{"id":"4a75c761-76aa-4fcf-b6f0-93c19e9f3e13","arxiv_id":"2411.15568","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"For all but a finite list of cases with m≥9, an element exists in F_{q^m} whose two rational-function values form a primitive pair with prescribed traces.","lead":"The paper proves a sufficient condition for finding an element in a finite field whose values under two rational functions are primitive generators, with the element and its inverse having specified trace values. It is a niche mathematical result with potential use in cryptographic constructions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The theorem's exception list is not proved: the proof only marks pairs as 'possible exceptional' when the sieve condition (3.3) fails, but failure of a sufficient condition is not a proof of nonexistence.","rationale":"The reader's verdict is REJECT, and I agree. The central assertion is the 'except' list in Theorem 1.1. The weakest link is not primarily the numerical constants in Lemmas 2.1-2.3, though those are worth checking; it is the logical step from 'the sieve could not find a suitable d' to 'this pair is exceptional'. The paper's own final paragraph in Section 4 says 'possible exceptional pairs', and the abstract also uses 'possible', but Theorem 1.1 states the list as actual exceptions. A concrete witness for (q,m)=(2,9) would falsify item (i) immediately under the paper's stated existential reading of A^m_p; if no witness is found, the burden remains on the authors to provide a nonexistence proof, since failure of (3.3) is only failure of a sufficient condition. I therefore keep the REJECT verdict, with the reason located in the missing nonexistence argument rather than in the character-sum bounds. Agreement with the reader is partial: the reader's formal weakest assumption was the bound constants, but the reader's rationale did flag the overclaim from 'possible exceptional pairs' to actual exceptions.","tokens_in":9103,"tokens_out":12511,"duration_ms":114008,"concrete_test":"Take the first listed exception: (q,m)=(2,9). Let f(x)=x^3 and g(x)=x^{-3}; each has degree sum 3 and both are in R because 3 does not divide 511=2^9-1. Enumerate all primitive epsilon in F_512 (or all nonzero elements, since gcd(3,511)=1). Check whether any epsilon has (epsilon^3,epsilon^{-3}) primitive and Tr(epsilon)=a, Tr(epsilon^{-1})=b for some a,b in F_2. If yes, item (i) of Theorem 1.1 is false under the paper's existential definition of A^m_p. If no, item (i) is still unproved unless the authors supply a nonexistence certificate instead of relying on failure of (3.3). This single check separates the overclaim from the sufficient-condition part.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4 ends with: 'The pairs for which we could not find such d are considered as possible exceptional pairs and thus we complete the proof of Theorem 1.1.' This is the logical hinge of Theorem 1.1. The computation verifies, for each listed pair, that the sufficient condition and then the sieving inequality (3.3) do not certify a positive value of N_{f,g,a,b}(q^m-1,q^m-1). That only shows the proof method fails there; it does not show that no epsilon exists, nor even that some f,g,a,b has zero count. To assert '(q,m) is not in A^m_p' one needs a separate argument: an upper bound showing N=0 for some or all choices, or a Galois or field-theoretic obstruction. None is given. The abstract honestly says 'possible exceptional pairs'; Theorem 1.1 drops the word 'possible'. There is also a formatting slip in the list ((4,12) appears twice), but the substantive problem is the missing nonexistence proof. Proposition 4.2 and the sufficient condition (3.2)/(3.3) can survive, but the central classification is logically unsupported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the existence of a nonzero element \\epsilon in F_{q^m} such that, for rational functions f,g in a certain class R, the pair (f(\\epsilon),g(\\epsilon)) is primitive and the traces of \\epsilon and \\epsilon^{-1} take prescribed values a,b in F_q. The authors derive