0
$\begingroup$

A book I am reading gives the following definitions:

A collection $\{L_i:i=0,1,2,...\}$ of halfgroupoids $L_i$ is called an extension chain if $L_{i+1}$ is an extension of $L_i$ for each $i$. If $G$ is a subhalfgroupoid of the halfgroupoid $H$, the maximal extension chain of $G$ in $H$ is defined as follows: $G_0 = G$ and, for each nonnegative integer $i$, (a) $G_i$ is a subhalfgroupoid of $H$; (b) $G_{i+1}$ is an extension of $G_i$; (c) if $x, y \in G_i$ and if $xy=z$ in $H$, then $xy=z$ in $G_{i+1}$. The halfgroupoid $K=\bigcup_{i=0}^\infty G_i$ may be characterized as follows: (i) G is a subhalfgroupoid of K; (ii) K is a closed subhalfgroupoid of H; (iii) if $G$ is a subhalfgroupoid of the closed subhalfgroupoid $L$ of $H$, then $K$ is a subhalfgroupoid of $L$. We call $K$ the subhalfgroupoid of $H$ generated by G. In particular, $G$ generates $H$ if $K=H$.

I am confused about (iii) of the chacterizing properties of $K$. If we generate the halfgroupoid $K$ from the maximal extension chain of $G$ in $H$, shouldn't any closed subhalfgroupoid $L$ of $H$ that has $G$ as a subhalfgroupoid be itself a subhalfgroupoid of $K$, rather than the other way around?

$\endgroup$

1 Answer 1

0
$\begingroup$

My intuition that the maximal extension chain of $G$ in $H$ should result in a 'maximal' subhalfgroupoid $K$ in $H$ is understandable.

The key to my misunderstanding is an accidental omission in the properties of the latter object - it is a 'maximal' subhalfgroupoid in $H$ generated by $G$. The definition of $L$ as closed in $H$, along with the fact that $G$ is a subhalfgroupoid of $L$, essentially generates K as a subhalfgroupoid of $L$. Observe that because $L$ is closed in $H$: $$\{(a,b \in G), \ (c \notin G), \ ((a,b)\alpha_H=c)\} \rightarrow a,b \in L \rightarrow \{(a,b)\alpha_L = c), (c \in L)\}$$

Thus we see that the definition of L 'self-recurses' in the same manner as the chain:

$$\{(c \in L), \ (x \in G \vee L), ((c, x)\alpha_H = d) \} \rightarrow \{ (c,x)\alpha_L = d), (d \in L)\}$$

Therefore the subhalfgroupoid generated by $G$ in $H$, $K$, is a subhalfgroupoid of any closed subhalfgroupoid $L$ of $H$.

tldr: $L$ will generally have $K$ as just one of its many maximally generated subhalfgroupoids.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .