Ecalle’s dimorphic transportation and Brown’s lifting
Hanamichi Kawamura · 2026
Read Ecalle’s dimorphic transportation and Brown’s lifting by Hanamichi Kawamura (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-17424.txt.
Introduction
Hanamichi Kawamura links two deep tools in the theory of multiple zeta values. Brown's lifting method solves double shuffle equations one level up. It builds full solutions from linearized ones, modulo products. Kawamura shows this method fits in Ecalle's dimorphic transportation frame. That frame moves structures between two unlike worlds while keeping their core.
A shared frame can make hard proofs shorter. It can also move results from one field to the other. Multiple zeta values sit at a crossroads of algebra and number theory. Tools that unify their study draw wide interest. Kawamura's link gives researchers a new path through old equations. It may spark fresh attacks on open cases.
Catalog record
| CREATOR | Hanamichi Kawamura |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-17424
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Ecalle’s dimorphic transportation and Brown’s lifting Hanamichi Kawamura Abstract . Brown’s lifting procedure shows that one can solve the double shuffle equations modulo products from solutions of linearized ones. In this paper, we reveal that it is described in the framework of Ecalle’s dimorphic transportation. 2020 Mathematics Subject Classification 11M32, 16B05. 1. Introduction We have been interested in the double shuffle structure in the framework \mathsf{DMR}_{0} as a group and \mathfrak{dmr}_{0} as a Lie algebra. They are originally formulated in a non-commutative power series (or polynomial) ring, but there is another setting which is constructed in the so-called flexion theory established by Ecalle. In particular, it is important to investigate a concrete solution of the double shuffle equation modulo products ( = an element of \mathfrak{dmr}_{0} ). Brown [ B ] got some systems of such solutions, and Matthes–Tasaka [ MT ] started to compare Brown’s and Ecalle’s solutions. Furthermore, recently Furusho–Hirose–Komiyama [ FHK ] revealed that one of Brown’s system \{\psi_{2n+1}\}_{n} can be exactly described as Ecalle’s polar singulates . On the other hand, Brown [ B ] and Ecalle [ E1 ] also independently obtained a method to get solutions \chi_{B}(f),\chi_{E}(f)\in\mathfrak{ds}_{\mathcal{Q}} of the double shuffle equations admitting some poles by lifting each solution f\in\mathfrak{ls}_{\mathcal{Q}} of the “linearlized” double shuffle equations. In this note, we reveal that Brown’s lifting \mathfrak{ls}_{\mathcal{Q}}\ni f\mapsto\chi_{B}(f)\in\mathfrak{ds}_{\mathcal{Q}} essentially coincides with the operator \mathsf{adari}(\mathsf{par}) , which is an example of Ecalle’s lifting. Theorem 1.1 . We have \mathsf{swap}\circ\mathsf{anti}\circ\chi_{B}=\mathsf{adari}(\mathsf{par})\circ\mathsf{swap}\circ\mathsf{anti}. While the original preprint [ B ] does not include the proof that \chi_{B}(f) gives a solution in \mathfrak{ds}_{\mathcal{Q}} for each f\in\mathfrak{ls}_{\mathcal{Q}} , Kimura–Tasaka’s report [ KT ] present that assertion as a open conjecture. We can give an affirmative answer to it by Theorem 1.1 and the machinery of Ecalle’s dimorphic transportation. Corollary 1.2 . For any f\in\mathfrak{ls}_{\mathcal{Q}} , we have \chi_{B}(f)\in\mathfrak{ds}_{\mathcal{Q}} . 