Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

ONT Re: Information = Comprehension x Extension -- Discussion




o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o

ICE.  Discussion Note 44

o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o

CvB = Christian von Baeyer
HT1 = Hugh Trenchard's initial post
JA  = Jon Awbrey's reply
HT2 = Hugh in response

Hx: ICE Discussion 22.  http://suo.ieee.org/ontology/msg05305.html
Hx: ICE Discussion 40.  http://suo.ieee.org/ontology/msg05330.html

JA: If I had to give names to the most distant horizons that I can see -- 
    from where I stand there appear to be several overlapping horizons --
    they would probably be terms like "semiotics" and "systems theory".

JA: Logic sits within semiotics as its "formal" branch, this being the word
    that Peirce used to connote its normative or "quasi-necessary" character,
    but explaining that will take another whole story to do.

JA: Systems theory, in the sense of mathematical systems theory, gives us the
    most general framework I can think of for describing change, or dynamics.

JA: One of the trickiest pieces of work that I have had to work out over
    the past fifteen years has been building an integrated framework for
    talking about both the dynamic and the symbolic aspects of what I've
    been calling "inquiry driven systems", in a single breath, as we say.

HT1: You may have seen from previous posts of mine that I am interested
     in identifying universal properties of information that may describe
     neural processes as well as they describe human socio-economic processes.

Yes, I think I am very copacetic with this way of doing things, at least, inasmuch
as I get the drift of your POV so far.  One of the lines of inquiry that I pursue in
my joint work with Susan Awbrey is concerned with "learning organizations" (LO's),
a la the work of Peter Senge et al., especially as seen from the standpoint of
"organizational behavior and development" (OBAD? -- I guess I'll have to work
on that one ...), and more especially as it affects the organizations that we
all know and love as our institutions of higher learning.  Vide these papers:

http://members.door.net/arisbe/menu/library/aboutcsp/awbrey/integrat.htm
http://WWW.SAGEPUB.CO.UK/JournalIssueAbstract.aspx?pid=105723&jiid=507153&jiaid=17772

[NB.  The second one is available via the Ingenta service if you have that.]

On the neural pan of the balance, a big share of the work that I found myself
forced to undertake under the head of "Differential Logic" came out of trying
to integrate connectionist, dynamical systems, and symbolic logical models of
intelligent systems.

HT1. I suggest that it is at this fundamental informational level that we may look
     for equivalencies in the complexity of these respective systems.  If it can be
     shown that there are equivalencies, I think we have support for an argument that
     a system such as the global human socio-economic network is conscious in a manner
     similar to that of the human brain.

JA: The word "conscious" is a tricky one.  If I want to talk about consciousness
    in the purest sense of the word, then I have to switch over to the language
    of phenomenology, and that's a whole different ball of awareness than where
    I currently have my focus of attention.  But we also use "conscious" to mean
    "cognizant" or even "intelligent", and there our operational definitions will
    usually be such that we attribute cognizance or intelligence to critters only
    if they do or say complex, interesting, or subtle enough things.  In that case,
    the resources that we can muster for describing complex, dynamic, law-governed
    systematic conduct will be critical to our success in understanding the beast.

HT2: I agree that the concept of consciousness is tricky, as there
     are a number of conceptions of the term.  I continue to suggest,
     however, that we may plot consciousness on a continuum which
     corresponds to the complexity/information content of a system.

Well, I'm perfectly capable, well, more or less capable
of operating with abstract definitions, and if you want
to define, well, prospectively define "consciousness_HT"
as something that is measured and plotted along a scale
of "complexity/information_content_HT" of a system, then
I can contemplate the consequences of doing that, one of
which consequences is putting the practical definition of
the construct "complexity/information_content_HT" on the
agenda for once and future contemplation.  There we find
general principles that come into play when we project to
turn any such construct into a systems dimension, measure,
property, or variable.

HT2: In this way we define consciousness as a property of complex systems, ...

At this point, just to keep cognizant of where we are, the "we define"
of both "consciousness_HT" and "complexity/information_content_HT" is
still very prospective, that is to say, "to be announced in future".

HT2: In this way we define consciousness as a property of complex systems,
     and any system of interacting agents is therefore conscious, but only to
     a degree which corresponds with the system's complexity.  Thus a system of
     interacting sub-atomic particles is conscious, but only to a very low degree.
     It isn't until we reach a very high level of complexity within a system that
     we see the degree of consciousness exhibited by animals or humans.  In these
     instances, the brain system consciousness corresponds to a highly complex
     set of interacting neural processes.

