>>65
If you accept that pi is transcendental, it isn't hard at all. The radicals are a subset of the algebraic numbers, which are (by definition) an algebraically closed field. If there was a polynomial with algebraic number coefficients which had pi as a root, then there are two possibilities:
(1) Pi is an algebraic number; contradiction, as pi is transcendental
(2) The algebraic numbers have a nontrivial algebraic extension; also a contradiction, as they are algebraically closed.
Of course, the hard part is the proof that pi is transcendental; if I recall correctly the idea is to show that i*pi (and therefore pi itself) must be transcendental because e is transcendental and e^(i*pi) = -1 is rational.