[Fis] Fw: "Mechanical Information" in DNA

Mark Johnson johnsonmwj1 at gmail.com
Thu Jun 9 16:31:20 CEST 2016

Dear all,

Is this a question about counting? I'm thinking that Ashby noted that Shannon information is basically counting. What do we do when we count something?

Analogy is fundamental - how things are seen to be the same may be more important than how they are seen to be different. 

It seems that this example of DNA is a case where knowledge advances because what was once thought to be the same (for example, perceived empirical regularities in genetic analysis) is later identified to be different in identifiable ways.

Science has tended to assume that by observing regularities, causes can be discursively constructed. But maybe another way of looking at it is to say what is discursively constructed are the countable analogies between events. Determining analogies constrains perception of what is countable, and by extension what we can say about nature; new knowledge changes that perception.

Information theory (Shannon) demands that analogies are made explicit - the indices have to be agreed. What do we count? Why x? Why not y? otherwise the measurements make no sense. I think this is an insight that Ashby had and why he championed Information Theory as analogous to his Law of Requisite Variety (incidentally, Keynes's Treatise on Probability contains a similar idea about analogy and knowledge). Is there any reason why the "relations of production" in a mechanism shouldn't be counted?  determining the analogies is the key thing isn't it?

One further point is that determining analogies in theory is different from measuring them in practice. Ashby's concept of cybernetics-as-method was: "the cyberneticist observes what might have happened but did not". There is a point where idealised analogies cannot map onto experience. Then we learn something new.

Best wishes,


-----Original Message-----
From: "Loet Leydesdorff" <loet at leydesdorff.net>
Sent: ‎09/‎06/‎2016 12:52
To: "'John Collier'" <Collierj at ukzn.ac.za>; "'Joseph Brenner'" <joe.brenner at bluewin.ch>; "'fis'" <fis at listas.unizar.es>
Subject: Re: [Fis] Fw:  "Mechanical Information" in DNA

Dear colleagues, 
It seems to me that a definition of information should be compatible with the possibility to measure information in bits of information. Bits of information are dimensionless and “yet meaningless.” The meaning can be provided by the substantive system that is thus measured. For example, semantics can be measured using a semantic map; changes in the map can be measured as changes in the distributions, for example, of words. One can, for example, study whether change in one semantic domain is larger and/or faster than in another. The results (expressed in bits, dits or nits of information) can be provided with meaning by the substantive theorizing about the domain(s) under study. One may wish to call this “meaningful information”. 
I am aware that several authors have defined information as a difference that makes a difference (McKay, 1969; Bateson, 1973). It seems to me that this is “meaningful information”. Information is contained in just a series of differences or a distribution. Whether the differences make a difference seems to me a matter of statistical testing. Are the differences significant or not? If they are significant, they teach us about the (substantive!) systems under study, and can thus be provided with meaning in the terms of  studying these systems. 
Kauffman et al. (2008, at p. 28) define information as “natural selection assembling the very constraints on the release of energy that then constitutes work and the propagation of organization.” How can one measure this information? Can the difference that the differences in it make, be tested for their significance? 
Varela (1979, p. 266) argued that since the word “information” is derived from “in-formare,” the semantics call for the specification of a system of reference to be informed. The system of reference provides the information with meaning, but the meaning is not in the information which is “yet meaningless”. Otherwise, there are as many “informations” as there are systems of reference and the use of the word itself becomes a source of confusion.
In summary, it seems to me that the achievement of defining information more abstractly as measurement in bits (H = - Σ p log(p)) and the availability of statistics should not be ignored. From this perspective, information theory can be considered as another form of statistics (entropy statistics). A substantive definition of information itself is no longer meaningful (and perhaps even obscure): the expected information content of a distribution or the information contained in the message that an event has happened, can be expressed in bits or other measures of information.

Loet Leydesdorff 
Professor, University of Amsterdam
Amsterdam School of Communication Research (ASCoR)
loet at leydesdorff.net ; http://www.leydesdorff.net/ 
Associate Faculty, SPRU, University of Sussex; 
Guest Professor Zhejiang Univ., Hangzhou; Visiting Professor, ISTIC, Beijing;
Visiting Professor, Birkbeck, University of London; 
From: Fis [mailto:fis-bounces at listas.unizar.es] On Behalf Of John Collier
Sent: Thursday, June 09, 2016 12:04 PM
To: Joseph Brenner; fis
Subject: Re: [Fis] Fw: "Mechanical Information" in DNA
I am inclined to agree with Joseph. That is why I put “mechanical information” in shudder quotes in my Subject line.
On the other hand, one of the benefits of an information approach is that one can add together information (taking care to subtract effects of common information – also describable as correlations). So I don’t think that the reductionist perspective follows immediately from describing the target information in the paper as “mechanical”. “Mechanical”, “mechanism” and similar terms can be used (and have been used) to refer to processes that are not reducible. “Mechanicism” and “mechanicist” can be used to capture reducible dynamics that we get from any conservative system (what I call Hamiltonian systems in my papers on the dynamics of emergence – such systems don’t show emergent properties except in a trivial sense of being unanticipated). I think it is doubtful at best that the mechanical information referred to is mechanicist.
John Collier
Professor Emeritus and Senior Research Associate
University of KwaZulu-Natal
From: Fis [mailto:fis-bounces at listas.unizar.es] On Behalf Of Joseph Brenner
Sent: Thursday, 09 June 2016 11:10 AM
To: fis <fis at listas.unizar.es>
Subject: [Fis] Fw: "Mechanical Information" in DNA
Dear Folks,
In my humble opinion, "Mechanical Information" is a contradiction in terms when applied to biological processes as described, among others, by Bob L. and his colleagues. When applied to isolated DNA, it gives at best a reductionist perspective. In the reference cited by Hector, the word 'mechanical' could be dropped or replaced by spatial without affecting the meaning.
----- Original Message ----- 
From: Bob Logan 
To: Moisés André Nisenbaum 
Cc: fis 
Sent: Thursday, June 09, 2016 4:04 AM
Subject: Re: [Fis] "Mechanical Information" in DNA
Thanks to Moises for the mention of my paper with Stuart Kauffman. If anyone is interested in reading it one can find it at the following Web site: 

[The entire original message is not included.]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listas.unizar.es/pipermail/fis/attachments/20160609/1465007f/attachment-0001.html>

More information about the Fis mailing list