MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I don't know whether this is known or not, but I was thinking of the following problem.

Let $n$ and $m$ be natural numbers. Are there $n$ polynomial $f_1,...,f_n\in \mathbb{C}[x]$, such that all of their intersection numbers are at least $m$?

It is also possible to state an even harder question in a more natural (albeit less basic) way: Are there $n$ homogeneous polynomials in 2 variables for which all of their intersection numbers in $\mathbb{P}^1_{\mathbb{C}}$ are at least $m$?

EDIT: Sorry, the comments made me realize I wanted a slightly different condition: that for every $j$ there's a unique $k\neq j$ such that $f_j(0)=f_k(0)$. (In particular, $n$ is assumed to be even.) I will allow the intersection at $x=0$ to not be of multiplicity $\geq m$, but I will ask that over $x\neq 0$ the intersection multiplicity will be $\geq m$.


The question was posed in a way that algebraic geometers would understand, because I suspect they are most likely to come up with a solution to this question. I wanted to emphasize, however, that intersection multiplicity is something that every high-school student can understand: If $f_1$ and $f_2$ are polynomials with coefficients in $x$, and they meet at say $x=3$, then their intersection multiplicity at $x=3$ is the greatest natural number $l$ such that $(x-3)^l$ divides the polynomial $f_1-f_2$.

share|cite|improve this question
It seems that if you make all your polynomials equal to 0, then you get the desired result (for the 1-variable case). – Zatrapilla Feb 12 '12 at 18:05
In fact, if you make $f_i(x)=x^m$, then all the pair-wise intersections are $2m$. – Zatrapilla Feb 12 '12 at 18:07
Hmmm, let me clarify in the body of the question. – James D. Taylor Feb 12 '12 at 18:09
Let's do a quick example: $0,x,1,1+x$ have the desired property at 0. If $m=2$, you're saying to multiply by $(x-1)^2$, say. $1(x-1)^2$ and $x(x-1)^2$ indeed intersect with multiplicity $2$ at $x=1$, but they would also intersect with multiplicity $1$ at some $x\neq 0,1$. So that's undesirable. – James D. Taylor Feb 12 '12 at 18:39
@Mahdi: Yes, that's indeed what I mean. – Florian Eisele Feb 12 '12 at 22:57
up vote 3 down vote accepted

Ok, so here's the long version of what I already said in the comments:

Start by picking $g_1,\ldots,g_n\in \mathbb C[x]$ such that $g_i\neq g_j$ whenever $i\neq j$ which satisfy the prescribed condition at $0$ (that's not hard). Since $g_i-g_j\neq 0$ for each $i\neq j$, there will be only finitely many $0\neq a\in\mathbb C$ such that $g_i(a)-g_j(a)=0$ for some pair $i,j$ with $i\neq j$. Call these values $a_1,\ldots,a_k$. Then define $$ f_i(x) := \left(\prod_{j=1}^k (x-a_j)^m\right)\cdot g_i(x) $$ Those $f_1,\ldots,f_n$ have pairwise intersection also in $a_1,\ldots,a_k$ (and possibly in $0$), but now with multiplicty at least $2m$ (except in $0$, where the intersection multiplicty is left as it was). The condition at zero is clearly preserved and we didn't add any further intersection points either.

share|cite|improve this answer
Just a small remark: we should only include the nonzero $a_i$'s, otherwise we will ruin the prescribed condition at zero. – Mahdi Majidi-Zolbanin Feb 12 '12 at 23:40
@Mahdi: Thanks! Just edited. – Florian Eisele Feb 12 '12 at 23:47

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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