2. Matthes–Tasaka’s setting and flexions 2.1. Some sets of functions We basically adopt conventions of Matthes–Tasaka [ MT ] and the author’s series [ Ka1 ] , [ Ka2 ] . The deck \mathcal{F} should be chosen to include the mutually-conjugate pair of flexion units \mathsf{Pa}\binom{u_{1}}{v_{1}}\coloneqq\frac{1}{u_{1}},\quad\text{and}\quad\mathsf{Pi}\binom{u_{1}}{v_{1}}\coloneqq\frac{1}{v_{1}}. The ground field \mathbb{K} is taken as \mathbb{Q} for comparing with Brown’s work, but we note here that all flexion-theoretic arguments done in this note work in any ground field \mathbb{K} with \mathrm{ch}(\mathbb{K})=0 . As we did in [ Ka2 ] , the information about \mathcal{F} and R is often omitted from several notations unless it matters. We denote by \mathsf{BIMU}^{\mathbf{v}\text{-}\mathsf{cst}} (resp. \mathsf{BIMU}^{\mathbf{u}\text{-}\mathsf{cst}} ) the subset of \mathsf{BIMU}(R) consisting of all bimoulds which are indendent of variables in the lower layer (resp. the upper layer). Definition 2.1 . (1) Let \mathcal{Q} be the direct sum \bigoplus_{k\in\mathbb{Z}}\mathcal{Q}_{k} of the space \mathcal{Q}_{k}\coloneqq\left\{\left.f=(f^{(r)})_{r}\in\prod_{r\geq 0}\mathbb{Q}(x_{1},\ldots,x_{r})~\right|~\deg(f^{(r)})=k-r~(r\geq 0)\right\}. Note that \mathbb{Q}(x_{1},\ldots,x_{0})\coloneqq\mathbb{Q} and the element 0 has any degree. (2) The subset \mathcal{G} (resp. \mathcal{L} ) of \mathcal{Q} is defined to be the set of all elements having 1 (resp. 0 ) as its r=0 component. (3) We prepare a copy \overline{\mathcal{Q}} (resp. \underline{\mathcal{Q}} ) of \mathcal{Q} and regard it as a subset of \mathsf{BIMU}^{\mathbf{v}\text{-}\mathsf{cst}}(\mathbb{Q}) (resp. \mathsf{BIMU}^{\mathbf{u}\text{-}\mathsf{cst}}(\mathbb{Q}) ) by the embedding determined by x_{i}\mapsto u_{i} (resp. x_{i}\mapsto v_{i} ). (4) Similarly we denote by \overline{\mathcal{G}} , \underline{\mathcal{G}} , \overline{\mathcal{L}} and \underline{\mathcal{L}} suitable copies of \mathcal{G} and \mathcal{L} being compatible with the definition of \overline{\mathcal{Q}} and \overline{\mathcal{L}} . By the identification above, we say that an element f=(f^{(r)})_{r}\in\overline{\mathcal{Q}} (or \underline{\mathcal{Q}} ) has a component f^{(r)} of length r , which can have terms of several degrees. Note that \overline{\mathcal{Q}} and \underline{\mathcal{Q}} are not completed with respect to this length filtration, but it does not matter since we always discuss after regarding them as subsets of \mathsf{BIMU} . By definition, we have \displaystyle\overline{\mathcal{G}}=\mathsf{MU}\cap\overline{\mathcal{Q}},\quad\overline{\mathcal{L}}=\mathsf{LU}\cap\overline{\mathcal{Q}}, (2.1) \displaystyle\underline{\mathcal{G}}=\mathsf{MU}\cap\underline{\mathcal{Q}},\quad\underline{\mathcal{L}}=\mathsf{LU}\cap\underline{\mathcal{Q}}. (2.2) The two types of musical accidentals \displaystyle f^{\sharp}(x_{1},\ldots,x_{r}) \displaystyle\coloneqq f(x_{1},x_{1}+x_{2},\ldots,x_{1}+\cdots+x_{r}), (2.3) \displaystyle f^{\flat}(x_{1},\ldots,x_{r}) \displaystyle\coloneqq f(x_{1},x_{2}-x_{1},\ldots,x_{r}-x_{r-1}) (2.4) in [ MT , §2.4] , are understood as maps \sharp\colon\underline{\mathcal{Q}}\to\overline{\mathcal{Q}} and \flat\colon\overline{\mathcal{Q}}\to\underline{\mathcal{Q}} , respectively. Definition 2.2 (Ihara bracket; [ MT , Definition 4.1] ) . Let f and g be elements of \underline{\mathcal{L}} . Then we define the Ihara bracket as the commutator \{f,g\}\coloneqq f\mathbin{\underline{\circ}}g-g\mathbin{\underline{\circ}}f of the Ihara action \mathbin{\underline{\circ}} given by the bilinearity and (f^{(r)}\mathbin{\underline{\circ}}g^{(s)})(x_{1},\ldots,x_{r+s})\\ \coloneqq\sum_{i=0}^{s}f^{(r)}(x_{i+1}-x_{i},\ldots,x_{i+r}-x_{i})g^{(s)}(x_{1},\ldots,x_{i},x_{i+r+1},\ldots,x_{r+s})\\ +(-1)^{r}\sum_{i=1}^{s}f^{(r)}(x_{i+r}-x_{i+r-1},\ldots,x_{i+r}-x_{i})g^{(s)}(x_{1},\ldots,x_{i-1},x_{i+r},\ldots,x_{r+s}). (2.5) The following assertion is obvious by definition. Lemma 2.3 . For any g\in\underline{\mathcal{L}} , its Ihara action is given by f\mapsto f\mathbin{\underline{\circ}}g=\mathsf{axit}((\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(f),f)(g)+\mathsf{mu}(f,g). Here we also introduce the framework of double shuffle equations used in [ B ] and [ MT ] . The symbol \mathbin{\mathcyr{sh}} denotes the usual shuffle product for variable sequences. We usually use the convention that an element of \mathcal{Q} behaves linearly for a linear combination of variable sequences. For example, we write f((x_{1},x_{2})\mathbin{\mathcyr{sh}}(x_{3}))=f(x_{1},x_{2},x_{3})+f(x_{1},x_{3},x_{2})+f(x_{3},x_{1},x_{2}). The harmonic product \ast is defined to be a similar \mathbb{Q}(x_{1},x_{2},\ldots) -bilinear product which determined by the rules \varnothing\mathbin{\mathcyr{sh}}\mathbf{x}=\mathbf{x}\mathbin{\mathcyr{sh}}\varnothing=\mathbf{x} and ((x_{i})\uplus\mathbf{x})\ast((x_{j})\uplus\mathbf{x}^{\prime})=(x_{i})\uplus(\mathbf{x}\ast((x_{j})\uplus\mathbf{x}^{\prime}))+(x_{j})\uplus(((x_{i})\uplus\mathbf{x})\ast\mathbf{x}^{\prime})+\frac{1}{x_{i}-x_{j}}(\mathbf{x}\ast\mathbf{x}^{\prime}), where \uplus denotes the concatenation of sequences. Definition 2.4 . We define the subspace \mathfrak{ls}_{\mathcal{Q}}\subseteq\underline{\mathcal{L}} to be the set of every function f\in\underline{\mathcal{L}} satisfying 1 1 1 As pointed out in [ KT ] , in the definition of the space \mathfrak{pdmr} in [ B ] , which should coincide with our \mathfrak{ds}_{\mathcal{Q}} , the parity condition (1) is missing. (1) f^{(1)}(x_{1})=f^{(1)}(-x_{1}) , (2) f^{(r+s)}((x_{1},\ldots,x_{r})\mathbin{\mathcyr{sh}}(x_{r+1},\ldots,x_{r+s}))=0 for r,s\geq 1 , (3) (f^{(r+s)})^{\sharp}((x_{1},\ldots,x_{r})\mathbin{\mathcyr{sh}}(x_{r+1},\ldots,x_{r+s}))=0 for r,s\geq 1 . Similarly, we denote by \mathfrak{ds}_{\mathcal{Q}}\subseteq\underline{\mathcal{L}} the subspace consisting of all function f satisfying (1), (3) and f^{(r+s)}((x_{1},\ldots,x_{r})\ast(x_{r+1},\ldots,x_{r+s}))=0 (2.6) for each r,s\geq 1 . Remark 2.5 . By definition, \mathfrak{ls}_{\mathcal{Q}} is equal to the intersection of \underline{\mathcal{L}} with \mathsf{ARI}_{\underline{\mathsf{al}}/\underline{\mathsf{al}}} . Moreover, the space \mathfrak{ds}_{\mathcal{Q}} is \underline{\mathcal{L}}\cap\mathsf{swap}(\mathsf{ARI}_{\underline{\mathsf{al}}/\underline{\mathsf{il}}}) , where \mathsf{il} denotes the \mathsf{Pi} -alternality ( = alternility). The latter coincidence is checked from Komiyama’s result [ Ko ] that the condition ( 2.6 ) is essentially equivalent to the definition of alternility using \mathsf{ganit}(\mathsf{pic}) . From the above remark, we can show Corollary 1.2 assuming Theorem 1.1 as follows. Since the operator \mathsf{anti} preserves alternality and alternility, it is sufficient to show that (\mathsf{swap}\circ\mathsf{anti}\circ\chi_{B})(f)\in\mathsf{ARI}_{\underline{\mathsf{al}}/\underline{\mathsf{il}}} for any f\in\mathfrak{ls}_{\mathcal{Q}} . Since f is an element of \mathsf{ARI}_{\underline{\mathsf{al}}/\underline{\mathsf{al}}} , so is (\mathsf{swap}\circ\mathsf{anti})(f) . Thus, from Theorem 1.1 , we obtain (\mathsf{swap}\circ\mathsf{anti}\circ\chi_{B})(f)\in\mathsf{adari}(\mathsf{par})(\mathsf{ARI}_{\underline{\mathsf{al}}/\underline{\mathsf{al}}}). The right-hand side is equal to \mathsf{ARI}_{\underline{\mathsf{al}}/\underline{\mathsf{il}}} , as shown in [ Ka2 , Theorem 3.5] . 