Friday, February 10, 2012

Conservation of information

This Entropy/Thermodynamics/Information Wikipedia article is as close as one can get to what the standard scientific treatise of the link between energy, information and entropy.

From the information is physical -> Szilards Engine.

The global entropy is not decreased, but information to energy conversion is possible..

This appears to put information on an equal footing as mass or energy, in the sense that it is conserved in general, but can be also transformed between energy and mass. It is tempting to follow the lead of the ID ers to intimate that informational entropy can either stay the same and increase but not decrease - My view is different, in that I believe that information can be generated through the use of energy in naturally occuring metabolic systems. .

My first postulate is that information can increase (be created) within a boundary if the entropy is increased outside the boundary..

My second postulate is that some of the energy within the boundary must necessarily increase the information if entropy is increased outside the boundary. This would mean that stars like the sun, that send a lot of waste heat outside of itself, are becoming more ordered on the inside, in proportion..

Thus random processes could not increase information, but metabolic processes can, by separating where the information is stored and expelling its increasing entropy to where it isn't..

To destroy information, the directionality of increased entropy must be directed back to within the boundary (making the rest of the universe slightly more ordered).

Neither the generation, nor the destruction of information is a perfect process, and there will be some leakage of energy to information..

5 comments:

Dr Clam said...

What I am really after is some sort of definition of what you mean by information in a biological context. You've said that you don't consider two fruit flies generating enough genetically-different descendants to cover the whole Earth to a depth of a hundred metres as an increase in information (I am paraphrasing), so what is this stuff that is not increased by sexual reproduction?

Marco Parigi said...

Information is equivalent to scientific knowledge (of the DNA as a turing machine in general). Whether the knowledge that phenotype A will be advantageous in situation Y and phenotype B will be advatageous in situation Z. Because the information is to do with each individual phenotype, and the traits we are talking about are Mendelian, it is equivalent to the information associated with the reproduction of each of the individual traits.
From this point, to each individual trait, if one is to put a *random* mutation on any one of these traits, it is almost certain that the trait would just fail to what the trait normally does.However, if the DNA has built up a database of mutations that may have a probability of relevance to particular environments, that is information that is already in the DNA, that got there before, in a way that more closely resembles our cycle of guessing, experimenting and generating theories and algorithms to account for that knowledge - All of which has to happen within a metabolising organism, as opposed to a distributed random non-metabolising aspect of mutation.

Marco Parigi said...

By the way, I've joined "ResearchGate", which is a kind of social network of scientists to discuss science. We are way more likely to get a knowledgable and interested third party discussion going there than here.

Chris Fellows said...

I remember this from our discussion years ago. I am a disbeliever in the existence of information as defined by you here, if I understand what you are saying.

Marco Parigi said...

I do understand *that* you disbelieve in information as I've defined it, but I still don't understand *why* you disbelieve it. I still think that it stems directly from the central dogma of evolution. Thus in my mind it is poisonous fruit eaten from the tree of parsimony. The dogma makes my definition necessarily false, and conversely, my definition if proved correct would disprove the central dogma.