a sufficient condition using multiplicative and additive character sums, apply a prime sieve, and then specialize to the case where the degree sums of f and g are both 3. The main theorem, Theorem 1.1, asserts that for every prime power q and every m \\ge 9, all pairs (q,m) belong to the set A^m_p except for an explicitly listed finite set of cases. The proof combines analytic estimates with a finite computation over m \\le 108.","tokens_in":9299,"tokens_out":17791,"duration_ms":156844,"significance":"If the main theorem were established, the paper would give a useful extension of primitive-pair results to rational functions with trace constraints, and the authors make a serious effort to combine standard tools: Cohen--Huczynska sieving, Weil-type character sum bounds, and estimates for the divisor function W(M). The sufficient condition itself, in the form of Theorem 3.3, is a reasonable technical contribution. However, the central classification is not proved: the list of ``possible exceptional pairs'' is only a list of cases where the sieve fails, and one of the character-sum bounds used to derive the sufficient condition is false. These problems affect the main theorem directly, so the paper cannot be accepted in its present form.","major_comments":[{"comment":"The proof of Theorem 1.1 does not establish the exception list. The text says that for the listed pairs no suitable d satisfying (3.3) could be found, and those pairs are ``considered as possible exceptional pairs.'' Failure of a sufficient condition is not a proof of nonexistence: to conclude (q,m) \\notin A^m_p one needs an independent upper bound showing that N_{f,g,a,b}(q^m-1,q^m-1)=0 for the relevant choices, or a field-theoretic obstruction. No such argument is given. The abstract's wording ``possible exceptional pairs'' is accurate, but Theorem 1.1 changes this to a definite assertion, and the logical gap is load-bearing for the main theorem.","section":"Section 4, end, and Theorem 1.1"},{"comment":"The claimed bound on |\\chi_{f,g}(d_1,d_2,u,v)| in Case IV is false as stated. Take a cubic h(x) with three distinct nonzero roots, choose c \\in F_{q^m}^* with \\chi_d(c)=1 for some prime d dividing q^m-1, and set f(x)=h(x), g(x)=c/h(x). Both rational functions have degree sum 3 and satisfy the defining condition of R. Then for d_1=d_2=d and u=v=0, \\chi_{f,g}(d,d,0,0)=\\sum_{\\epsilon\\in F_{q^m}\\setminus S} \\chi_d(h(\\epsilon))\\chi_d(c/h(\\epsilon)) = \\chi_d(c)(q^m-|S|)=q^m-|S|, which contradicts the bound (m_1+m_2-1)q^{m/2} printed in that subsection. Since these Case IV estimates feed directly into the lower bound (3.1) and hence into the sieve inequality (3.3), the general sufficient condition is not established.","section":"Section 3, Case IV"},{"comment":"The finite verification from m'=109 down to m=9 is not reproducible from the text. The proof refers to a ``trial and error'' determination of m', to lower bounds q_{m,i}, and to a search for d satisfying (3.3), but it does not give the algorithm, the intermediate tables, or the code used. The listed exception set is the substantive content of Theorem 1.1, and without the computational data being available, the classification cannot be checked independently.","section":"Section 4, finite verification"}],"minor_comments":[{"comment":"The pair (4,12) appears twice in the list; Theorem 1.1 lists q=4 for m=12 only once.","section":"Section 4, list of possible exceptional pairs"},{"comment":"The abstract says ``possible exceptional pairs'' while Theorem 1.1 says ``except the following cases''; these should be aligned, since the proof only identifies possible exceptions.","section":"Abstract and Theorem 1.1"},{"comment":"The displayed formulas appear to have lost exponent formatting: they should read q^{m/2-2} \\ge 2^{2(2+w(q^m-1))} and (x+1)^{1/2-2/m} \\ge 2^{4+2.77\\log x/\\log\\log x}. As printed, the right-hand sides appear as ordinary products, which is not consistent with the preceding bounds.","section":"Section 4, Eq. (4.4) and following display"},{"comment":"In Lemma 4.1, D depends on M=q^m-1 through the