Suppose T is an operator (defined on a Hilbert space, say). Does the "algebra generated by $T$" have a commonly understood meaning? I couldn't find a satisfactory answer after some internet searching. It seems to me that there are two ways one could define the algebra $A$ generated by $T$ (without including additional structure on $A$, like closure), namely $A=\{p(T): p$ is a polynomial}, or $A=\{p(T): p$ is a polynomial with $p(0)=0$}. Both definitions contain only the operators which are linear combinations of powers of $T$. In the first definition, we allow for a power of $0$, so that the identity operator belongs to the algebra. Is this the standard definition, or is it something else?
2026-04-08 21:27:11.1775683631
On
Definition of algebra generated by an operator.
220 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail At
2
There are 2 best solutions below
2
On
One one speaks about things generated by a set (linear subspaces, subgroups, subrings, and so on) there is usually a category in the background and then the generated object is supposed to be the smallest "acceptable" thing containing the set of generators.
If the background category is that of
rings, you would only take polynomials without constant term,
unital rings, then take all polynomials,
C$^*$-algebras, then take (non commutative) polynomials in $T$ and $T^*$, with or without constant term, and close it.
The algebra generated by $T$ is usually defined to be the intersection of all algebras containing $T$. If you work out this definition, you will find that the result is one of the algebras you mention, depending on whether you require algebras to have a unit.
More genreally, if $A$ is an algebra and $S \subseteq A$ is a subset, then the subalgebra generated by $S$ is the intersection of all subalgebras $B \subseteq A$ containing $S$. Again, you have to specify whether or not you require all algebras to be unital. (On this point, different authors use different conventions.)