Entering the Observer

In reply to Pedro C. Mariju�n who wrote on May 24:

>1. ENTERING THE OBSERVER. In most branches of science the observer is left
>out the picture (eg, mechanics, Shanon's communication theor., economics)
>notwithstanding the curious, and very different, idealized properties
>he/she may be endowed. But in info science the observer intrudes into the
>picture quite crudely (e.g., meaning)... and so the mess. My contention is
>that if he/she enters with the usual, idealized properties, it means the
>end of the interesting business �the different reductionist approaches to
>info finally win the day. But if the observer is admitted within sort of a
>'limited prehension' framework (superseding the 'principle of rationality'
>so dear for physicists--Anderson, Gell-Mann), a few curious consequences
>may follow. I have briefly explored this issue regarding the info dynamics
>of the sciences and the emergence of interdisciplinary overlappings,
>somehow it becomes an expansion on what is contemporarily called in AI & AL
>'swarm intelligence', or a new version of Ortega y Gasset's
>perspectivism... I was reminded the theme by some recent comments (Gyuri
>and Juan) about info and the limitations of the observer in physics itself,
>and by Koichiro's comments on book-keeping. But apart from us, limited
>human observers, at least another couple of paradigmatic observers have to
>enter into the info science picture: cells and firms... Economist
>Schumpeter 'creative destruction' in economic systems becomes quite an
>interesting piece of reflection about the structural peculiarities inherent
>in these informational observers.

Dear Pedro
Your theme "entering the observer" addresses several interesting
items, among them the horizon of observers and the risk of
reductionism that the observer introduces in info sciences. Let me
try to join these two points vs the degree of complexity addressed.
I agree with you that the horizon of observers is wide: from cells
to firms, including insects (swarm intelligence) and humans. And
perhaps the spectrum could be widened even more with your other
theme "molecular recognition" (assuming we can consider
molecule as an observer. Would you ?).
Now, regarding the risk of reductionist approaches to info entering
the game, I also agree with you about the reality of that risk,
especially when pointing at the notion of meaning. But with so wide
an horizon of covered items, I feel the risk can be managed. And
one can estimate that there is place for several different models of
observers, so avoiding being taxed of reductionism. The question
will then be about the coverages of the proposed models and about
the completeness of used hypothesis.
For cells or insects, I believe we can look at them as
"living machines" and find some ways to modelize them as
observers with little risk of reductionism.
When we reach the level humans (sorry for the gap...), then
comes for sure a risk of reductionism: modelization of human
as observer is not possible today. I mean globally. Our
ignorance about the nature of human brings us to address
only limited and very partial possible models. This because we
do not understand the nature of reflected consciousness, of
emotions, of free will, and so. Philosophy, neurosciences,
psychology, cognisciences and others are still turning around
these subjects, looking for a way in. The "hard problem"...
But coming back to the level of insects, a simple modelization can
do it. Like the one proposed for a "meaning generator system"
(Information and Meaning. FIS 2002). For instance, pheromone
smell can be looked at as a signal creating a meaningful
information in an ant. The meaningful information being
"pheromone smell has some positive link with vital constraints to
be satisfied, so trail is to be followed".
Now, observer/meaning at the level of human or firms address
degrees of complexity that have little to do with insects. Other
models are needed, � priori. But it is interesting (and risky
reductionism wise) to investigate a possible usage of previous
model by transfering the complexity increase to the constraints
of the meaning generator system. More precisely, by trying to
imagine how some new constraints to be satisfied came up during
the course of evolution. This approach is to be looked as an
hypothesis to be validated (not validated as of today).
And there may come in also some form of the Schumpeter
"creative destruction" you propose to introduce. A constraint
could be satisfied only at the expense of a new (cheaper)
constraint, coming in as more acceptable for the system
(group survival vs individual survival, pleasure/reality and
Freudian repressions, ...).

Many are the Foundations of Information Science......

Regards

Christophe Menant
Received on Wed May 29 16:15:16 2002