> <snip ... (I am ignoring my own allergy to the idea that 1st person aspects
> can be faithfully represented by Turing algorithms.) ...>
I take the opportunity of that statement to insist on a key point which is admittedly not obvious.
The fact is that I am also totally allergic to the idea that 1st person aspects can be represented. Comma.
And that is the main reason I appreciate the computational hypothesis; it prevents the existence of any
such representation. This is a consequence of two things:
1) The first person possesses an unbreakable umbilical chord with truth (it is related to what is called "knowledge incorrigibility");
2) by Tarski theorem the concept of truth on a machine cannot be represented in any way in the machine.
In particular if we define (a little bit contra Rafe Champion(for-list) knowledge of p by [provability of p] + [truth of p] (more or less Theaetetus' definition),
although provability and knowledge can be shown to be "third person" equivalent, they can also be shown first person NOT equivalent.
I even think that the comp hyp, thanks to incompleteness, is the most powerful vaccine ever find against any attempt to reduce a first person to any third person notion, and this in a third person communicable way. This makes me optimistic for the long run ...