primes p_i \\le 2\\nu dividing M, but the proof of Proposition 4.2 treats D as a constant when taking logarithms. This can be repaired by bounding D uniformly over all primes \\le 2\\nu, but as written the monotonicity claim is not fully justified.","section":"Section 4, Proposition 4.2"},{"comment":"The phrase ``for some prescribed values a,b'' is ambiguous. The sufficient condition in Section 3 is independent of a,b and of the particular rational functions, suggesting that Theorem 1.1 is intended as a universal statement; the definition should say so explicitly.","section":"Section 1, definition of A^m_p"}],"recommendation":"reject","confidential_remarks":"I have no conflict of interest. For this journal's standards, the absence of reproducible data for the finite verification is especially important because the exception list is the main content of Theorem 1.1. In addition, the gap between ``possible exceptional pairs'' and the definite assertion of Theorem 1.1 needs a substantive new argument rather than a local revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the sufficient condition and the sieving inequality are legitimate and probably correct, but Theorem 1.1 is not proved. The proof shows only that the sieve fails to certify existence for the listed (q,m) pairs; it does not show that no such ε exists. The abstract says 'possible exceptional pairs'; the theorem drops 'possible'. That is the difference between a classification and a list of unresolved cases.\n\nThe new combination is real. As far as I know, primitive pairs of the form (f(ε), g(ε)) for rational functions in R with both Tr(ε) and Tr(ε^{-1}) prescribed has not been studied. The character-sum estimates from Lemmas 2.1–2.3 are applied in a standard way, and the lower bound (3.1)–(3.2) and the sieve inequality (3.3) look right to me. The 'all but finitely many' argument in Proposition 4.2 is plausible, though the constant m' = 109 comes from 'trial and error' and I did not verify the arithmetic.\n\nThe main soft spot is the missing nonexistence proof. For the listed pairs, all the computation shows is that the sieve fails to force a positive count. That says nothing about whether an ε exists. You need an upper bound or a field-theoretic obstruction. Without that, Theorem 1.1 overclaims. The computational section is also under-specified: 'we try to calculate suitable d' is not a reproducible algorithm, and the list has a duplicate (4,12). The bounds for m up to 108 are asserted without code or tables, so the finite verification cannot be checked from the text. These are fixable but they are real gaps.\n\nWho gets value: finite-field people working on primitive elements and trace conditions, and anyone using such existence results in cryptography. The paper is not worthless, but in its current form the central claim is not supported.\n\nMy take on the reader's verdict: reject is right if it means 'not acceptable as is', too harsh if it means 'no useful content'. I would send this to a serious referee, with the instruction that the exception list has to be either proved or relabeled as unresolved, and the computations made available. If the authors do that, the sufficient condition alone is a publishable contribution.","headline":"The paper's sufficient condition is solid, but Theorem 1.1's exception list is only a list of sieve failures, not a classification of non-existence.","tokens_in":9820,"tokens_out":3307,"would_cite":false,"duration_ms":23762,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["12E20","11T23"],"pacs":[],"model":"deepseek-v4-flash","headline":"For $m\\ge 9$, every pair $(q,m)$ outside a short explicit list admits a nonzero $\\varepsilon$ whose images under two rational functions are primitive and whose traces are prescribed.","keywords":["finite fields","primitive elements","primitive pairs","rational functions","prescribed traces","character sums","Kloosterman sums","sieve methods"],"falsifier":"In $\\mathbb{F}_{2^{17}}$, take $f(x)=x^3+x$ and $g(x)=x^3+x^2$, and exhaustively check the 131071 nonzero $\\varepsilon$ for each of the four