18
$\begingroup$

Let $S=\mathbb{C}[x_1,x_2,\dots,x_n]$ be a polynomial ring. Let $n \geq 3$. Let $h_a$ denotes the complete homogeneous symmetric polynomial of degree $a$. $$ h_a=\text{ sum of all monomials of degree } a.$$ For example: for $n=3$ and $a=2$, one has: $$h_2=x_1^2+x_2^2+x_3^2+x_1x_2+x_1x_3+x_2x_3.$$ Question: Is it true that $h_a$ is an irreducible element in $\mathbb{C}[x_1,x_2,\dots,x_n]$.

The $h_a$ was introduced by Sir Issac Newton in seventeenth centuary along with many other symmetric polynomials such as Power sum symmetric polynomials and elementary symmetric polynomials.

It is known that $p_a=x_1^a+\cdots+x_n^a$ is an irreducible element in $\mathbb{C}[x_1,x_2,\dots,x_n]$ for $n \geq 3$. I am interested to know similar result for the complete homogeneous symmetric polynomial.

Thank you
Neeraj Kumar.

$\endgroup$
9
  • 2
    $\begingroup$ You can reduce to the case $n=3$ using an induction argument, but I can't immediately see how to solve $n=3$. $\endgroup$
    – Will Sawin
    May 26, 2012 at 15:51
  • $\begingroup$ You should have a look to McDonald's book about symmetric and Hall polynomials $\endgroup$ May 26, 2012 at 16:14
  • 7
    $\begingroup$ View the polynomial as a monic polynomial in $x_n$. The leading term is $1$, the constant term is the polynomial with $a$ the same and $n$ one less. By induction on $n$, this is irreducible, so each factor must have constant term $1$ or that polynomial. But since the polynomial is homogeneous, the factors are homogeneous, thus degree $0$ or degree $a$, thus no nontrivial factors. $\endgroup$
    – Will Sawin
    May 26, 2012 at 16:30
  • 1
    $\begingroup$ For the case $n=3$ it would suffice to prove that the projective curve $h_a(x,y,z)=0$ is smooth, in other words that $h_a$ has no common zero with all its partial derivatives in $\mathbf{P}^2(\mathbf{C})$. $\endgroup$ May 26, 2012 at 16:51
  • 1
    $\begingroup$ @Will Sawin: Thank you for your explanation! And I also agree with Patricia that answers should not be hidden as comments. $\endgroup$ May 26, 2012 at 20:08

5 Answers 5

17
$\begingroup$

EDIT : prompted by Will Sawin's comment, the argument now works for every $n \geq 3$. Thanks !

The polynomial $h_a(x_1,\ldots,x_n)$ is irreducible for every $a \geq 1$ and $n \geq 3$.

Recall that if $h_a = FG$ with $F$ and $G$ non constant then $F$ and $G$ have to be homogenous. By Bézout's theorem, the hypersurfaces $F=0$ and $G=0$ intersect in the projective space $\mathbf{P}^{n-1}(\mathbf{C})$ since $n \geq 3$. This gives a singular point on the hypersurface $h_a=0$. So it suffices to prove that $h_a,\frac{\partial h_a}{\partial x_1},\ldots,\frac{\partial h_a}{\partial x_n}$ have no common zero in $\mathbf{C}^n \backslash \{0\}$. This fact is true for every $a \geq 1$ and $n \geq 2$, and we prove this by induction.

For $a=1$ it is easy. For $n=2$ it amounts to the fact that the polynomial $T^a+\cdots+T+1 = (T^{a+1}-1)/(T-1)$ has distinct roots.

In general, we have $$h_a = \sum_{a_1+\cdots+a_n=a} x_1^{a_1} \cdots x_n^{a_n}$$ so that $$\frac{\partial h_a}{\partial x_i} = \sum_{a_1+\cdots+a_n=a-1} (a_i+1) x_1^{a_1} \cdots x_n^{a_n}.$$ Note that $\sum_{i=1}^n \frac{\partial h_a}{\partial x_i} = (a+n-1) h_{a-1}$. Moreover $h_a=x_i h_{a-1}+R$ for some polynomial $R$ not depending on $x_i$, so that $$\frac{\partial h_a}{\partial x_i}=h_{a-1}+x_i \frac{\partial h_{a-1}}{\partial x_i}.$$ If $x=(x_1,\ldots,x_n)$ is a common zero of $h_a$ and all its partial derivatives then $h_{a-1}(x)=0$ and $x_i \frac{\partial h_{a-1}}{\partial x_i}(x)=0$ for all $i$. By induction, we must have $x_i=0$ for some $i$. Assume for example $x_n=0$. Then $(x_1,\ldots,x_{n-1}) \in \mathbf{C}^{n-1}$ provides in fact a common zero of $h_a(x_1,\ldots,x_{n-1})$ and all its partial derivatives, so applying the induction hypothesis for $n-1$ we get $x=0$.

