Let's consider the magma $(G,*)$ with infinite elements.

Now I define $\operatorname{left}(G)$ the set of all the left translations

$$\operatorname{left}(G):\{L_a:a \in G ,L_a(b)=a*b\}$$

And $iter(a)$ as the set formed by the left traslation by $a\in G$ and closed under function composition or in other words the iterations of the left translations.

$$\operatorname{iter}(a):\{L_a^{n}:n \in \Bbb N\setminus\{0\}\}$$

What are the weakest conditions that $(G,*)$ must satisfies if we want that exist an collection injective functions $\mathcal F_a:\operatorname{iter}(a)\rightarrow > \operatorname{left}(G)$? $\mathcal F_a$ are defined in this way

$$\mathcal F_a(L_a^1)=L_a$$ $$\mathcal F_a(L_a^n)=L_{a'}$$ $$\mathcal (\mathcal F_a(L_a^n))(x)=L_a^n(x)$$

I think that this is equivalent to the this statement

$$\operatorname{iter}(a)\subseteq \operatorname{left}(G)$$.

If $(G,*)$ is associative then this always holds because $L_a(L_a(x))=L_{a*a}(x)$ and in general $\mathcal F_a(L_a^n)=L_{a^n}$

But this assumption is too stroong for my needs

A weaker condition is that if $*$ is right invertible (exists a $R_a^{-1}$ such that $R_b^{-1}(a*b)=(a*b)\setminus_* b=a$) then we can define the functions $\beta_a$

$$\beta_a(n)*b=a_1*\cdots(a_n*b)$$ $$\beta_a(n)=(a_1*\cdots(a_n*b))\setminus_* b$$


and these functions MUST be always constants (for every $b$) so we can define the injections $\mathcal F_a$ :


and thus $ \mathcal F_a$ exists

$$\mathcal F_a(L_a^n)=L_{\beta_a(n)}$$

$1$-$\beta_a$ not depend from $b$ and is always costant, $\mathcal F_a$ exists and $\operatorname{iter}(a)\subseteq \operatorname{left}(G)$ are three equivalent statements?

$3$-When $\beta_a$ satisfies that weak condition?

$3$-If $(G,*)$ is not commutative, not associative and not left invertible but is right invertible is possible that this bijection from $\operatorname{iter}(a)\rightarrow \operatorname{left}(g)$ exists? Or maybe there can be a surjection?


the image of $\operatorname{iter}(a)$ by$\mathcal F_a$, $\mathcal F_a[\operatorname{iter}(a)]$ should be always a commutative subsemigroup of $(\operatorname{left}(G),\circ)$, commutative submonoid if $*$ has a left unit and a commutative subgroup if $*$ is right invertible.

Update: here a related question.

Existence of an operation $\cdot$ such that $(a*(b*c))=(a\cdot b)*c$

Seems to me that the conditiont that the User Goos found is really similar to inclusion condition.

Anyways I think that the questions here are still meaningfull.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.