HT2: Thus with this definition we can do away with associating consciousness
     with critters "which do complex, interesting, or subtle things".

Only if your prospective definition of "consciousness_HT"
and "complexity/information_content_HT" can be actualized
independently of any attributes like the ones I mentioned.

HT2: Under my theoretical framework consciousness is a very objective thing,
     and, as I say, corresponds to the measure of the complexity of a system.

Again, the project to define "the measure of the complexity of a system"
is a rather big promissory note at this point.  We had some discussion of
the issues that affect its feasibility a while back.  My general attitude
would be that we not take too simplistic a view of complexity.  For example:
Can we assume that there is a "unique measure of complexity of a system"?

As I currently view the scene, many different measures of complexity
come to mind, of such a diverse and heterogeneous nature that it is
a good question whether they can be sensibly related to one another.

HT2: Thus I speculate that if we can measure the complexity/information content
     of the human socio-economic network, we may find that it exhibits a nearly
     equivalent complexity or information content to that of a human brain.

That another couple of big if's.  And miles to go before we sleep ...

HT2: In this way, we can make the grand leap of faith and induce
     that the human socio-economic system is similarly conscious
     to an individual human brain.

If all it requires is a leap of faith, then I could have done that
right away and saved myself a lot of trouble, but the question is:
What do we do on Monday morning?  What difference does our faith
make to what we do the rest of the week?

HT2: As I've noted before, I am suggesting "complexity" is essentially synonymous
     with the information content of the system -- but it is at this point that
     I hit a wall because I do not know how to establish the information content
     of respective complex systems.  So it is that I am watching to see if you
     are about to deliver the "mother lode" from which we can say "here is the
     universal measure of information from which we can examine neural processes
     in the human brain, which measure we also apply to the human socio-economic
     network;  match our findings against one another and see if we have a "fit",
     so to speak.

These are suggestions worth pursuing, but the first wall I hit
is the complex nature of complexity, especially as we encounter
it in our environment and in ourselves.  I think that I referred
to the "2-edged" nature of qualities like complexity, diversity,
entropy, variety the last time that we discussed this problem.

But even that little complication was only with respect to one natural class
of complexity measures that we frequently find ourselves having to consider,
namely, those we might call "measures of distributional uniformity" (MODU's).

Another type of complexity measure, one that comes up especially often in
graph-theoretic and network-theoretic models, is the "arity" of relations
that we invoke in our models.  (AKA "adicity", "adinity", or "valency".)

HT1: My problem is of course a lack of a full appreciation of how information
     is described and analyzed.  So you can see my great interest in what you
     are doing.  I am appreciating the manner in which you have reverted to
     some basics for us to start with, and are moving toward the more complex
     aspects of your arguments.  I look forward to the next instalment!

JA: One of the things that I am trying to stress on several of these threads is the
    difference between "nominal" information, what we get if all we do is count the
    bits in a string, and "real" information, what we get from considering the bits
    in the midst of a choice process where they have the power to make a difference.

HT2: Far be it to question your judgment on Bayesian probabilities,
     as we've discussed before, but let me resort again to my favourite
     popular resource: 'Information', by Christian von Baeyer:

CvB: | Bayes' theorem answers the following question:  Suppose you know, or assume
     | you know, the probability that a certain conclusion follows from an initial
     | hypothesis.  Suppose further that a new bit of information is obtained, and
     | added to the hypothesis.  How do you compute the updated probability that
     | the conclusion is true, based on the combination of the old hypothesis
     | with the new information.

HT2: Von Baeyer does actually go on to explain this in more detail, but yes,
     it is still nevertheless a popular treatment of Baeysian probabilities.
     All I am suggesting is that this sounds an awful lot like what you are
     doing, and maybe it bears another look.  Again, I can only go by my
     intuition, because I haven't studied this stuff enough to be conversant
     of its applications -- but could you just humour me a little, and take
     another look at the application of Baeysian probabilities to your
     ultimate goals?

Let me stick a flag in this Bayesian issue, for a separate careful treatment
at the place where it comes up in the larger analysis of inquiry processes.
I've already given my "bird's eye view" (BEV) of it, and a few references,
and if that was not understandable, then I'll try to do better next time.

JA: It's still something of a work in progress, though.

o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
http://www.cs.bsu.edu/homepages/mighty/history.html
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o