Re: [Fis] Limited info
At 18:41 21/06/2006, you wrote: Pedro -- OK, I think I see your basic point. If so, then we do agree because I have concluded (tentatively) that, in the context of Universal disequibilibrium, the principle of least action can be explained by the maximum entropy production principle [e.g., the fastest action would require the hardest work, and the shortest path for entropic energy flows (heat, light, sound) would be sought in the interest of Universal equilibration]. STAN Maybe you are right, Stan, but my impression is that, if we are truly a la recherche de l'information perdue, we cannot follow that entropic path only . Playing with the Proustian metaphor, there are two paths which have to be intertwined: du côté de chez Swann le côté de Guermantes. The dissipation of structures via diminishing supra-atomic distinctions the creation of new structures via atomic bonds implying diminishing intra-atomic distinctions... Which path does predominate? It depends entirely on the existing boundary conditions. That's the general trick of life to navigate easily in both directions: a fantastic multiplication of boundaries by way of organs, organelles, compartments, membranes, etc. Besides, both ways of information counting are very different, the entropic and the atomic internal energy (enthalpic), notwithstanding that Gibb's and other free energy expressions unite them algebraically. In this sense, the problem raised by Hans days ago, on the numbering discrepancy implicit in Schrodinger's equation, looks a very intriguing point. As said, my hunch concerning the informational quest for unification, is that the principle of least action is more general and more easily translatable to a form similar to least informational description than any acceptation derived exclusively of the second law... and perhaps more amenable to dialog with string theories too (which seemingly can deal with gravity and are cosmologically and ontologically quite creative). Information physics is indeed a very fundamental region within the whole information science enterprise. If there is any possibility in the future, we should devote a complete real conference or seminar to it. At the time being, Andrei's patience should be overstretched by all this continuous handweaving! best regards Pedro ___ fis mailing list fis@listas.unizar.es http://webmail.unizar.es/mailman/listinfo/fis
Re: [Fis] Quantum Information - Probability Functions and Information
Dear Steven, I was not able to reply you earlier. But I think that I should do this even so late after your Email. You posted problems which are very important (at least for me). 1. Quantum probability functions are either directly equivalent to probability functions in Shannon\'s information theory or they are not. Which is it? Quantum probability functions are not equivalent probability functions in Shannon\'s information theory. They equivalent to quantum (von Neumann) information functions. But we can generalize classical probability and consider contextual probability (I did this last years). I have never tried to proceed to contextual information theory, but it is possible. Such a classical probabilistic information theory will cover both CI and QI and some new information theories which are neither classical (noncontextual) nor quantum. 2. If there are new physical mechanisms discovered in quantum mechanics Personally I do not know. The common viewpoint is that QM is really about completely new physics, but if you ask people working in Bohmian mechanics, SED and other random field models, they would reply that QM is a special representation of classical random fields, I am at the latter position. then I am with Penrose - recall my earlier report of his observation concerning cricket balls. The mechanisms must exist independent of scale. And that implies to me that a clear mechanistic integration with information theory is possible and required. I would like to say that mechanism is indepent of scale, but its representation, e.g., the QM-representation of laws of nature, is dependent on transition from one scale to another. Therefore I think that quantum-like descriptions can be useful not only in quantum physics, but everywhere we have a tarnsition from one scale to another. 3. It seems to me that the problem here is the parallel postulate and its equivalent by extension to computation. This is the reason probabilities come into it at all. Perhaps we need to be reminded that probabilities are the result of observations of the statistical behavior of individuals. Individuals have an ontological status while probability functions only have epistemological status. But probability is just a special way of encoding of the onthological properties of individuals. In this sense probabilities (as well as information) are objective. This was the viewpoint of Richard von Mises. 4. Recurring laws of probability do appear to be stable laws, but they are founded upon the aggregation of individual behavior. Their ontological status is derived from the behavior of individuals, not by their own account. Well I agree, but probabilistic laws represent in a special way ontological laws. The Kolmogorov equation for probabilities represent Brownian motion as well. All the best, Andrei ___ fis mailing list fis@listas.unizar.es http://webmail.unizar.es/mailman/listinfo/fis
