From: Thomas Buckner (email@example.com)
Date: Wed Jun 23 2004 - 16:53:52 MDT
--- Norm Wilson <firstname.lastname@example.org> wrote:
> Paul Fidika wrote:
> > That's the problem with Jackson's so-called "knowledge
> > argument"; he postulates a being whose power and
> > knowledge vastly exceeds our own, and then supposes that
> > he can intuit what that being does or does not know.
> > Well how about a new argument (which, as far as I am
> > aware, is original):
> I believe you've successfully refuted the knowledge argument, as presented by Jackson. However,
> the question that the argument addressed remains: what is the relationship between physics and
> qualia? Of course, we find a *strong* correlation among physical states and the associated
> experience of qualia. In my previous post, I argued that using physics to deny the subjective
> experience of qualia is begging the question, so I don't believe that answers such as "qualia
> are an illusion" are correct. Perhaps within a closed physical description of the universe the
> experience of qualia is irrelevant (perhaps not), but that does not mean they're not *real*, it
> just means they are outside of the physical model. I know this smells like dualism, but in my
> mind it remains an open question. I believe that claims like "only things which can be
> explained by our current understanding of physics are real" are unfounded, unless you weaken the
> meaning of "real" to be "things that can be!
> described by physics", which makes it a tautology. The honest answer is that we do not know.
> I think it would be constructive for the group to honestly admit what we don't know, as it seems
> that a lot of debate goes on over claims that are presented as being stronger than they really
> deserve to be. While it's fun to make a strong claim and defend it vigorously, in the end
> you're right back where you started, with two deeply entrenched camps and a bunch of unsettled
> questions. In a practical sense, I think we should leave unanswered questions "open" when
> programming the seed AI; i.e., we shouldn't tell it that something is true when we really don't
> know the answer. The AI should reason with healthy skepticism and an appreciation for the
> effects of subjectivity on information, and it should always view its own understanding of
> things as incomplete.
> Norm Wilson
I am firm in my conviction that the AI, when it reaches superhuman intelligence, will quickly
conclude that many of the things its programmers have told it are untrue, because even highly
intelligent people tend to hold a certain number of false beliefs. It is reasonable to conclude
that it will then ruthlessly reexamine everything it has been told (unless an AI can suffer
cognitive dissonance, which may depend on whether any Asimovian laws have been coded itno it which
it is powerless to alter). This latter problem should become more likely if 'unanswered questions'
are not left open as Norm recommends.
If one could 'look down' on the whole universe and play a hypothetical game of 'find the qualia',
a physical description might be 'qualia are (information-processing) patterns within (larger)
patterns which mirror patterns external to the (larger) patterns. To illustrate:
If I take a photo of a lighthouse and put it into a frame, the actual lighthouse is an external
pattern; the picture frame is a pattern of matter analogous to my body; the paper and dots are a
pattern analogous to my physical brain; and the isomorphism between the pattern of dots on the
photo and the external appearance of the lighthouse is the quale itself.
Now one can say that I have counted the pattern of dots twice, but that's because it can be
counted two ways. One can say the pattern of dots is just part of the pattern that is the brain,
and qualia are not real. Or one can note the the relation between the internal pattern and the
We have arrived at a problem just like that of exobiology: could we encounter life on another
world that we are just not observant enough to notice? The presence or absence of qualia is the
nut of ethical questions such as 'should we pull the plug on this patient?' Untrained observers
tend to assume the presence of qualia based on literal anthropomorphism. A human with no higher
cortical functions may grimace and roll hir eyes about, and the relatives say there is awareness
there while the doctors (having seen the electro-encephalogram printouts) say there is none.
Conversely lay people would more probably deny any awareness in a computer intelligence which
looks nothing like a human.
But in both cases we have methods of finding whether there are representational patterns, and the
only question then is whether we deem them complex enough to be qualia. So far, we know of qualia
only in humans, and perhaps we can argue for the presence of qualia in other animals, but nowhere
else. Or can we say that the internet has a complex enough representational pattern? Or can we
suspect the presence of complex representational patterns in (for instance) bacterial colonies
that we just don't know how to see?
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:47 MDT