Simplify x^p^(1/p) to r?



Gary O wrote:
> Q: why doesn't this return x?
>    ratsimp(x^p^(1/p))
> 
> If maxima doesn't know that x^y^z => x^(y*z), can I teach it somehow?

But what is 2^3^4?  I think ^ is right associative, so 2^3^4 is not 
2^12, but 2^81.

Ray