[FOM] The Lucas-Penrose Thesis

Apostolos Syropoulos asyropoulos at gmail.com
Mon Oct 9 14:00:14 EDT 2006


2006/10/8, Eray Ozkural <examachine at gmail.com>:
> On 10/7/06, Apostolos Syropoulos <asyropoulos at gmail.com> wrote:
> > 2006/10/6, Eray Ozkural <examachine at gmail.com>:
> > > To know its programming is exactly to have access to this
> > > sequence of symbols. It implies nothing else. However, this is
> >
> > To have access to its symbols does not imply that a program does understand
> > the meaning we have assigned to them.
>
> Some quick comments.
>
> I think there is a high probability that you are contradicting with the
> usual notion of "semantics" that is employed in formal programming
> language design. I have snipped the other part, but I think it ought
> to be fairly obvious that understanding a language does not mean
> reading the mind of a person who utters nonsense.

Exactly. But the point is that a program cannot understand anything
by itself. It is the machine which executes the program that may understand
the meaning of words. By your example one may say that a sequence of
meaningful English or Greek sentences are able to understand another
sequence of such words. So programs by themselves do not understand anything.

> On the other hand, one can conceive of analyzers that will correct
> simple errors (i.e. fault tolerant). This issue has also been pursued
> theoretically (CAs, graph automata, etc.), but I think ultimately

Again you are making the same mistake: you are talking about
machines not about programs per se.

> Also, human programmers are in no way "oracles". Such machines
> have been clearly depicted as "hypothetical" by Turing himself who

And who say so? Copeland who has studied extensively Turing's writings
would really disagree with such a statement. In addition, I did not say that
humans are oracles, but that have the role of oracles in the debugging process.

A.S.
-- 
Apostolos Syropoulos
366, 28th October Str.
GR-671 00 Xanthi, GREECE
http://obelix.ee.duth.gr/~apostolo


More information about the FOM mailing list