{"id":"b6349839-eddc-4011-8c05-c4b89491e244","arxiv_id":"1908.05093","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Quadratic split quaternion polynomials factorize exactly under conditions that can be read off from intersections with the null quadric, and every such polynomial with linearly independent coefficients or vanishing norm admits a factorization.","lead":"This paper settles when a quadratic polynomial with split quaternion coefficients can be written as a product of two linear factors. It adds a geometric picture of these conditions in projective space and covers degenerate cases that earlier work left open.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Counterexample falsifies §3.3 reduction: P=(i+j)t^2+t+i has non-invertible leading coefficient and PP*=3t^2+1, yet the transformed polynomial is factorable while P is not in the paper's sense.","rationale":"The reader's weakest assumption correctly located the reduction in §3.3, but treated it as an unproved claim needing a proof. The problem is stronger: the reduction is false. The counterexample P=(i+j)t^2+t+i has all the properties the reduction assumes (non-invertible leading coefficient, PP*≠0, a point off the null quadric), its reciprocal transform Q has invertible leading coefficient and is factorable, yet P cannot equal a(t-h1)(t-h2) because the coefficient 1 of t is not in the left ideal S a. This invalidates the central claim of a complete discussion of factorization for quadratic split quaternion polynomials as stated. The issue is directly load-bearing for the headline result and is not resolved by minor edits; either the factorization notion must be broadened (and the non-existence criteria re-examined, since non-monic linear factors can exist without a right zero) or the non-invertible non-vanishing-norm cases must be excluded and treated separately. For these reasons the verdict should move from CONDITIONAL to REJECT.","tokens_in":18606,"tokens_out":28108,"duration_ms":276727,"concrete_test":"Verify the counterexample: compute (i+j)(-i-j)=0, PP*=(a t^2+t+i)(a^* t^2+t-i)=3t^2+1, and with s=x0+x1 i+x2 j+x3 k compute s(i+j)=(x2-x1)+(x0+x3)i+(x0+x3)j+(x1-x2)k, so 1∉S(i+j). Also expand (s+i-j)(s-2i+j)=s^2-i s+1-k, so Q=s^2P(1/s)=i s^2+s+i+j factors as i(s+i-j)(s-2i+j). This settles that factorability of the transformed polynomial does not imply factorability of P under the paper's definition.","verdict_should_be":"REJECT","load_bearing_attack":"The completeness claim depends on the reduction in §3.3 (just before Theorem 3.14) that a non-invertible leading coefficient can be made invertible by a parameter transformation and that factorizability is preserved. This is false for the paper's definition P=a(t-h1)(t-h2). Let a=i+j∈S, so aa*=0, and set P=a t^2+t+i. A direct calculation gives PP*=3t^2+1≠0, and P(0)=i has norm 1, so the curve is not contained in N. The substitution t=1/s gives Q(s)=s^2P(1/s)=i s^2+s+a, whose leading coefficient i is invertible. Since after multiplying by i^{-1} one obtains M(s)=s^2-i s+1-k=(s+i-j)(s-2i+j), Q is factorable by the paper's own criterion. However, if P=a(t-h1)(t-h2), the t-coefficient would be -a(h1+h2)∈S a. But S a={x+y i+y j-x k : x,y∈R}; in particular 1∉S a, while the t-coefficient of P is 1. Hence P has no factorization of the required form. The Möbius reparametrization destroys the monic structure of the factors: h1=-i+j is a zero divisor, so (1-h1 t) cannot be normalized to t-h' under left multiplication by a. Thus the reduction is not merely unproved; it is invalid, and the claimed complete classification is false.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies factorization of quadratic left polynomials P=at^2+bt+c over the split quaternions into products a(t-h1)(t-h2). After reductions to monic polynomials with zero real part of the linear coefficient, it gives inequality criteria for factorizability in the dependent-coefficient case (Theorems 3.5 and 3.6), recalls and expands a proof for independent coefficients (Theorem 3.7), develops a geometric interpretation via remainder polynomials, interpolation lines, and the null quadric (Theorem 3.12 and Theorem 3.14), and treats polynomials with vanishing norm or non-invertible leading coefficient (Theorems 3.15 and 3.17, Corollaries 3.18 and 3.19). The paper claims in Section 4 to have presented a complete discussion of factorizability of