\documentclass[12pt,final,notitlepage,onecolumn]{article}%
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{hyperref}
\usepackage{graphicx}%
\setcounter{MaxMatrixCols}{30}
%TCIDATA{OutputFilter=latex2.dll}
%TCIDATA{Version=5.50.0.2960}
%TCIDATA{CSTFile=LaTeX article (bright).cst}
%TCIDATA{Created=Sat Mar 27 17:33:36 2004}
%TCIDATA{LastRevised=Wednesday, June 03, 2026 14:37:34}
%TCIDATA{}
%TCIDATA{}
%TCIDATA{}
%TCIDATA{BibliographyScheme=Manual}
%TCIDATA{}
%BeginMSIPreambleData
\providecommand{\U}[1]{\protect\rule{.1in}{.1in}}
%EndMSIPreambleData
\voffset=-2.5cm
\hoffset=-2.5cm
\setlength\textheight{24cm}
\setlength\textwidth{15.5cm}
\begin{document}
\begin{center}
\textbf{Rep\#1: Deformations of a bimodule algebra}
Darij Grinberg
[corrected 2026-06-03, not proofread]
\bigskip
\end{center}
The purpose of this short note is to generalize Problem 2.24 in \cite{1}. We
begin with a few definitions:
\begin{quote}
\textbf{Definition 1.} In the following, a \textit{ring} will always mean a
(not necessarily commutative) ring with unity. Ring homomorphisms are always
assumed to respect the unity. For every ring $R$, we denote the unity of $R$
by $1_{R}$. All $K$-algebras (for a commutative ring $K$) are understood to be
$K$-modules and simultaneously rings with a $K$-bilinear multiplication.
\textbf{Definition 2.} Let $A$ and $B$ be two rings, and let $\rho
:A\rightarrow B$ be a ring homomorphism. Then, for any $a\in A$ and any $b\in
B$, we shall denote the product $\rho\left( a\right) b$ by $ab$ and denote
the product $b\rho\left( a\right) $ by $ba$. Thus, $B$ becomes a left
$A$-module and a right $A$-module, and -- even better -- an $\left(
A,A\right) $-bimodule (that is, the two actions of $A$ commute: $\left(
ab\right) a^{\prime}=a\left( ba^{\prime}\right) $ for all $a\in A$, $b\in
B$ and $a^{\prime}\in A$).
\textbf{Definition 3.} Let $B$ be a ring. Then, we denote by $B\left[ \left[
t\right] \right] $ the ring of formal power series over $B$ in the
indeterminate $t$, where $t$ is supposed to commute with every element of $B$.
Formally, this means that we define $B\left[ \left[ t\right] \right] $ as
the ring of all sequences $\left( b_{0},b_{1},b_{2},...\right) \in
B^{\mathbb{N}}$ (where $\mathbb{N}$ means the set $\left\{ 0,1,2,...\right\}
$), with addition defined by%
\[
\left( b_{0},b_{1},b_{2},...\right) +\left( b_{0}^{\prime},b_{1}^{\prime
},b_{2}^{\prime},...\right) =\left( b_{0}+b_{0}^{\prime},b_{1}+b_{1}%
^{\prime},b_{2}+b_{2}^{\prime},...\right)
\]
and multiplication defined by%
\[
\left( b_{0},b_{1},b_{2},...\right) \cdot\left( b_{0}^{\prime}%
,b_{1}^{\prime},b_{2}^{\prime},...\right) =\left( \sum_{\substack{\left(
i,j\right) \in\mathbb{N}^{2};\\i+j=0}}b_{i}b_{j}^{\prime},\sum
_{\substack{\left( i,j\right) \in\mathbb{N}^{2};\\i+j=1}}b_{i}b_{j}^{\prime
},\sum_{\substack{\left( i,j\right) \in\mathbb{N}^{2};\\i+j=2}}b_{i}%
b_{j}^{\prime},...\right) ,
\]
and we denote a sequence $\left( b_{0},b_{1},b_{2},...\right) \in
B^{\mathbb{N}}$ by $\sum\limits_{i=0}^{\infty}b_{i}t^{i}$. For every
$m\in\mathbb{N}$, the element $b_{m}\in B$ is called the $t^{m}$%
\textit{-coefficient of the power series }$\left( b_{0},b_{1},b_{2}%
,...\right) =\sum\limits_{i=0}^{\infty}b_{i}t^{i}$ (or the
\textit{coefficient of the power series }$\left( b_{0},b_{1},b_{2}%
,...\right) =\sum\limits_{i=0}^{\infty}b_{i}t^{i}$ \textit{before }$t^{m}$).
