@def mintoclevel=2 @def maxtoclevel=3

Algebraic Curves Lecture 16:

The 16th lecture of algebraic curves lectured by Karl Christ
  1. General and uniform positions

(Missed Monday due to flight)

General and uniform positions

Theorem: Let C⊆PrC\subseteq \mathbb P^r, r≥3r\geq 3, CC is non-degenerate irreducible of degree dd. Let HH be a general hyperplane in Pr\mathbb P^r. Then any k≤rk\leq r points in H∩CH\cap C will span a linear space of dimension k−1k-1.
Example: Take C⊆P3C\subseteq \mathbb P^3. Then no three points in C∩HC\cap H are colinear for a general plane H⊆P3H\subseteq \mathbb P^3.
Slogan: For a general hyperplane HH the subsets containing kk elements in H∩CH\cap C are indistinguishable

Let's let

I={(p,H)∣p∈H∩C}⊆C×U\begin{aligned} I = \{(p,H) | p\in H \cap C\} \subseteq C\times U \end{aligned}

where U⊆(Pr)∗U\subseteq (\mathbb P^r)^* is the set of hyperplanes in Pr\mathbb P^r meeting CC transversely. Now consider the map π2:I→U\pi_2:I\to U, it is a dd-sheeted covering of UU. Fix H0∈UH_0\in U and consider the action of π1(U,H0)\pi_1(U,H_0) (the fundamental group of UU based at H0H_0) on π2−1(H0)\pi^{-1}_2(H_0).

Lemma: π1(U,H0)\pi_1(U,H_0) acts as the full symmetric group SdS_{d} on π2−1(H0)\pi^{-1}_2(H_0).

Proof.   Introduce the sets

I~(2)={(p1,p2,H) ∣ p1≠p2,p1,p2∈C∩H}⊂C×C×(Pr)∗\begin{aligned} \tilde{I}(2) = \{(p_1,p_2,H) ~\mid~ p_1\neq p_2, p_1,p_2 \in C\cap H\}\subset C\times C\times (\mathbb P^r)^* \end{aligned}
and
I(2):=I~(2)∩(C×C×U).\begin{aligned} I(2) := \tilde{I}(2) \cap (C\times C\times U). \end{aligned}
Note that passing from I~(2)\tilde I(2) to I(2)I(2) simply requires the plane to meet the curve transversely. Now let π1:I~(2)→(C×C)∖Δ\pi_1:\tilde I(2)\to (C\times C)\setminus \Delta be the projection map (here we remove the diagonal since we required that p1≠p2p_1\neq p_2 for a tuple in I~(2)\tilde I(2).

The set π1−1(p1,p2)\pi_1^{-1}(p_1,p_2) is the set of all hyperplanes passing through p1p_1 and p2p_2, and it is isomorphic to Pr−2\mathbb P^{r-2}. Claim: both I~(2)\tilde I(2) and I(2)I(2) are both irreducible and are in particular connected. > **Side Remark:** if f:X→Yf:X\to Y is some dominant map of varieties and you know that XX is irreducible, you can immediately conclude that YY is irreducible. However, if YY is irreducible, concluding that XX is irreducible is much harder. A sufficient condition is the following: (1) YY is irreducible (2) the fibers of ff are all irreducible (3) the fibers have the same dimension and (4) ff is proper. The space C×C∖ΔC\times C\setminus \Delta is irreducible -- the diagonal is a codimension 2 subset in this case. Furthermore, π1\pi_1 satisfies (1)-(4) in the remark above, so I~(2)\tilde I(2) is irreducible. A similar argument tells me that I(2)I(2) is irreducible. Now we return our attention to the map and fiber

π2:I→U,π2−1(H0).\begin{aligned} \pi_2:I\to U, \hspace{1cm} \pi_2^{-1}(H_0). \end{aligned}
This means π1(U,H0)\pi_1(U,H_0) acts 2-transitively on π2−1(H0)\pi^{-1}_2(H_0). Let G⊆SdG\subseteq S_d be the transitive subgroup of SdS_d given as the image of the fundamental group π1(U,H0)→Aut⁡(π2−1(H0))=Sd\pi_1(U,H_0)\to \operatorname{Aut}(\pi_2^{-1}(H_0)) = S_d. To show that G=SdG = S_d it suffices to show that GG contains a transposition. Indeed, if it has a transposition τ∈G\tau \in G, then

  1. I can get any other transposition τ′\tau' by conjugation with a permutation σ\sigma such that (τ1,τ2)↦(τ1′,τ2′)(\tau_1,\tau_2) \mapsto (\tau_1', \tau_2'). Then στσ−1=τ′.\sigma\tau\sigma^{-1} = \tau'.

  2. The transpositions generate SdS_d, so G=SdG = S_d.

□\square
 

©Isaac Martin. Last modified: March 18, 2024.