{"id":"44975a72-69a8-4692-9d6d-3b09e750de52","arxiv_id":"1908.04303","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Generalized P and CP transformations of order two in the three-Higgs-doublet model form two equivalence classes for P and a single equivalence class for CP.","lead":"This paper classifies the generalized parity and CP symmetries of a three-Higgs-doublet model, showing there are two inequivalent classes of P and one class of CP. The result provides a reference classification for constructing three-Higgs-doublet models with exact discrete symmetries.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Completeness proof depends on an unpublished computer check of Eq. (5.6); without it, the classification is not independently verifiable.","rationale":"The paper's central claim is a complete classification of order-two generalized P and CP transformations in the 3HDM under the stated bilinear conditions. I checked the internal consistency of the classification: the eight diagonal sign matrices in Table I indeed satisfy the d-tensor condition (6.17) for the triples listed in Table II, and the explicit flavour equivalences (e.g., S4 = R(U(0,0,π/2)) S3 R^T) are correct. The hand-written parts of the proof are plausible and the SO(8) analysis in Appendix C is rigorous. However, the proof of equation (5.6), which is essential for reducing the geometric orbit-space condition to the algebraic equations solved in Section 6, is delegated to an unspecified computer program. No code or output is provided, so the completeness part of the claim cannot be independently checked by a reader. The reader's stated weakest assumption focuses on the scope of the classification (K0 mixing, order-four CP), which the paper explicitly acknowledges and is not an internal flaw; the more load-bearing concern is the verifiability of (5.6). I therefore agree with the conditional verdict but not with the specific framing in the weakest_assumption field, hence 'partial'. A successful independent recomputation of the polynomial identities in Appendix A would settle the concern and justify upgrading the verdict.","tokens_in":28073,"tokens_out":38091,"duration_ms":347605,"concrete_test":"Use a computer algebra system to recompute D_abc = d_{a'b'c'} C_{a'a} C_{b'b} C_{c'c} for the eight special cases (a)-(h) of Appendix A, with C a general 8×8 orthogonal symmetric matrix satisfying (4.14),(5.2), and check that the resulting polynomial identities in L0..L3 force D_abc = d_abc. If the identities do not force D=d, the completeness proof of Table I is invalid; if they do, the conditional can be upgraded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central classification (Table I, Section 6) is derived from equations (6.1)-(6.2), which in turn follow from (5.2) and (5.6). Equation (5.6) — the statement that C preserves the d-tensor — is proven in Appendix A only by comparing polynomials for special cases (a)-(h) 'with the help of a computer program'; no code, output, or explicit polynomial identities are provided. This step is load-bearing: if the computer check is insufficient or erroneous, the reduction of the determinant-preservation condition to the algebraic condition (6.17) fails, and the claimed list of eight solutions (or the equivalence classes) could be wrong. The hand-proven parts (special case (a) for (5.2), the SO(8) analysis in Appendix C, and the explicit flavour equivalences for S4, S5, S6, S7, S8) are consistent and check out, but they do not remove the need for a reproducible verification of (5.6). The scope restriction (K0 fixed, order-two) is explicit in (4.11)-(4.14) and acknowledged at the end of Section 8, so it is not an internal gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript classifies, within the bilinear formalism, all generalized parity and generalized CP transformations of the 3HDM that are linear on the eight bilinears, leave K0 fixed, preserve the allowed bilinear domain, and square to the identity (conditions (4.11)-(4.14)). The authors derive that such transformations are described by symmetric orthogonal 8x8 matrices C satisfying the d-tensor preservation condition (5.6), reduce C to diagonal form using flavour transformations, obtain exactly eight solutions (S1)-(S8) listed in Table I, and show that these fall into two equivalence classes for generalized P (representatives S1 and S8) and one class for generalized CP (representative S3). They also derive the conditions for the potential to be invariant under these transformations (7.4)-(7.6), recover the known criterion that CP invariance is equivalent to the existence of a conventional basis with all parameters real, and generalize the standard CP transformation on bilinears to the nHDM with the explicit matrix (E26).","tokens_in":28288,"tokens_out":10414,"duration_ms":107673,"significance":"If correct, the classification is a useful and nontrivial result for multi-Higgs-doublet model building. It provides a complete set of candidate generalized discrete symmetries within the stated class, gives explicit invariance conditions for the potential, and settles the equivalence-class structure under flavour transformations. The derivation