1

I'd like to show that $\alpha$ adjoined to $\mathbb{Q}$ is always a field, but I'm having trouble doing so.

For closure under addition, for example,

$(a_{0}+a_{1}\alpha +a_{2}\alpha ^{2}+...+a_{n-1}\alpha ^{n-1})+(b_{0}+b_{1}\alpha +b_{2}\alpha ^{2}+...+b_{n-1}\alpha ^{n-1})$

is clearly able to be expressed in the desired form.

Closure under multiplication isn't so simple, on the other hand. I'm not sure how to show that the product

$(a_{0}+a_{1}\alpha +a_{2}\alpha ^{2}+...+a_{n-1}\alpha ^{n-1})\times(b_{0}+b_{1}\alpha +b_{2}\alpha ^{2}+...+b_{n-1}\alpha ^{n-1})$

can also be expressed in the desired form and thus in $\mathbb{Q}[\alpha]$.

As for the multiplicative inverse, I'm also not sure how to prove that $\frac{1}{(a_{0}+a_{1}\alpha +a_{2}\alpha ^{2}+...+a_{n-1}\alpha ^{n-1})}$ is always able to be rationalized, and it seems like a pain to do so.

At this point, we have only been made aware of the definition of a field, so I suspect that a proof of this would follow a line of reasoning not necessarily based on just the definition of a field. The one hint given says to consider polynomials, but that doesn't help me out much. Help and hints would be appreciated.

Greg Martin
  • 78,820
PythonCZX
  • 103
  • 1
    You have not used the fact that $\alpha$ is algebraic. In general, $\mathbb Q[\alpha]$ is not field if $\alpha$ is not algebraic. But $\mathbb Q(\alpha)$ is a field. – GEdgar Jul 10 '22 at 19:25
  • 1
    Remember that with your notation there exists a polynomial $p$ of degree $n$ such that $p(\alpha) = 0$. Therefore $\alpha^n$ may be expressed as a number in $\mathbb{Q}(\alpha)$. Now think about how this can be generalized. – Hans Engler Jul 10 '22 at 19:26
  • 1
    Probable duplicate of https://math.stackexchange.com/questions/1631401/if-alpha-is-an-algebraic-element-and-l-a-field-does-the-polynomial-ring-l – lhf Jul 10 '22 at 19:38

1 Answers1

1

The set $\mathbb{Q}[\alpha]$ is the image of the ring homomorphism $\varphi\colon\mathbb{Q}[x]\to\mathbb{C}$ defined by $$ \varphi(b_0+b_1x+\dots+b_nx^n)=b_0+b_1\alpha+\dots+b_n\alpha^n $$ and therefore it is a subring.

Verifying that this is a homomorphism is exactly what you're attempting to do in order to show $\mathbb{Q}[\alpha]$ is a subring. However, this homomorphism is very useful for the rest of the proof.

More generally, if $\tau\colon R\to S$ is a homomorphism of commutative rings and $s\in S$ is a fixed element, there exists a unique ring homomorphism $\tau_s\colon R[x]\to S$ such that

  1. $\tau_s(r)=r$, for $r\in R$;
  2. $\tau_s(x)=s$.

The homomorphism is defined by $\tau_s(b_0+b_1x+\dots+b_nx^n)=\tau(b_0)+\tau(b_1)s+\dots+\tau(b_n)s^n$. In your case $\tau$ is the embedding $\mathbb{Q}\to\mathbb{C}$ and $s=\alpha$.

Now you surely know that $\alpha$ is the root of an irreducible polynomial $f(x)\in\mathbb{Q}[x]$ (the minimal polynomial). Let $c\in\mathbb{Q}[\alpha]$, $c\ne0$. Then $$ c=\varphi(g(x)) $$ for some polynomial $g(x)\in\mathbb{Q}[\alpha]$ and, since $c\ne0$, we can deduce that $g(x)$ is not divisible by $f(x)$. Since $f(x)$ is irreducible, there are polynomials $p(x),q(x)\in\mathbb{Q}[\alpha]$ such that $$ g(x)p(x)+f(x)q(x)=1 $$ because Bézout's identity carries over to $\mathbb{Q}[x]$, which is a Euclidean domain. Thus $$ 1=\varphi(1)=\varphi(g(x)p(x)+f(x)q(x))=g(\alpha)p(\alpha)+f(\alpha)q(\alpha)=g(\alpha)p(\alpha)=cp(\alpha) $$ and so $c^{-1}=p(\alpha)\in\mathbb{Q}[\alpha]$.

egreg
  • 238,574