{"id":"e8000f3e-63c0-459b-8c02-08122488edcb","arxiv_id":"2506.05254","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For every prime p up to 1021, every multiple n of p, and every m at least 2, the irreducibility assumption implies G_{m,p}(c0) is not an algebraic unit for any root c0 of G_{m,n}.","lead":"The paper proves new cases of a conjecture about Misiurewicz parameters, the special numbers in quadratic dynamics where the critical orbit is eventually periodic. Under a widely believed irreducibility assumption, it shows that for every small prime period the difference of two such parameters is never an algebraic unit.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.3 hinges on Magma base-case computations that are not shipped; if any trace valuation or 2-specialness check is wrong, the induction for P_{m,p} collapses.","rationale":"The paper's analytic arguments—Theorem 2.10, Lemma 3.8, Theorem 3.1, and the induction in Theorem 4.1—are detailed and appear coherent; I found no internal inconsistency there. The central claim is explicitly conditional on irreducibility, which is a standard and acknowledged hypothesis rather than a hidden flaw. The single most load-bearing unresolved point is that the proof of 2-specialness for all m≥2 rests on a large body of Magma computations that the preprint neither ships nor documents beyond a partial table. Since these finite computations are not independently checkable from the text, the theorem as written is conditional on their correctness as much as on irreducibility. This is exactly the concern the reader raised, and it supports the CONDITIONAL verdict. I recommend no change to the verdict, but the authors should release the code or data (or a certificate) to resolve it. A successful independent recomputation would move the paper toward ACCEPT.","tokens_in":17039,"tokens_out":22376,"duration_ms":218696,"concrete_test":"Independently recompute v2(tr(P_{m,p})) for all primes 3≤p≤1021 and each m=3..10 (m=2 has a closed form, 2^{2p}-2^{p+1}) using equation (22) with an independent computer algebra system such as Sage or PARI, and confirm inequality (23). Separately verify the 2-specialness claims in the exceptional cases of Theorem 4.1 that are actually needed for Theorem 1.3, especially P_{6,11}, and also P_{3,3}, P_{3,5}, P_{4,3}, P_{4,5}, P_{4,7}, P_{5,3}, P_{5,5}, P_{5,7}, P_{6,7}. If all pass, the computational base is sound.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 1.3 reduces to showing that P_{m,p} is 2-special for every prime 3≤p≤1021 and every m≥2. The inductive step (Theorem 4.1) requires the trace bound v2(tr(P_{m,p})) < m + 3p/2, displayed as inequality (23), for all intermediate m. The paper asserts that Magma verified this bound for all 2≤m≤10 and all primes 3≤p≤1021, using equation (22), but no code or full data is provided; Table 1 shows only a small subset of the valuations. Theorem 4.1 also relies on 27 direct Magma checks of 2-specialness in exceptional cases, including P_{6,11}, a polynomial whose degree is about 65,000. If any of these computations is incorrect, the induction producing 2-special P_{m,p} for all m collapses, and with it the resultant inequality that drives Proposition 2.2. The irreducibility hypotheses are explicit and therefore not a hidden flaw; the computational base is the unverified load-bearing step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"Benedetto and Goksel prove, under explicit irreducibility assumptions on Misiurewicz polynomials, that G_{m,p}(c0) is not an algebraic unit for every root c0 of G_{m,n}, whenever p is a prime at most 1024, n is a positive multiple of p with n != p, and m >= 2. The proof uses the notion of p-specialness from their earlier work, a trace formula for the multiplier polynomials P_{m,p}, a coefficient/trace relation comparing (P_{m,n})^2 with P_{m+1,n}, and an induction over m showing that P_{m,p} is 2-special for all m. The base of the induction is a set of Magma computations of 2-adic trace valuations and 27 exceptional 2-specialness checks.","tokens_in":17266,"tokens_out":13166,"duration_ms":150225,"significance":"If correct, Theorem 1.3 gives infinitely many parameter pairs (m,n) for each prime period p <= 1024 for which the difference of two Misiurewicz parameters is not an algebraic unit, conditional on the stated irreducibility conjecture. This is a substantive extension of the authors' earlier cases ell = 1,2 and provides a dynamical analogue of Li's theorem on singular moduli. The proof is arithmetic and has no fitted parameters; the irreducibility conditions are explicitly isolated rather than hidden. The main obstacle to verification is the computational base, whose completeness and reproducibility are essential to the induction.","major_comments":[{"comment":"The induction proving Theorem 1.3 rests on two asserted Magma computations: the verification of v2(tr(P_{m,p})) < m + 3p/2 for all primes 3 <= p <= 1021 and all integers 2 <= m <= 10, and