is algebraic and parameter-free, and it reproduces the known THDM/3HDM CP-real-basis theorem as a consistency check. The nHDM formula (E26) is an explicit, easily applicable result. The main caveat is the reproducibility of the computer-assisted proof of (5.6), which is load-bearing for the central classification. The scope restriction to order-two, K0-preserving, linear transformations is explicit in Section 4 and acknowledged in Section 8, so the completeness claim is well defined, but it should not be overinterpreted as covering order-four transformations such as those of ref. [32].","major_comments":[{"comment":"The proof of (5.6) is load-bearing for the classification, since it leads to (6.2) and then to condition (6.17), which selects the eight solutions in Table I. However, the verification that D_abc = d_abc for the special cases (a)-(h) is stated to be done \"with the help of a computer program\", and neither the program, its output, nor the explicit polynomial identities are provided in the manuscript. The same issue affects the assertion (A27)-(A28) used in Section 7 to derive the invariance conditions (7.4). As a result, the completeness of the eight-solution list and the necessity part of the potential-invariance conditions cannot be independently checked by the reader. Please supply the program as supplementary material, display the coefficient identities case by case, or give a hand proof of (5.6).","section":"Appendix A, Eqs. (A21)-(A26); used in Section 6"}],"minor_comments":[{"comment":"The sentence referring to \"appendix C\" for the standard P and CP transformations in the nHDM is incorrect; the relevant material is presented in Appendix E.","section":"Section 4, text after Eq. (4.10)"},{"comment":"The text says \"We consider then six more special cases (b) ... (e)\", but only four cases, (b), (c), (d), and (e), are listed; please correct the count.","section":"Appendix A, after Eq. (A12)"},{"comment":"The notation c_8 c^(8) c^(8)T in Eq. (5.22) is confusing because the scalar eigenvalue c_8 and the eigenvector c^(8) are denoted too similarly; please use a distinct symbol for the eigenvalue.","section":"Section 5.2, Eq. (5.22)"},{"comment":"The set-builder notation for I_n^as is hard to parse for small k, since for k=2 the displayed sequence \"k^2-2k+2, k^2-2k+4, ..., k^2-2\" degenerates to a single term; rewriting the set as {k^2+2j-1 | k=1,...,n-1, j=1,...,k} would be unambiguous.","section":"Appendix E, Eq. (E24)"}],"recommendation":"major_revision","confidential_remarks":"The central result appears credible and the hand-checked parts of the derivation are consistent. The sole blocking issue is the unpublished computer check behind Eq. (5.6), which is load-bearing for the classification. If the authors provide the code or explicit polynomial identities as supplementary material, I would be willing to accept the paper. The manuscript is within the scope of the journal and the authors' prior bilinear-formalism work is appropriately cited."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Maniatis and Nachtmann finish the classification of order-two generalized P and CP transformations in the 3HDM, and the answer is clean: two parity classes, one CP class, with explicit bilinear matrices. The nHDM result—standard CP acts as diag(±1) with minus entries on the antisymmetric generalized Gell-Mann directions—is simple and correct, and the consistency check reproducing the known real-basis CP theorem is reassuring.\n\nThe bilinear method is standard and the derivation is mostly careful. Sections 5 and 6, plus Appendix C, reduce the problem to eight candidate matrices and then show the flavour equivalences. That part is hand-checkable and I think it holds up. The explicit equivalence transformations in Section 6 are a nice touch.\n\nThe soft spot is exactly what the stress-test flags. Equation (5.6), the claim that any admissible transformation preserves the d-tensor, is proven in Appendix A only by comparing polynomials for eight special cases \"with the help of a computer program.\" No code, no output, no explicit polynomial identities, not even the case (a) displayed for (5.6) the way it is for (5.2). That step is load-bearing: without it, the reduction to the algebraic condition (6.17) fails and the classification has a hole. The authors do show the hand-checkable case (a) for (5.2), and the 3HDM-to-THDM embeddings are plausible, but a referee cannot verify (5.6) from the text alone. This is a genuine reproducibility gap, not a manufactured one.\n\nA smaller point: the completeness is explicitly scoped to transformations that keep K0 fixed, act linearly on the bilinears, and square to the identity—stated in (4.11)-(4.14) and acknowledged at the end. That is not a hidden assumption, but the abstract's \"complete answer\" could over-promise to casual readers.\n\nWho is this for? Multi-Higgs model builders, especially 3HDM people who need to know which generalized P or CP symmetries a potential can support. They will cite this paper. It is not a framework-changer, but it is a useful reference result with a clear method.