The element $b_{0}\in B$ is also called the \textit{constant term} of the
power series $\left( b_{0},b_{1},b_{2},...\right) =\sum\limits_{i=0}%
^{\infty}b_{i}t^{i}$. We embed the ring $B$ into $B\left[ \left[ t\right]
\right] $ by identifying each $b\in B$ with the power series $\left(
b,0,0,0,\ldots\right) \in B\left[ \left[ t\right] \right] $. (The latter
power series is known as a \textit{constant power series}, even though it is
not constant as a sequence!)
\textbf{Definition 4.} Let $B$ be a ring, and let $\left( b_{m}%
,b_{m+1},...,b_{n}\right) $ be a sequence of elements of $B$. Then, we denote
by $\overset{\longleftarrow}{\prod\limits_{i=m}^{n}}b_{i}$ the product
$b_{n}b_{n-1}...b_{m}$ (this product is supposed to mean $1$ if $m>n$).
\end{quote}
Now comes the generalization of Problem 2.24 (a) in \cite{1}\footnote{Problem
2.24 in \cite{1} is recovered from this generalization by setting
$B=\operatorname*{End}V$.}:
\begin{quote}
\textbf{Theorem 1.} Let $K$ be a commutative ring. Let $A$ and $B$ be two
$K$-algebras, and let $\rho:A\rightarrow B$ be a $K$-algebra homomorphism. As
in Definition 2, we thus conclude that $B$ becomes an $\left( A,A\right) $-bimodule.
Assume that%
\begin{equation}
\left(
\begin{array}
[c]{c}%
\text{for every }K\text{-linear map }f:A\rightarrow B\text{ which satisfies}\\
\left( f\left( aa^{\prime}\right) =af\left( a^{\prime}\right) +f\left(
a\right) a^{\prime}\text{ for all }a\in A\text{ and }a^{\prime}\in A\right)
,\\
\text{there exists an element }s\in B\text{ such that}\\
\left( f\left( a\right) =as-sa\text{ for all }a\in A\right)
\end{array}
\right) \label{1ass}%
\end{equation}
(where we are using this $\left( A,A\right) $-bimodule structure on $B$ to
make sense of $af\left( a^{\prime}\right) $, $f\left( a\right) a^{\prime}%
$, $as$ and $sa$).
Let $B\left[ \left[ t\right] \right] $ be the ring of formal power series
over $B$ in the indeterminate $t$, where $t$ is supposed to commute with every
element of $B$.
Here and in the following, we let $1$ denote the unity $1_{B}$ of the ring $B$.
Let $\overline{\rho}:A\rightarrow B\left[ \left[ t\right] \right] $ be a
$K$-linear map such that any $a\in A$ and any $a^{\prime}\in A$ satisfy
\begin{equation}
\overline{\rho}\left( aa^{\prime}\right) =\overline{\rho}\left( a\right)
\overline{\rho}\left( a^{\prime}\right) , \label{1rhoaa'}%
\end{equation}
and such that for every $a\in A$,
\begin{equation}
\text{the constant term of the power series }\overline{\rho}\left( a\right)
\text{ equals }\rho\left( a\right) .\ \ \ \ \ \ \ \ \ \ \label{1rho0}%
\end{equation}
Then, there exists a power series $b\in B\left[ \left[ t\right] \right] $
with constant term $1$ such that for every $a\in A$, the power series
$b\overline{\rho}\left( a\right) b^{-1}\in B\left[ \left[ t\right]
\right] $ equals the (constant) power series $\rho\left( a\right) $. (Note
that $b^{-1}$ is well-defined, since $b$ has constant term $1$ and thus has an
inverse\footnote{Here we are using the fact that any power series in $B\left[
\left[ t\right] \right] $ with constant term $1$ has an inverse. This is
well-known when $B$ is commutative, but the same proof applies in general (as
long as one remembers that an element of a monoid that has a left inverse and
a right inverse must always have an inverse).}.)