choices $a,b\\in\\mathbb{F}_2$. If some choice yields no $\\varepsilon$ with $f(\\varepsilon)$ and $g(\\varepsilon)$ both primitive and the two trace equations satisfied, Theorem 1.1 is false for the non-exceptional pair $(2,17)$; if all four choices succeed, the theorem survives this test.","tokens_in":8885,"feed_emoji":"🔢","tokens_out":16236,"duration_ms":117695,"temperature":0.7,"pith_summary":"The paper establishes a sufficient condition for a finite field $\\mathbb{F}_{q^m}$ to contain a nonzero element $\\varepsilon$ whose images under two rational functions $f,g$ are both primitive, while the traces of $\\varepsilon$ and $\\varepsilon^{-1}$ equal prescribed values $a,b\\in\\mathbb{F}_q$. The condition is expressed in terms of the degree sums of $f$ and $g$ and the number $W(q^m-1)$ of squarefree divisors of $q^m-1$. Using it, the authors prove that such an $\\varepsilon$ exists in all but finitely many fields $\\mathbb{F}_{q^m}$ over $\\mathbb{F}_q$. When both degree sums are $3$, they sharpen this to a classification: for every prime power $q$ and every $m\\ge 9$, the guarantee holds outside an explicit list of exceptional pairs.","feed_headline":"Two fixed traces still allow primitive pairs for m≥9, barring a few","feed_subtitle":"A sieve over finite fields shows that fixing two traces still leaves room for both images to be primitive.","key_machinery":"The counting function $N_{f,g,a,b}(e_1,e_2)$ counts nonzero elements $\\varepsilon$ for which $f(\\varepsilon)$ is $e_1$-free, $g(\\varepsilon)$ is $e_2$-free, and both trace conditions hold, where an element is $e$-free when it is not a $d$-th power for any non-trivial divisor $d$ of $e$. Expansion through the characteristic functions (2.1) and (2.2) turns this count into sums of mixed character sums $\\chi_{f,g}(d_1,d_2,u,v)$. Lemmas 2.1, 2.2 and 2.3 bound those sums, producing the sufficient condition (3.1) and its simpler version (3.2). The sieving inequality of Theorem 3.3 then upgrades counts from a divisor $d$ to the full $q^m-1$, and Lemma 4.3 bounds $w(n)$ to make the numerical check for $m_1=m_2=3$ finite.","core_discovery":"The central claim is Theorem 1.1: with degree sums $m_1=m_2=3$, for every prime power $q$ and $m\\ge 9$ except the pairs listed in the theorem, and for every pair $f,g$ in the class $\\mathcal{R}$ (rational functions not of the form $c\\,x^j h(x)^d$ with $d>1$ dividing $q^m-1$), there is a nonzero $\\varepsilon\\in\\mathbb{F}_{q^m}$ such that $(f(\\varepsilon),g(\\varepsilon))$ is a primitive pair and $\\operatorname{Tr}_{\\mathbb{F}_{q^m}/\\mathbb{F}_q}(\\varepsilon)=a$, $\\operatorname{Tr}_{\\mathbb{F}_{q^m}/\\mathbb{F}_q}(\\varepsilon^{-1})=b$ for any prescribed $a,b\\in\\mathbb{F}_q$. The paper also proves Proposition 4.2: for $m\\ge 5$, all but finitely many fields $\\mathbb{F}_{q^m}$ contain such an element. The element $\\varepsilon$ itself is not required to be primitive; only its two rational-function values are required to generate $\\mathbb{F}_{q^m}^*$.","pith_inferences":["Beyond the paper, the same characteristic-function expansion with an extra multiplicative condition suggests a sufficient condition for $\\varepsilon$ itself to be primitive in addition to having primitive images.","A natural extension to $k$ rational functions should replace $W(q^m-1)^2$ by $W(q^m-1)^k$ in the sieve, preserving an all-but-finitely-many conclusion.","The exceptional list records pairs where the sieve does not prove existence, not pairs where existence is disproved; direct computation over those small fields would show which cases genuinely fail."],"forward_implications":["For every prime power $q$ and $m\\ge 9$ outside the listed exceptional pairs, the existence of $\\varepsilon$ is guaranteed for all admissible $f,g$ of degree sums $3$ and all prescribed traces $a,b$, with no search required.","For $m\\ge 109$ the numerical criterion (4.5) holds for every prime power $q$, so the theorem is unconditional in that range.","For the listed exceptional pairs the method leaves existence undecided; each case is small enough to be settled by direct