Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Let $p_1,...,p_{n+2}$ and $q_1,...,q_{n+2}$ be two sets of distinct points in general position in $\Bbb P^n$ (there may however be overlaps between the two sets.). Then there exists $\phi\in Aut(\mathbb{P}^n)$ taking $p_i\mapsto q_i$.

I've been having trouble with what should be the extremely simple proof of this fact. Michael Artin proves it for $n=2$ (though the same idea should work for any $n$) here following the line of reasoning I was trying to use. My confusion arises when he says to adjust the $p_i$ by a factor of $1/c_1$. This sounds like a perfectly fine thing to do in $\Bbb P^n$, but I don't see how it leads to the conclusion that $q=p_0+p_1+p_2$. If there is a typo, and he meant to write that we adjust each $p_i$ by $1/c_i$, then I understand how he reaches the conclusion, but I don't see how we are allowed to separately scale three tuples of coordinates before adding them together, i.e. Artin seems to be jumping back-and-forth between projective and affine coordinates without justification. Do I misunderstand the nature of $PGL(n)$ in some way?

share|cite|improve this question
Gah, yes! Thanks :) – Tabes Bridges Dec 3 '12 at 2:08
It looks like it may be a typo, and should rather say "adjust $p_i$ by $1/c_i.$" – Andrew Dec 3 '12 at 2:10
Which makes sense, since he only needs to work with some coordinate representative of the $p_i$'s,i.e., any particular choice can be scaled without affecting the fact that the coordinates represent $p_i.$ – Andrew Dec 3 '12 at 2:12
Yes; the weird part is that he seems to make this adjustment and then add the coordinates together as one might do in an affine situation. – Tabes Bridges Dec 3 '12 at 2:16
I see; we are absorbing the coefficients into the basis, thereby making a change of basis which is invisible under the scaling action. – Tabes Bridges Dec 3 '12 at 3:08

Based on Andrew's comments, I'm going to answer myself. I was misunderstanding Artin's proof. On one side of the linear map he works in terms of arbitrary $p_0,p_1,p_2,q\in\Bbb P^2$ in general position; on the other side he fixes $e_0,e_1,e_2\in\Bbb P^2$ the projections of the standard basis vectors, and for the final point chooses $\epsilon:=e_0+e_1+e_2$.

Since the given quadruples are in general position, the lifts of $p_0,p_1,p_2$ to $\Bbb A^3$ form a basis, so we can write $q=c_0p_0 + c_1p_1 + c_2p_2$. At this point, Artin says to adjust each $p_i$ by a factor of $1/c_i$, which I misinterpreted as invoking the projective scaling action; rather, we simply absorb the $c_i$ into the chosen representative affine coordinates for the $p_i$, and we map $q\mapsto\epsilon=e_0+e_1+e_2=(1:1:1)$, regardless of the values of $c_i$.

share|cite|improve this answer

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.