Abstract
Richard Hain and Makoto Matsumoto constructed a category of universal mixed elliptic motives, and described the fundamental Lie algebra of this category: it is a semi-direct product of the fundamental Lie algebra \({{\mathrm{Lie}}}\pi _1(\mathsf {MTM})\) of the category of mixed Tate motives over \(\mathbb {Z}\) with a filtered and graded Lie algebra \(\mathfrak {u}\). This Lie algebra, and in particular \(\mathfrak {u}\), admits a representation as derivations of the free Lie algebra on two generators. In this paper we study the image \(\mathscr {E}\) of this representation of \(\mathfrak {u}\), starting from some results by Aaron Pollack, who determined all the relations in a certain filtered quotient of \(\mathscr {E}\), and gave several examples of relations in low weights in \(\mathscr {E}\) that are connected to period polynomials of cusp forms on \({{\mathrm{SL}}}_2(\mathbb {Z})\). Pollack’s examples lead to a conjecture on the existence of such relations in all depths and all weights, that we state in this article and prove in depth 3 in all weights. The proof follows quite naturally from Ecalle’s theory of moulds, to which we give a brief introduction. We prove two useful general theorems on moulds in the appendices.
Résumé
Richard Hain et Makoto Matsumoto ont construit une catégorie de motifs elliptiques mixtes universels et décrit l’algèbre de Lie fondamentale de cette catégorie: il s’agit d’un produit semi-direct de l’algèbre de Lie fondamentale \(\mathrm{Lie}\,\pi _1\)(MTM) de la catégorie des motifs de Tate mixtes sur \(\mathbb {Z}\) par une algèbre de Lie graduée et filtrée \(\mathfrak {u}\). Cette algèbre de Lie, et en particulier l’algèbre \(\mathfrak {u}\), admet une représentation par dérivations sur l’algèbre de Lie libre à deux générateurs. Dans cet article nous étudions l’image \(\mathscr {E}\) de cette représentation de \(\mathfrak {u}\) à partir de résultats d’Aaron Pollack, qui a déterminé toutes les relations dans un certain quotient filtré de \(\mathscr {E}\), et donné plusieurs exemples de relations en petit poids dans \(\mathscr {E}\) qui sont liées aux polynômes de périodes de formes paraboliques sur \({{\mathrm{SL}}}_2(\mathbb {Z})\). Les exemples de Pollack conduisent à une conjecture sur l’existence de telles relations en toute profondeur et en tout poids, que nous énonçons dans cet article et prouvons en profondeur 3 en tout poids. La démonstration découle naturellement de la théorie des moules d’Écalle, à laquelle nous donnons une brève introduction. Nous démontrons deux théorèmes utiles sur les moules dans les appendices.
Similar content being viewed by others
Notes
This result follows from combining standard results from the literature. Indeed, Goncharov constructed a Hopf algebra \({\mathcal {MZ}}\) of motivic multiple zeta values, and it follows from work of Brown [2] that \({{\mathrm{Lie}}}\pi _1(\mathsf {MTM})\) is isomorphic to the dual of the quotient of \({\mathcal {MZ}}\) modulo products. Goncharov also showed that elements of \({\mathcal {MZ}}\) satisfy the associator relations, which by duality gives an injective map \({{\mathrm{Lie}}}\pi _1(\mathsf {MTM}) \hookrightarrow \mathfrak {grt}\) into the Grothendieck-Teichmüller Lie algebra. Finally, in [7], Enriquez constructed an explicit injective map from \(\mathfrak {grt}\) into \({\mathrm {Der}}\,{{\mathrm{Lie}}}[a,b]\). The injectivity of the map \({{\mathrm{Lie}}}\pi _1(MTM)\hookrightarrow \mathrm{Der}\,{{\mathrm{Lie}}}[a,b]\) also appears as Theorem 3.1 in the recent preprint [4], where it has a very different proof based on substantial work by Hain and Levin-Racinet.
The same statement appears without proof in the online lecture notes [5] by Brown.
Note added in proof: Hain and Matsumoto have just released a new version of the preprint [11] containing a proof of this conjecture in all depths.
Ecalle uses \({{\mathrm{ARI}}}\) for bimoulds which are functions of two sets of variables \(u_i\) and \(v_i\); in the framework of this article we never use actual bimoulds ( i.e., functions that are non-trivial in both families), but there are situations in which they are very useful.
This result is also given in §4 of the preprint [8] by B. Enriquez from January 2013, with a brief indication in lieu of proof. Brown (cf. [4, 5]) studies a Lie algebra denoted \(\mathfrak {pls}\) (polar linearized double shuffle) which is isomorphic to \({{\mathrm{\mathrm{{swap}}}}}\Bigl ({{\mathrm{ARI}}}^{\text {sing}}_{\text {al}/\text {al}}\Bigr )\). In the recent preprint [4], Brown mentions without proof the analogous result to Theorem 3.3 for \(\mathfrak {pls}\) (see Definition 4.5), i.e., \(\mathfrak {pls}\) is closed under the “linearized Ihara bracket”. He also gives results that are analogous to Theorem 3.5 and Corollary 3.6.
References
Baumard, S.: Aspects modulaires et elliptiques des relations entre multizêtas. Ph.D. thesis, June 2014. http://tel.archives-ouvertes.fr/tel-01017022
Brown, F.: Mixed Tate motives over \({\mathbb{Z}}\). Ann. Math. 175(2), 949–976 (2012)
Brown, F.: Depth-graded motivic multiple zeta values (2013). arXiv:1301.3056. http://www.ihes.fr/~brown
Brown, F.: Zeta elements in depth 3 and the fundamental Lie algebra of a punctured elliptic curve (2015). arXiv:1504.04737. http://www.ihes.fr/~brown
Brown, F.: Anatomy of the motivic Lie algebra, online lecture notes (2013). http://www.ihes.fr/~brown
Ecalle, J.: The flexion structure and dimorphy. In: Asymptotics in Dynamics, Geometry and PDEs. Generalized Borel Summation, Publications of the Scuola Normale Superiore 12(2), 27–211 (2011)
Enriquez, B.: Elliptic associators. Selecta Math. New Ser. 20, 491–584 (2014)
Enriquez, B.: Analogues elliptiques des nombres multizêtas. Bulletin de la SMF (2013, to appear). arXiv:1301.3042
Goncharov, A.B.: Multiple polylogarithms, cyclotomy and modular complexes. Math. Res. Let. 5, 497–516 (1998)
Goncharov, A.B.: The dihedral Lie algebras and Galois symmetries of \(\pi _1^{(\ell )}(\mathbb{P}^1-(\{0,\infty \}\cup {\mu }_N))\). Duke Math. J. 110(3), 397–487 (2001)
Hain, R.: Letter to P. Deligne, Durham, winter (2009)
Hain, R., Matsumoto, M.: Universal mixed elliptic motives (2015). arXiv:1512.03975
Ihara, Y.: Some arithmetic aspects of Galois actions in the pro-\(p\). In: Arithmetic fundamental groups and noncommutative algebra, Proceedings of Symposia in Pure Mathematics. 70 (2002)
Ihara, K., Kaneko, M., Zagier, D.: Derivation and double shuffle relations for multiple zeta values. Compos. Math. 142(2), 307–338 (2006)
Martin, F., Royer, E.: Formes modulaires et périodes, SMF, Paris. Sém. Congrès 12 (2005)
Luque, J.-G., Novelli, J.-C., Thibon, J.-Y.: Period polynomials and Ihara brackets. J. Lie Th. 17(2), 229–239 (2007)
Pollack, A.: Relations between derivations arising from modular forms, senior thesis (2009)
Racinet, G.: Doubles mélanges des polylogarithmes multiples aux racines de l’unité. Publ. Math. de l’IHES. 95 (2002)
Salerno, A., Schneps, L.: A mould-theoretic proof of the double shuffle Lie algebra structure (2015). arXiv:1510.05535
Schneps, L.: On the Poisson bracket on the free Lie algebra in two generators. J. Lie Th. 16(1), 19–37 (2006)
Schneps, L.: Double shuffle and Kashiwara-Vergne Lie algebras. J. Algebra 367, 54–74 (2012)
Schneps, L.: ARI, GARI, Zig and Zag: an introduction to Ecalle’s theory of multiple zeta values (2015). arXiv:1507.01534
Acknowledgments
The first author thanks Clément Dupont for useful discussions. We also thank Richard Hain for pointing out an ambiguity in an early version of this paper, as well as the anonymous referee for pointing out that the original proof was more complicated than necessary.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: Proof of Theorem 3.3
This appendix is devoted to the proof of Theorem 3.3, which makes essential use of the \({{\mathrm{\mathrm{{swap}}}}}\) operator and some of the basic notions of Ecalle’s theory of moulds, in particular the \({{\mathrm{ari}}}\)-bracket.