quadratic split quaternion polynomials.","tokens_in":18978,"tokens_out":8805,"duration_ms":84045,"significance":"If the classification were correct, it would be a useful complete algebraic-geometric characterization, with direct applications to mechanism science and hyperbolic kinematics. The paper has real strengths: the proofs are mostly constructive, the generic factorization algorithm is used transparently, and the remainder-polynomial viewpoint is elegant and gives a unified interpretation for the invertible-leading-coefficient cases. However, the central completeness claim is false as stated: the reduction in Section 3.3 for non-invertible leading coefficients is invalid, and the explicit counterexample below contradicts Corollary 3.18. The geometric interpretation for monic polynomials with invertible leading coefficient remains valuable, but the claimed coverage of all quadratic split quaternion polynomials is not established.","major_comments":[{"comment":"The reduction to an invertible leading coefficient is invalid. The paper asserts that whenever the norm polynomial does not vanish and the curve has a point not on the null quadric, a parameter transformation can make the leading coefficient invertible and that factorizability of the transformed polynomial implies factorizability of the original. This is false. Consider P=(i+j)t^2+t+i. Here a=i+j has aa*=0, and a direct computation gives PP*=3t^2+1, so P is not contained in N. The substitution t=1/s gives Q(s)=s^2P(1/s)=i s^2+s+i+j, whose leading coefficient i is invertible; after left multiplication by i^{-1} the polynomial becomes s^2-i s+1-k=(s+i-j)(s-2i+j), so Q is factorable. But P itself is not factorable in the paper's sense: if P=(i+j)(t-h1)(t-h2), the coefficient of t is -(i+j)(h1+h2), which lies in S(i+j)={x+y i+y j-x k : x,y∈R}. The coefficient of t in P is 1, and 1 is not in S(i+j) because an element of S(i+j) with zero i- and j-coefficients has the form x-x k and hence is never 1. Thus the asserted preservation of factorizability under re-parametrization is false, and Corollary 3.18 is contradicted by this example. The claimed completeness of the classification for non-invertible leading coefficients therefore collapses.","section":"Section 3.3, just before Theorem 3.14"},{"comment":"Theorem 3.14 is stated without proof. The text only says that its content is visualized in Figure 1 and then discusses the figure qualitatively. Since the theorem is presented as the geometric unification of the dependent-coefficient cases and is used in the completeness discussion, a proof is required. Moreover, if the intended proof relies on the reduction criticized in the previous comment, the theorem is not sound for non-invertible leading coefficients.","section":"Theorem 3.14"},{"comment":"The necessity direction for the case bb*<0 is not proved. After presenting a candidate factorization under the stated inequalities, the proof says only that 'a detailed inspection of the equation system (10)' shows that no solutions exist when the conditions are violated. This is a load-bearing step in an iff statement. The omitted verification must be supplied, or the theorem must be weakened to a one-directional statement.","section":"Theorem 3.6, third bullet (bb* < 0)"},{"comment":"Corollary 3.18 claims that every quadratic split quaternion polynomial with linearly independent coefficients admits a factorization. The counterexample from the first comment has linearly independent coefficients (i+j, 1, and i are independent over R) and does not admit a factorization in the paper's sense. Hence the corollary as stated is false, and the Section 4 claim of a 'complete discussion of factorizability of quadratic polynomials over the split quaternions' is not supported.","section":"Corollary 3.18 and Section 4"}],"minor_comments":[{"comment":"Theorem 3.7 is attributed to [10], but the surrounding text says 'In [8] the authors showed...' and discusses an improved proof of the statement. The citation should be checked and made consistent with the bibliography.","section":"Theorem 3.7 and references"},{"comment":"The figures would be easier to verify if each row had a short caption explicitly identifying which theorem or inequality case it illustrates; currently the reader must infer the correspondence from surrounding text.","section":"Section 3.3, Figure 1 and Figure 2"},{"comment":"The