the 27 direct checks of 2-specialness of P_{m+1,n} in the exceptional cases listed in Theorem 4.1. Neither the code nor the full data is shipped; Table 1 contains only a small subset of the trace valuations, and the 27 checks are asserted without stating a verification protocol. Because any error in these computations collapses the induction producing 2-special P_{m,p} for all m, and with it the resultant bound driving Proposition 2.2, I cannot verify the base of the induction from the manuscript as it stands. Please include the Magma code or a complete machine-readable table of the valuations, and explain how the 27 checks, including the large-degree case P_{6,11}, were performed.","section":"Section 4, Eq. (23) and Theorem 4.1"},{"comment":"The displayed difference (2/3)(2^{m-1}-m-3) - (2m+2) = (2/3)(2^{m-1}-4m-6) is asserted to be negative for m >= 6. This is backwards: for m = 6 the expression equals 4/3, and it is positive for all m >= 6. The subsequent conclusion that inequality (19) holds for all m >= 6 requires the difference to be nonnegative, so the written proof contains a sign error at a load-bearing step. The intended argument is clear, but the text must be corrected.","section":"Section 4, proof of Theorem 4.1"},{"comment":"Equation (22) is invoked for all 2 <= m <= 10, but Theorem 2.10 is stated for m >= 3. The m = 2 case is supplied separately in Remark 2.11, but the proof of Theorem 1.3 does not explicitly say that it uses Remark 2.11 for the m = 2 trace computations. Please state this explicitly so the reader can see how the claimed base cases are covered.","section":"Section 4, proof of Theorem 1.3, Eq. (22)"}],"minor_comments":[{"comment":"The word 'polymomial' should be 'polynomial'.","section":"Definition 2.1"},{"comment":"The caption says the entries are 'relative to m+p', but the text explains that the relevant bound used in the proof is m + 3p/2. Please state explicitly that all tabulated entries, including the boldface ones where the sharper bound v2 <= m+p fails, do satisfy the required inequality (23).","section":"Table 1 and Section 5"},{"comment":"The parenthetical remark says that [3, Theorem 1.7] covers 'the six cases above with n=1,2'; this count is correct because m=5 has no n=1,2 cases, but the sentence could be phrased more clearly to avoid confusion with the eight total n=1,2 combinations over m=2,3,4,5.","section":"Theorem 4.1, exceptional cases"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is conditional on an explicit irreducibility conjecture, and its proof relies on a substantial computational base that is not yet reproducible from the manuscript. If the authors supply the missing code or complete data and fix the sign error in Theorem 4.1, I would be inclined to accept. The paper is well within the scope of the journal, and I saw no evidence of problematic citation or duplication issues."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a real extension, not a repackaging. The trace relation in Theorem 3.1 and the induction via 2-specialness are new, and they push the non-unit result from periods 1 and 2 to every prime p up to 1021. The theorem is conditional on irreducibility, which is stated clearly, so the conditional status is not a hidden flaw.\n\nWhat the paper does well: the proof structure is transparent. The new trace relation is a solid piece of algebra, and the propagation argument (Corollary 3.2 plus the trace-bound induction) is elegant. The paper is honest about what depends on computation and what depends on the irreducibility conjecture. The base cases are finite and checkable, and the authors give a table with enough entries to make the pattern plausible.\n\nSoft spots: the load-bearing computational base is not shipped. Inequality (23) is verified by Magma for all primes 3<=p<=1021 and 2<=m<=10, but only a subset of the valuations appears in Table 1, and no code or full data is available. Theorem 4.1 also relies on 27 direct Magma checks of 2-specialness, including P_{6,11}, which is huge. If any of those computations is wrong, the induction that produces 2-special P_{m,p} for all m collapses, and with it Theorem 1.3. This is the one place a referee should push. The irreducibility hypotheses are explicit, so they are fine, but they mean the main theorem is conditional on an open conjecture. The paper itself notes that the full data would take too much space, which is understandable, but for a statement that depends on thousands of machine computations, shipping code or a data file is the norm.\n\nWho it is for: arithmetic dynamicists working on dynamical Andre-Oort, Misiurewicz parameters, and algebraic units. It is written clearly and the math is serious. The result is conditional, but the conditional statement is new and meaningful. If the computations check out, the theorem extends the program in a substantial way.