Consider the given system of polynomial equations, where all the coefficients are in $\mathbb{C}$: $$\begin{cases} y^n=P(x)\\ Q(x,y)=0\end{cases}$$ I would like to establish that either this system has solutions $(x,y)\in\mathbb{C}$ for all $x \in \mathbb{C}$, or either has solutions for countably (or even better, finitely) many $x\in\mathbb{C}$. I am not completely sure this is true but haven't found any counterexamples yet.

So far, I have found that if we have solutions for infinitely many $x$, then we have solutions for infinitely many $y$.

Any ideas?


You can determine the resultant $R(x)=Res_y(y^n-P(x),Q(x,y))$ of both polynomials eliminating $y$. If it is the zero polynomial, then the polynomials have a common factor, which means that for any $x$ you find at least one $y$ so that $(x,y)$ a solution (of the common factor and thus of the system).

If the resultant is not zero, then $R(x)$ is a polynomial. Only at the roots of this polynomial you will then find at least one $y$ so that $(x,y)$ is a solution of the system.


I misunderstood the question and will edit this later with (hopefully) a response.

Given any $x \in \mathbb{C}$, there exists $n$ possible $y \in \mathbb{C}$ such that $y^n = P(x)$ (possibly with multiplicity, if say $P(x) = 0$). However, determining whether the system has a solution (and how many) depends completely on the structure of $Q(x,y)$. Consider the examples below.

If $Q(x,y) = y^n - P(x)$, then any pair $(x,y)$ from above satisfies the system, so we get uncountably many solutions.

If $Q(x,y) = y^n - P(x) +1$, then the first relation forces $y^n - P(x) = 0$, so that $Q(x,y) = 1$ for all $(x,y)$ satisfying the first. Hence the system has no solutions at all in this case.

  • $\begingroup$ Yes, that's precisely my point: I want to establish the duality between "for any $x\in\mathbb{C}$, there is some solution $(x,y)$" and "there are only countably many $x$, such that there is some solution $(x,y)$". You gave examples of these two cases. Can you show there are no other cases? $\endgroup$
    – Lukas
    Mar 31 '14 at 19:02
  • $\begingroup$ Oh, I misunderstood. I will edit my response. $\endgroup$ Mar 31 '14 at 19:08

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.