From: Ben Goertzel (ben@goertzel.org)
Date: Thu Aug 11 2005 - 23:23:38 MDT
> Theorem 7: A sentient cannot consistently make the
> value judgement that 'the cognitive mechanisms behind
> value judgements are bad', since the sentient would be
> contradicting itself - he needs to believe in the
> validity of those very cognitive mechanisms to make
> any value judgements in the first place.
But, a sentient could consistently conclude
X = "the only value of my cognitive mechanisms is their ability to correctly
conclude X"
So your chain of deductions becomes basically trivial at this point, due to
this version of Socrates maxim
"The only thing I know is that I know nothing"
;-)
Or, less amusingly but more plausibly, it could be the case that all
sentients are necessarily inconsistent because consistent sentience is not
computationally feasible...
Anyway, this argument you've presented is hardly an airtight sequence of
deductions, and certainly doesn't deserve the mathematical label "Theorem"
throughout!!
And even if these loopholes were closed somehow, all you would have shown is
that any sentient is able to judge whether something in its mind actually
exists (i.e. was hypothesized via well-functioning cognitive methods). This
leads to a universal "morality" which distinguishes existence from
nonexistence instead of "good from bad" in any meaningful sense....
Bleh....
I think Eliezer's "Collective Volition" is an ultimately unworkable and
ill-defined idea, but it's sure a lot more interesting and carefully thought
out than this series of non-Theorems you're presenting!!
-- Ben G
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:51 MDT