{"id":"40e1d20d-c7a4-44e7-abf3-67b4913c0038","arxiv_id":"2504.14293","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"An injective Lie algebra morphism exists from the double shuffle Lie algebra ds into the Kashiwara-Vergne Lie algebra krv, compatible with the known injections from grt.","lead":"The paper proves that the double shuffle Lie algebra embeds into the Kashiwara-Vergne Lie algebra, completing a triangle of known injections that starts from the Grothendieck-Teichmüller Lie algebra. The result settles a proof gap left by an earlier argument that relied on an unproven claim by Écalle.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Corollary 1 relies on a false uniqueness assertion for derivations, so the formula for D(t01) and the push-invariance step leading to Theorem 1 are not established.","rationale":"The reader's weakest assumption identifies the same load-bearing flaw: the uniqueness assertion in Corollary 1 is plainly false, and the formula for D(t01) derived from it feeds directly into Corollary 2 and Lemma 5. I checked that the issue is not merely cosmetic: D(t12)=0 gives D(t01)=-D(t02), so the desired first line of (4) reduces to an identity [f(t01,-t12),t01] = -[f(t02,-t12),t02] that is neither stated nor proved. The pullback derivation in Corollary 2 needs both image formulas, and the push-invariance of f(y,-z) depends on that pullback. Lemma 5 then transports this push-invariance to f(z,-y), which is the g used in Theorem 1 and is needed for the divergence computation and the claim that D_{g,h} lands in krv. Since the central construction is not established, the manuscript does not prove Theorem 1 as written. This is an argument-level gap, not a disagreement with the intended result; the theorem may well be true and the gap may be repairable. I therefore leave the reader's REJECT verdict unchanged rather than moving to a harsher or more lenient one.","tokens_in":6713,"tokens_out":13776,"duration_ms":120037,"concrete_test":"Audit Corollary 1 by replacing the false uniqueness with the correct statement: if E=D'-D, then E(a)=0 and E([a,b])=0 force E(b)=λa. Check whether the remaining hypotheses of Theorem 2 (D(a) push-invariant and odd, D(t02)=[f(t02,-t12),t02]) force E(t02)=0 and hence λ=0. If they do not, test the missing identity directly for the lowest-degree nontrivial element of ds (degree 5) using the mould-theoretic construction cited in Theorem 2 from [11, Thm 5.6.1]: compute [f(t01,-t12),t01] and -[f(t02,-t12),t02]; if they differ, the first line of (4) is false and the proof of Theorem 1 fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Corollary 1 claims that a derivation annihilating [a,b] is uniquely determined by its value on a. This is false: D(a)=0, D(b)=a defines a nonzero derivation with D([a,b])=0. The proof of D'=D therefore only shows that D and D' agree on a and on [a,b]; the difference E=D'-D could still satisfy E(a)=0 and E(b)=λa. Consequently D is not shown to commute with the involution ι, and the first line D(t01)=[f(t01,-t12),t01] is not justified. From D(t12)=0 one only gets D(t01)=-D(t02), so the first line of (4) would require the unproved identity [f(t01,-t12),t01] = -[f(t02,-t12),t02]. Corollary 2 pulls back both D(t01) and D(t02) to obtain the special derivation with x↦[f(x,-z),x] and y↦[f(y,-z),y], and uses it to conclude that f(y,-z) is push-invariant. Without D(t01), that pullback is not available. Lemma 5 and the divergence argument in Theorem 1 depend on this push-invariance. Thus Theorem 1 is not supported by the proof as written, even though the statement may be true.