\end{quote}
\textit{Proof of Theorem 1.} First, we endow the ring $B\left[ \left[
t\right] \right] $ with the $\left( t\right) $-adic topology. This
topology is defined as the product topology on $B^{\mathbb{N}}$ (where each
copy of $B$ has the discrete topology), recalling that $B\left[ \left[
t\right] \right] =B^{\mathbb{N}}$ as a set. Equivalently, this topology is
defined in such a way that for every $f\in B\left[ \left[ t\right] \right]
$, the family
\[
\left( f+t^{0}B\left[ \left[ t\right] \right] ,\ f+t^{1}B\left[ \left[
t\right] \right] ,\ f+t^{2}B\left[ \left[ t\right] \right]
,\ ...\right)
\]
is a basis of open neighbourhoods of $f$. This topology makes $B\left[
\left[ t\right] \right] $ a topological ring, since $t^{i}B\left[ \left[
t\right] \right] $ is a two-sided ideal of $B\left[ \left[ t\right]
\right] $ for every $i\in\mathbb{N}$. Note that $t^{i}B\left[ \left[
t\right] \right] $ is the set of all power series whose first $i$
coefficients (starting with the constant term) are $0$.
For any $k$ elements $u_{1}$, $u_{2}$, $...$, $u_{k}$ of $B$, and for every
$a\in A$, let us denote by $\overline{\rho}_{u_{1},u_{2},...,u_{k}}\left(
a\right) $ the element%
\[
\left( \overset{\longleftarrow}{\prod_{i=1}^{k}}\left( 1-u_{i}t^{i}\right)
\right) \cdot\overline{\rho}\left( a\right) \cdot\left(
\overset{\longleftarrow}{\prod_{i=1}^{k}}\left( 1-u_{i}t^{i}\right) \right)
^{-1}\in B\left[ \left[ t\right] \right]
\]
(here, the inverse $\left( \overset{\longleftarrow}{\prod\limits_{i=1}^{k}%
}\left( 1-u_{i}t^{i}\right) \right) ^{-1}$ exists because
$\overset{\longleftarrow}{\prod\limits_{i=1}^{k}}\left( 1-u_{i}t^{i}\right)
$ is a power series with constant term $1$). Clearly, $\overline{\rho}%
_{u_{1},u_{2},...,u_{k}}:A\rightarrow B\left[ \left[ t\right] \right] $ is
a $K$-linear map for any $k$ elements $u_{1}$, $u_{2}$, $...$, $u_{k}$ of $B$.
Now, we are going to recursively construct a sequence $\left( u_{1}%
,u_{2},u_{3},...\right) \in B^{\left\{ 1,2,3,...\right\} }$ of elements of
$B$ such that every $n\in\mathbb{N}$ satisfies%
\begin{equation}
\left( \overline{\rho}_{u_{1},u_{2},...,u_{n}}\left( a\right) \equiv
\rho\left( a\right) \ \operatorname{mod}\ t^{n+1}B\left[ \left[ t\right]
\right] \ \ \ \ \ \ \ \ \ \ \text{for every }\left. a\in A\right. \right)
. \label{1proof}%
\end{equation}
In fact, we first notice that the congruence (\ref{1proof}) is automatically
satisfied for $n=0$ (note that the product $\overset{\longleftarrow
}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right) $ is an empty product
when $n=0$), because in the case $n=0$, we have $\overline{\rho}_{u_{1}%
,u_{2},...,u_{n}}\left( a\right) =\left( \text{empty product}\right)
\cdot\overline{\rho}\left( a\right) \cdot\left( \text{empty product}%
\right) ^{-1}=\overline{\rho}\left( a\right) \equiv\rho\left( a\right)
\ \operatorname{mod}\ tB\left[ \left[ t\right] \right] $ (since
(\ref{1rho0}) says that the constant term of the power series $\overline{\rho
}\left( a\right) $ equals $\rho\left( a\right) $). Now, we are going to
construct our sequence $\left( u_{1},u_{2},u_{3},...\right) \in B^{\left\{
1,2,3,...\right\} }$ by recursion: Let $m$ be a positive integer. Assume that
we have constructed some elements $u_{1}$, $u_{2}$, $...$, $u_{m-1}$ of $B$
such that (\ref{1proof}) holds for $n=m-1$. Then, we are going to construct a
new element $u_{m}$ of $B$ such that (\ref{1proof}) holds for $n=m$.