\n\nMy recommendation: if this lands on your desk, send it to a referee with experience in bilinear methods and computational algebra. Ask the authors to make the computer code available or provide hand-checkable polynomial identities for (5.6), ideally both. If they can satisfy that, the classification is solid and publishable. If not, the result stays conditional.","headline":"A clean, mostly hand-checkable classification of order-two generalized P and CP in the 3HDM, but the load-bearing proof of (5.6) rests on an unpublished computer check that a referee should ask to see.","tokens_in":28799,"tokens_out":2199,"would_cite":true,"duration_ms":22965,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.30.Er","12.60.Fr"],"model":"deepseek-v4-flash","headline":"The paper proves that every generalized parity or CP transformation in the three-Higgs-doublet model that acts linearly on the bilinears and squares to the identity falls into exactly three equivalence classes: two parity classes and one…","keywords":["three-Higgs-doublet model","generalized CP transformations","generalized parity transformations","bilinear formalism","Higgs potential symmetries","n-Higgs-doublet model","SU(n) generators","discrete symmetries"],"falsifier":"A direct computer search over real symmetric $8\\times 8$ matrices $C$ with $C^2=1$ satisfying $d_{a'b'c'}C_{a'a}C_{b'b}C_{c'c}=d_{abc}$ would settle the completeness claim: any solution whose eigenvalue pattern is not one of the eight rows of table I, or any explicit transformation satisfying (4.11)–(4.14) that mixes $K_0$ with the $K_a$, would refute the claimed exhaustive classification.","tokens_in":27877,"feed_emoji":"⚡️","tokens_out":12831,"duration_ms":106357,"temperature":0.7,"pith_summary":"The paper asks which discrete space-time symmetries a model with three Higgs doublets can possess. Working with the eight real bilinears that encode the gauge-invariant products of the doublets, it classifies all generalized parity (P) and CP transformations that are linear, preserve the allowed space, and return to the identity when applied twice. The central result is that, up to Higgs-basis changes, there are exactly two classes of such parity transformations and exactly one class of CP transformations; in particular every generalized CP transformation is equivalent to the standard one. This matters because it settles, for the scalar sector of the 3HDM, which discrete symmetries a potential may be forced to obey, and it gives the explicit parameter conditions for each class. The paper further shows that in an $n$-doublet model the standard CP transformation on the bilinears is a diagonal matrix of $\\pm 1$ entries, with minus signs exactly on the $n(n-1)/2$ antisymmetric generator directions.","feed_headline":"All generalized P and CP symmetries in the 3HDM fit in three classes","feed_subtitle":"The result tells model-builders exactly which discrete space-time symmetries three Higgs doublets can enforce.","key_machinery":"The central object is the eight-component vector $K$ of bilinears, obtained by decomposing the hermitian matrix $K=\\phi\\phi^\\dagger$ (with $\\phi$ the $3\\times 2$ matrix of doublet fields) in a basis of SU(3) generators. A generalized P or CP transformation is a real $8\\times 8$ matrix $C$ acting linearly on $K$, preserving the length of $K$ and the allowed orbit space, and squaring to the identity; the constraints force $C$ to be symmetric and orthogonal and to satisfy the SU(3) $d$-symbol condition $d_{a'b'c'}C_{a'a}C_{b'b}C_{c'c}=d_{abc}$. Flavour transformations act by $C\\to R(U)C\\,R(U)^T$ with $R(U)\\in\\mathrm{SO}(8)$, and the proof uses these to bring $C$ to diagonal form, reducing the classification to an eigenvalue problem plus the $d$-symbol condition, whose solutions are the eight matrices of table I.","core_discovery":"For the 3HDM with only Higgs and gauge fields, the complete set of generalized P and CP transformations satisfying the conditions (4.11)–(4.14) consists of the eight matrices (S1)–(S8). Under flavour transformations these form exactly two equivalence classes of parity transformations, represented by the identity and by $\\mathrm{diag}(1,1,1,-1,-1,-1,-1,1)$, and one equivalence class of CP transformations, represented by $\\mathrm{diag}(1,-1,1,1,-1,1,-1,1)$. A potential is invariant under such a transformation precisely when its bilinear parameters satisfy $\\xi=C\\xi$, $\\eta=C\\eta$, and $E=CE\\,C^T$; the known criterion that a potential is CP-invariant exactly when a basis with all real parameters exists is recovered from these conditions. In the nHDM the standard CP transformation acts on the $n^2-1$ bilinears as the diagonal matrix (E26), whose $-1$ entries sit precisely on the $n(n-1)/2$ antisymmetric generalized SU($n$) generator directions; for $n=4$ this matrix has determinant $+1$, so the reflection picture familiar from the two- and three-doublet cases does not extend.","pith_inferences":["The same flavour-equivalence strategy could be applied to transformations requiring four applications to close, such as the order-four CP of ref. [32]; extending the equations to $C^4=1$ would likely produce additional inequivalent classes and new potential constraints.","The count of antisymmetric