Throughout this post, is a field and is the ring of matrices with entries in
We have already seen the concepts of nil and nilpotency in a ring in this blog several times, but let’s see it here again. Let be a ring, and a subset of We say that is nil if every element of is nilpotent, i.e. for every there exists a positive integer such that We say that is nilpotent if there exists a positive integer such that i.e. for any It is clear that is nilpotent if and only if for all positive integers that are large enough.
If is nilpotent, then is clearly nil too but, as the following example shows, the converse is not true.
Example. Let be the standard -basis for and let Then and so is nil. But is not nilpotent because for any positive integer
Notice that the set in the above Example is not multiplicatively closed because We are now going to show that if is both nil and multiplicatively closed, then is nilpotent. In fact, we show that But first a couple of useful lemmas.
Lemma 1. Any finite sum of nilpotent left ideals of a ring is nilpotent.
Proof. i) By induction, we only need to prove that for two ideals. So let be two nilpotent left ideals of So for some positive integers We claim that So we need to show that if then
The product on the left-hand side is a sum of terms of the form where for each either or So we just need to show that Since there are of either at least of are in or at least of are in We will assume that at least of are in since the argument for the other case is similar. So we can write
where each is in and each is in Notice that there might be nothing before or after or between and for some but that won’t cause any problem for our argument. Now, for all because and is a left ideal. We also have that Thus
and so
Next is to generalize this basic fact that if is nilpotent, then
Lemma 2. If is a nilpotent subset of then
Proof. We will assume since there is nothing to prove in this case. Let be the -vector space of all vectors with entries in and let
So is the annihilator of in Clearly each is a -vector subspace of and Since is nilpotent, there exists the smallest positive integer such that So and
Claim. for all
Proof. By contradiction. Let be the largest positive integer such that That means if for some then Notice that in fact we have because But then implies that and so contradicting the maximality of This completes the proof of the Claim.
So, by the Claim, Therefor
and so which gives because
We are now ready to prove the main result.
Theorem. Let be a nil, multiplicatively closed subset of Then
Proof. Since there is nothing to prove if we will assume that Let be the -vector subspace of generated by We are done if we show that because Notice that is also a ring because is multiplicatively closed.
The proof of is by induction over If then for some and so
Suppose now that Let and define the map by for all Clearly is an onto -linear map. Also, because if then and if then where is the smallest positive integer such that So, as -vector spaces, and hence Therefore since is the -vector space generated by and is a nil multiplicatively closed subset of because is so, we have by our induction hypothesis. So we have shown that for all Now, for some positive integer and so because Each is clearly a left ideal of and we just showed that they are all nilpotent. Thus, by Lemma 1, is nilpotent and so, by Lemma 2,
The Theorem in fact holds in any finite-dimensional -algebra not just To prove that, we need the following simple yet important lemma.
Lemma 3 (Cayley). Every -dimensional -algebra with identity can be embedded into
Proof. Since the ring of matrices is just the ring of -linear maps Define the map by for all and see that is an injective ring homomorphism.
Corollary. Let be an -dimensional -algebra with identity, and let be a nil, multiplicatively closed subset of Then
Proof. By Lemma 3, is a subset of and so, by the Theorem,
Note. The reference for this post is Chapter 1, Section 3, of the book Polynomial Identities in Ring Theory by Louis Rowen.
Exercise. Try to fully understand the proof of Lemma 3.