RE: Reply to Barham, Holgate, Mariju�n, Collier - info in the wings

Thanks Juan,

Yes I'll try out your food experiment on my retriever this weekend.

My guess is you're right - she's a tad smarter than Maturana's frog
or your average amoeba.

<I really would like to see an organized discussion of the concept of purpose
<or intentionality (I suppose they are the same).

I think we agree if there is no purposive interaction there is probably not an
informational situation. But let's not wander down the semantic track
for the momement. I am happy to suspend the word 'purpose' and bracket
'consciousness' for the life of this conference. Quotidian informational
behaviour is notoriously serendipitous and unconscious.

But how far back down the evolutionary process do we have to go before we stop
speaking of decisions and talk about automatic mapping?

<I think that only a neural information system can make decisions. Do you
<agree?

I agree with the 'neural system' bit but if we are defining information in
terms of decision-making we need to leave 'information' out of the equation -
is it necessarily a 'system'? That is what we are trying to discover.

The most interesting point in your paper for me was:

'Given a complex system, structural order alone does not represent
information-information appears only when structural order leads
to specific change elsewhere.'

That is what agency in an 'informational experience' is all about.

John Collier made the nice analogy :

'the information about spots in the butterfly genes is there, but
not much use to the butterfly -- it has to get this information to
the wings for it to be useful'

i.e. the experience of 'having spots' only becomes informational
for the butterfly when the genetic 'spots' (structural order) lead
to a specific change in the wings. The physical spots ('info-mation')
become 'in-for-mation' - data in context for an observing system
(say, for a predatory creature)in an environmental interaction.

In the same way this email (structure) does not become a message
for somebody unless there is an interactive change in the recipient's
mind etc beyond the 'limbic' reaction of hitting the delete key.

I think informational agency possibly operates across the spectrum of
interactions from protein binding to courtship behaviour and database
searching - but that needs cross-disciplinary research to justify.

We should continue to compare notes.

Cheers,

John H

-----Original Message-----
From: Juan G Roederer [mailto:jgr@gi.alaska.edu]
Sent: Thursday, 20 June 2002 16:27
To: Multiple recipients of list FIS
Subject: Reply to Barham, Holgate, Mariju�n

Dear FISers:

This is in reply to some comments by James Barham, John Holgate (both 19
June) and Pedro Mariju�n (June 5).

James wrote:

>>I think you go a little too far when you say (in paragraph 3) that "there
is no energy coupling between sender and receiver" in a true information-use
scenario. Rather, I would say, the informational signal must be very low in
energy in relation to the energies involved in the stability of the sender
and receiver.>>

""Well, maybe my use of "coupling" needs a little clarification, or perhaps
a better term should be used. This issue is crucial, however, because I find
the role of energy (its fluxes or balance) to be a fundamental discriminator
between force-field driven and information-based interactions (in addition
to the distinct roles of initial conditions and continuity). The key
question is: Where does the energy involved in the interaction come from? In
a purely physical interaction between two bodies, that energy is provided
by, or through, the interaction process itself, whereas in an
information-driven interaction the energy needed for the interaction must be
delivered (or absorbed) locally from separate sources at both ends and in
the "transmission line". In the first case, the unique correspondence
between "pattern" at the source (e.g., in my example of the asteroid, the
details of its odd shape) and the "change" effected in the recipient body
(the satellite's orbit) is mediated by the peculiarities of energy coupling
"one-by-one" (i.e., by the details of the gravitational potential); in other
words, the energy itself is the controlling factor. In the second case, the
correspondence between pattern (in my example of the dog, the obstacle) and
the change (the dog changing its course) is controlled by information, not
energy.""

>>(2) This is really a minor quibble---and, as I say, I think your examples
are great---but without a little elaboration, someone might take away the
impression that true information use is tied intrinsically to brains, since
you use the dog as your example of the transition from force-type
interactions to information-type interactions. I believe that the
information-type interaction is an essential feature of life as such, and
that the bacterium is interacting with its environment (and indeed its
internal parts are interacting with each other) according to essentially the
same principles that the dog uses in circling around the obstacle. I assume
that you will agree with this (no?)>>

""Indeed, I do. To clarify this, I have now added in the text of my paper
the example of a paramecium or an amoeba guided by a salinity gradient in
its movement. (And I'd love to also find a biomolecular example!)""

>>(3) Finally, you write (also in paragraph 3) "Note that this class of
interactions must evolve; they cannot arise spontaneously in the abiotic
world.". First, if we are naturalists, we must assume that such interactions
DID evolve spontaneously in the abiotic world, even if we have no idea at
present how this was possible.>>

""Why? Couldn't there have been a sort of transition period involving
prebiotic processes of replication, etc. of macromolecules, that later
started organizing their environment in ways that depended less and less on
energy-driven processes and more and more on patterns (trigger-dependent
processes)? Maybe I should say in the sentence you quoted:"...they cannot
appear spontaneously in their final form in the abiotic world""

John wrote:

>>2. Your remarks about 'purpose' bring us back to the problem of Maturana's
frog. The frog's interaction with the environment by trapping an insect with
its tongue is hard-wired - not a deliberate act based on spatial perception.
There is no intentionality - we simply ascribe it to the interaction. To
assume purpose here is nursery teleology>>

""True, the neural system involved is mostly hard-wired, but why do you say
that it is 'not a deliberate act based on spatial perception'? The frog's
action is indeed more "automatic" (robotic) than us eating ice cream, but I
would say that the purpose here is also hard-wired: a genetic command to eat
whenever the opportunity arises! (This is what also sits in our own limbic
system, see section 6 of my FIS paper, but it can be overruled by other
real-time input).

I really would like to see an organized discussion of the concept of purpose
or intentionality (I suppose they are the same). Anyone out there who has a
good definition for "purpose of a given action"? I think it's not enough to
say "the end-state of a system which without that action could never be
achieved" (as I think I did in my paper), because that really also includes
all physical interactions between inanimate systems which don't have a
"purpose" in the commonly accepted sense.""

>>IMO my dog does not make decisions (although I agree with Fred Dretske
that it can 'know' where it has left its ball)>>

""I propose a couple of experiments with Imo. Place him before three bowls
with food. In the first experiment, put the same food and same amount in
each one. In the second experiment, put different amounts of the same food;
in the third, put three different kinds of food. I would predict that in
each case, Imo will definitely make a decision. In the first case, it will
be most probably random. In the second it will be "limbic" or hard-wired
(going to the bowl with the largest amount, if he was hungry); in the third
experiment the choice will be cortical, i.e., going to the bowl with the
food he likes best according to past experience.

This has brought me to Pedro's (and Peter's) references to decision-making.
I think that only a neural information system can make decisions. Do you
agree? Here again I'd like to see a good definition of "decision-making".
Decision is an action that obviously appears when there are alternatives. I
see decisions that are purely random (see Imo's first experiment); decisions
that are based on the distant past (inherited choice, Imo's second case);
decisions that are based on the ontological past (based on experience or
learned information); and, only in humans, decisions based on the future
(based on long-term predictions-see section 7 of my FIS paper).""

Sorry for the long reply!

Juan

*****************************************
* *
* Juan G. Roederer *
* Geophysical Institute *
* University of Alaska-Fairbanks *
* 903 Koyukuk Drive *
* Fairbanks AK 99775-7320, USA *
* *
* Phone: +1 (907) 474-5099 *
* Home: +1 (907) 479-6834 *
* Fax: +1 (907) 474-7290 *
* e-mail: JGR@gi.alaska.edu *
* *
*****************************************