Fwd: Re: [Fis] Re: What is the definition of information ?Fwd: Re: [Fis] Re: What is the definition of information ?
From: John Collier <collierj@ukzn.ac.za>
Date: Fri 02 Sep 2005 - 21:18:05 CEST
Iam sending this to the whole list, since Ithink the current discussion is losing perspective, and
is starting to go in circles. History is important.
John
Professor John Collier
attached mail follows:
There is no universally accepted notion of information. The commonsense notion is highly equivocal,
and there
Information is commonly understood as knowledge or facts acquired or derived from, e.g., study,
instruction or observation (Macmillan Contemporary Dictionary, 1979). On this notion, information is
presumed to be both meaningful and veridical, and to have some appropriate connection to its object;
it is concerned with representations and symbols in the most general sense (MacKay 1969).
Information might be misleading, but it can never be false. Deliberately misleading data is
misinformation. The scientific notion of information abstracts from the representational idea, and
includes anything that could potentially serve as a source of information. The most fundamental
notion of information, attributed to a number of different authors, is "a distinction that
makes a difference" (MacKay 1969), or "a differnece that makes a difference" (Bateson
1973: 428). Information theory, then, is fundamentally the rigorous study of distinctions and their
relations, inasmuch as they make a difference.
MacKay, Donald M., Information, Mechanism and Meaning. Cambridge, MA: MIT Press, 1969.
The above definitions are in line with your remarks. Shannon's approach does not define information,
but the information capacity of a channel. A different way to define information that does deal with
specific amountsd of information is algorithmic complexity theory:
# Chaitin, Gregory J., Randomness and Mathematical Proof. Scientific American 232, No. 5 (May 1975):
47-52. http://www.cs.auckland.ac.nz/CDMTCS/chaitin/sciamer.html
In neither the Shannon nor the algorithmic case is meaning relevant to information. If you want to
extend the theory to include semantic values, I would suggest that you consider information as a
sign along the lines of Peircean semiotics. In this case a difference in information is a difference
in a sign that makes a difference to its meaning.
I hope this helps.
John
Professor John Collier
--------------------------------------------------------------------
_______________________________________________
|
This archive was generated by hypermail 2.1.8 on Fri 02 Sep 2005 - 21:18:47 CEST