paper uses 'vectorial' for Re(h)=0 but later speaks of the 'vector part'; these terms should be used consistently, and the definition of Im(h) could be cross-referenced when first used.","section":"Section 2.1, notation"}],"recommendation":"reject","confidential_remarks":"The counterexample in the main report is simple and verifiable, and it directly falsifies a stated corollary and the paper's main completeness claim. The algebraic results for invertible leading coefficients and the geometric remainder-polynomial interpretation still appear to be sound and could form the basis of a substantially revised paper that either restricts the classification claims or proves a correct treatment of non-invertible leading coefficients. As written, the central claim is false."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does useful work on quadratic split quaternion polynomials, and the degenerate non-invertible leading coefficient results are genuinely new. But the advertised complete classification is not correct as stated: the reduction in §3.3 is false, and there is a concrete counterexample.\n\nThe monic invertible case is mostly a repackaging of Abrate and Cao, and the paper says so. The new material is Theorems 3.15 and 3.17 with Corollaries 3.18 and 3.19, covering vanishing norm and non-invertible leading coefficients. The proofs there are constructive and look right; the geometric interpretation in Section 3.3 is nice and the figures help. This part is worth having.\n\nThe problem is the bridge between Sections 3.3 and 3.4. Just before Theorem 3.14 the paper asserts that if the norm polynomial does not vanish, a parameter transformation can make the leading coefficient invertible, and factorizability of the transformed polynomial implies factorizability of the original. That implication is false. Take P=(i+j)t^2+t+i. Here a=i+j has zero norm, PP*=3t^2+1, and P(0)=i has norm 1, so the curve is not entirely in N. The substitution t=1/s gives Q(s)=s^2P(1/s)=i s^2+s+i+j; after multiplying by i^{-1} we get a monic quadratic that factors as (s+i-j)(s-2i+j). But if P had a factorization (i+j)(t-h1)(t-h2), the t-coefficient would be -(i+j)(h1+h2). Every element of (i+j)S has its real part equal to its k-coefficient (a direct calculation gives {x+y i+y j+x k}); the coefficient 1 does not have that shape. So P is not factorable in the paper's sense. This is not a minor technicality; it breaks the completeness claim in Section 4.\n\nThe other issues are smaller. Theorem 3.14 is stated without proof. The necessity direction of Theorem 3.6 for bb*<0 is dismissed as 'a detailed inspection' of the system; that should be shown. Both are likely repairable. The citation practice is honest: the monic results are attributed to earlier work.\n\nBottom line: This paper is for people working on factorization of motion polynomials and split quaternion algebra. The degenerate-case theorems are a solid contribution, and the geometric viewpoint is useful. But the 'complete discussion' is not true as written. The authors need to either prove a correct reduction, add hypotheses that make it true, or drop the completeness claim. I would send it to review with the counterexample attached; the flaw is specific and probably fixable, and the paper deserves referee time rather than a desk rejection.","headline":"Genuinely new degenerate-case results, but the claimed complete classification is false as stated: the §3.3 reduction is contradicted by a simple counterexample.","tokens_in":19489,"tokens_out":9425,"would_cite":false,"duration_ms":78804,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["12D05","16S36","51M09","51M10","70B10"],"pacs":[],"model":"deepseek-v4-flash","headline":"Quadratic split quaternion polynomials either factor explicitly or provably cannot; the test is geometric.","keywords":["split quaternions","polynomial factorization","skew polynomial ring","null quadric","left and right rulings","zero divisors","projective geometry","motion polynomials"],"falsifier":"The most direct falsifier is to search symbolically for a factorization of $P=t^2+bt+\\lambda+\\mu b$ with $bb^*=0$, $\\lambda+\\mu^2\\ne0$, and $\\lambda\\ge0$; Theorem 3.6 says none exists, so any factorization found would collapse the classification. Independently, a non-invertible-leading-coefficient polynomial whose norm does not vanish but that resists the Section 3.3 