\n\nRecommendation: send to peer review. The referee should ask for the code or full trace data before accepting, but the paper deserves a serious referee. The missing data is fixable, and the central argument holds up if the computational base is sound.","headline":"Conditional but genuinely new extension of the non-unit result to all small prime periods; the unshipped Magma base cases are the only real worry.","tokens_in":17763,"tokens_out":3927,"would_cite":true,"duration_ms":46095,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37P15","11R09","37P20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Differences of Misiurewicz parameters are never algebraic units for prime periods up to 1024, assuming irreducibility.","keywords":["Misiurewicz parameters","multiplier polynomials","algebraic units","p-special polynomials","arithmetic dynamics","irreducibility","singular moduli","quadratic maps"],"falsifier":"Compute the $2$-adic valuations of the relevant traces independently for all primes $3 \\le p \\le 1021$ and $2 \\le m \\le 10$; if any value violates the bound claimed in the paper, the induction in the proof of Theorem 1.3 collapses. Alternatively, find $m$ and a prime $p \\le 1024$ with irreducible $G_{m,p}$ and $G_{m,n}$ for which $G_{m,p}(c_0)$ actually is an algebraic unit, which would disprove the theorem itself.","tokens_in":16863,"feed_emoji":"🔁","tokens_out":14499,"duration_ms":134507,"temperature":0.7,"pith_summary":"Misiurewicz parameters are the complex numbers $c$ for which the orbit of the critical point $0$ under $z^2+c$ lands on a periodic cycle after finitely many steps; they are the dynamical analogue of singular moduli on modular curves. The paper attacks the conjecture that a difference of two Misiurewicz parameters can never be an algebraic unit. The main theorem shows that for every $m \\ge 2$, every prime $p \\le 1024$, and every $n$ that is a multiple of $p$ different from $p$, the value $G_{m,p}(c_0)$ at a root $c_0$ of $G_{m,n}$ is not an algebraic unit, provided the two Misiurewicz polynomials $G_{m,p}$ and $G_{m,n}$ are irreducible over $\\mathbb{Q}$. The proof works by proving that the associated multiplier polynomials are $2$-special — a condition on the $2$-adic valuations of their coefficients that forces the relevant resultants to be nontrivial — and then propagating this property from $m$ to $m+1$. If the widely believed irreducibility conjecture is eventually confirmed, the non-unit statement would hold unconditionally for all these pairs.","feed_headline":"For prime periods up to 1024, Misiurewicz differences are never units","feed_subtitle":"For quadratic maps, the dynamical analogue of singular moduli also forbids unit differences — assuming the usual irreducibility holds.","key_machinery":"The central objects are the Misiurewicz polynomials $G_{m,n} \\in \\mathbb{Z}[c]$ (whose roots are the parameters $c$ for which $0$ is strictly preperiodic of type $(m,n)$ under $z^2+c$) and their associated multiplier polynomials $P_{m,n}(x) \\in \\mathbb{Z}[x]$ (whose roots are the multipliers of the cycles at those parameters). The load-bearing notion is $p$-specialness: a monic integer polynomial whose second coefficient has $p$-adic valuation strictly above $v_p(2)$ and whose remaining non-leading coefficients have even larger $p$-adic valuation. The proof's engine is an induction that makes $P_{m,p}$ $2$-special for every $m$: a trace formula expresses the coefficient sum of $P_{m,p}$ in closed form; a coefficient identity compares the tail coefficients of $(P_{m,n})^2$ and $P_{m+1,n}$; and an induction step transmits $2$-specialness from $m$ to $m+1$ once a valuation bound on the trace holds. The base of the induction is $m=2$, where an explicit formula gives $\\operatorname{tr}(P_{2,p}) = 2^{2p} - 2^{p+1}$, plus a small set of direct computer checks.","core_discovery":"The paper's central claim is Theorem 1.3: fix $m \\ge 2$, let $p$ be a prime at most $1024$, let $n$ be a positive multiple of $p$ different from $p$, and let $c_0$ be a root of $G_{m,n}$. If $G_{m,p}$ and $G_{m,n}$ are irreducible over $\\mathbb{Q}$, then $G_{m,p}(c_0)$ is not an algebraic unit. The argument works by showing that the multiplier polynomial $P_{m,p}$ — whose roots are the multipliers of the periodic cycles of the maps $z^2+c$ as $c$ runs over roots of $G_{m,p}$ — is $2$-special in the sense of the paper's Definition 2.3: the $2$-adic valuation of its second coefficient exceeds $v_2(2)$, and all other non-leading coefficients have even larger valuation. By the authors' earlier result, $2$-specialness of $P_{m,p}$ forces the resultant with every cyclotomic polynomial to exceed $1$ in absolute value, and that implies the non-unit conclusion. The main new work is an induction on $m$: a trace formula and a coefficient-matching identity show that the required valuation bounds propagate from $m$ to $m+1$ once they hold for $m \\le 10$, and the base cases are verified by an explicit formula for $m=2$ together with a finite computation for all primes $3 \\le p \\le 1021$ and $2 \\le m \\le 10$.","pith_inferences":["The