Break before next stage

From: John Collier <[email protected]>
Date: Sat 27 Jul 2002 - 20:09:05 CEST

Folks,

We are now going to take an official break until Monday, September 2.
Advice will follow then about the next set of papers and moderators.

I want to thank everyone for making my job as moderator so easy.

I will just review the recent exchanges, without going into detail.

First, I asked some questions that were designed to get at whether or not a
purely computational model is sufficient for information science. I take it
that the overall response was "no", though this should not preclude
rigorous formalization in some weaker sense than Turing effectiveness.
Since the information processing model is restricted to the same set of
cases as Turing effectiveness, I think it is fair to say that information
theory is not to founded on an information processing model alone. Taking
this step has costs and benefits. The main cost is that information theory
may be undecidable, so we can't always tell, even with our best theory,
whether we have a case of information or not. On the other hand, this
openness allows information theory to apply to systems that are not
mechanical or algorithmic in the strictest sense of being Truing computable.

This raises the issue of what we need in addition to information
processing. A number of participants have held that meaning is essential to
information, though the details are quite different. I therefore raised the
question of whether the foundations of information science are in semiotics
(or perhaps better, semiology), the theory of signs. This led to a focus on
abduction as central to information theory. Now the relation between
information science and semiotics could be of several kinds: 1) semiotics
could be more basic than information, 2) information could be more basic
than semiotics, 3) the two could go hand in hand, neither more fundamental
than the other, or 4) the two could be independent but bear on certain
aspects of the other. I don't see that which of 1-4 holds has been
resolved, or even much addressed as yet.

Andcution is clearly not a purely formal notion, so if it is essential to
information science, or even required for a large part of information
science, then it seems that there are further limits on the possibility of
rigorous formalization of information science. In this case our ideal
information theory may be not just undecidable, but irreducibly incomplete.
If this is so, then no single formalization can capture all of information
science. Again, there are costs and benefits to making this move.

Pedro gave us an example of how abduction might be introduced at the
molecular level by making a micro-meso-macro distinction. This is inherent
in the information systems described in my own paper as required for
information expression, so personally I agree with this, and have used the
very words in other papers. Edwina objected to this categorization on the
grounds that purpose cannot be constrained to the macro. I think we have a
misunderstanding here. The terms macro, micro and meso are, as I understand
Pedro's use, relative to what Stan Salthe calls the "focal level" of a
system, and not absolute. This is consistent with the usage of Ingarden et
al in Information Dynamics and Open Systems. Edwina's objection seems to me
to assume an absolute standard of macroscopicity, etc. I don't think this
is a good restriction, and I am unclear what it would mean in any case.

We are still some way away from having consensus on how information can get
meaning, but I think there is some agreement that although the relatively
micro might contain information (or something weaker -- protoinformation,
or potential information), and that the meso level is where the causal
aspects of information are significant, a relatively macro level is
required for the sort of integration necessary to make the causal aspects
meaningful. I think that even if this is right, we are still some way from
giving sufficient conditions, though Christophe Menant's paper addressed
this issue from a similar perspective with his notion of "system
constraint". Althought this meaning of constraint is rather special, it is
more neutral than "abduction" and "meaning". I think it is best to try to
use neutral terms insofar as possible (but we also have to make sure that
we don't lose something important by using overly dry terminology). In any
case, I would like to see some effort towards relating the notion of
abduction, or perhaps meaning, to that of system constraint.

So with that personal preference stated, I will wish you all a good holiday.

Cheers,

John

----------
Dr John Collier john.collier@kla.univie.ac.at
Konrad Lorenz Institute for Evolution and Cognition Research
Adolf Lorenz Gasse 2 +432-242-32390-19
A-3422 Altenberg Austria Fax: 242-32390-4
http://www.kli.ac.at/research.html?personal/collier
Received on Sat Jul 27 20:09:56 2002

This archive was generated by hypermail 2.1.8 : Mon 07 Mar 2005 - 10:24:46 CET