computation.","The general sufficient condition (3.2) applies to arbitrary degree sums $m_1,m_2$, so larger degree sums merely delay the threshold $q^{m/2-2}$ rather than requiring new machinery."],"supporting_citations":[{"why":"Supplies Lemma 2.1, the Weil-type bound on multiplicative character sums over rational functions that fixes the constants in Cases II and IV.","marker":"[11]"},{"why":"Supplies Lemma 2.2, the mixed multiplicative-additive character-sum bound used when both additive parameters are present.","marker":"[5]"},{"why":"Supplies Lemma 2.3, the Kloosterman-sum bound used to control the case with both additive parameters nonzero.","marker":"[12]"},{"why":"Provides the sieving inequality in Lemma 3.1 that lifts counts from a divisor d to q^m-1.","marker":"[13]"},{"why":"Aids Lemma 3.2, bounding the difference between N_{f,g,a,b}(pd,d) and theta(p)N_{f,g,a,b}(d,d) during sieving.","marker":"[9]"},{"why":"Supplies Lemma 4.1, the bound on W(M) used to prove the all-but-finitely-many statement.","marker":"[8]"},{"why":"Supplies Lemma 4.3, the bound on w(n) that makes the m >= 109 numerical cut work.","marker":"[1]"}],"fun_headline_variants":["Primitive pairs with two fixed traces, few exceptions for m≥9","Even with traces set, primitive images persist for m≥9","Prescribed traces, primitive values: m≥9 except finite list","For m≥9, rational functions yield primitive pairs with fixed traces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The character-sum estimates quoted in Section 3 must hold with exactly the constants used there, because the sufficient condition and the resulting list of exceptional pairs are built on those constants.","fun_headline_variants_meta":{"raw":{"variants":["Primitive pairs with two fixed traces, few exceptions for m≥9","Even with traces set, primitive images persist for m≥9","Prescribed traces, primitive values: m≥9 except finite list","For m≥9, rational functions yield primitive pairs with fixed traces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000606,"raw_usage":{"total_tokens":2867,"prompt_tokens":1026,"completion_tokens":1841,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":642,"completion_tokens_details":{"reasoning_tokens":1766}},"tokens_in":642,"tokens_out":1841,"duration_ms":14374,"temperature":1.0,"reasoning_tokens":1766,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:09:47.519268+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In $\\mathbb{F}_{2^{17}}$, take $f(x)=x^3+x$ and $g(x)=x^3+x^2$, and exhaustively check the 131071 nonzero $\\varepsilon$ for each of the four choices $a,b\\in\\mathbb{F}_2$. If some choice yields no $\\varepsilon$ with $f(\\varepsilon)$ and $g(\\varepsilon)$ both primitive and the two trace equations satisfied, Theorem 1.1 is false for the non-exceptional pair $(2,17)$; if all four choices succeed, the theorem survives this test.","supporting_citations":[{"cited_title":"Laishram, R","cited_arxiv_id":null,"evidence_quote":"Supplies Lemma 2.1, the Weil-type bound on multiplicative character sums over rational functions that fixes the constants in Cases II and IV."},{"cited_title":"Castro and C","cited_arxiv_id":null,"evidence_quote":"Supplies Lemma 2.2, the mixed multiplicative-additive character-sum bound used when both additive parameters are present."},{"cited_title":"Lidl and H","cited_arxiv_id":null,"evidence_quote":"Supplies Lemma 2.3, the Kloosterman-sum bound used to control the case with both additive parameters nonzero."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the sieving inequality in Lemma 3.1 that lifts counts from a divisor d to q^m-1."},{"cited_title":"Gupta, R","cited_arxiv_id":null,"evidence_quote":"Aids Lemma 3.2, bounding the difference between N_{f,g,a,b}(pd,d) and theta(p)N_{f,g,a,b}(d,d) during sieving."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies Lemma 4.1, the bound on W(M) used to prove the all-but-finitely-many statement."},{"cited_title":"Booker, S","cited_arxiv_id":null,"evidence_quote":"Supplies Lemma 4.3, the bound on w(n) that makes the m >= 109 numerical cut work."}],"review_version":1}