Definitions
If we break the tuple \(w=(u_1,\ldots ,u_r)\) into three parts
we write
If \(A,B\in {{\mathrm{ARI}}}\), we define the operator \({{\mathrm{arit}}}(B)\) acting on A by
If \(A,B\in \overline{{{\mathrm{ARI}}}}\) and the tuple \((v_1,\ldots ,v_r)\) breaks into pieces
then setting
we let
For two moulds A and B, let \({{\mathrm{mu}}}(A,B)\) be the product defined by
When \(A={{\mathrm{ma}}}(F)\) and \(B={{\mathrm{ma}}}(G)\) for polynomials F, G in a, b, the multiplication \({{\mathrm{mu}}}\) coincides with ordinary multiplication of polynomials (cf. [22, (3.2.13)]):
Let
so \({{\mathrm{lu}}}\) is the corresponding Lie bracket. For \(A,B\in {{\mathrm{ARI}}}\) or \(\overline{{{\mathrm{ARI}}}}\), we set
Recall that is the mould in the \(u_i\)’s defined by
Let ; explicitly,
For any mould \(A\in \overline{{{\mathrm{ARI}}}}\), let \(\check{A}\) be the mould defined by
Let \(\overline{{{\mathrm{ARI}}}}^{\text {sing}}_{\text {al}}\) denote the space of alternal rational-valued moulds A in the variables \(v_i\) such that \(\check{A}\) is polynomial-valued.
The heart of the proof of Theorem 3.3 [1, lemme 4.40] consists in the following proposition.
Proposition 5.1
The space \(\overline{{{\mathrm{ARI}}}}^{\text {sing}}_{\text {al}}\) is a Lie algebra under the \({{\mathrm{ari}}}\)-bracket.
We first need a lemma.
Lemma 5.2
Let M be an element of \(\overline{{{\mathrm{ARI}}}}^{\text {sing}}_{\text {al}}\). Then the mould \(\check{M}\) satisfies the relation
Proof
Let us write the simplest of the alternality relations, which gives
Hence, multiplying by \(v_1\) and evaluating at \(v_1=0\), we get
which is well-defined since both numerators at play are polynomials by the hypothesis on M. The result follows immediately.
Proof of Proposition 5.1
Let \(M,N\in \overline{{{\mathrm{ARI}}}}^{\text {sing}}_{\text {al}}\). By linearity, we may assume that M and N are respectively concentrated in depths \(r\geqslant 1\) and \(s\geqslant 1\). Now, alternality is preserved by the \({{\mathrm{ari}}}\)-bracket (cf. [19, Theorem 3.1] for a complete proof), so it only remains to show that \({{\mathrm{ari}}}(M,N)\) is an element of \(\overline{{{\mathrm{ARI}}}}^{\text {sing}}\).
The essential step of the proof consists in determining the poles of \({{\mathrm{arit}}}(M)\cdot N\) by reducing their values modulo the subspace of polynomials in the \(v_i\)’s. We will use the simplifying notation . The fact that M is concentrated in depth r and N in depth s also simplifies the defining formula for \({{\mathrm{arit}}}(M)\cdot N\), as follows:
Consequently,
with
\(\square \)
or, equivalently,
In light of equalities (5.2) and (5.3), there are three types of poles which can occur in the rational function :
-
(1)
the poles \(\frac{1}{v_{r+1}}\) and \(\frac{1}{v_{s}}\), which come from only one term and thus do not immediately cancel out;
-
(2)
the poles of the form \(\tfrac{1}{v_i-v_{i+r+1}}\);
-
(3)
the poles of the form \(\tfrac{1}{v_i-v_{i+r}}\cdot \)
Let us deal first with Case (2). The corresponding pole \(\tfrac{1}{v_i-v_{i+r+1}}\) can only appear in the term ; we then want to check that this pole is compensated by the corresponding difference of the \(\check{M}\)’s. Let us compute this difference for
it gives
which is indeed zero, hence the compensation.
For Case (3), we need to distinguish three subcases, according to whether \(i=1\), or \(i=s\), or i belongs to \(\mathopen {[\![} 2,s-1 \mathclose {]\!]}\):
(a) \(i=1\): the pole is multiplied by
if \(v_1=v_{r+1}=a\), this gives
which is zero thanks to Lemma 5.2.
(b) \(i=s\): the pole is multiplied by
if \(v_s=v_{r+s}=a\), this gives
which is zero for the same reason as above.
(c) \(i\in \mathopen {[\![} 2,s-1 \mathclose {]\!]}\): the pole \(\tfrac{1}{v_i-v_{i+r}}\) comes from , and is multiplied by
if \(v_i=v_{i+r}=a\), this gives
which is zero again.
In consequence, the only remaining poles in are those in \(\frac{1}{v_{r+1}}\) and \(\frac{1}{v_s}\) from Case (1), and more precisely we can write
It remains to show that the above expression cancels in the bracket \({{\mathrm{ari}}}(M,N)\), using the definition by equation (5.1). The possible poles for the rational function come from the sum
paying attention to the exchange of r and s when we switch M and N. The discussion above shows that it suffices to check that the poles \(\frac{1}{v_{r}}\) and \(\frac{1}{v_s}\), \(\frac{1}{v_{r+1}}\) and \(\frac{1}{v_{s+1}}\) cancel out.
Let us note that the alternality of \({{\mathrm{ari}}}(M,N)\), together with the basic fact that any alternal mould M satisfies
[22, Lemma 2.5.3], imply that it is enough to deal with the case of \(\frac{1}{v_{r}}\) and \(\frac{1}{v_{s}}\), the other two being deduced from them by applying the involution corresponding to the symmetry around \(\tfrac{r+s+1}{2}\). And the pole \(\frac{1}{v_s}\) is deduced from \(\frac{1}{v_{r}}\) since the total expression for \({{\mathrm{ari}}}(M,N)\) is antisymmetric in M and N, and thus in r and s.
The only case left to check is that of the pole \(\frac{1}{v_{r}}\). The corresponding factor is
which is clearly zero when \(v_r=0\). This concludes the proof of Proposition 5.1.
We can now complete the proof of Theorem 3.3. We use the following elementary result of mould theory [22, Lemma 2.4.1]:
Let \(A,B\in {{\mathrm{ARI}}}^{\text {sing}}_{\underline{\text {al}}/\underline{\text {al}}}\). Then
by the part (v) of Theorem 3.1. Consider the mould \({{\mathrm{\mathrm{{swap}}}}}(C)\in \overline{{{\mathrm{ARI}}}}\). Again by part (v) of Theorem 3.1, \({{\mathrm{ARI}}}_{\underline{\text {al}}/\underline{\text {al}}}\) is contained in \({{\mathrm{ARI}}}_{{{\mathrm{\mathrm{{push}}}}}}\), so the moulds A, B and C are all \({{\mathrm{\mathrm{{push}}}}}\)-invariant. Thus (5.4) holds, i.e.,
Both \({{\mathrm{\mathrm{{swap}}}}}(A)\) and \({{\mathrm{\mathrm{{swap}}}}}(B)\) lie in \(\overline{{{\mathrm{ARI}}}}^{\text {sing}}_{\text {al}}\) by definition. Thus by Proposition 5.1, we also have \({{\mathrm{\mathrm{{swap}}}}}(C)\in \overline{{{\mathrm{ARI}}}}^{\text {sing}}_{\text {al}}\) and thus \(C\in {{\mathrm{ARI}}}^{\text {sing}}\). But we also have
so \(C\in {{\mathrm{ARI}}}^{\text {sing}}_{\underline{\text {al}}/\underline{\text {al}}}\), which concludes the proof of Theorem 3.3. \(\square \)
Appendix 2: Proof of Theorem 3.5
This section is devoted to the proof of Theorem 3.5, which is stated as Proposition 6.2 below.