2.2. Flexions By definition and the identification we gave in the previous subsection, we see that \sharp=(\mathsf{swap}\circ\mathsf{anti})|_{\underline{\mathcal{Q}}} and \flat=(\mathsf{anti}\circ\mathsf{swap})|_{\overline{\mathcal{Q}}} hold. Moreover, we prepare the definition of the \mathsf{ila} -bracket \displaystyle\mathsf{preila}(A,B) \displaystyle\coloneqq\mathsf{ilat}(B)(A)+\mathsf{mu}(A,B)\coloneqq\mathsf{axit}(B,(\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(B))(A), (2.7) \displaystyle\mathsf{ila}(A,B) \displaystyle\coloneqq\mathsf{preila}(A,B)-\mathsf{preila}(B,A), (2.8) for A,B\in\mathsf{LU} . Lemma 2.6 . Let f,g\in\underline{\mathcal{L}} . Then we have \{f,g\}=\mathsf{anti}(\mathsf{ila}(\mathsf{anti}(g),\mathsf{anti}(f))). Proof. Since \mathsf{anti} is an anti-homomorphism with respect to the \mathsf{mu} -product [ Ka1 , Lemma 2.8] , we have \displaystyle\mathsf{anti}(\mathsf{ila}(\mathsf{anti}(g),\mathsf{anti}(f))) (2.9) \displaystyle=\begin{multlined}(\mathsf{anti}\circ\mathsf{ilat}(\mathsf{anti}(f))\circ\mathsf{anti})(g)-(\mathsf{anti}\circ\mathsf{ilat}(\mathsf{anti}(g))\circ\mathsf{anti})(f)\\ +\mathsf{anti}(\mathsf{mu}(\mathsf{anti}(g),\mathsf{anti}(f)))-\mathsf{anti}(\mathsf{mu}(\mathsf{anti}(f),\mathsf{anti}(g)))\end{multlined} \displaystyle=(\mathsf{anti}\circ\mathsf{ilat}(\mathsf{anti}(f))\circ\mathsf{anti})(g)-(\mathsf{anti}\circ\mathsf{ilat}(\mathsf{anti}(g))\circ\mathsf{anti})(f)+\mathsf{mu}(f,g)-\mathsf{mu}(g,f), (2.12) while the Ihara bracket is computed by Lemma 2.3 as \displaystyle\{f,g\} \displaystyle=f\mathbin{\underline{\circ}}g-g\mathbin{\underline{\circ}}f (2.13) \displaystyle=\mathsf{axit}((\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(f),f)(g)-\mathsf{axit}((\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(g),g)(f)+\mathsf{mu}(f,g)-\mathsf{mu}(g,f). (2.14) Hence it is sufficient to show the equality (\mathsf{anti}\circ\mathsf{ilat}(\mathsf{anti}(f))\circ\mathsf{anti})(g)=\mathsf{axit}((\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(f),f)(g) (2.15) because this identity itself and the one after interchanging f\leftrightarrow g yield the desired equality. Since \mathsf{anti}\circ\mathsf{amit}(\mathsf{anti}(A))\circ\mathsf{anti}=\mathsf{anit}(A) holds for any A\in\mathsf{LU} , we have \displaystyle(\mathsf{anti}\circ\mathsf{ilat}(\mathsf{anti}(f))\circ\mathsf{anti})(g) (2.16) \displaystyle=(\mathsf{anti}\circ\mathsf{amit}(\mathsf{anti}(f))\circ\mathsf{anti})(g)+(\mathsf{anti}\circ\mathsf{anit}((\mathsf{neg}\circ\mathsf{pari})(f))\circ\mathsf{anti})(g) (2.17) \displaystyle=\mathsf{anit}(f)(g)+\mathsf{amit}((\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(f))(g) (2.18) \displaystyle=\mathsf{axit}((\mathsf{neg}\circ\mathsf{pari}\circ\mathsf{anti})(f),f)(g), (2.19) and it completes the proof. ∎ Let us here review the auxilirary operation \mathbin{\underline{\circ}_{\mathsf{ari}}} and the bracket \{-,-\}_{\mathsf{ari}} treated in [ MT ] , by describing how they related to the original flexion structure instead of directly introducing their definition. As refered in [ MT , §3.2] ,