In fact, applying (\ref{1proof}) to $n=m-1$ (we can do this since we have
\textit{assumed} that (\ref{1proof}) holds for $n=m-1$), we obtain%
\[
\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( a\right) \equiv\rho\left(
a\right) \ \operatorname{mod}\ t^{m}B\left[ \left[ t\right] \right]
\ \ \ \ \ \ \ \ \ \ \text{for every }\left. a\in A\right. .
\]
In other words, every $a\in A$ satisfies%
\[
\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( a\right) -\rho\left(
a\right) \in t^{m}B\left[ \left[ t\right] \right] .
\]
Denoting the power series $\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left(
a\right) -\rho\left( a\right) $ by $q\left( a\right) $, we thus have
$q\left( a\right) \in t^{m}B\left[ \left[ t\right] \right] $ for every
$a\in A$. Hence, the first $m$ coefficients of the power series $q\left(
a\right) $ are $0$ for every $a\in A$. In other words,
\begin{equation}
q_{0}\left( a\right) =q_{1}\left( a\right) =...=q_{m-1}\left( a\right)
=0\ \ \ \ \ \ \ \ \ \ \text{for every }a\in A,\label{1proof.pa0}%
\end{equation}
where $q_{i}\left( a\right) $ denotes the $t^{i}$-coefficient of the power
series $q\left( a\right) $ for every $i\in\mathbb{N}$. Thus, for every $a\in
A$, we have%
\begin{align*}
q\left( a\right) & =\sum_{i=0}^{\infty}q_{i}\left( a\right) t^{i}%
=\sum_{i=0}^{m-1}\underbrace{q_{i}\left( a\right) }_{\substack{=0\\\text{(by
(\ref{1proof.pa0}))}}}t^{i}+q_{m}\left( a\right) t^{m}+\sum_{i=m+1}^{\infty
}\underbrace{q_{i}\left( a\right) t^{i}}_{\substack{\equiv
0\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] \\\text{(since
}i\geq m+1\text{ yields}\\t^{i}\equiv0\operatorname{mod}t^{m+1}B\left[
\left[ t\right] \right] \text{)}}}\\
& \equiv\sum_{i=0}^{m-1}0t^{i}+q_{m}\left( a\right) t^{m}+\sum
_{i=m+1}^{\infty}0=q_{m}\left( a\right) t^{m}\operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right] .
\end{align*}
Hence,%
\begin{align}
\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( a\right) &
=\underbrace{\left( \overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left(
a\right) -\rho\left( a\right) \right) }_{=q\left( a\right) \equiv
q_{m}\left( a\right) t^{m}\operatorname{mod}t^{m+1}B\left[ \left[
t\right] \right] }+\,\rho\left( a\right) \nonumber\\
& \equiv q_{m}\left( a\right) t^{m}+\rho\left( a\right)
\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] .\label{1proof4}%
\end{align}
However, the map $q:A\rightarrow B\left[ \left[ t\right] \right] $ (which
sends each $a\in A$ to $q\left( a\right) $) is $K$-linear (by its
definition, since the maps $\rho:A\rightarrow B$ and $\overline{\rho}%
_{u_{1},u_{2},...,u_{m-1}}:A\rightarrow B\left[ \left[ t\right] \right] $
are $K$-linear). Thus, the map $q_{m}:A\rightarrow B$ (which sends each $a\in
A$ to $q_{m}\left( a\right) $) is $K$-linear as well (since $q_{m}%
=\operatorname*{coeff}_{m}\circ\,q$, where $\operatorname*{coeff}_{m}:B\left[
\left[ t\right] \right] \rightarrow B$ is the map that takes every power
series to its $t^{m}$-coefficient, and thus $q_{m}$ is $K$-linear because both
$\operatorname*{coeff}_{m}$ and $q$ are $K$-linear).
Now, any $a\in A$ and $a^{\prime}\in A$ satisfy%
\begin{align*}
& \overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( a\right) \cdot
\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( a^{\prime}\right) \\
& =\left( \overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}%
t^{i}\right) \cdot\overline{\rho}\left( a\right) \cdot\left(
\overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right)
\right) ^{-1}\right) \\
& \ \ \ \ \ \ \ \ \ \ \cdot\left( \left( \overset{\longleftarrow
}{\prod_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right) \right) \cdot\overline{\rho
}\left( a^{\prime}\right) \cdot\left( \overset{\longleftarrow}{\prod
_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right) \right) ^{-1}\right) \\
& =\left( \overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}%
t^{i}\right) \right) \cdot\underbrace{\overline{\rho}\left( a\right)
\cdot\overline{\rho}\left( a^{\prime}\right) }_{\substack{=\overline{\rho
}\left( aa^{\prime}\right) \\\text{(by (\ref{1rhoaa'}))}}}\cdot\left(
\overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right)
\right) ^{-1}\\
& =\left( \overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}%
t^{i}\right) \right) \cdot\overline{\rho}\left( aa^{\prime}\right)
\cdot\left( \overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}%
t^{i}\right) \right) ^{-1}\\
& =\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( aa^{\prime}\right) .
