How to show that that every finitely generated variety is locally finite

441 Views Asked by At

Revised Question:

From what I read, an algebra is locally finite if every finitely generated sub-algebra is finite, and a variety is locally finite if every member is locally finite (where a variety is the class of algebras closed under homomorphisms, sub-algebras, and direct products (direct product is denoted as $\prod_{i \in I}\mathbb{A}_i$)).

A variety is finitely generated if it is generated by a finite set of finite algebras.

Now, I read that there's a result (from universal algebra) that every finitely generated variety is locally finite.

The result says that given a variety generated by a set of finite algebras $\{\mathbb{A}_1,\mathbb{A}_2,...,\mathbb{A}_k\}$, we have that for any algebra $\mathbb{B}$ in the variety, there exists a homomorphism $\mathbb{C} \to \mathbb{B}$, for some algebra $\mathbb{C}$ in the variety, where $\mathbb{C} \cong \prod_{i\in I} \mathbb{C}_i$.

The proof uses the fact that $\textbf{Hom}(\mathbb{C},\mathbb{A}_i)$ is supposed to be a finite set - which is then supposed to imply that there are only finitely many maps from $\mathbb{C} \rightarrow \prod_{i \in I} \mathbb{C}_i\rightarrow \mathbb{C}_i$. Given this, there will be a finite subset $J$ of $I$ such that $\mathbb{C}$ embeds in $\prod_{j \in J}\mathbb{C}_j$, where $\prod_{j \in J}\mathbb{C}_j$ is a finite product of finite algebras. From this embedding of $\mathbb{C}$ into the finite product of finite algebras $\prod_{j \in J}\mathbb{C}_j$, we have that $\mathbb{C}$ is finite, and in turn, $\mathbb{B}$ is finite as well

But the proof did not go into any details as to why (1) $\textbf{Hom}(\mathbb{C},\mathbb{A}_i)$ is finite and how (2) this fact implies that $\mathbb{C} \to \prod_{i \in I}\mathbb{C}_i \to \mathbb{C}_i$ ... Any help for (1) and (2) ?

1

There are 1 best solutions below

5
On BEST ANSWER

First, let me include the full proof from the book you're most likely using. It'll be easier than rewriting it.

The book is Bergman's "Universal Algebra".

enter image description here

enter image description here

First, you stated that $\mathbb{C}$ is isomorphic to $\prod_{i\in I}\mathbb{C}_i$. This is not true, it's only isomorphic to a sub-algebra of $\prod_{i\in I}\mathbb{C}_i$.

Since $X$ generates $\mathbb{C}$, $C = \bigcup_{n=1}^\infty X_n$ where $X_0 = X$, $X_{n+1} = X_n\cup \{ p(a_1, ..., a_k): a_i\in X_n, p\text{ is } k\text{-ary}\}$.

Now if two homomorphisms $f, g:\mathbb{C}\to\mathbb{A}_i$ are equal on $X$, one can prove (and this is easy) by induction that they are equal on each of $X_n$, so that they are equal on all of $C$, so $f=g$.

Therefore the map $f\mapsto f|_X, \text{Hom}(\mathbb{C}, \mathbb{A_i})\to A_i^X$ which sends a homomorphism to its restriction to $X$ is injective. So there can be at most $|A_i^X|<\infty$ of such homomorphisms.

I hope this clears it up, I invite you to a further discussion if it's not clear.