directions, $n(n-1)/2$, suggests that the geometric intuition “CP is a reflection” fails specifically when $n(n-1)/2$ is even; this may change how spontaneous CP violation is diagnosed in models with four or more doublets.","One direct test of the classification is numerical: randomly generate symmetric orthogonal $8\\times 8$ matrices with $\\pm 1$ eigenvalues and check the $d$-symbol condition; any solution outside table I would signal an error in the proof, while agreement would lend confidence to the completeness claim."],"forward_implications":["Any 3HDM potential that is invariant under some generalized CP transformation is, in a suitable basis, invariant under the standard CP transformation, with the parameter restrictions $\\xi_2=\\xi_5=\\xi_7=0$, $\\eta_2=\\eta_5=\\eta_7=0$, and the corresponding zero pattern in $E$.","There is a genuinely distinct generalized parity class: a potential invariant under it must satisfy $\\xi_1=\\xi_2=\\xi_4=\\xi_5=0$, $\\eta_1=\\eta_2=\\eta_4=\\eta_5=0$, and the matching $E$ conditions, in the basis of the representative (S8).","The standard parity transformation is represented by the identity matrix on the bilinears, so every 3HDM potential built from Higgs and gauge fields is automatically invariant under standard parity.","In the nHDM the standard CP transformation is always a diagonal $\\pm 1$ matrix on the bilinears, with exactly $n(n-1)/2$ minus signs; consequently the determinant is $+1$ when $n(n-1)/2$ is even, so for $n=4$ and higher the standard CP action is no longer a reflection in bilinear space."],"supporting_citations":[{"why":"Supplies the 3HDM bilinear formalism and the potential parametrization in terms of $\\xi$, $\\eta$, and $E$ that the whole classification is built on.","marker":"[14]"},{"why":"Gives the geometric view of generalized CP transformations as reflections in the two-doublet case, the pattern the 3HDM classification extends.","marker":"[5]"},{"why":"Provides the two-doublet bilinear identities used in the special-case proofs that establish equations (5.2) and (5.6).","marker":"[9]"},{"why":"Supplies the numbering scheme of generalized SU($n$) generators used for the nHDM result (E26).","marker":"[15]"},{"why":"Contains the known criterion that CP invariance is equivalent to the existence of a basis with all real parameters, which the paper reproduces in appendix D.","marker":"[4]"},{"why":"Describes the order-four CP transformation that lies outside the paper's assumptions, marking the boundary of the claimed completeness.","marker":"[32]"}],"fun_headline_variants":["3HDM: two P classes, one CP class under flavour","Complete classification of P and CP in 3HDM","nHDM standard CP: explicit diagonal ±1 matrix","For four doublets, standard CP is not a reflection"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification assumes the symmetry is a linear map on the eight Higgs-bilinear variables that keeps the overall scale direction fixed, preserves the allowed space, and returns to the identity when applied twice; any symmetry that mixes the scale direction in, or that needs four applications to close, would be missed.","fun_headline_variants_meta":{"raw":{"variants":["3HDM: two P classes, one CP class under flavour","Complete classification of P and CP in 3HDM","nHDM standard CP: explicit diagonal ±1 matrix","For four doublets, standard CP is not a reflection"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000815,"raw_usage":{"total_tokens":3600,"prompt_tokens":1004,"completion_tokens":2596,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":620,"completion_tokens_details":{"reasoning_tokens":2527}},"tokens_in":620,"tokens_out":2596,"duration_ms":19676,"temperature":1.0,"reasoning_tokens":2527,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:47:57.306376+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct computer search over real symmetric $8\\times 8$ matrices $C$ with $C^2=1$ satisfying $d_{a'b'c'}C_{a'a}C_{b'b}C_{c'c}=d_{abc}$ would settle the completeness claim: any solution whose eigenvalue pattern is not one of the eight rows of table I, or any explicit transformation satisfying (4.11)–(4.14) that mixes $K_0$ with the $K_a$, would refute the claimed exhaustive classification.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the geometric view of generalized CP transformations as reflections in the two-doublet case, the pattern the 3HDM classification extends."},{"cited_title":"(C23) Therefore, the r.h.s","cited_arxiv_id":null,"evidence_quote":"Provides the two-doublet bilinear identities used in the special-case proofs that establish equations (5.2) and (5.6)."},{"cited_title":"(4.2) For the bilinears we find from (4.1) Kα(x) Ps −→K′ α(x) =Kα(x′)","cited_arxiv_id":null,"evidence_quote":"Contains the known criterion that CP invariance is equivalent to the existence of a basis with all real parameters, which the paper reproduces in appendix D."},{"cited_title":"Exploring the Quark Flavour Puzzle within the 3 Higgs Model","cited_arxiv_id":"1705.09743","evidence_quote":"Describes the order-four CP transformation that lies outside the paper's assumptions, marking the boundary of the claimed completeness."}],"review_version":1}