2
$\begingroup$

Consider the algebra $ T $ of trigonometric polynomials (i.e. functions like $ \sum_{n=-N}^{N}c_{n}e^{inz} $ where $ c_n \in \mathbb{C} $) on the closed disc $ D $ of radius $ \frac{1}{3} $. $ T $ is self-adjoint, it separates points in the disc, and it vanishes nowhere on the disc. Thus, by Stone-Weierstrass theorem, any continuous function on $ D $ can be approximated uniformly by a trigonometric polynomial on $ D $. This contradicts with the fact that some continuous functions on $ D $ are not holomorphic. Could you help me find what I did wrong? Thanks very much

$\endgroup$
1
  • 4
    $\begingroup$ Why is $T$ self-adjoint? $\endgroup$
    – Jose27
    Feb 18, 2015 at 6:41

1 Answer 1

2
$\begingroup$

To expand on @Jose27's comment:

$\overline{e^{inz}}$ is not $e^{-inz}$, since $z$ is not assumed to be real.

In fact with $z=x+iy$, $$ e^{inz} = e^{in(x+iy)} = e^{inx}e^{-ny} $$ so $$ \overline{e^{inz}} = e^{-inx}e^{-ny} $$ but $$ e^{-inz} = e^{-in(x+iy)} = e^{-inx}e^{ny}. $$ In other words, $\overline{e^{inz}}$ is not a member of $T$ (it's not holomorphic when $n\neq 0$).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.