Re: Fuzzy vs Probability

From: Stephen Tattum (
Date: Mon Jan 17 2005 - 05:03:58 MST

Mike Wilson Wrote -

>There are two basic problems with philosophy. One is that it lacks
>a strong selective mechanism, and thus while 10% of it is highly
>useful the other 90% of it is useless junk.

There are two basic poroblems with saying this - the first and a large
part of the reason why 'philosophy' does get a bad rap - is that it
encompasses so much. I am not saying that any AGI projects need to take
into account any metaphysics, aesthetics or comparisons of milk and wine
- 90% of 'philosophy' is useless. Secondly philosophy has the same
selective mechanism as science, if a philosophical conjecture is
experimentally disproven then it will never again see the light of day,
its just that philosophical ideas may be discussed for a long time
before they can be experimentally disproven - scientific conjectures are
made expressly for the purpose of being experimentally tested - or
sometimes just mathematically tested, like in the case of string theory,
which a lot of people do view as philosophy, this doesn't make it
useless though. As for ideas that obviously can never be experimentally
tested, that is why discussion needs to take place, because its the only
tool we have for coming to any conclusions and scientific conjectures
are subject to this exact same process before they are or can be
experimentally tested.

>The second is
>that most philosophy takes the correctness of intuitive human notions
>about various aspects of cognition, as well as the basic validity of
>introspection, pretty much as a given.

You say this as though philosophy and science are two diametrically
opposed disciplines
I don't know what philosophy you have been reading but you don't seem
to have a very good idea of what its about - taking something as a given
is the domain of other disciplines - thats why they have particular
names and delve into particular areas of knowledge, philosophy as a
discipline doesn't really have a strict set area of knowledge that one
must subscribe to to be a philosopher and it certainly doesn't rest on
any assumptions - other than the one that we can communicate to come to
aggreement or disagreement about stuff - there are even those who say we
shouldn't bother because its useless -but never more useless than their
view though. Sorry I just realised you were talking about 'most'
philosophy - is that more or less vague a term than philosophy as a

>By contrast cognitive science
>has been slowly but steadily proceeding from trying to answer
>mysterious philosophical questions to explaining where the illusion
>of mystery comes from and why we ask these badly formed questions
>in the first place.

 I was talking about philosophy of mind, which is undeniably important,
even if we don't model an AGI exactly around human intelligence, and any
philosopher of mind worth reading operates on the edge of cognitive
science's discoveries. The trouble is that philosophy is on the scene
before science and I think that trained philosophers in the same
position as scientists are often better at knowing where the directions
of progress should be, what the right questions to ask are, and what
kind of experiments should be done - philosophers make great scientists
in this respect - they're usually just better at performing though
experiments than real ones.

>Much of AGI consists of untested theories that are supported at
>best by several steps of inference from experimental psychology,
>neuroanatomy or decision theory. More often, they are just
>plausible sounding guesses based on someone's intuitive ideas on
>cognition works, possibly aided by some compelling metaphors to
>intelligent seeming processes people are familiar with.

This is exactly why philosophers can provide many valuable insights
into what direction any AGI project should take - they can't be any
worse than the people already doing so - except that they don't have the
expertise in particular fields - but they are trained to understand
problems - and to find innovative ways of solving them. Sometimes its
better to think about a problem when you are not working entirely inside
that system, this is what philosophers can do for the AGI project.

>normative probability and decision theory (including Bayes) and
>the failings of the dualistic viewpoint (and other philosophical
>delusions inconsistent with the integrated causal model and the
>specific findings of evolutionary psychology) are not in this

I find the inclusion of a reference to the dualistic viewpoint typical
of your misunderstanding - dualism is the most criticised doctrine in
the history of philosophy and rightly so (the same goes for the John
Searle references in your story - except his Chinese room is only the
second most criticised) I don't know anyone that believes this and I
actually have some philosophers as friends. I'm not sure normative
probability and decision theory should be taken as facts though - I know
very little about either so I'll try and get back to you as to why I
think not - or that I think they should be taken as facts...

>The hostility to philosophers is quite frankly because as a rule
>philosophers have proven to have a vastly inflated expectation
>of how relevant their work is to reality, in particular how the
>human brain works and how to build AGI.

This is exactly what I'm talking about - unneccessary hostility - and
hostility that has I'm sure been aimed at anyone else who so arrogantly
believes themselves to be unquestionably right, or presents
unquestionable facts. So calling yourself a philosopher automatically
conjures up peoples expectations of this 90% (which I think isn't
neccessarily accurate but won't debate it just now) of philosophers who
tend to be idiots. End the discrimination - jettison your

>Numerous philosophers have
>nonetheless made important contributions, but all those that have
>done so took the time to intensively study relevant science and
>maths before speculating. If you don't do this your fate will be
>probably resemble that of the philosopher in this story;

What do you think interests me about SL4 if I haven't at least read
some of the relevant material? It is entirely unlikely that your
preconcieved 'philosopher' would accidentally stumble here but thats the
way your first reply treated me. I felt somewhat patronised that your
first reply mimicked my initial post, but from a memetic point of view I
feel honoured - does memetics have any place in a theory of AGI
everybody? What about cybernetics and systems thinking like Lovelock's
Gaia theory and non-linear mathematics? A great contemporary philosophy
site which may surprise you is Principia Cybernetica:-

Steve Tattum



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:50 MDT