%%%% THIS FILE IS AUTOMATICALLY GENERATED
%%%% DO NOT EDIT THIS FILE DIRECTLY,
%%%% ONLY EDIT THE SOURCE, tom-63/document.tex.
%%%%
%% Standard package list
\documentclass[letterpaper]{article}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage[english]{babel}
\usepackage[top=3cm, bottom=3cm, left=3.5cm, right=3.5cm]{geometry}
\usepackage[onehalfspacing]{setspace}
\usepackage{amsmath,amssymb,amsthm,wasysym}
\usepackage{nicefrac,booktabs}
\usepackage{mathptmx}
\usepackage{cite}
\usepackage[colorlinks=true]{hyperref}

%% Various helpers for Tom's papers
\newcommand{\gs}{\textnormal{gs}}
\newcommand{\ord}{\textnormal{ord}}
\newcommand{\Exp}{\textnormal{Exp}}
\newcommand{\Log}{\textnormal{Log}}
\newcommand{\lcm}{\textnormal{lcm}}
\newcommand{\range}{\textnormal{range}}
\newcommand{\NR}{\textnormal{NR}}
\newcommand{\Mod}[1]{\left(\textnormal{mod}~#1\right)}

\newcommand{\ap}[2]{\left\langle #1;#2 \right\rangle}
\newcommand{\summ}[1]{\sum_{k=1}^m{#1}}
\newcommand{\bt}[1]{{{#1}\mathbb{N}}}
\newcommand{\fp}[1]{{\left\lbrace{#1}\right\rbrace}}
\newcommand{\intv}[1]{{\left[1,{#1}\right]}}


%% Lifted from http://stackoverflow.com/questions/2767389/referencing-a-theorem-like-environment-by-its-name
%% This lets me do things like "Theorem A" and have the references work properly.
\makeatletter
\let\@old@begintheorem=\@begintheorem
\def\@begintheorem#1#2[#3]{%
  \gdef\@thm@name{#3}%
  \@old@begintheorem{#1}{#2}[#3]%
}
\def\namedthmlabel#1{\begingroup
   \edef\@currentlabel{\@thm@name}%
   \label{#1}\endgroup
}
\makeatother
% end lift

\newtheoremstyle{namedthrm}
  {}{}{}{}{}{}{ }  % This last space needs to be there
  {\bf\thmname{#1} \thmnote{#3}.}
%% End reference hack


%% Document start
\date{}
\begin{document}

%% Content start


\newtheorem{cor}{Corollary}
\newtheorem{lemma}{Lemma}
\newtheorem*{thm}{Theorem}
\newtheorem{q}{Problem}
\theoremstyle{definition}
\newtheorem*{defn}{Definition}
\newtheorem*{ack}{Acknowledgements}

\title{A Semigroup Union of Disjoint Locally Finite Subsemigroups Which is Not Locally Finite}
\author{T. C. Brown}
\date{}
\maketitle
\begin{center}{\small {\bf Citation data:} T.C. {Brown}, \emph{A semigroup union of disjoint locally finite subsemigroups which
  is not locally finite}, Pacific J. Math. \textbf{22} (1967), 11--14.}\bigskip\end{center}

\begin{abstract}
The semigroup $S$ of the title is the free semigroup $F$ on four generators factored by the congruence generated by the set of relations $\{w^2 = w^3: w \in F\}$. The following lemma is proved by examining the elements of a given congruence class of $F$:

\textsc{Lemma.} If $x,y \in S$ and $x^2 = y^2$, then either $xy = x^2$ or $yx = x^2$.

From the Lemma it then easily follows that the (disjoint) subsemigroups $\{y \in S: y^2 = x^2\}$ of $S$ are locally finite.
\end{abstract}

This note answers in the negative a question raised by Shevrin in~\cite{shevrin1965}.

\begin{thm} \label{thm: 1}
There exists a semigroup $S$ with disjoint locally finite subsemigroups $S_e$ such that $S = \bigcup S_e$ and $S$ is not locally finite.
\end{thm}

Let $F$ be the free semigroup with identity on four generators. Let $\sim$ denote the smallest congruence on $F$ containing the set $\{(x^2,x^3): x \in F\}$. That is, for $w,w' \in F$, $w \sim w'$ if and only if a finite sequence of ``transitions", of either of the types $ab^2c \to ab^3c$ or $ab^3c \to ab^2c$, transforms $w$ into $w'$.