bound $1024$ is a computational cutoff, not a conceptual one: the induction that proves $2$-specialness is uniform in $p$, so redoing the finite trace-valuation check for larger primes would extend the theorem to any desired range.","The $p$-special machinery is defined for any prime, and the same trace-valuation comparison might work for an odd prime $q$: determining whether the $q$-adic valuations of the same traces obey an analogous bound would reveal whether the non-unit phenomenon is purely $2$-adic in this family.","The structural similarity with the singular-moduli non-unit theorem suggests a general principle for one-dimensional arithmetic families: special points are never connected by unit differences, a phenomenon that might extend to higher-degree dynamical families once the appropriate multiplier polynomials are understood."],"forward_implications":["For every $m \\ge 2$ and every prime $p \\le 1024$, once $G_{m,p}$ and $G_{m,n}$ are irreducible, the value $G_{m,p}(c_0)$ at a root of $G_{m,n}$ is never an algebraic unit, settling the dynamical unit question for all such pairs.","If in addition $G_{m,p}$ stays irreducible over the field $\\mathbb{Q}(c_0)$, then the actual difference between a root of $G_{m,n}$ and a root of $G_{m,p}$ is not an algebraic unit, giving the exact dynamical analogue of the singular-moduli non-unit theorem.","This is a substantial extension of the authors' earlier result, which covered only periods $1$ and $2$; the new theorem covers every prime period up to $1024$.","Should the irreducibility conjecture for all $G_{m,n}$ be proved, the conditional non-unit statement becomes unconditional, with the range of $p$ limited only by the verified trace computations."],"supporting_citations":[{"why":"Supplies the definition of p-specialness, the result that p-special polynomials have nontrivial cyclotomic resultants, and the earlier non-unit theorem for periods 1 and 2.","marker":"[3]"},{"why":"Earlier answer to the dynamical unit question in many cases; gives the valuation bound for iterates used in Proposition 2.5.","marker":"[2]"},{"why":"Provides irreducibility of G_{m,1} used in the proof of Theorem 1.3 and the integrality and valuation property of iterates.","marker":"[7]"}],"fun_headline_variants":["Prime periods to 1024: Misiurewicz differences can't be units","Under irreducibility, Misiurewicz gaps resist being units","Misiurewicz parameters: no unit differences for primes ≤1024","Theorem: Misiurewicz differences are non-units for primes ≤1024","Dynamical singular moduli forbid unit gaps for prime periods"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole induction rests on a large finite computer calculation, only partially reproduced in the paper, that bounds the $2$-adic valuation of certain traces for every prime up to $1021$ and every $m$ from $2$ to $10$; if any of those computed values were wrong, the chain that makes all multiplier polynomials $2$-special would break.","fun_headline_variants_meta":{"raw":{"variants":["Prime periods to 1024: Misiurewicz differences can't be units","Under irreducibility, Misiurewicz gaps resist being units","Misiurewicz parameters: no unit differences for primes ≤1024","Theorem: Misiurewicz differences are non-units for primes ≤1024","Dynamical singular moduli forbid unit gaps for prime periods"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000373,"raw_usage":{"total_tokens":2016,"prompt_tokens":993,"completion_tokens":1023,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":609,"completion_tokens_details":{"reasoning_tokens":922}},"tokens_in":609,"tokens_out":1023,"duration_ms":10517,"temperature":1.0,"reasoning_tokens":922,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:23:05.003620+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the $2$-adic valuations of the relevant traces independently for all primes $3 \\le p \\le 1021$ and $2 \\le m \\le 10$; if any value violates the bound claimed in the paper, the induction in the proof of Theorem 1.3 collapses. Alternatively, find $m$ and a prime $p \\le 1024$ with irreducible $G_{m,p}$ and $G_{m,n}$ for which $G_{m,p}(c_0)$ actually is an algebraic unit, which would disprove the theorem itself.","supporting_citations":[{"cited_title":"Benedetto and V","cited_arxiv_id":null,"evidence_quote":"Supplies the definition of p-specialness, the result that p-special polynomials have nontrivial cyclotomic resultants, and the earlier non-unit theorem for periods 1 and 2."},{"cited_title":"Benedetto and V","cited_arxiv_id":null,"evidence_quote":"Earlier answer to the dynamical unit question in many cases; gives the valuation bound for iterates used in Proposition 2.5."},{"cited_title":"Goksel,On the orbit of a post-critically finite polynomial of the formx d +c, Funct","cited_arxiv_id":null,"evidence_quote":"Provides irreducibility of G_{m,1} used in the proof of Theorem 1.3 and the integrality and valuation property of iterates."}],"review_version":1}