(Lax, Functional Analysis) Why does he say this space is poor in isometries when there continuum many?

256 Views Asked by At

In chapter 5.3 of the book Functional Analysis by Peter D. Lax, the author says

It is a fact of life that some Banach spaces are very rich in isometries; others are very poor. (...) Among the poor ones are the function spaces with the max norm. Here is an example due to Schur:

He immediatly tells us to consider the space of sequences of complex numbers that converge to $0$ with the max norm $||(x_n)|| = max |x_n|$. Then he tells us that for any sequence $(b_n)$ of complex numbers of absolute value $1$ that the operator given by $(Ux)_n = b_n x_n$ is an isometry. And that an operator given by a permutation of the terms of the sequences is an isometry. He finally says that compositions of these kinds of operators are the only ones.

This is a continuum-sized set of isometries. Why then does the author say that this space is very poor in isometries?

2

There are 2 best solutions below

4
On BEST ANSWER

The gold standard for a "rich" collection of isometries is $\ell^2$. One way of quantifying precisely what is meant by "rich" here is, for example, that the isometry group of $\ell^2$ acts transitively on its unit sphere. On the other hand this is very far from true of the isometry group of $\ell^{\infty}$ which has many more orbits acting on its unit sphere.

When it comes to spaces this large just considering the cardinality is not a good way to measure how "rich" the isometry group is.

1
On

Notice that for any Banach space adding a vector to everything is an isometry, so "continuum of isometries" is not worth much.