We use the notation and terminology of Sect. 3, with one further definition: for any polynomial \(F\in {{\mathrm{Lie}}}[a,b]\) of homogeneous depth r, we define \({{\mathrm{da}}}(F)\) to be the mould given by
As we did for \({{\mathrm{Da}}}(F)\), if \(F=\sum _r F^r\) is any Lie polynomial broken up into its depth-graded parts, we define
Let \({{\mathrm{lu}}}(A,B)\) be as in the beginning of Appendix 1. For any polynomial U, let \({{\mathrm{Darit}}}_U\) be the operator on moulds defined by
Proposition 6.1
Let U be a push-invariant polynomial in \({{\mathrm{Lie}}}[a,b]\), and let \(D_U\) be the associated derivation. Then for any Lie polynomial F in a and b, we have
Proof
By additivity, we may assume that U is of homogeneous depth r and F is of homogeneous depth s. We will use induction on the depth of F to prove the proposition; the only difficult part is the base case \(s=1\).
So assume that F is of depth 1, i.e., \(F=C_n={{\mathrm{\text {ad}}}}(a)^{n-1}(b)\). Here, since \(D_U([a,b])\) is zero, we have the explicit formula
Applying \({{\mathrm{ma}}}\) to both sides of this formula, and using the facts that \({{\mathrm{ma}}}(C_{n-i-1})\) is equal to \((-1)^{n-i-2}u_1^{n-i-2}\) and that for a Lie polynomial P of homogeneous depth r we have
by the part (3) of Theorem 3.1, and using the identity
we find that
so, with a little rewriting of indices, we find that the left-hand side of (6.1) equals
or, adding up the sums to obtain a closed expression,
Let us now compute the right-hand side of (6.1). We have \({{\mathrm{ma}}}(F)=(-1)^nu_1^{n-1}\), so \({{\mathrm{da}}}(F)=(-1)^nu_1^{n-2}\); thus
Also, we have
since \({{\mathrm{da}}}(F)\) is a depth 1 mould and therefore the only possible decomposition \(w=abc\) with \(c\ne \varnothing \) is \(a=\varnothing \), \(b=(u_1,\ldots ,u_r)\) and \(c=u_{r+1}\), and the only possible decomposition with \(a\ne \varnothing \) is \(a=u_1\), \(b=(u_2,\ldots ,u_r)\) and \(c=\varnothing \). We add (6.3) and (6.4) to get the right-hand side of (6.1), obtaining
which is equal to (6.2). This settles the base case where F is of depth 1.
Now assume that (6.1) holds up to depth \(s-1\), and let F be a Lie polynomial of depth s. Then F is a linear combination of Lie brackets, so by additivity, we may assume that F is a single Lie bracket; thus \(F=[G,H]\) for Lie brackets G, H that are of depth \(<s\). We saw in Appendix 1 that
so by definition of \({{\mathrm{da}}}\) we also have
Furthermore, \({{\mathrm{Darit}}}_U\) is a derivation for the \({{\mathrm{lu}}}\)-bracket since both \({{\mathrm{arit}}}(B)\) and are — this is obvious for \({{\mathrm{lu}}}\) but difficult for \({{\mathrm{arit}}}\), cf. [22, Prop. 2.2.1] —, so using the induction hypothesis for G and H, we have
This proves the proposition. \(\square \)
Proposition 6.2
The map
given by \(D_U\mapsto {{\mathrm{Da}}}(U)\) is a Lie algebra morphism, i.e.,
Proof
It is easily seen that
Using only this identity, the Jacobi relation and (6.1), we compute
where W is the polynomial such that
(obtained by multiplying the right-hand mould by and inverting the injective map \({{\mathrm{ma}}}\)). However, the above calculation shows that
where \(P=D_U(V)-D_V(U)\), that is \(D_P=[D_U,D_V]\). Thus the two derivations \({{\mathrm{Darit}}}_W\) and \({{\mathrm{Darit}}}_P\) take the same values on all elements F of \({{\mathrm{Lie}}}[a,b]\), so \(W=P\), i.e.,
so under the map \(D_U\mapsto {{\mathrm{Da}}}(U)\) of (6.6), we have
This proves Proposition 6.2. \(\square \)
Rights and permissions
About this article
Cite this article
Baumard, S., Schneps, L. On the derivation representation of the fundamental Lie algebra of mixed elliptic motives. Ann. Math. Québec 41, 43–62 (2017). https://doi.org/10.1007/s40316-015-0040-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40316-015-0040-8