About Stability of Irreducibility for Germs of Holomorphic Functions

Huayi ZengThanks for helpful discussion with Sorin Popescu, Yusuf Mustopa, and Luis E. Lopez State University of New York at Stony Brook (hzeng@math.sunysb.edu)

April 11, 2005

Abstract
This survey is about irreducibility for germs of a holomorphic functions f   . I will show that when the dimension of the domain U   of this holomorphic function f   is greater than 2, the irreducibility of germs are not necessary to be stable.
That means, if the germ of f   at point p   is irreducible in the stalk of holomorphic functions at p   , this does NOT means there exists an open neighborhood V ⊂ U   of this point p   , such that for any point q ∈ V   , the germ of f   at q   is irreducible at the stalk of holomorphic functions at q  

1 Introduction

Let U   be an open set in C n   which contains 0, f   be a holomorphic function defined on U   , f p   is the germ of f   at point p ∈ U   .
For any two holomorphic functions g , h   defined on U   ,if g 0 , h 0   are relatively prime with each other, then with the help of resultants, we know that g , h   are relatively prime with each other nearby. Precisely to say, that means their exists an open neighborhood V ⊂ U   of 0, such that for any point q ∈ V   , g q   and h q   are relatively prime with each other. In this sense, we can say that Being co-prime is a stable property.
Can we say Irreducibility is a stable property?In the case of dimension 2, the answer is positive, and the proof is easy. But in the case of dimension 3, I will present a polynomial as counter-example.

2 Proof for the Case of Dimension 2

Statement: For any holomorphic function f = f ( z 1 , z 2 )   on U ⊂   C 2   ( 0 ∈ U   ), and the germ of f at origin is irreducible, then their exists an open neighborhood V ⊂ U   of 0, such that for any point q ∈ V   , f q   is irreducible.(Remark:If f ( p ) ≠ 0   , the f   is irreducible at p   . So we only need to care about zero points of f   .) Proof: Without the loss of generality, we can assume f ( 0 , z 2 )   is not identically 0 near the origin, and f ( 0 , 0 ) = 0   .
let w = z 2 d + e 1 ( z 1 ) z 2 d − 1 + ⋯ + e d − 1 ( z 1 ) + e d ( z 1 )   be a Weierstrass polynomial of f   near 0.
Because w   is irreducible at 0, so w   and ∂ w ∂ z 2   are relatively prime near 0. Then the resultant of w   and ∂ w ∂ z 2   is not zero. Then the common zero loci of w   and ∂ w ∂ z 2   are discrete near 0.
From above, we know that their exists an open set V ( 0 ∈ V ⊂ U )   , such that in U   , (0,0) is the only zero point of w   which is POSSIBLE to be singular.(since for other points in q ∈ U   , ∂ w ∂ z 2 ( p ) ≠ 0   ).We can conclude that at any zero point p ( p ≠ 0 )   of w   in V   , w   is a local complex parameter near p   . Since w   is a local complex parameter near p   , then the germ of w   at p   is irreducible.
Finally, because w   is a Weierstrass polynomial of f   at 0, then we know that in V   , the irreducibility of f   is as the same as t that of w   . □  

3 A Counter Example in Dimension 3

In the case of dimension 3, the statement should be:
Statement: For any holomorphic function f = f ( z 1 , z 2 , z 3 )   on U ⊂   C 3   ( 0 ∈ U   ), and the germ of f at origin is irreducible, then their exists an open neighborhood V ⊂ U   of 0, such that for any point q ∈ V   , f q   is irreducible.
But unfortunately, this statement is not true.In this section, I will present, a polynomial of three variables, as a counter example.
This polynomial is f = z 3 2 − z 1 z 2 2   .

3.1 Irreducibility of f   at origin

Obviously, near 0, f   is a Weierstrass polynomial of itself(we choose z 3   as the polynomial variable).Now, we will show the irreducibility at origin by means of contradiction.
If f   is not irreducible at origin, then its Weierstrass polynomial is decomposable at origin as a Weierstrass Polynomial.Assume that,near origin, f = ( z 3 − g ( z 1 , z 2 ) ) ( z 3 − h ( z 1 , z 2 ) )   , here g , h   are holomorphic functions of variable z 1 , z 2   near 0, and g(0,0)=h(0,0)=0.
From the factorization f = ( z 3 − g ( z 1 , z 2 ) ) ( z 3 − h ( z 1 , z 2 ) )   , we know that g + h = 0 , g h = − z 1 z 2 2   , which implies g 2 = z 1 z 2 2   near 0.
But if g 2 = z 1 z 2 2   near 0. Then for some ɛ ∈   C whose norm is small enough, g 2 ( z 1 , ɛ ) = ɛ 2 z 1   near 0. But just from elementary knowledge of functions of one complex variable, we know this is not possible.
From argument above, we know f   is irreducible at origin.

3.2 Further Argument

At point p = ( z , 0 , 0 ) ( z ≠ 0 )   , we know that f ( p ) = 0   , and easily we can factorize f   as f = ( z 3 + z 2 r ) ( z 3 − z 2 r )   near p   , here r   is a one-variable holomorphic function such that r 2 = z 1   near ( z , 0 , 0 )   (Because z is not 0, so we can take square-root of z 1   near by.).
From the argument in 3.2, we know that, in any neighborhood U   of origin, there EXISTS some point p   such that f   is not irreducible at p   . This fact can destroy our statement at the beginning of this section.