information and information structures

Wed Apr 11 09:26:28 EDT 2007

Dear David:

When you speak of information systems are you referring to specific kinds of
nodes and link relations? If so what are they? You seem to have something
specific in mind.

I would be inclined to define an information system as an information
structure which is maintained over time. In the static case (Delta t = 0),
we would only have an information structure or a fingerprint of the
information system. 

An information structure is a network or a two-dimensional probability
distribution (matrix). 

(Shannon-type) information is only an uncertainty or a one-dimensional
probability distribution. 

Is this sufficiently abstract to counter your general theory? :-) I agree
that it depends of the research question and the research design whether
these distinctions are important or not. 

Best wishes,  Loet 

In contrast, my general theory is designed to define all the possible kinds
of links (i.e., relations) between pieces of information, in such a way that
we can sort out the important structures. Change over time may or may not be
important, depending on the structure, as well as the application of the
analysis.

Cheers, David

At 12:46 AM 4/10/2007, you wrote:

Dear David, 

Yes, the time dimension makes it urgent because in information systems the
values  of he variables change and the eigenvectors (representing the
underlying structure) may change. Repeating the analysis for different years
then leads to comparative statics. For a dynamic analysis one needs a
calculus. Information theory offers a calculus (Bar-Hillel, 1955; Theil,
1972), while the measurement results for each moment of time can also be
fully appreciated. 

With best wishes, 

Loet

On 4/9/07, David E. Wojick <dwojick at hughes.net> wrote: 

Dear Loet,

I will look into this with interest. For my part, I agree that most of the
information structures I am talking about can be represented as matrices,
although I prefer to visualize them as linear, tree-like or network-like
arrays, depending on the case. Each structure is defined by some relation or
set of relations among the pieces of information. So far so good. 

Given my definition of information, each piece of information is
theoretically an atomic proposition. However, since the number of atomic
propositions in a sentence is typically roughly equal to half the number of
words, we usually work at a courser scale. Pieces of information may be
sentences, whole documents, or even collections of documents. 

As I explain, pieces of information can be related according to their
propositions, their physical expressions, their referents (what the
propositions are about), or a combination. For any given body of information
there will usually be a large number of important relation types, so there
are in fact many matirices of interest. And yes many of these change over
time. In this context the matrices all exist, whether we know it or not.
That is, how the information is related is a fact about the information,
independent of our analysis. 

Important relations (or matrices) in common use range from alphabetical
order applied to some aspect of the expression, to the so-called topic,
which is usually an aspect of what the information is about. Many popular
categorization schemes merge and confuse what are actually different
relations. 

Probability does not enter into it so far as I can tell, so I am interested
how this may relate to your matrices? Perhaps that has to do with the change
over time?

Best wishes, David

Dear David,

In reaction to your paper entitled "Outline of a new model of information
contents and structure," let me shortly react.

I would be inclined to think about information and information structures in
terms of the dimensionality of the probability distribution. For example, a
structure requires a network and thus a two-dimensional matrix can contain
this information. An information system would additionally require that the
structure is extended along the time axis and this would lead to matrices at
each moment of time and thus a three-dimensional cube of information would
be required. Unstructured information can be considered as a vector. One can
extend beyond a three-dimensional array towards a hyper-cube of information.

In The  <http://www.leydesdorff.net/evolcomm/index.htm> Evolution of
Communication Systems, Int. J. Systems Research and Information Science 6
(1994) 219-30. I elaborated this scheme as follows: 

Table I Organization of concepts in relation to degrees of freedom in the
probability distribution

  _____  

                first           second          third           fourth

                dimension       dimension       dimension       dimension 

operation       variation       selection       stabilization
self-organization 

nature          entropy;        extension;      localized       identity or

                disturbance     network         trajectory      regime 

character       probabilistic;  deterministic;  reflexive;      globally 

of               uncertain       structural      reconstructiv   organized; 

operation                                                       resilient 

appearance      instantaneous   spatial;        historically    hyper-cycle
in                   

                and volatile    multi-variate   contingent      space and
time 

unit of         change in       latent          stabilities     virtual 

observation     terms of        positions       during          expectations

                relations                       history 

type of         descriptive     multi-variate   time-series     non-linear 

analysis        registration    analysis        analysis        dynamics

  _____  

The complexity of the declared information system (the data) thus determines
the type of analysis which is possible. (In my book The Sociology of
Communication (2003), this scheme is discussed on pp. 99 ff.)

With best wishes,

Loet

  _____  

Loet Leydesdorff

Amsterdam School of Communications Research (ASCoR)

Kloveniersburgwal 48, 1012 CX Amsterdam

Tel.: +31-20- 525 6598; fax: +31-20- 525 3681

loet at leydesdorff.net ; http://www.leydesdorff.net/

  _____  

From: ASIS&T Special Interest Group on Metrics [
<mailto:SIGMETRICS at listserv.utk.edu> mailto:SIGMETRICS at listserv.utk.edu] On
Behalf Of David E. Wojick 

Sent: Sunday, April 08, 2007 7:12 PM

To: SIGMETRICS at listserv.utk.edu

Subject: Re: [SIGMETRICS] The communication of meaning in social systems;
preprint version available 

<http://web.utk.edu/~gwhitney/sigmetrics.html> 

Dear Loet,

It is delightful to get something like this on a holiday. I take it this is
what the phenomenology of meaning looks like these days. Not that I pretend
to understand phenomenology, so please correct me if I am wrong. I also take
it that the interpretation of the parameters in the very interesting
equations, as well as the technical concepts being used, is to be found in
the cited references. 

Since I have also presented a theory of the nature of information here, I
thought it appropriate that I speculate upon the difference between this
body of work and my own.

CF: http://www.bydesign.com/powervision/Mathematics_Philosophy_Science/in
formation.html

My work derives from the tradition of analytical philosophy and mathematical
logic begun by Russell and Wittgenstein. I suggest that it is looking at
meaning in a very narrow sense, as exemplified by the atomic proposition.
The phenomenological tradition is looking at meaning in a very broad sense,
what it is to be meaningful if you like. 

The human condition is rich enough to accommodate both approaches and so I
do not see any disagreement here between us. The question is if there is any
connection?

Best regards,

David

<http://web.utk.edu/~gwhitney/sigmetrics.html> 

The  <http://www.leydesdorff.net/meaning0704/index.htm> communication of
meaning in social systems

pdf-version <http://www.leydesdorff.net/meaning0704/meaning0704.pdf> 

Abstract

The sociological domain is different from the psychological one insofar as
meaning can be communicated at the supra-individual level (Schï¿½tz, 1932;
Luhmann, 1984). The computation of anticipatory systems enables us to
distinguish between these domains in terms of weakly and strongly
anticipatory systems with a structural coupling between them (Maturana,
1978). Anticipatory systems have been defined as systems which entertain
models of themselves (Rosen, 1984). The model provides meaning to the
modeled system from the perspective of hindsight, that is, by advancing
along the time axis towards possible future states. Strongly anticipatory
systems construct their own future states (Dubois, 1998a and b). The
dynamics of weak and strong anticipations can be simulated as incursion and
hyper-incursion, respectively. Hyper-incursion generates ï¿½horizons of
meaningï¿½ (Husserl, 1929) among which choices have to be made by incursive
agency. 

Loet Leydesdorff & Sander Franse

  _____  

Amsterdam School of Communications Research (ASCoR)

Kloveniersburgwal 48, 1012 CX Amsterdam

Tel.: +31-20- 525 6598; fax: +31-20- 525 3681

loet at leydesdorff.net ; http://www.leydesdorff.net/

Now available: The
<http://www.universal-publishers.com/book.php?method=ISBN&book=1581129378>
Knowledge-Based Economy: Modeled, Measured, Simulated . 385 pp.; US$ 18.95 

The
<http://www.universal-publishers.com/book.php?method=ISBN&book=1581126956>
Self-Organization of the Knowledge-Based Society ; The
<http://www.universal-publishers.com/book.php?method=ISBN&book=1581126816>
Challenge of Scientometrics

--

"David E. Wojick, Ph.D." <WojickD at osti.gov>

Senior Consultant -- The DOE Science Accelerator
http://www.osti.gov/innovation/scienceaccelerator.pdf

http://www.osti.gov/innovation/

A strategic initiative of the Office of Scientific and Technical
Information, US Department of Energy

(540) 858-3150 

391 Flickertail Lane, Star Tannery, VA 22654 USA

http://www.bydesign.com/powervision/resume.html provides my bio and client
list.

http://www.bydesign.com/powervision/Mathematics_Philosophy_Science/ presents
some of my own research on information structure and dynamics.

-- 
Loet Leydesdorff 
Amsterdam School of Communications Research (ASCoR)
Kloveniersburgwal 48, 1012 CX Amsterdam
Tel.: +31-20- 525 6598; fax: +31-20- 525 3681 
loet at leydesdorff.net ; http://www.leydesdorff.net/ 
--------------------------------------- 
Now available: The Knowledge-Based Economy: Modeled, Measured, Simulated,
385 pp.; US$ 18.95; 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20070411/667306a7/attachment.html>