The equivalence classes of $F$ with respect to $\sim$ are taken as the elements of $S$, and multiplication in $S$ is defined in the natural way.

There is given in~\cite{dean1965} a sequence on four symbols in which no block of length $k$ is immediately repeated, for any $k$. Thus the left initial segments of this sequence give elements of $F$ containing no squares. Since no transition of the form $ab^2c \to ab^3c$ or $ab^3c \to ab^2c$ can be applied to an element of $F$ contining no squares, the equivalence classes containing these elements consist of precisely one element each; thus the semigroup $S$ is infinite, and hence not locally finite.

In what follows, the symbols $\alpha, \alpha_1,\alpha_2, \dots$ refer to transformations (on elements of $F$) of the form 
\[ ab \to ayb, \quad \textup{where $a \sim ay$,} \quad \textup{and} \quad a,b,y \in F. \]
The symbols $\beta,\beta_1,\beta_2,\dots$ refer to transformations of the type
\[ axb \to ab, \quad \textup{where $a \sim ax$,} \quad \textup{and} \quad a,b,x \in F. \]

Note that $ab^2c \to ab^3c$ is an $\alpha$, and $ab^3c \to ab^2c$ is a $\beta$.

\begin{lemma} \label{lemma: 1}
If $w,w' \in F$, and $w\beta\alpha = w'$, then there are $\alpha_1,\beta_1$ such that $w\alpha_1\beta_1 = w'.$
\end{lemma}

\begin{proof}
Let 
\[ w = axb, \quad w\beta = ab, \quad \textup{where $a \sim ax$.} \]
Let 
\[ w\beta = AB, \quad w\beta\alpha = AyB, \quad \textup{where $A \sim Ay$.} \]
There are two cases:

(i) $A$ is contained in $a$. That is,
\[ a = Aa' \quad \textup{and} \quad w' = wb\alpha = a\beta \alpha = Aa'b\alpha = Aya'b. \]
Here let
\[ w\alpha_1 = axb\alpha_1 = Aa'xb\alpha_1 = Aya'xb. \]
Now since
\[ Aya' \sim Aa' = a \sim ax = Aa'x \sim Aya'x, \]
we may let
\[ w\alpha_1\beta_1 = Aya'xb\beta_1 = Aya'b = w'. \]

(ii) $A$ is not contained in $a$. That is,
\[ b = b_1b_2, \quad A = ab_1, \quad A \sim Ay, \]
and 
\[ w' = w\beta \alpha = ab\alpha = ab_1b_2\alpha = Ab_2\alpha = Ayb_2 = ab_1yb_2. \]
Since 
\[ axb_1 \sim ab_1 = A \sim Ay = ab_1y \sim axb_1y, \]
we may let 
\[ w\alpha_1 = axb\alpha_1 = axb_1b_2\alpha_1 = axb_1yb_2, \]
and
\[ w\alpha_1\beta_1 = axb_1yb_2\beta_1 = ab_1yb_2 = w'. \qedhere \]
\end{proof}

\begin{lemma} \label{lemma: 2}
If $w,w' \in F$, $w\gamma_1\gamma_2\cdots \gamma_m = w'$, where each $\gamma_i$ is either an $\alpha$ or a $\beta$, then there are $\alpha_1,\dots,\alpha_n, \beta_1,\dots,\beta_k$ such that $w\alpha_1 \cdots \alpha_n \beta_1 \cdots \beta_k = w'$.
\end{lemma}

\begin{proof} This follows immediately from Lemma~\ref{lemma: 1} by induction.
\end{proof}

\begin{lemma} \label{lemma: 3}
The word $ab\alpha$ contains a left initial segment which is equivalent to $a$.
\end{lemma}

\begin{proof} Let $ab = AB$, $ab\alpha = AyB$, where $A \sim Ay$. Again there are two cases:

(i) $A$ is contained in $a$. That is, $a = Aa'$, $ab\alpha = Aa'b\alpha = Aya'b$. Since $A \sim Ay$, the left initial segment $Aya'$ of $ab\alpha$ is equivalent to $a$.

