For the sake of brevity let me use the following terms. A subset $X$ of $\bar{\mathbb{Q}}$ will be called "small" if for any number field $K$, the intersection $X\cap K$ is finite. Similarly, a set $Y\subset \bar{\mathbb{Q}}$ will be called "large" if for any number field $K$, $Y$ contains all but finitely many elements of $K.$ My question is whether the following claim is true.

Claim: Let $f\in \bar{\mathbb{Q}}[x, y]$ be a polynomial which is non constant in $y,$ and let $X\subset \bar{\mathbb{Q}}$ be a "small" set. Then the set $Y=\lbrace y\in \bar{\mathbb{Q}}$: there exists $x\in X$ such that $f(x,y)=0\rbrace$ is not "large".


Not if $f(x,y)$ cuts out a curve of genus at least $2$.

Consider such a polynomial $f(x,y)$. Choose an ordering of number fields, such as by discriminant. Consider the set $X = \{ x \in \overline{\mathbb Q} :$ there exists $y \in \overline{\mathbb Q}$ such that $f(x,y)=0$ and $\mathbb Q(y)$ comes before $\mathbb Q(x)$.

Then $Y$ contains the set of all $y$ such that there exists $x \in \overline{\mathbb Q}$ such that $f(x,y)=0$ and $\mathbb Q(x)$ comes after $\mathbb Q(y)$.

It is sufficient to prove that the first set is small. Then by symmetry, and the fact that for all but finitely many $y$ there is at least one such $x$, the second set is large.

But this follows from Mordell's conjecture. Given a number field $K$, there are only finitely many number fields that come before $K$ or a subfield of $K$. For each such field $L$, there are finitely many pairs $(x,y) \in K \times L$ with $f(x,y)=0$, because all such pairs are $KL$-points of the curve $f(x,y)=0$. Adding those all together, only finitely many values of $x$ occur.


Your Answer

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

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