Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $A \subseteq B$ be integral domains and let $\phi:A \rightarrow \Omega$ be a homomorphism of $A$ into the infinite algebraically closed field $\Omega$. Let $x \in B$ and suppose that $x$ is transcendental over $A$, i.e. it is the root of no non-zero polynomial with coefficients in $A$.

I would like to see whether my understanding about extending $\phi$ to $A[x]$ is correct. Let $g(x) \in A[x]$ be a monic polynomial of positive degree. Then it will be non-zero because $x$ is transcendental over $A$. Denote by $g^{\phi}$ the polynomial of $\Omega[z]$ obtained by taking its coefficients to be the images of the coefficients of $g$ under $\phi$. Then $g^{\phi}$ will also be non-zero and it will have a finite number of roots. Since $\Omega$ is infinite, we can find a $\xi \in \Omega$ such that $g^{\phi}(\xi) \neq 0$. Then we can extend $\phi$ to $A[x]$ by sending $x$ to $\xi$. Is this argument correct?

Now, if $A$ is a field and we are interested in extending $\phi$ to $A(x)$, the above argument is not valid anymore, because we can not guarantee that a non-zero element of $A(x)$ will be sent to a non-zero element of $\Omega$. Is this argument correct?


share|cite|improve this question
The notation is a bit of a mess, because we usually mean $A[x]$ to be a ring of polynomials with an indeterminate $x$. As it turns out, since $x$ is transcendental, your $A[x]$ is isomorphic to that ring of polynomials over $A$, however, so the existence of extensions of $\phi$ follows from universal properties of $A[x]$ – Thomas Andrews Aug 13 '12 at 14:53
For your last paragraph: if a nonzero element of $A(x)$ is sent to $0\in\Omega$, what can you say about the kernel of $\phi$? (Remember that $A(x)$ is itself a field.) – M Turgeon Aug 13 '12 at 14:56
@MTurgeon: The kernel of any field homomorphism must be the zero ideal. – Manos Aug 13 '12 at 14:58
@Manos Then what can you conclude about your last paragraph? – M Turgeon Aug 13 '12 at 14:59
Dear Manos: If $\phi:A\to B$ is a commutative ring morphism, $X$ an indeterminate, and $b$ an element of $B$, then there is a unique morphism $A[X]\to B$ which coincides with $\phi$ on $A$ and maps $X$ to $b$. – Pierre-Yves Gaillard Aug 13 '12 at 15:54
up vote 3 down vote accepted

Gathering up some of the comments, including P-Y Gaillard's and Th. Andrews':

First, we might demonstrate that a "transcendental" behaves just like an "indeterminate". The polynomial ring $A[X]$ in an indeterminate $X$ over ring $A$ is really the free $A$-algebra on one generator, meaning that, given any ring hom $f:A\rightarrow B$ and given $b\in B$, there is a unique extension $F:A[X]\rightarrow B$ such that $F(X)=b$. (Note, no "solving" of equations is required: $X$ can go anywhere, $0$ or otherwise.)

Thus, letting $x$ be the "transcendental", the identity map $A\rightarrow A$ extends uniquely to $A[X]\rightarrow A[x]$ extending $X$ to $x$. Now observe that the "transcendental" condition is exactly equivalent to the kernel being trivial, so $A[X]$ is in natural bijection with $A[x]$.

Thus, given a ring hom $A\rightarrow B$, and given any $b\in B$, there is a (unique) extension $A[x]\rightarrow B$ sending $x$ to $b$.

Such maps need not extend to the field of fractions $A(x)$, however, as obvious examples show: for example, with $B=A$ and the identity map on $A$, $x$ can be sent anywhere at all, say to $b\in B=A$, but then $x-b$ maps to $0$, so this map cannot be extended to the fraction $1/(x-b)$.

share|cite|improve this answer
Thank you for your nice answer :-) – Manos Aug 14 '12 at 13:45

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.