Question: Let $f(x)\in \mathbb{C}[x]$. Does there always exist some $\alpha \in \mathbb{C}$ such that $g(x):=f(x)-\alpha$ has distinct roots?
My intuition leads me to believe this is true. For a simple example, if $f(x)=x^n$ then we can set $\alpha=1$ so that the roots of $g(x)$ are the $n$th roots of unity. Proceeding via contradiction, if there were some $f(x)$ such that for all $\alpha\in \mathbb{C}$, $g(x)$ does not have distinct roots, what goes wrong? Taking this route, such a $g(x)$ would be inseparable, so for any $\alpha$ we see that $g(x)$ would always have a common root with its derivative. Can anyone find a contradiction? Is there a simple direct proof? Is the statement incorrect and can someone provide a counterexample?
Yes, such $\alpha$ exists. Note that if $w$ is a double root for the polynomial $g(z)=f(z)-\alpha$ then $g(z)=(z-w)^2q(z)$ and $g'(z)=(z-w) (2q(z)+(z-w)q'(z))$ which implies that $g'(w)=f'(w)=0$. Hence just take $\alpha$ such that $g(w)=f(w)-\alpha\not=0$ when $f'(w)=0$, that is $$\alpha\in \mathbb{C}\setminus\{f(w) : f'(w)=0\}$$ (we are assuming that the degree of $f$ is greater than one).