Classical Case
Finding pullback
Question: Suppose X \subset k^m and Y \subset k^n are irreducible algebraic sets (with Zarisky topology). Let f: X \to Y be given by(x_1, \ldots, x_m) \mapsto (f_1(x), \ldots, f_n(x)).
What is the induced map on coordinate rings?
Answer: The induced maps f^{\sharp}: A(Y) = k[y_1, \ldots, y_n]/ I \to A(X) = k[x_1. \ldots, x_m]/ J is given by pre-composition with f so it is
y_i \mapsto f_i.
Question: On the other hand, suppose we have a map \varphi: A(Y) \to A(X). How do we recover f such that \varphi = f^{\sharp}?
Answer: f: X \to Y can be defined by
(x_1, \ldots, x_m) \mapsto \left(\varphi (y_1)(x), \ldots, \varphi(y_n)(x)\right).
Answer: The induced maps f^{\sharp}: A(Y) = k[y_1, \ldots, y_n]/ I \to A(X) = k[x_1. \ldots, x_m]/ J is given by pre-composition with f so it is
y_i \mapsto f_i.
Recovering from Pullback
Question: On the other hand, suppose we have a map \varphi: A(Y) \to A(X). How do we recover f such that \varphi = f^{\sharp}?
Answer: f: X \to Y can be defined by
(x_1, \ldots, x_m) \mapsto \left(\varphi (y_1)(x), \ldots, \varphi(y_n)(x)\right).
Affine Scheme Case.
In the examples below, we first define the maps on points on \mathbb{C}^n. Then we interpret the map as the maps on coordinate rings as before. Afterwards, we recover the map on \text{Spec} \mathbb{C}^n.
Example. Parabola
Let f: \mathbb{C} \to \mathbb{C} be given by x \mapsto y = x^2. Then the induced map on coordinate rings is f^{\sharp}: \mathbb{C}[y] \to \mathbb{C}[x] given by y \mapsto x^2.
We claim that the induced map on Spec is
f: \text{Spec} \mathbb{C} \to \text{Spec} \mathbb{C}
(x - a) \mapsto (y - a^2).
(x - a) \mapsto (y - a^2).
Indeed f(x-a) = (f^{\sharp})^{-1} (x-a). Now f^{\sharp}(y - a^2) = (x^2 - a^2) \in (x-a) so (f^{\sharp})^{-1} (x-a) \ni (y- a^2), but the latter is maximal.
Generalization
Suppose f: k^m \to k^n is given by x \mapsto (f_1(x), \ldots, f_n(x)), where f_1, \ldots, f_n \in k[x_1, \ldots, x_m]. Then as before, it induces f^{\sharp}: k[y_1, \ldots, y_n] \to k[x_1, \ldots, x_m] be given by
y_i \mapsto f_i.
Note: We could have also just started with f^{\sharp}.
For any ideal I \subset k[x_1, \ldots, x_m] and J \subset k[y_1, \ldots, y_n] such that I \supset \phi(J) we have a morphism
f: \text{Spec} k[x_1, \ldots, x_m]/ I \to k[y_1, \ldots, y_n]/J.
Claim: f sends points (x_1 - a_1, \ldots, x_m - a_m) to (y_1 - f_1(a), \ldots, y_n - f_n(a)).
Proof: As (y - f_1(a), \ldots, y - f_n(a)) is a maximal ideal, it suffices to show that it is in the preimage under f^{\sharp} of (x_1 - a_1, \ldots, x_m - a_m). Indeed,
f^{\sharp}(y_1 - f_1(a), \ldots, y_n - f_n(a)) = (f_1(x) - f_1(a), \ldots, f_n(x) - f_n(a)) \in (x_1 - a_1, \ldots, x_n - a_n)
as the latter is the kernel of the evaluation at a map.
Thus we see that the induced map on Spec generalizes the induced map on classical algebraic varieties.
No comments:
Post a Comment