From: Marc Geddes (email@example.com)
Date: Sat Aug 13 2005 - 02:19:10 MDT
--- Ben Goertzel <firstname.lastname@example.org> wrote:
> I do understand your basic point, I think.
> You're saying that anything in objective reality
> could be contained within
> the subjective reality of a hypothetical
> sufficiently powerful mind.
> And you're saying that anything contained in a mind
> is constructed according
> to certain cognitive principles, including
> principles of deductive and
> inductive reasoning.
> This second point is a bit confusing, because
> induction and deduction are
> procedures of inference not methods of
> *representation*. However, one can
> say that inductive and deductive reasoning are the
> processes a mind would
> use to figure out a compact representation of some
> entity. In that sense
> one can perhaps associate an entity with a series of
> inductive and deductive
> reasoning steps. Is this what you mean, or am I
> missing something?
That's (roughly) what I meant yes.
> give me an example of how you intend inductive and
> deductive processes to be
> used for knowledge representation?
> Next, you're saying that chains of
> deductive/inductive reasoning can be
> meaningfully assessed via various valuation methods.
Um... yes. I'm saying that valuation (value
judgements) are *implicit* in the very processes of
induction and deduction themselves.
> And you're saying that
> in order for a mind to trust in its own reasoning at
> all, it must accept the
> proposition in the former sentence -- if not, then
> its reasoning would be
> untrustworthy from its own point of view.
> So, ultimately, you're arguing that if any mind
> trusts its own reasoning
> processes, then this mind can assess the "value" of
> an entity in objective
> reality via assessing the value of the chain of
> cognitive processes it uses
> to reconstruct this entity within its own mind (or
> would use to do so, if it
> had the memory/processing power).
> I don't think this kind of argumentation deserves
> the label "Theorem", which
> should really be reserved for statements proved with
> mathematical rigor.
> But that is a stylistic point and if you don't agree
> with me, that's your
> right, obviously.
O.K. I was using 'theorem' in an informal sense to
simply mean 'reasoning step'.
> I don't really see what this has to do with ethics
> or morality, though.
> Indeed, as you say, I can try to assess the value of
> the cognitive processes
> I would use to reconstruct the Holocaust within my
> mind if I had more memory
> and processing power. But, I don't see how this is
> much deeper than just
> saying (for example) "Yes, there is a universal
> morality, because we can
> model the universe as a computer, so everything in
> the universe is made of
> bits, and we can define the value of X as the number
> of bits in X."
> -- Ben G
The value of X is not *the number of bits in X*. That
would be a very crude definition of value!
I'd say that the value of X would be *how close the
inductive and deductive procedures used to represent X
in ones mind came to 'optimal' induction and deduction
in the given domain of knowledge which is X is classed
Let me give an example. Suppose a super-intelligence
was trying to objectivity determine the aesthetic
value of a painting. To simplify the example I'll
consider only the aesthetic value of the *design* (i.e
the visual-spatial patterns). I'll also only look at
Assume the super-intelligence can form a complete
conscious representation of the painting (i.e the
design is captured in complete detail i.e. to the
limits of artistic resolution)
To recognize spatial patterns in the first place the
sentient has to carry out the Inductive procedure of
To apply my system, the sentient determines how close
the inductive procedures used to recognize all the
patterns making up the painting came to the 'optimally
efficient'. An 'optimal' procedure for inducting
spatial patterns is one that that would be 'optimally
efficient' in the sense of the most compact
representation. The 'objective aesthetic value' would
simply be this figure.
For example: A painting with spatial patterns
displaying a high level of 'symmetry' could be more
'optimally represented' because the inductive
procedures used to recognize the patterns making up
the painting enable the sentient to compress the data
more readily (i.e Patterns with high levels of
Symmetry are 'good' for induction).
In fact 'Symmetry' is needed in *some* degree for any
induction to occur at all. The degree to which
perceptual data about spatial patterns can compressed
depends on the amount of symmetry. So one could
conjecture that 'Symmetry' is an aesthetic property of
spatial patterns which is 'objectively good'.
i.e The properties of spatial patterns are 'good'
which enable cognitive procedures to recognize spatial
patterns in the first place.
One could run through all the knowledge domains
looking for this tautological sounding property and
thus a way to objectivity assign value could be
i.e Property X in domain Y is 'objectivity good' when
it is a neccessery condition the cognitive recognition
and represention procedures require to carry out
recognition and representation in the first place.
Example: Property 'Symmetry' in domain 'Spatial
Patterns' is objectivity good because *some* degree of
symmetry is required for any induction (data
compression) to occur at all.
For instance, check out my table of :
'Universal Utilities' that I posted on the SL4 Wiki.
I ran through my 7 'core' knowledge domains
(Mind, Morality, Models, Meaning, Mentality, Measure
and Matter) and for each domain I searched for general
'values' which take the form explained above. I was
thus able to come up with what I believe to be 21
'objectivity good' values which form my Universal
The table is at the end of this web-page:
--- Please vist my website: http://www.riemannai.org Science, Sci-Fi and Philosophy --- THE BRAIN is wider than the sky, For, put them side by side, The one the other will include With ease, and you beside. -Emily Dickinson 'The brain is wider than the sky' http://www.bartleby.com/113/1126.html Send instant messages to your online friends http://au.messenger.yahoo.com
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:51 MDT