5 052 (200 9)","cited_arxiv_id":null,"evidence_quote":"Gives the determinant representation of the on-shell Bethe-vector norm used in the five-vertex model partition function."},{"cited_title":"Motegi, K","cited_arxiv_id":null,"evidence_quote":"Provides the half-filling norm and long-time asymptotic results that the special-root evaluations extend."},{"cited_title":"Motegi and K","cited_arxiv_id":null,"evidence_quote":"Supplies the factorized norm and overlap formulas, including the Grothendieck-polynomial setting, used for the free energy and vanishing results."}],"review_version":1}