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims a proof that the double shuffle Lie algebra ds injects into the Kashiwara-Vergne Lie algebra krv via the map f ↦ D_{g,h} with g = f(-x-y,-y), completing a commutative triangle with the known injections grt → ds and grt → krv. The argument proceeds by invoking a mould-theoretic derivation D on Lie[a,b] satisfying certain properties (Theorem 2), deriving formulas for D(t01), D(t02), D(t12) (Corollary 1), using these to prove push-invariance of f(y,-z) (Corollary 2, Lemma 5), constructing a special tangential derivation with the correct divergence condition, and finally checking the Lie morphism property.","tokens_in":6967,"tokens_out":6836,"duration_ms":58448,"significance":"If the main theorem is correct, it is a significant structural result: it closes the long-sought injection ds → krv and completes the commutative triangle grt → ds → krv. The paper is clearly written and honestly reports the previous incomplete proof and the reliance on work of Ecalle and of Enriquez–Furusho. The strategy is attractive and, except for the critical gap discussed below, the calculations are coherent. However, the central proof as written is invalid because of a false uniqueness assertion in Corollary 1; this gap cascades through Corollary 2, Lemma 5, and the construction of the special derivation in Theorem 1. The claimed theorem is therefore not established by this manuscript.","major_comments":[{"comment":"The proof of Corollary 1 uses the assertion that 'a derivation annihilating [a,b] is uniquely determined by its value on a.' This is false. For instance, the derivation E of Lie[a,b] defined by E(a)=0, E(b)=a satisfies E([a,b])=[E(a),b]+[a,E(b)]=0, so E annihilates [a,b] while vanishing on a but is not zero. Consequently, comparing D and D' = ιDι on a and [a,b] only shows that E = D-D' satisfies E(a)=0 and E([a,b])=0, which permits E(b) to be a scalar multiple of a. The claimed conclusion D' = D, hence D commutes with ι, is not established. The first line of (4), namely D(t01)=[f(t01,-t12),t01], depends directly on this commutativity and is therefore unsupported.","section":"Corollary 1"},{"comment":"Because the formula for D(t01) in (4) is not proven, the pullback derivation on Lie[x,y] displayed in (7) is not available as written. From D(t12)=0 and t01+t02+t12=0 one obtains only D(t01) = -D(t02) = -[f(t02,-t12),t02]. To arrive at the two displayed formulas in (7) one would need the nontrivial identity [f(t01,-t12),t01] = -[f(t02,-t12),t02], and no proof of this identity is given. Thus the asserted push-invariance of f(y,-z) in Corollary 2 is unsupported, and the subsequent Lemma 5 and the construction of the special derivation in Theorem 1 lack their required starting point.","section":"Corollary 2"},{"comment":"The proof of Theorem 1 depends on the push-invariance of f(z,-y) (Lemma 5), which in turn depends on Corollary 2. Since Corollary 2 is not established, the construction of the partner h via Theorem 2.1 of [12] and the verification of the divergence condition are not justified. Thus the main theorem is not proven within this manuscript. The statement may be true, but the presented chain of reasoning contains a load-bearing error that is not a local typo: the false uniqueness assertion is essential to the derivation of D(t01).","section":"Theorem 1"}],"minor_comments":[{"comment":"The displayed Poisson bracket in the proof of Theorem 1 reads '{f, f′}− Df (f′)− Df ′(f ) + [f, f′]', which conflicts with the definition given in §1 and with the line immediately following, where α(D_f(f') - D_f'(f) + [f,f']) is used. This appears to be a typographical error in the minus signs.","section":"Theorem 1 proof, equation (14)"},{"comment":"The commutative diagram is garbled in the text (the arrows appear as '/d32/d32' etc.), making it hard to read. Please redraw it with standard commutative-diagram syntax.","section":"Remark after the main theorem"},{"comment":"The proof that the image of a push-orbit under ǫ breaks into r+1 push-orbits of length r is terse and would benefit from an explicit example, especially because the argument is used in Lemma 5.","section":"Lemma 4"},{"comment":"The sentence 'is an injective Lie algebra morphism.' appears as a fragment. The paragraph describing the composition of maps ds → ds' → Der → Der should be expanded to show explicitly that the composed map sends f to the derivation with z ↦ 0 and y ↦ [y, f(z,-y)], and why this equals D_{g,h}.","section":"End of Theorem 1 proof"}],"recommendation":"reject","confidential_remarks":"The paper addresses an important open question and the intended result is likely true. However, the proof as written contains a clear false assertion in Corollary 1 that is load-bearing: all subsequent push-invariance and divergence arguments depend on the unproved formula for D(t01). This is not a routine gap that can be repaired by a short local patch, since no alternative argument for D(t01) or for the pullback derivation (7) is provided. I recommend rejection, although I would look favorably on a corrected version. The introduction also relies on the uncirculated preprint [6]; while this is disclosed transparently, the editor may wish to verify the status of that reference."