re-parameterization would expose the unproved gap.","tokens_in":18430,"feed_emoji":"📐","tokens_out":8668,"duration_ms":82209,"temperature":0.7,"pith_summary":"Quadratic polynomials over the split quaternions are the simplest case where zero divisors complicate factorization, and this paper aims to say exactly which ones factor into linear factors. It derives necessary and sufficient conditions, with inequality tests for dependent coefficients and a geometric line criterion in general, and proves that polynomials with linearly independent coefficients always factor, as do all polynomials with vanishing norm. This matters for kinematics because factoring a split quaternion polynomial decomposes a rational motion into lower-degree motions, so a complete quadratic test is a building block for motion factorization. The geometric formulation via the null quadric turns the many case distinctions into one picture.","feed_headline":"Complete factorization test found for quadratic split quaternions","feed_subtitle":"A geometric criterion on the null quadric decides factorizability, including zero divisors and non-invertible leading coefficients.","key_machinery":"The load-bearing object is the null quadric $N$ defined by $hh^*=0$ in the projective space $\\mathbb{P}(S)$, together with its two rulings: the left rulings $L=\\{[r]: rh^*=0\\}$ and right rulings $R=\\{[r]: h^*r=0\\}$. Remainder polynomials $R_{ij}=P-M_{ij}$, for real quadratic factors $M_{ij}$ of the norm polynomial $PP^*$, parameterize lines through the points where the curve $P$ meets $N$; a root of such a remainder gives a linear factor. Theorem 2.8 supplies explicit parametrizations of the affine two-planes solving $g=xh$, which is used to construct zeros and factorizations in null-line cases.","core_discovery":"The central claim is that factorizability of a quadratic split quaternion polynomial $P=at^2+bt+c$ is fully characterized. After normalization to $P=t^2+bt+c$ with $\\operatorname{Re}(b)=0$, factorizability is equivalent to existence of a right zero (Lemma 3.1) and is decided by explicit inequalities when $1,b,c$ are dependent, while independent coefficients always factor (Theorem 3.7). Geometrically, $P$ parameterizes a line segment or conic in the projective space over the split quaternions; a degree-one remainder polynomial obtained from a real quadratic factor of the norm polynomial has a root exactly when the corresponding interpolation line is real. The vanishing-norm case always factorizes (Corollary 3.19), and non-invertible leading coefficients are handled by a parameter reduction whenever the curve is not contained in the null quadric.","pith_inferences":["The same line-intersection picture should extend to higher-degree split quaternion polynomials: factorizability is likely tied to the existence of a real linear factor among the remainders obtained from real factors of the norm polynomial.","A practical algorithm can be read off the geometric criterion: factor the real norm polynomial into quadratics, compute the corresponding remainder lines, and test whether any has independent coefficients and a real zero, avoiding the case distinctions of the inequalities.","The vanishing-norm corollary suggests that zero-divisor structure alone does not obstruct factorization for quadratics; the obstructions all live in the non-vanishing-norm dependent-coefficient cases.","The unproved reduction before Theorem 3.14 can likely be closed by a fractional-linear re-parameterization sending any parameter value with $P(t)\\notin N$ to infinity, since the new leading coefficient is then proportional to $P(t)$ and hence invertible."],"forward_implications":["Corollary 3.18: every quadratic split quaternion polynomial with linearly independent coefficients admits a factorization.","Corollary 3.19: every quadratic split quaternion polynomial with vanishing norm admits a factorization.","For monic polynomials with dependent coefficients, factorizability reduces to checking the inequalities in Theorems 3.5 and 3.6.","Geometrically, a monic quadratic without a real factor factorizes exactly when one of its remainder polynomials is a real line with independent coefficients (Theorem 3.12).","The geometric criterion unifies cases that require separate inequality conditions, and extends to polynomials with non-invertible leading coefficient whose curve is not contained in the