%! TEX root = AG2.tex % vim: tw=50 % 11/10/2024 10AM % Introductory reading % B. Hassett % M. Reid % Commutative Algebra texts % Atiyah + MacDonald % Matsumura Commutative Algebra, Commutative % Ring Theory % AG texts: % Hartshorne, Algebraic Geometry, 1977 % Gortz-Wedhorn Algebraic Geometry I, 2010 % Ravi Vakil, The rising sea: Foundations of % algebraic geometry \newpage \setcounter{section}{-1} \section{Motivation} In Part II \courseref{AG} we defined an affine algebraic variety by: \begin{definition*}[Affine algebraic variety] We fix an algebraically closed field $k$ and defined affine $n$-space $\Abb^n_k \defeq k^n$, and for an ideal $I \le k[x_1, \ldots, x_n]$ we defined \[ Z(I) \defeq \{(a_1, \ldots, a_n) \in \Abb^n \st f(a_1, \ldots, a_n) = 0 ~\forall f \in I\} \subseteq \Abb^n_k .\] We define a topology on $\Abb^n$ by taking the closed sets to be the sets of the form $Z(I) \subseteq \Abb^n$. \end{definition*} We will introduce \emph{schemes}. Why schemes? Why not varieties? \begin{enumerate}[(1)] \item With varieties, we always work with algebraically closed fields. For example, take $k = \Rbb$, $I = (x^2 + y^2 + 1) \subseteq \Rbb[x, y]$. Then $Z(I) = \emptyset$. \item Number theory? We study Diophantine equations, i.e. $I \subseteq \Zbb[x_1, \ldots, x_n]$. So $Z(I) \subseteq \Zbb^n$. \item Even if $k$ is algebraically closed, we lose information when we pass from $I$ to $Z(I)$. For example, $I = (x^2) \subseteq \Cbb[x]$. Then $Z(I) = \{0\} = Z(x)$. Recall Hilbert's Nullstellensatz, which states that the set of polynomials in $k[x_1, \ldots, x_n]$ vanishing on $Z(I)$ for $I \subseteq k[x_1, \ldots, x_n]$ is the \emph{radical} of $I$. But it is natural ideals like $(x^2)$. $(y - x^2, y - \alpha) \subseteq \Cbb[x, y]$ ($\alpha \in \Cbb$). \begin{center} \includegraphics[width=0.6\linewidth]{images/9deb3be75ff44e0e.png} \end{center} If $x = 0$, $(x^2 - y, y - \alpha) = (x^2, y)$. \end{enumerate} \subsection{Categorical philosophy} (Read definition on Wikipedia or see \courseref{CT}). Let $\mathbf{Sets}$ be the category of sets. $\mathbf{Sets}$ is the category with objects being all sets, with morphisms between objects maps of sets. If $X$ and $Y$ are sets, we write $\Hom(X, Y)$ for the set of maps between $X$ and $Y$. Note that there is a bijection \begin{align*} \Hom(\{*\}, X) &\to X \\ (f : \{*\} \to X) &\mapsto f(*) \end{align*} Let's use this philosophy to understand points on affine algebraic varieties. $\Abb_k^0$ is a point. If $X$ is an affine variety, then the points of $X$ should be in $1-1$ correspondence with $\Hom(\Abb_k^0, X)$. Giving a morphism between affine varieties is easy. Denote by $A(X)$ (sometimes $k[X]$) the coordinate ring ($A(X) = k[x_1, \ldots, x_n] / I(X)$, where $I(X)$ is the ideal of functions that vanish on $X$). This is a $k$-algebra. We showed that if $X$ and $Y$ are affine varieties, then $\Hom(X, Y) = \Hom_{\text{$k$-alg}}(A(Y), A(X))$ (see Part II \courseref{AG} or handout). So \[ \Hom(\Abb_k^0, X) = \Hom_{\text{$k$-alg}}(k[x_1, \ldots, x_n] / I(X), k) .\] Note that giving a $k$-algebra homomorphism \[ k[x_1, \ldots, x_n] \to k \] can be done by specifying the images of $x_i$ can be done by specifying the images of $x_i$, say $x_i \mapsto a_i$ such that for any $f \in I(X)$, $f(a_1, \ldots, a_n) = 0$. So there is a $1-$ correspondence between such $k$-algebra homomorphisms and points of $X$. If $k$ is algebraically closed, the maximal ideals of $k[x_1, \ldots, x_n]$ are precisely the ideals of the form $(x_1 - a_1, \ldots, x_n - a_n)$ for $(a_1, \ldots, a_n) \in \Abb^n$ (a form of Hilbert Nullstellensatz) and the maximal ideals of $A(X)$ are of the form $(x_1 - a_1, \ldots, x_n - a_n) \bmod I(X)$ with $(a_1, \ldots, a_n) \in X$. Thus we have a $1-1$ correspondence between points of $X$ and maximal ideals of $A(X)$. \begin{center} \includegraphics[width=0.6\linewidth]{images/2b4760156e344021.png} \end{center} Now suppose $k$ is \emph{not} algebraically closed. Let's consider $k$-algebra homomorphisms \[ k[x_1, \ldots, x_n] / I(X) = A(X) \to L \] where $L$ is an extension of $k$. These are given by $x_i \mapsto a_i$ with $f(a_1, \ldots, a_n) = 0$ for all $f \in I(X)$. Thus \[ \Hom_k(A(X), L) = \{(a_1, \ldots, a_n) \in \Abb_L^n \st f(a_1, \ldots, a_n) = 0 ~\forall f \in I(X)\} .\]