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The thing to know: the advertised injection ds -> krv is not established by this manuscript. The strategy is new and credible—pulling back a derivation from mould theory via t01, t02, t12 and using push-invariance to get a special derivation that satisfies divergence. But there is a load-bearing error early on.\n\nThe paper honestly recounts that [12] was incomplete because it rested on an unproved Écalle result, and that Enriquez and Furusho have an uncirculated proof. So the goal is sharp.\n\nCorollary 1 claims that a derivation annihilating [a,b] is uniquely determined by its value on a. That is false. The derivation with D(a)=0 and D(b)=a is a nonzero derivation killing the bracket. Therefore the argument that D and D' agree on a and [a,b] does not force D'=D; the difference could send b to a scalar multiple of a. So D is not shown to commute with ι, and the first line D(t01)=[f(t01,-t12),t01] is not justified. From D(t12)=0 alone one only gets D(t01)=-D(t02); the displayed formula is extra and unsupported.\n\nThis matters because Corollary 2 pulls back that formula to get the special derivation with x->[f(x,-z),x] and y->[f(y,-z),y], which is then used in Lemma 5 and the divergence argument in Theorem 1. Without D(t01), that pullback is unavailable. So the central construction—the h, the special-derivation property, and divergence—does not follow from the written proof.\n\nWhat is good: the paper is clear, the background is honest, the earlier gap is acknowledged, and the references to mould theory are parameter-free and published. The heavy self-citation is not circular here; it points to real previous derivations. The result may well be true, and the new route might be repairable—one could try to prove invariance under ι by a different argument, or derive D(t01) directly from the mould-theoretic existence theorem. But as it stands, Theorem 1 is not supported.\n\nWho this is for: specialists in Grothendieck-Teichmüller and mould theory. A general reader will not get much from it. I would still send it to a serious referee, because the claim is important within the field and the mistake is local; a sound referee report could push the author to fix or clearly justify that line. I would not accept it for publication as is.","headline":"The paper targets a real open problem and gives a plausible new route, but the proof has a concrete false uniqueness step in Corollary 1 that breaks the derivation of the key formula.","tokens_in":7508,"tokens_out":1853,"would_cite":false,"duration_ms":17415,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["17B01","17B40","11M32"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims an injective Lie algebra morphism from the double shuffle Lie algebra into the Kashiwara-Vergne Lie algebra, completing a commutative triangle of injections with the Grothendieck-Teichmüller Lie algebra.","keywords":["double shuffle Lie algebra","Kashiwara-Vergne Lie algebra","Grothendieck-Teichmüller Lie algebra","special derivation","divergence condition","push-invariance","mould theory","free Lie algebra"],"falsifier":"Compute the derivation $D$ given by Theorem 2 for an explicit homogeneous $f\\in ds$ and check whether $D(t_{01})=[f(t_{01},-t_{12}),t_{01}]$, the equality used to define the special derivation; a failure would break the construction. A minimal test of the suspect uniqueness step is already available: $D(a)=0$, $D(b)=a$ is a nonzero derivation of $\\mathrm{Lie}[a,b]$ with $D([a,b])=0$, so the stated justification in Corollary 1 cannot stand without added hypotheses.","tokens_in":6482,"feed_emoji":"🔗","tokens_out":12657,"duration_ms":104419,"temperature":0.7,"pith_summary":"This paper claims that the double shuffle Lie algebra -- the Lie algebra of elements of the