RE: [Fis] Molecular-Experimental sciences

RE: [Fis] Molecular-Experimental sciences

From: Kevin Kirby <[email protected]>
Date: Mon 28 Nov 2005 - 11:18:37 CET

Dear FIS discussants,

Ana's posting on chemical informatics and XML is very relevant. Many of us
remember her very clear FIS talk last summer in Paris. What is particularly
relevant to our recent online discussions is the layer of logic and
representation underlying the use of XML technologies in the natural
sciences and elsewhere: the role of Semantic Web description and logic
languages such as RDF and OWL. (This is a wonderful echo of the artificial
languages and ontologies of the 17th century, e.g. Dalgarno and Wilkins.)
If, as many of our group have voiced, there is something insufficient, or
off-target, in contemporary relatives of first-order logic in the domain of
life science (OWL can be viewed as a fragment of first-order logic), does
that mean that these schemes can/should be "upgraded" to something that
could be more valuable to the chemists and biologists who use them? To a
logic or formalism that captures what is distinctive about biochemistry,
instead of treating this field uniformly with the same tools one would use,
say, in financial informatics? Is this what we have been talking about?
Ana's raising of this issue gives us the chance to make this precise I think.

And, if this is the place for it, I would very much like to hear a few more
sentences on the application of category theory here, if indeed this is
where it would be deployed. It is such nice mathematics, simultaneously
concrete (arrows and diagrams) and fabulously abstract (capturing the key
idea of universal construction). In my filing cabinet I have an old brown
copy of Robert Rosen's 1958 paper from the Bulletin of Mathematical
Biophysics, casting the theory of "Metabolic-Repair" systems into the
language of category theory! (As a student in the 1980s I reported on
this paper in one of Michael Conrad's classes.) So for nearly half a
century many have found the urge to employ category theory to capture what
is unique in living systems. So, at what level can we use this kind of
thing? Are molecular bionetworks captured by commutative diagrams, or is
this too literal an interpretation? And could they play a role in the
knowledge representation and inference systems of bioinformatics?

Praxis aside, there is a deep philosophical issue here of the interface
between the formal and the material. Would mapping complex "molecular
bionetworks" onto formal graphical diagrammatic structures lose something,
well, "vital"? Surely their essence is not purely formal? Now, in the
emerging ambient philosophy, the formal is deprecated, the embodied is
valorized. There is new respect for the physical in natural computing (and
it is the properties of biological matter that I believe give it its
computing power, not any formal algorithms). Is there something inescapably
material, essentially non-formal, in biological networks, that requires a
new formalism-- and why isn't such a statement the oxymoron it appears to be?

Take the recent conversation on vagueness with Stan and Loet. My first
reaction to reading sentences like "I am not as vague as I was when I was a
blastodisc…" was to suspect a category mistake: certainly it is expressions
(e.g. utterances) that can be vague or not vague, not things (e.g.
organisms). But my first reaction is an old prejudice, a mindset that is
less useful, based on this form / substance distinction-- that vagueness is
something linguistic. But just as the vagueness of an utterance is
something only meaningful as it emerges into a context, into a situation in
the world, so too, I wonder: could we take vagueness as example of
something that hovers on the edge of the logical and the material?

This is the crux of it. To take a phrase from Gordana's paper (in Mind and
Machines 2003), we are dealing with Theory Materialized.

-- Kevin

__________________________________
Kevin G. Kirby
Department of Computer Science
Northern Kentucky University, Highland Heights KY 41099
<mailto:kirby@nku.edu>kirby@nku.edu (859) 572-6544 ST 340
Received on Mon Nov 28 11:10:19 2005


This archive was generated by hypermail 2.1.8 on Mon 28 Nov 2005 - 11:10:20 CET