\end{align*}
Since
\begin{align*}
& \underbrace{\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( a\right)
}_{\substack{\equiv q_{m}\left( a\right) t^{m}+\rho\left( a\right)
\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] \\\left(
\text{by (\ref{1proof4})}\right) }}\cdot\underbrace{\overline{\rho}%
_{u_{1},u_{2},...,u_{m-1}}\left( a^{\prime}\right) }_{\substack{\equiv
q_{m}\left( a^{\prime}\right) t^{m}+\rho\left( a^{\prime}\right)
\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] \\\left(
\text{by (\ref{1proof4}), applied to }a^{\prime}\text{ instead of }a\right)
}}\\
& \equiv\left( q_{m}\left( a\right) t^{m}+\rho\left( a\right) \right)
\cdot\left( q_{m}\left( a^{\prime}\right) t^{m}+\rho\left( a^{\prime
}\right) \right) \\
& =\underbrace{q_{m}\left( a\right) q_{m}\left( a^{\prime}\right) t^{2m}%
}_{\substack{\equiv0\operatorname{mod}t^{m+1}B\left[ \left[ t\right]
\right] \text{ (since }2m\geq m+1\\\text{yields }t^{2m}\equiv
0\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] \text{)}%
}}+\underbrace{q_{m}\left( a\right) \rho\left( a^{\prime}\right)
t^{m}+\rho\left( a\right) q_{m}\left( a^{\prime}\right) t^{m}%
}_{\substack{=\left( q_{m}\left( a\right) \rho\left( a^{\prime}\right)
+\rho\left( a\right) q_{m}\left( a^{\prime}\right) \right) t^{m}%
\\=\left( q_{m}\left( a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right)
\right) t^{m}\\\text{(since we are denoting }q_{m}\left( a\right)
\rho\left( a^{\prime}\right) \text{ by }q_{m}\left( a\right) a^{\prime
}\\\text{and denoting }\rho\left( a\right) q_{m}\left( a^{\prime}\right)
\text{ by }aq_{m}\left( a^{\prime}\right) \text{)}}}+\underbrace{\rho\left(
a\right) \rho\left( a^{\prime}\right) }_{\substack{=\rho\left( aa^{\prime
}\right) \\\text{(since }\rho\text{ is a }K\text{-algebra}%
\\\text{homomorphism)}}}\\
& \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \left( \text{since }t\text{
commutes with every element of }B\right) \\
& \equiv0+\left( q_{m}\left( a\right) a^{\prime}+aq_{m}\left( a^{\prime
}\right) \right) t^{m}+\rho\left( aa^{\prime}\right) \\
& =\left( q_{m}\left( a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right)
\right) t^{m}+\rho\left( aa^{\prime}\right) \operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right]
\end{align*}
and%
\[
\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left( aa^{\prime}\right) \equiv
q_{m}\left( aa^{\prime}\right) t^{m}+\rho\left( aa^{\prime}\right)
\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right]
\ \ \ \ \ \ \ \ \ \ \left( \text{by (\ref{1proof4})}\right) ,
\]
this equality yields%
\[
\left( q_{m}\left( a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right)
\right) t^{m}+\rho\left( aa^{\prime}\right) \equiv q_{m}\left( aa^{\prime
}\right) t^{m}+\rho\left( aa^{\prime}\right) \operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right] .
\]
In other words,%
\[
\left( q_{m}\left( a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right)
\right) t^{m}\equiv q_{m}\left( aa^{\prime}\right) t^{m}\operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right] .