(ii) $A$ is not contained in $a$. That is, $b = b_1b_2$, $A = ab_1$, $ab\alpha = ab_1b_2\alpha = ab_1yb_2$. Here, $a$ itself is a left initial segment of $ab\alpha$, and is certainly equivalent to $a$.
\end{proof}

\begin{lemma} \label{lemma: 4}
If $x,y \in F$ and $x^2 \sim y^2$, then either $y \sim xa$ for some $a \in F$, or $x \sim yb$ for some $b \in F$.
\end{lemma}

\begin{proof}
By Lemma~\ref{lemma: 2}, thee are $\alpha_i$ and $\beta_j$ such that $xx\alpha_1 \cdots \alpha_m \beta_1 \cdots \beta_n = yy$. Let $w = xx\alpha_1\cdots \alpha_m= yy\beta_n^{-1} \cdots \beta_1^{-1}$. By Lemma~\ref{lemma: 3}, $w$ contains a left initial segment $A$ equivalent to $x$. Similarly, since each $\beta_i^{-1}$ is an $\alpha$, $w$ also contains a left initial segment $B$ equivalent to $y$. Depending on which segment contains the other, either $B = Aa$ for some $a$, or $A = Bb$ for some $b$. In the first case, $y \sim B = Aa \sim xa$; in the second, $x \sim A = Bb \sim yb$.
\end{proof}

\begin{lemma} \label{lemma: 5}
In this lemma, ``=" will denote equality in $S$. Let $e$ be an idempotent element of $S$: $e = e^2$. Let $S_e = \{x \in S: x^2 = e\}$. Then $S_e$ is a locally finite subsemigroup of $S$.
\end{lemma}

\begin{proof} First, we note that $z \in S_e$ implies $ez = ze = e$. For $ez = z^2z = z^2 = e$, and similarly $ze = e$. Now let $x,y \in S_e$, that is, $x^2 = y^2 = e$. By Lemma~\ref{lemma: 4}, either $y = xa$ or $x = yb$. In the first case, we obtain 
\[ xy = x^2a = x^3a = x^2y = ey = e. \]
In the second case, we obtain similarly that $yx = e$. Thus $x, y \in S_e$ implies $xy = e$ or $yx = e$. In either case, $(xy)^2 = e$, that is, $xy \in S_e$. Thus $S_e$ is a semigroup.

Now let $x_1,\dots,x_n \in S_e$, and let $\langle x_1,\dots,x_n \rangle$ denote the subsemigroup of $S_e$ generated by $x_1,\dots,x_n$. If $n = 1$, then $\langle x_1 \rangle$ is clearly finite; so suppose $n > 1$. Then every element of $\langle x_1,\dots,x_n \rangle$ may be expressed as a product of not more than $n$ of the $x_i$'s. For any product $z$ of more than $n$ $x_i$'s must contain some $x_i$ twice: $z = ax_ibx_ic$, where $a,b,c \in S_e$. Since either $x_ib = e$ or $bx_i = e$, it follows that $x_ibx_i = e$ and $z = aec = ec = e = x_1x_1$. This shows that $\langle x_1,\dots,x_n \rangle$ is finite, and hence that $S_e$ is locally finite.
\end{proof}

The theorem follows immediately from Lemma~\ref{lemma: 5}, since clearly $e \ne e'$ implies that $S_e$ and $S_{e'}$ are disjoint, and
\[ S = \cup S_e. \]


\bibliographystyle{amsplain}
\bibliography{tom-all}
\end{document}