0
$\begingroup$

Let $A_3 := K[[X_1,X_2,X_3]]$ be a three-variable formal power series ring over a field $K$. We consider a linear equation

$(\sharp) \phantom{aa} a_1(X_1,X_2,X_3)Y_1 + \ldots + a_n(X_1,X_2,X_3)Y_n = 0$ $\phantom{a}$ with $\phantom{a}$ $a_1,\ldots,a_n \in A_3$.

We shall denote by $S(\sharp)$ the set of whole solutions of $(\sharp)$ such that $Y_1,\ldots,Y_n \in A_3$.

Definition. We shall define ``specialisation'' as substituting $X_3$ with some element $F(X_1,X_2) \in K[[X_1,X_2]]$.

After specialisation with $F(X_1,X_2)$, we obtain a linear equation

$(\sharp)_F \phantom{aa} a_1(X_1,X_2,F(X_1,X_2))Y_1 + \ldots + a_{n}(X_1,X_2,F(X_1,X_2))Y_n = 0$.

We shall denote by $S((\sharp)_F)$ the set of whole solutions of $(\sharp)_F$ such that $Y_1,\ldots,Y_n \in K[[X_1,X_2]]$.

By inserting $F(X_1,X_2)$ for $X_3$, we obtain a natural map $S(\sharp) \to S((\sharp)_F)$.

Q. For a given linear equation $(\sharp)$, does there always exist some specialisation $X_3 = F(X_1,X_2)$ such that the natural map $S(\sharp) \to S((\sharp)_F)$ is a surjection?

$\endgroup$
3
  • 1
    $\begingroup$ Is the same statement for $2$ variables obvious or well-known? $\endgroup$
    – Luc Guyot
    Sep 20, 2017 at 16:13
  • $\begingroup$ I think 2 variables case may be false. The number of variables must be at least 2 after specialisation. $\endgroup$
    – Pierre
    Sep 20, 2017 at 21:02
  • $\begingroup$ The Question turns out to be false. For example, X_1Y_1_1 + X_2Y_2 + X_3Y_3 = 0 is a counter-example. For any specialisation, always we have a solution (Y_1,Y_2,Y_3) = (f,g,1) which does NOT come from those of (#). This is given by Jopseph. $\endgroup$
    – Pierre
    Sep 22, 2017 at 12:43

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.