\]
Hence, for every $i\in\left\{ 0,1,...,m\right\} $, the $t^{i}$-coefficient
of the power series $\left( q_{m}\left( a\right) a^{\prime}+aq_{m}\left(
a^{\prime}\right) \right) t^{m}$ equals the $t^{i}$-coefficient of the power
series $q_{m}\left( aa^{\prime}\right) t^{m}$. Applying this to $i=m$, we
see that the $t^{m}$-coefficient of the power series $\left( q_{m}\left(
a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right) \right) t^{m}$ equals
the $t^{m}$-coefficient of the power series $q_{m}\left( aa^{\prime}\right)
t^{m}$. But the $t^{m}$-coefficient of the power series $\left( q_{m}\left(
a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right) \right) t^{m}$ is
$q_{m}\left( a\right) a^{\prime}+aq_{m}\left( a^{\prime}\right) $, and the
$t^{m}$-coefficient of the power series $q_{m}\left( aa^{\prime}\right)
t^{m}$ is $q_{m}\left( aa^{\prime}\right) $. Hence, $q_{m}\left( a\right)
a^{\prime}+aq_{m}\left( a^{\prime}\right) $ equals $q_{m}\left( aa^{\prime
}\right) $. In other words,%
\[
q_{m}\left( aa^{\prime}\right) =q_{m}\left( a\right) a^{\prime}%
+aq_{m}\left( a^{\prime}\right) =aq_{m}\left( a^{\prime}\right)
+q_{m}\left( a\right) a^{\prime}.
\]
Note that we have proved this for all $a\in A$ and $a^{\prime}\in A$. Since
$q_{m}$ is a $K$-linear map, the condition (\ref{1ass}) (applied to $f=q_{m}$)
thus yields that there exists an element $s\in B$ such that%
\[
\left( q_{m}\left( a\right) =as-sa\text{ for all }a\in A\right) .
\]
Consider this $s$.
Now, let $u_{m}$ be the element $-s$. Then, we conclude that%
\begin{equation}
q_{m}\left( a\right) =u_{m}a-au_{m}\text{ for all }a\in A \label{1u_m}%
\end{equation}
(since $u_{m}=-s$ yields $s=-u_{m}$ and thus $q_{m}\left( a\right)
=as-sa=a\left( -u_{m}\right) -\left( -u_{m}\right) a=u_{m}a-au_{m}$). Now,
we must prove that (\ref{1proof}) holds for $n=m$. In fact, every $a\in A$
satisfies%
\begin{align*}
& \overline{\rho}_{u_{1},u_{2},...,u_{m}}\left( a\right) \\
& =\left( \underbrace{\overset{\longleftarrow}{\prod_{i=1}^{m}}\left(
1-u_{i}t^{i}\right) }_{=\left( 1-u_{m}t^{m}\right) \cdot
\overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right)
}\right) \cdot\overline{\rho}\left( a\right) \cdot\left(
\underbrace{\overset{\longleftarrow}{\prod_{i=1}^{m}}\left( 1-u_{i}%
t^{i}\right) }_{=\left( 1-u_{m}t^{m}\right) \cdot\overset{\longleftarrow
}{\prod\limits_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right) }\right) ^{-1}\\
& =\left( 1-u_{m}t^{m}\right) \cdot\underbrace{\left(
\overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right)
\right) \cdot\overline{\rho}\left( a\right) \cdot\left(
\overset{\longleftarrow}{\prod_{i=1}^{m-1}}\left( 1-u_{i}t^{i}\right)
\right) ^{-1}}_{\substack{=\overline{\rho}_{u_{1},u_{2},...,u_{m-1}}\left(
a\right) \equiv q_{m}\left( a\right) t^{m}+\rho\left( a\right)
\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] \\\left(
\text{by (\ref{1proof4})}\right) }}\cdot\underbrace{\left( 1-u_{m}%
t^{m}\right) ^{-1}}_{\substack{\equiv1+u_{m}t^{m}\operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right] \\\text{(since }\left(
1-u_{m}t^{m}\right) \left( 1+u_{m}t^{m}\right) \\=1-u_{m}^{2}t^{2m}%
\equiv1\operatorname{mod}t^{m+1}B\left[ \left[ t\right] \right] \text{)}%
}}\\
& \equiv\left( 1-u_{m}t^{m}\right) \left( q_{m}\left( a\right)
