If A is a communative, associative k-Algebra, it is finitely generated as an Algebra, if there exists, $a_1,...,a_n$, so that the morphism:
$$
\phi_A: K[X_1,...,X_n] \mapsto A\\
f \to f(a_1,...,a_n)
$$
is surjectiv.
A k-Algebra is a vector space, with a map
$$
A \times A \mapsto A
$$
so that for, $ x,y,z \in A$ and $a,b \in k$
$$
(x+y)*z=xz+yz\\
x*(y+z)=xy+zy\\
(ax)*(by)=(ax)(by)
$$
So what i don't unterstand is that by definition k does not have to be in A(?).
But that would mean the morphismen $\phi$ would not work.
So what am I missing?
-
$\begingroup$ What is the codomain of $\phi$? $\endgroup$– RandallCommented Jul 13 at 12:16
-
$\begingroup$ oh, sorry, it is A, edited my post $\endgroup$– jojo mathCommented Jul 13 at 12:21
-
1$\begingroup$ So the confusion is $ \phi_A $, how it is defined, because if $k$ is not in $A$, then what happens with a constant polynomial? For example if A is the Polynomials generated by monoms, with degree >1 over $\mathbb{Q}$, then 5 is in $\mathbb{Q}[X_1,...X_n]$, but 5 is not in $A$ $\endgroup$– jojo mathCommented Jul 13 at 13:47
-
$\begingroup$ I deleted my comment because I realized I made a silly mistake in what I was trying to say. Christopher Nicol's answer is what I wanted to get at. The only complaint is the case of the $k$-algebra $0$ where $\phi_A$ is zero and does not induce an inclusion on $k$. $\endgroup$– ShrugsCommented Jul 13 at 14:17
1 Answer
Usually an algebra is a ring with unit. In this situation you have an element $1 \in A$. And your canonical morphism from $k$ to $A$ is
$ \alpha \longrightarrow \alpha \cdot 1. $
So the axioms you used to define a $k$-algebra imply we haceva ring morphism. As $k$ is a field and our morphism is non $0$, it is injective. So we have a canonic inclusion of $k$ in $A$, and now it is straightforward that $\phi_A$ maps constant polynoms to the image of $k$ by my morphism.
However if you don't have a unit in your ring, you may not be able to find a canonical copy of $k$ in $A$. For instance assume $A=k^n$ with the multiplication being $0$. In this context you won't be able to find a canonical image of $k$ in $A$. But this pathological cas is not important in your context. Indeed we can take the existence of $\phi_A$ as a definition of finitely generated algebra. So $A$ is a quotient of a ring with unit, thus $A$ has a $1$.
-
$\begingroup$ Thanks, I will add the existence of a 1 to my definition $\endgroup$ Commented Jul 13 at 14:36