null quadric."],"supporting_citations":[{"why":"Supplies the earlier quadratic formulas for generalized quaternions that the paper compares with and partially extends.","marker":"[1]"},{"why":"Provides the prior split-quaternion quadratic factorization results that the new inequality criteria and geometric criteria refine.","marker":"[2]"},{"why":"Motivates factorization as decomposition of rational motions into revolute-joint motions.","marker":"[5]"},{"why":"Gives the Hamiltonian quaternion analogue whose zero set differs because norms are non-negative.","marker":"[6]"},{"why":"Supplies the zero-factor correspondence (Lemma 3.1) and the generic factorization algorithm used throughout.","marker":"[8]"},{"why":"Is the source of the theorem that independent coefficients always admit a factorization (Theorem 3.7).","marker":"[10]"},{"why":"Provides the dense-class factorization theory and algorithms that this quadratic classification complements.","marker":"[11]"},{"why":"Supplies the split quaternion geometry of the null quadric and its rulings, including Lemma 2.2.","marker":"[12]"}],"fun_headline_variants":["Geometric test nails quadratic split quaternion factorization","Inequality conditions decide split quaternion factorization","Zero divisors included in new split quaternion factorization rule","Conic geometry characterizes split quaternion polynomial roots","Explicit test for factorizing quadratic split quaternions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"For polynomials whose norm polynomial does not vanish, the classification assumes a parameter change can make the leading coefficient invertible; if any such polynomial resists that change, the non-invertible-leading-coefficient cases are not fully covered.","fun_headline_variants_meta":{"raw":{"variants":["Geometric test nails quadratic split quaternion factorization","Inequality conditions decide split quaternion factorization","Zero divisors included in new split quaternion factorization rule","Conic geometry characterizes split quaternion polynomial roots","Explicit test for factorizing quadratic split quaternions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00041,"raw_usage":{"total_tokens":2019,"prompt_tokens":735,"completion_tokens":1284,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":351,"completion_tokens_details":{"reasoning_tokens":1210}},"tokens_in":351,"tokens_out":1284,"duration_ms":8925,"temperature":1.0,"reasoning_tokens":1210,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:24:52.436544+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The most direct falsifier is to search symbolically for a factorization of $P=t^2+bt+\\lambda+\\mu b$ with $bb^*=0$, $\\lambda+\\mu^2\\ne0$, and $\\lambda\\ge0$; Theorem 3.6 says none exists, so any factorization found would collapse the classification. Independently, a non-invertible-leading-coefficient polynomial whose norm does not vanish but that resists the Section 3.3 re-parameterization would expose the unproved gap.","supporting_citations":[{"cited_title":"Quadratic formulas for generalized quaternions","cited_arxiv_id":null,"evidence_quote":"Supplies the earlier quadratic formulas for generalized quaternions that the paper compares with and partially extends."},{"cited_title":"Quadratic formulas for split quaternions","cited_arxiv_id":"1905.08153","evidence_quote":"Provides the prior split-quaternion quadratic factorization results that the new inequality criteria and geometric criteria refine."},{"cited_title":"Quadratic formulas for quaternions","cited_arxiv_id":null,"evidence_quote":"Gives the Hamiltonian quaternion analogue whose zero set differs because norms are non-negative."},{"cited_title":"Kempe’s universality theorem for rational space curves","cited_arxiv_id":null,"evidence_quote":"Is the source of the theorem that independent coefficients always admit a factorization (Theorem 3.7)."},{"cited_title":"Factorization of motion polynomi- als","cited_arxiv_id":null,"evidence_quote":"Provides the dense-class factorization theory and algorithms that this quadratic classification complements."},{"cited_title":"The geometry of quadratic quater- nion polynomials in Euclidean and non-Euclidean planes, in: Cocchiarella, L","cited_arxiv_id":null,"evidence_quote":"Supplies the split quaternion geometry of the null quadric and its rulings, including Lemma 2.2."}],"review_version":1}