t^{m}+\rho\left( a\right) \right) \left( 1+u_{m}t^{m}\right) \\
& =q_{m}\left( a\right) t^{m}+q_{m}\left( a\right) u_{m}t^{2m}%
+\rho\left( a\right) +\rho\left( a\right) u_{m}t^{m}\\
& \ \ \ \ \ \ \ \ \ \ -u_{m}q_{m}\left( a\right) t^{2m}-u_{m}q_{m}\left(
a\right) u_{m}t^{3m}-u_{m}\rho\left( a\right) t^{m}-u_{m}\rho\left(
a\right) u_{m}t^{2m}\\
& \equiv q_{m}\left( a\right) t^{m}+\rho\left( a\right) +\underbrace{\rho
\left( a\right) u_{m}}_{=au_{m}}t^{m}-\underbrace{u_{m}\rho\left( a\right)
}_{=u_{m}a}t^{m}\\
& \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \left(
\begin{array}
[c]{c}%
\text{here we have removed all addends that contain a }t^{2m}\text{ or }%
t^{3m}\text{ factor,}\\
\text{because }2m\geq m+1\text{ yields }t^{2m}\equiv0\operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right] \text{ and}\\
\text{because }3m\geq m+1\text{ yields }t^{3m}\equiv0\operatorname{mod}%
t^{m+1}B\left[ \left[ t\right] \right]
\end{array}
\right) \\
& =q_{m}\left( a\right) t^{m}+\rho\left( a\right) +au_{m}t^{m}%
-u_{m}at^{m}\\
& =\left( u_{m}a-au_{m}\right) t^{m}+\rho\left( a\right) +au_{m}%
t^{m}-u_{m}at^{m}\ \ \ \ \ \ \ \ \ \ \left( \text{by (\ref{1u_m})}\right) \\
& =\rho\left( a\right) \operatorname{mod}t^{m+1}B\left[ \left[ t\right]
\right] .
\end{align*}
Hence, (\ref{1proof}) holds for $n=m$.
Thus we have shown that, if we have constructed some elements $u_{1}$, $u_{2}%
$, $...$, $u_{m-1}$ of $B$ such that (\ref{1proof}) holds for $n=m-1$, then we
can define a new element $u_{m}$ of $B$ in a way such that (\ref{1proof})
holds for $n=m$. This way, we can recursively construct elements $u_{1}$,
$u_{2}$, $u_{3}$, $...$ of $B$ which satisfy the congruence (\ref{1proof}) for
every $n\in\mathbb{N}$. Now, define a power series $b\in B\left[ \left[
t\right] \right] $ by $b=\lim\limits_{n\rightarrow\infty}%
\overset{\longleftarrow}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right) $
(this power series $b$ is well-defined since the sequence $\left(
\overset{\longleftarrow}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right)
\right) _{n\in\mathbb{N}}$ is a Cauchy sequence with respect to the $\left(
t\right) $-adic topology on the ring $B\left[ \left[ t\right] \right]
\ \ \ \ $\footnote{This is because for every $k\in\mathbb{N}$, there exists
some $j\in\mathbb{N}$ such that
\[
\left( \overset{\longleftarrow}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}%
t^{i}\right) \equiv\overset{\longleftarrow}{\prod\limits_{i=1}^{m}}\left(
1-u_{i}t^{i}\right) \operatorname{mod}t^{k}B\left[ \left[ t\right]
\right] \text{ for every }n\in\mathbb{N}\text{ and }m\in\mathbb{N}\text{
satisfying }n\geq j\text{ and }m\geq j\right)
\]
(namely, take $j=k$; then, any $n\geq j$ satisfies%
\begin{align*}
\overset{\longleftarrow}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right)
& =\left( \overset{\longleftarrow}{\prod\limits_{i=j+1}^{n}}%
\underbrace{\left( 1-u_{i}t^{i}\right) }_{\substack{\equiv
1\operatorname{mod}t^{k}B\left[ \left[ t\right] \right] \text{,}%
\\\text{since }i\geq j+1=k+1>k\\\text{yields }t^{i}\equiv0\operatorname{mod}%
t^{k}B\left[ \left[ t\right] \right] }}\right) \cdot\left(
\overset{\longleftarrow}{\prod\limits_{i=1}^{j}}\left( 1-u_{i}t^{i}\right)
\right) \\
& \equiv\left( \overset{\longleftarrow}{\prod\limits_{i=j+1}^{n}}1\right)
\cdot\left( \overset{\longleftarrow}{\prod\limits_{i=1}^{j}}\left(
1-u_{i}t^{i}\right) \right) =\overset{\longleftarrow}{\prod\limits_{i=1}%
^{j}}\left( 1-u_{i}t^{i}\right) \operatorname{mod}t^{k}B\left[ \left[
t\right] \right] ,
\end{align*}
and similarly any $m\geq j$ satisfies%
\[
\overset{\longleftarrow}{\prod\limits_{i=1}^{m}}\left( 1-u_{i}t^{i}\right)
\equiv\overset{\longleftarrow}{\prod\limits_{i=1}^{j}}\left( 1-u_{i}%
t^{i}\right) \operatorname{mod}t^{k}B\left[ \left[ t\right] \right] ,
\]
so that any $n\geq j$ and $m\geq j$ satisfy $\overset{\longleftarrow
}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right) \equiv
\overset{\longleftarrow}{\prod\limits_{i=1}^{m}}\left( 1-u_{i}t^{i}\right)
\operatorname{mod}t^{k}B\left[ \left[ t\right] \right] $).} and therefore
converges\footnote{since $B\left[ \left[ t\right] \right] $ is complete
for the $\left( t\right) $-adic topology}). Then, $b$ is a power series with
constant term $1$ (since $\overset{\longleftarrow}{\prod\limits_{i=1}^{n}%
}\left( 1-u_{i}t^{i}\right) $ has constant term $1$ for each $n\in
\mathbb{N}$), and every $a\in A$ satisfies%
\begin{align*}
b\overline{\rho}\left( a\right) b^{-1} & =\left( \lim
\limits_{n\rightarrow\infty}\overset{\longleftarrow}{\prod\limits_{i=1}^{n}%
}\left( 1-u_{i}t^{i}\right) \right) \cdot\overline{\rho}\left( a\right)
\cdot\left( \lim\limits_{n\rightarrow\infty}\overset{\longleftarrow
}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right) \right) ^{-1}\\
& =\lim\limits_{n\rightarrow\infty}\left( \overset{\longleftarrow
}{\prod\limits_{i=1}^{n}}\left( 1-u_{i}t^{i}\right) \cdot\overline{\rho
}\left( a\right) \cdot\left( \overset{\longleftarrow}{\prod\limits_{i=1}%
^{n}}\left( 1-u_{i}t^{i}\right) \right) ^{-1}\right) \\
& \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \left(
\begin{array}
[c]{c}%
\text{since taking reciprocals is a continuous map on}\\
\text{the set of all power series in }B\left[ \left[ t\right] \right]
\text{ with constant term }1
\end{array}
\right) \\
& =\lim\limits_{n\rightarrow\infty}\overline{\rho}_{u_{1},u_{2},...,u_{n}%
}\left( a\right) =\rho\left( a\right)
\end{align*}
(because of (\ref{1proof})\ \ \ \ \footnote{In fact, for every $i\in
\mathbb{N}$, there exists some $k\in\mathbb{N}$ such that every $n\in
\mathbb{N}$ satisfying $n\geq k$ satisfies $\overline{\rho}_{u_{1}%
,u_{2},...,u_{n}}\left( a\right) \equiv\rho\left( a\right)
\operatorname{mod}t^{i}B\left[ \left[ t\right] \right] $ (namely, set
$k=i$; then, every $n\in\mathbb{N}$ satisfying $n\geq k$ satisfies
$\overline{\rho}_{u_{1},u_{2},...,u_{n}}\left( a\right) \equiv\rho\left(
a\right) \operatorname{mod}t^{n+1}B\left[ \left[ t\right] \right] $ by
(\ref{1proof}) and thus also satisfies $\overline{\rho}_{u_{1},u_{2}%
,...,u_{n}}\left( a\right) \equiv\rho\left( a\right) \operatorname{mod}%
t^{i}B\left[ \left[ t\right] \right] $ because $t^{n+1}B\left[ \left[
t\right] \right] \subseteq t^{i}B\left[ \left[ t\right] \right] $ (since
$n+1\geq n\geq k=i$)).}). This proves Theorem 1.
\begin{thebibliography}{9} %
\bibitem {1}Pavel Etingof, Oleg Golberg, Sebastian Hensel, Tiankai Liu, Alex
Schwendner, Elena Udovina and Dmitry Vaintrob, \textit{Introduction to
representation theory}, July 13, 2010.\newline%
\texttt{http://math.mit.edu/\symbol{126}etingof/replect.pdf}
\end{thebibliography}
\end{document}