-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathCommRingLec07.tex
More file actions
executable file
·94 lines (82 loc) · 4.28 KB
/
Copy pathCommRingLec07.tex
File metadata and controls
executable file
·94 lines (82 loc) · 4.28 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
\stepcounter{lecture}
\setcounter{lecture}{7}
\sektion{Lecture 7}
\subsektion{Noetherian Descent}
Certain theorems about commutative rings can be proven by a reduction to the noetherian
case (the Hilbert basis theorem is secretly being used). If the statement $\F$ you want
to prove only involves a finite number of elements of $R$, say $a_1,\dots, a_n$. Then
look at the ring $R_0$ generated by $1, a_1,\dots, a_n$. It is the homomorphic image of
the ring $\ZZ[x_1,\dots, x_n]$. By the Hilbert basis theorem, this ring is noetherian,
so homomorphic images are also noetherian.
Here is a typical application. A non-zero commutative ring satisfies the strong rank
property ($R^m\hookrightarrow R^n$ implies $m\le n$). The strong rank property can be
rephrased as $R^{n+1}\not\hookrightarrow R^n$ (module theoretically). This can be
formulated as, ``a homogeneous system of $n$ equations and $n+1$ unknowns has a a
nontrivial solution''. It suffices to solve this system in $R_0$, so apply descent. Now
we just have to prove it in a noetherian ring. $R^{n+1}\hookrightarrow R^n$ can be
easily contradicted in the noetherian case. Think of $R^{n+1}$ as $X_1=R^n\oplus R$;
think of this $R$ as generated by $x$, so the image of $R^{n+1}$ is an image of $R^n$
direct sum with another module. Now repeat by embedding $R^{n+1}$ into the copy of
$R^n$. Then the module generated by $x,x_1,\dots, x_n$ gives an infinite ascending
chain.
\subsektion{\S 5 Artinian Rings}
\begin{definition}
The \emph{(Krull) dimension} of a commutative ring $R$ is defined as $\dim R =
\sup\{$lengths of chains of primes$\}$.
\end{definition}
In particular, a zero dimensional ring is one in which every prime ideal is maximal:
$\spec R=\Max R$.
\begin{definition}[von Neumann]
An element $a\in R$ is called \emph{von Neumann regular} if there is some $x\in R$
such that $a=axa$.
\end{definition}
\begin{definition}[McCoy]
A element $a\in R$ is \emph{$\pi$-regular} if some power of $a$ is von Neumann
regular.
\end{definition}
\begin{definition}
A element $a\in R$ is \emph{strongly $\pi$-regular} (in the commutative case)
if the chain $aR\supseteq a^2R\supseteq a^3R\supseteq \cdots$ stabilizes.
\end{definition}
A ring $R$ is von Neumann regular (resp.\ (strongly) $\pi$-regular) if every element of
$R$ is.
\begin{theorem}[5.2]
For a commutative ring $R$, the following are equivalent.
\begin{enumerate}
\item $\dim R=0$.
\item $R$ is rad-nil (i.e. $\rad R = \nil R$) and $R/\rad R$ is von Neumann regular.
\item $R$ is strongly $\pi$-regular.
\item $R$ is $\pi$-regular.
\item[] \hspace{-7ex} And any one of these implies
\item $\C(R)=U(R)$; any non-zero-divisor is a unit.
\end{enumerate}
\end{theorem}
\begin{proof}
$1\Rightarrow 2\Rightarrow 3\Rightarrow 4 \Rightarrow 1$ and $4\Rightarrow 5$. We will
not do $1\Rightarrow 2\Rightarrow 3$ here.
($3\Rightarrow 4$) Given $a\in R$, there is some $n$ such that $a^n R = a^{n+1}
R=a^{2n}R$, which implies that $a^n = a^n x a^n$ for some $x$.
($4\Rightarrow 1$) Is $\p$ maximal? Let $a\not\in \p$. Since $a$ is $\pi$-regular, we
have $a^n=a^{2n}x$, so $a^n(1-a^nx)=0$, so $1-a^nx\in \p$. It follows that $a$ has an
inverse mod $\p$.
($4\Rightarrow 5$) Using $1-a^nx=0$, we get an inverse for $a$.
\end{proof}
\begin{example}
Any local rad-nil ring is zero dimensional, since $2$ holds.
In particular, for a ring $S$ and $\m\in \Max S$, $R=S/\m^n$ is zero dimensional
because it is a rad-nil local ring.
\end{example}
\begin{example}[Split-Null Extension]
For a ring $A$ and $A$-module $M$, let $R=A\oplus M$
with the multiplication $(a,m)(a',m')=(aa',am'+a'm)$ (i.e.\ take the multiplication on
$M$ to be zero). In $R$, $M$ is an ideal of square zero. ($A$ is called a
\emph{retract} of $R$ because it sits in $R$ and can be recovered by quotienting by
some complement.) If $A$ is a field, then $R$ is a rad-nil local ring, with maximal ideal $M$.
\end{example}
\[\def\x#1#2{\left\{\parbox{#1}{\hfuzz=1pt \centering #2}\right\}}
\xymatrix @!0 @R=4pc @C=4pc{
\x{3.5pc}{reduced rings} \ar@{-}[dr] & &
\x{5pc}{zero-dimensional rings} \ar@{-}[dr]\ar@{-}[dl]& &
\x{4.5pc}{noetherian rings} \ar@{-}[dl]\\
& \x{6pc}{von Neumann regular rings} & & \x{3.3pc}{artinian rings}
}\]