[time 785] Re: [time 784] Falsifiability


Stephen P. King (stephenk1@home.com)
Sat, 18 Sep 1999 15:46:22 -0400


Dear Hitoshi,

        Here is a quote from my last post to Lance as a tentative answer to
your question:

**

[SPK]
> > I
> > understand well the problem that the concept of artificial intelligence
> > has with most, and I find myself in agreement with Searle and Penrose on
> > most issues, but if you could understand the work of Peter Wegner, which
> > is a carefully worked out application of the Chu space concept, you will
> > find that it works!
[LF]
> "Works" is a strange term. What does it work for? What work does it do?
[SPK]
        I am using the word "work" in terms of pragmatic utility! Can
the model
make testable predictions? Can the model allow us to have an intuitive
understanding of the various problems that are addressed?

[SPK]
> > If a model works, that is all that matters. We must not
> > forget that all
> > we can apprehend is merely a finite model of the Universe in itself.
[LF]
> How do you know this? Is this a "falsifiable" assertion? And what does it
> mean?
[SPK]
        Well, I should be able to apply my requirement to my own
assertions! I
have only time to give a sketch argument. Have you looked at the
Bremermann's Limit link? http://pespmc1.vub.ac.be/ASC/Bremer_limit.html

Robert Fung has a discussion of it at
http://www.bestweb.net/~ca314159/BREMMANN.HTM.

        The idea is that it is impossible to encode an arbitrary amount
of
information given a finite amount of mass/energy. If we consider that
the "writing down" process of involved in constructing a model as a
describable as a process itself, we can compare the amount of
matter/energy needed for each.
        What is a model? It is a representation that allow some system
to infer
information about the "original" and communicate this to others so that
multiple observers can participate and a consensus can be reached. The
Time List is an example of this! :-) The model of a process of
constructing a model is more complex and thus requires more matter that
the original model. (Sorry that I do not have the specific proof of this
assertion here!)
        If we were to assume that finite entities, such as ourselves,
are
composed of matter and are able to comprehend more than a finite model
of the Universe, which is by definition infinite (since it must include
all descriptions, e.g. models, of itself) since it is "the totality of
existence", we will generate a contradiction!
        If we equate the act of comprehension with the ability to
construct a
representation that allows for a meaningful set of inferences about the
properties of the original, then it is obvious that the Universe itself
cannot be modeled in a complete sense by a finite entity since that
would imply (by a diagonalization argument) that the model of the whole
Universe is encodable in a finite amount of matter/energy and this is in
contradiction with the falsifiable Bremermann's limit!
**

Hitoshi Kitada wrote:
>
> Dear Lance and Stephen,
>
> The word "falsifiable" is a difficult one for me. I tried to find it in my
> on-line English-Japanese dictionary, but there is no such word. Instead I
> found the word "falsify" that means: "A falsifies B" is "A proves that B is
> wrong." From this I imagine "falsifiable" means: "A theory T is falsifiable"
> is "There is a possibility that someone is able to prove that the theory T is
> wrong." Is this interpretation correct? Or are there any other and/or deeper
> meanings in the word "falsifiable."

        The "A" in your staement to me implies a physical experiment, the "B"
implies a particular measurement that has a restricted interpretation.
 
> Also I am not sure what is thought as falsifiable in this correspondences
> between you. I.e., I am not sure what is the issue of this arguments. For me,
> the objects that each of you address the word "falsifiable" look different
> between you. Could you explain in the easier and/or more understandable
> words?

        We will need to discuss this further!
 
> Best wishes,
> Hitoshi

later,

Stephen
 
> ----- Original Message -----
> From: Lancelot Fletcher <lance.fletcher@freelance-academy.org>
> To: Stephen P. King <stephenk1@home.com>
> Cc: <time@kitada.com>; Paul Hanna <phanna@ghs.org>
> Sent: Saturday, September 18, 1999 7:17 AM
> Subject: [time 781] RE: [time 779] Re: [time 778] Re: Noumenon and Phenomenon
> => Mind-Body problem.
>
> > Dear Stephen,
> >
> > > Here I go again! ;-) We are getting back to the problem
> > > that Vaughan
> > > Pratt addresses in his paper ratmech.ps
> > > (http://boole.stanford.edu/chuguide.html#ratmech) The relation between
> > > noumenon and phenomenon is reflected in the mind-body problem to which a
> > > solution is addressed. I am unhappy that the paper is very difficult to
> > > comprehend as it is written in the language of theoretical computer
> > > science, but I am working hard to translate it over for our use.
> > > (BTW, I
> > > understand Lance's objections, but we do need some falsifiability in our
> > > models of the world!)
> >
> > I hope you won't think I am being unfair, but the claim that a model or
> > theory is falsifiable is a strong claim. Indeed, it might be a falsifiable
> > claim with respect to Pratt's paper. Therefore, as proof, I request that
> > you cite at least one falsifiable proposition contained in Pratt's paper
> > or, if you are unable to do so, explain how it makes sense to describe
> > Pratt's theory as falsifiable.
> >
> > Lance Fletcher, President
> > The Free Lance Academy Foundation
> > http://www.freelance-academy.org
> > lance@interactive.net
> >



This archive was generated by hypermail 2.0b3 on Sat Oct 16 1999 - 00:36:41 JST