free Lie algebra on two variables satisfying the double shuffle relations -- injects into the Kashiwara-Vergne Lie algebra, whose elements are tangential derivations satisfying a special-derivation condition and a divergence equation. The construction sends each double shuffle element $f$ to a derivation $D_{g,h}$ with $g=f(-x-y,-y)$ and an auxiliary Lie polynomial $h$, and the paper proves that this derivation satisfies exactly the two conditions defining Kashiwara-Vergne elements. If the proof is correct, it completes the known commutative triangle of injections linking the Grothendieck-Teichmüller, double shuffle, and Kashiwara-Vergne Lie algebras, a connection of independent interest in the study of associators and multiple zeta values. The argument replaces a previously unproved ingredient with a direct proof using mould-theoretic theorems.","feed_headline":"Double shuffle Lie algebra injects into Kashiwara-Vergne","feed_subtitle":"The injection would complete the known triangle of Lie algebra embeddings in associator theory.","key_machinery":"The carrying device is the mould-theoretic derivation of Theorem 2: a unique derivation $D$ of the free Lie algebra on $a,b$ with $D(a)$ push-invariant, $D([a,b])=0$, $D(t_{02})=[f(t_{02},-t_{12}),t_{02}]$, and $D(a)$ of odd degree. Its action on the elements $t_{01},t_{02},t_{12}$ is pulled back along the isomorphism $x \\mapsto t_{01}$, $y \\mapsto t_{02}$ to produce the special derivation on $x,y$. The push-operator, which cyclically permutes the blocks of $x$'s between successive $y$'s, detects invariance; the partner formula $h=-g'$ converts push-invariance of $g=f(-x-y,-y)$ into the relation $D_{g,h}(x+y)=0$, and push-constancy of $g_y-g_x$ is shown equivalent to the divergence trace equation.","core_discovery":"Theorem 1 is the paper's central claim: for every homogeneous $f\\in ds$ of degree $n\\geq 3$, with $g=f(-x-y,-y)$, there exists a Lie polynomial $h$ such that the tangential derivation given by $x \\mapsto [x,h]$, $y \\mapsto [y,g]$ satisfies $D_{g,h}(x+y)=0$ and the divergence condition $$\\operatorname{tr}(h_{xx}+g_{yy})=c\\operatorname{tr}\\bigl((x+y)^n-x^n-y^n\\bigr),$$ and the map $f \\mapsto D_{g,h}$ is an injective Lie algebra morphism into $krv$. The proof proceeds by attaching to $f$ a derivation of the free Lie algebra on letters $a,b$ with prescribed action on the three elements $t_{01},t_{02},t_{12}$ built from a Bernoulli-series operator, then pulling this action back to $x,y$; push-invariance of $g$ makes the derivation special, and a push-constancy identity for $f_y$ yields the trace condition. The paper also states that the map is compatible with the known injections $grt\\to ds$ and $grt\\to krv$, so the three injections form a commutative triangle.","pith_inferences":["If the gap in Corollary 1 is repaired, the same mould-theoretic derivation may yield an explicit closed form for the partner $h$ rather than an existence proof, making the injection easier to compute.","The paper's remark that the simpler derivation $x \\mapsto [x,-f(x,-z)]$, $y \\mapsto [y,-f(y,-z)]$ probably also satisfies the divergence condition suggests a testable shorter route; proving that would give a more direct injection.","A natural extension is to check whether the same push-invariance machinery transfers to the linearised or elliptic variants of the Kashiwara-Vergne algebra mentioned in the paper, which would broaden the triangle to those settings.","Because the divergence constant is read off from a single coefficient $(f|x^{n-1}y)$, low-degree computer checks on explicit $f\\in ds$ could verify the construction numerically before the proof gap is settled."],"forward_implications":["Every double shuffle element would yield a Kashiwara-Vergne derivation, so the double shuffle relations directly produce solutions of the divergence equation with explicit constant $c=(f|x^{n-1}y)$.","The commutative triangle of injections means the Grothendieck-Teichmüller Lie algebra embeds compatibly into both the double shuffle and Kashiwara-Vergne algebras, so invariants can be transferred along either route.","The construction gives a proof that elements of $ds$ induce special derivations without relying on the previously