Re: [Fis] Entropy, Information and Quantum Coherence

Dear Koichiro and colleagues,

On Jun 4, 2004, at 5:10 AM, Koichiro Matsuno wrote:

>
>    Information basically assumes the contrast between before and after
> the act of measurement or experiencing.

I have snipped this first line from a longer and interesting post for
comment. A sometimes explicit, but often ambiguously implicit,
difficulty in discussing the physics of information has to do with the
definition of information itself. I would like to suggest a hierarchy
of information concepts, and labels for my categories (I don't really
want to claim these categories as MINE, but I also don't want to impose
them on others without their approval), that might improve the clarity
of our future comments.

As I see it, Shannon's familiar equation, H = - Sum p log(p), is a
measure of POTENTIAL INFORMATION content. The actual information
encoded by the observed data depends on the configuration of the parts,
which is not represented in Shannon's formula.

The ACTUAL INFORMATION encoded by the observed data exists in the form
of patterns in the data. The extent of structure in the data is always
<= H.

The PERCEIVED INFORMATION, which is the form of information described
above by Koichiro, results from an interaction between the state of the
observer and the observed data. I think that this is an interesting
and important kind of information, but its inherent contingency on the
state of the observer makes it a particularly difficult (if not
entirely intractable) platform for the development of a general theory
of information physics. Note that neither POTENTIAL INFORMATION nor
ACTUAL INFORMATION embody such a contingency, which is the reason I
turn to them in the context of discussions about the relationship
between information and entropy.

I am not claiming that PERCEIVED INFORMATION is unreal; rather I see it
as pattern imposed on the observer by the act of observation. In the
context of human communication, as with this post, the goal of clear
communication might be to recreate the geometry of a neural path in the
brain of the sender, which encodes an idea, in the brain of the
receiver. There is a great deal of information re-encoding along the
way (e.g., my writing these words, which become digitally encoded by my
computer, which are decoded by your computer, and re-encoded into your
neural paths), which makes communication error-prone; but I think that
by far the most sensitive junction in human communication is the
process of re-encoding by the receiver due to the strong dependence of
the outcome of this process on the state of the receiver (e.g., as a
result of past experience). In principle, I see no reason why we can't
develop objective measures of PERCEIVED INFORMATION, even for abiotic
objects. A rock rolling down hill can be marked by the event (e.g.,
it can be re-shaped), and these induced patterns would be the
information perceived by the rock. A good geologist may be able to
infer this information by studying the rock, but he/she would never be
able to learn everything about the hill or the rolling event because
the PERCEIVED INFORMATION is extremely unlikely to faithfully record
all of the ACTUAL INFORMATION encountered by the rock as it moved over
the surface of the hill during its roll.

I am very interested in criticisms of this framework, and if it is
palatable to most, then I hope use of the terms POTENTIAL, ACTUAL, and
PERCEIVED INFORMATION will help to clarify our future posts.

Regards,

Guy Hoelzer

Department of Biology
University of Nevada Reno
Reno, NV 89557

Phone: 775-784-4860
Fax: 775-784-1302

_______________________________________________
fis mailing list
fis@listas.unizar.es
http://webmail.unizar.es/mailman/listinfo/fis
Received on Fri Jun 4 19:59:49 2004