$\endgroup$
2
  • 2
    $\begingroup$ There is no reason this argument should fail for $n>3$. $\endgroup$
    – Will Sawin
    May 26, 2012 at 19:26
  • 1
    $\begingroup$ @Will Sawin : Thanks for pointing this out ! The same argument should prove that the hypersurface $h_a(x_1,\ldots,x_n)=0$ is smooth for any $n \geq 3$, which implies that $h_a$ is irreductible. Alternatively, your argument reduces the general case to the case $n=3$. $\endgroup$ May 26, 2012 at 19:35
8
$\begingroup$

You were probably mainly interested in the case of $a\le n$. Here's a quick counterexample for $a>n$, letting $n=1$ and $a=2$. Notice then that $h_2 = (h_1)^2$.

If there were a factorization with $a\le n$, it would need to involve nonsymmetric polynomials. But since $S$ is a unique factorization domain, this would mean for any nonsymmetric factor, we would also need all the elements of its orbit as factors. While I was editing this, Will Sawin came up with an elegant proof, whereas I was just grinding through cases, so I will omit those cases.

Since Will mentions still needing the base case of $n=a=3$, notice that we would need a nonsymmetric factor, which means a nonsymmetric factor with either (a) terms of the form $x_ix_j^2$, with nontrivial orbit, yielding too high a degree, or (b) it has terms of the form $x_i$ with some included and some omitted, contradicting the product including all $x_i^3$ terms, or (c) it has some terms $x_ix_j$ with $i\ne j$ and omits others, again necessitating other factors in the orbit, giving too high a degree for the product, or (d) it has some terms $x_i^2$ and not others, giving the same contradiction.

$\endgroup$
1
  • $\begingroup$ It is well known in symmetric functions that $h_1,…,h_n$ are algebraically independent in ${\bf Q}[x_1,…,x_n]$ and that the symmetric functions $h_{\lambda }=h_{\lambda_1}⋯h_{\lambda_k}$ for $\lambda = (\lambda_1,\dots ,\lambda_k)$ with $\lambda_1\ge \lambda_2\ge \cdots \ge\lambda_k$ and $\sum \lambda_i = d$ form a vector space basis for the degree $d$ homogeneous piece of the ring of symmetric functions in $n$ variables. Hence my claim about needing nonsymmetric factors if there were going to be a nontrivial factorization. $\endgroup$ May 26, 2012 at 17:39
2
$\begingroup$

To go further in Patricia's direction, take $n=2$ and $a=3$. Then $h_a=(x+y)(x^2+y^2)$.

$\endgroup$
4
  • $\begingroup$ For generally, for $a>2$ odd, we have that $h_a = \sum_{i=0}^{a} x^i y^{a-i}$ is reducible since it is normed in $x$ and has $-y$ as a zero. For $a$ even I think it will be irreducible ... $\endgroup$ May 26, 2012 at 16:27
  • $\begingroup$ One might want to go a little bit further, as Neeraj specified $n\geq 3$. $\endgroup$
    – Will Sawin
    May 26, 2012 at 16:28
  • $\begingroup$ Oh, right. For $n \geq 3$ these trivial examples vanish. $\endgroup$ May 26, 2012 at 16:31
  • 2
    $\begingroup$ On $a$ even: We are working over $\mathbb C$. Every polynomial in two variables, homogenous in degree at least two, is reducible. $\endgroup$
    – Will Sawin
    May 26, 2012 at 16:38
2
$\begingroup$

This is most relevant to Patricia's answer: There is a paper by Schinzel where he addresses the question of whether you have to factorize into symmetric factors (the answer is yes, under some conditions). I have no access to the paper, but the math review is useful

$\endgroup$
3
  • $\begingroup$ A related paper of Schinzel seems to be *Reducibility of a special symmetric form, Acta Mathematica Universitatis Ostraviensis, vol. 14 (2006), issue 1, pp. 71-74 (dml.cz/handle/10338.dmlcz/137486) $\endgroup$ May 26, 2012 at 16:58
  • $\begingroup$ The main theorem of that paper requires, in our notation, $n>a+1$, the case Patricia Hersh already handled. $\endgroup$
    – Will Sawin
    May 26, 2012 at 17:00
  • $\begingroup$ This makes me curious for $n\le a+1$ if there is any more general result (i.e. not just for homogeneous symmetric functions) about when irreducibility in the ring of symmetric functions implies irreducibility in ${\bf Q}[x_1,\dots ,x_n]$. But then I guess there are examples like the square of the Vandermonde determinant that limit how much one can hope for. $\endgroup$ May 26, 2012 at 19:52
1
$\begingroup$

We can extend Patricia's argument somewhat. For $n=3$ and $a=4$, symmetrically it is $s_1^4-3s_1^2s_2+s_2^2+2s_1s_3$ which is irreducible (view it as a polynomial in s_2, complete the square, constant term is not a perfect square) so there are no symmetric factors, so it must have 3 asymmetric factors, which must be linear, then another linear symmetric factor. But there are no linear symmetric factors!

$\endgroup$

Your Answer

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

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