unproved flexion result, replacing that dependence by mould-theory theorems.","The formula for $h$ as the partner of $g$ makes the morphism algorithmic in principle: once $g$ is known, $h=-g'$ is determined, and the divergence check reduces to computing one coefficient of $f$."],"supporting_citations":[{"why":"supplies the definition of the Kashiwara-Vergne Lie algebra and the known injection from the Grothendieck-Teichmüller Lie algebra used in the commutative triangle.","marker":"[1]"},{"why":"supplies the push-constancy identity for the coefficient polynomial $f_y$ that the divergence proof starts from.","marker":"[3]"},{"why":"supplies the known injection from the Grothendieck-Teichmüller Lie algebra into the double shuffle Lie algebra, the third side of the triangle.","marker":"[7]"},{"why":"contains the mould-theory theorems assembled as Theorem 2, the main machinery that constructs the derivation $D$.","marker":"[11]"},{"why":"provides the partner formula $h=-g'$, the equivalence between push-constancy and the trace condition, and the earlier incomplete attempt at the injection.","marker":"[12]"},{"why":"is the original source for the derivation $D$ and its four defining properties used in Theorem 2.","marker":"[13]"}],"fun_headline_variants":["Double shuffle Lie algebra embeds in Kashiwara-Vergne","Injection from double shuffle to Kashiwara-Vergne proven","Completing the Lie algebra triangle in associator theory","Double shuffle injects into Kashiwara-Vergne: triangle closed","ds embeds in krv: new injection ties grt to both"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the assertion in Corollary 1 that a derivation of the free Lie algebra on $a,b$ that annihilates $[a,b]$ is uniquely determined by its value on $a$; that assertion is false in general, so the proof of Theorem 1 as written depends on an invalid uniqueness step.","fun_headline_variants_meta":{"raw":{"variants":["Double shuffle Lie algebra embeds in Kashiwara-Vergne","Injection from double shuffle to Kashiwara-Vergne proven","Completing the Lie algebra triangle in associator theory","Double shuffle injects into Kashiwara-Vergne: triangle closed","ds embeds in krv: new injection ties grt to both"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000439,"raw_usage":{"total_tokens":2192,"prompt_tokens":869,"completion_tokens":1323,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":485,"completion_tokens_details":{"reasoning_tokens":1232}},"tokens_in":485,"tokens_out":1323,"duration_ms":9072,"temperature":1.0,"reasoning_tokens":1232,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:53:59.969754+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the derivation $D$ given by Theorem 2 for an explicit homogeneous $f\\in ds$ and check whether $D(t_{01})=[f(t_{01},-t_{12}),t_{01}]$, the equality used to define the special derivation; a failure would break the construction. A minimal test of the suspect uniqueness step is already available: $D(a)=0$, $D(b)=a$ is a nonzero derivation of $\\mathrm{Lie}[a,b]$ with $D([a,b])=0$, so the stated justification in Corollary 1 cannot stand without added hypotheses.","supporting_citations":[{"cited_title":"Alekseev, C","cited_arxiv_id":null,"evidence_quote":"supplies the definition of the Kashiwara-Vergne Lie algebra and the known injection from the Grothendieck-Teichmüller Lie algebra used in the commutative triangle."},{"cited_title":"Baumard, Thesis, 2014","cited_arxiv_id":null,"evidence_quote":"supplies the push-constancy identity for the coefficient polynomial $f_y$ that the divergence proof starts from."},{"cited_title":"Enriquez, H","cited_arxiv_id":null,"evidence_quote":"supplies the known injection from the Grothendieck-Teichmüller Lie algebra into the double shuffle Lie algebra, the third side of the triangle."},{"cited_title":"Salerno, L","cited_arxiv_id":null,"evidence_quote":"contains the mould-theory theorems assembled as Theorem 2, the main machinery that constructs the derivation $D$."},{"cited_title":"Schneps, Double shuffle and Kashiwara-Vergne Lie algebras, J","cited_arxiv_id":null,"evidence_quote":"is the original source for the derivation $D$ and its four defining properties used in Theorem 2."}],"review_version":1}