fredrik

Sorry, i didnt notice the op was suspended. I responded for one reason: The topic of Smolin is deep and worthy discussion/commwnt, but it was obscured by a terrible formatting of the originaö post without own comments. So i just aimed to stand up for and comment on the OT. I suggest for tou all to read smolins book. Or gooogle his perimeter talks with roberto unger. You find mp4 versions of essentially the same content as the book. /Fredrik

Yes i even like that phrase better, good point. The same applies to "nature of law". If we instead call it "physics of law" the hint becomes more clear - are there in fact physical constraints on law and inference processes i nature? (Think information capacity and limits of complexity of the physical observer) /Fredrik

I also think lets not forget also the "nature of law" which is really the other side of the coin here. Smolin well argued what "nature of law" is NOT. But so far in my reading of the book there arent much about alternative view except its effectiv laws that evolves. Which somehow seems like unavoidably correct, but i think there is more to be said. I would expect evolution of law not to take place in a fixed configuration space as that resorts to the same fallacy as was critiqued? Doesnt it? I think We must also see how hypothesis space itself can be understood from the intrinsic perspective. /Fredrik

I bought Smolins book when it came out but only now, 3 years after i started to read it. If you take the inference perspective of science seriously the argument against "timeless" law is imho very compelling. The timeless law can be nothing but a metaphysical fantasy or at best a rational expectation in the light of past experience. To extrapolate that to a constraint on the future seems like a logical jump. I think we need a new understanding of the scientific method as well as falsification alone is a simplistic view that is on par with the view of timeless law, thats either forever true - or wrong. Look at biological evolution, thats not how nature evolves. I think these Stubborn paradigms root back to the scientific paradigm. If we are lost on our ever more inflated map, we need someone to question the way we devise the map. I have the last part of the book left to read. But while i agree with the initial analysis i see other possibilities. Apparently smolin does not like the information interpretation om qm, which sutprised me. As information connects tighly to the scientific inferencprodcess which i see as the main argument against timeless law simply because there is no real finitr physical inference process that can be attached infinite confidence. I draw the conclusion that smolin doesnt make that parallell, which is where he looses me as well. His cosmological natural selection is imo not radical enough as it takes existing parameter space as input. I think we should be able todo much better. Its way too much baggage and thus have not so much explanatory power? /Fredrik

I too think these are very generic symptoms, that in theory could have a wide range of causes, ranging from simple stress/depression, to bacterial or viral infections or tumours. And it's probably not rational to jump into worrying about worst case scenarious without further examinations and considerations of the more natural causes, like stress, depression etc. As we know the state of the brain and mind, can actually give symptoms on the physical body as well by feedback mechanisms of the brain controlling body functions, even if it's subconscios. Although anything can happen to all us us, it's probably no rational to worry about this based on weak generic symptoms. But some people are somewhat hypocondric, and then it's easy to overinterpret generic harmless symptoms. That said, if you are seriously worried about a specific disease, infection or tumour, I think the doctor should take is serious and make a test. To doubt something since it's statistically unlikely, is a rational first response bit if no other cause is found, I'd expect that you can demand to get a test, to at least release yourself for the pain of worrying, incease you've been increasinly worried, that itself may increase the symptoms. /Fredrik

I would be more worried for someone who didn't raise the question. I think some questions, like this one, more represents "direction" and "arrow of evolution" rather than an inconsistent state. After all we tend to perceive nature as as succession of moments, each moment encodes the implicit question "what next". Next question please. /Fredrik

These comments relate to the point I tried to make as well. As I think of it, each observer or subsystem of the universe interaction with another and has a complexity limit that constraints the amount of resolvable and distinguishable information it can encode. It follows that there is a minimum amount of "evidence" that can be counted, and that hypotetically smaller fragments than this would be indistinguishable from zero TO THIS observer. The effect I would expec that the physical action of this observer is then invariant with respect to these hypotetical indistinguishable fragments. Ie. the systems responds as if they weren't there. It's what I tried to say here:

Like has been said it probably depends on what organisms you are talking about. Since not all are the same. That said, doing like you suggest with gelatin works just fine in many cases. I've used this technique on yeast cells. However you might want more nutrition than you get from plain sugar to get good growth. For yeast cells (S.Cerevisae) a very good medium is to use brewers wort made from malted barley which contains an excellent mix of sugars, amino acids and even good lipids. It's a complete growth medium for yeasts. You boil the wort to sterilise it of course. Then you add gelatin in the right amount. Edit: I trust the experts advice that there are downsides with gelatin, but as you say, you got to start somewhere, and for this gelatin works just fine at least for the yeasts I've play with. There is no reason why you couldn't try it. /Fredrik

I agree it's an interesting question, and about real numbers are not in nature etc. So to the question, what can we do then? Here I disagree a bit about us beeing trapped. There is an undeniable utility of real numbers as a means of "continous counting" when it comes to sets so large that the continous view in a certain sense gets simpler and are accurate for most practical purposes, even though in a fundamental sense the the continuum is not quite one-2-one with reality. But what I think, there are IMO good reasons also to think that there is more here, and there may be a connection between "counting", in particualr "counting evidence" and probability theory and which at a deep level connects to physics. In particular in the foundation of QM, the physical basis of probability is still somewhat unclear. The frequentists interpretation and ensembles of systems are not quite satisfactory, although in some domains it's good enough. Another issue is to understand what "information" and relative information means, given that this seems to be a fundamental concept to QM. And certaainly here we have a deep connection between QM - information - counting. One possible subjective bayesian view is that the "counting" could be seen as one system "counting" expected states of fellow systems in it's environment, and that this counting, when combining it with idea of information capacity bounds etc, may yield predictions that is related to constrained cardinality of state spaces. So one alterantive, to beeing trapped in real numbers, is to try to make a reconstruction of the counting, and during the course get a reconstruction also of information theory that is not based on continuum probability. After all, if we are to develop a theory of counting, and counting evidence to be used in desicion making in particular, the real numbers are IMO not a natural starting point. The integers are more natural and thus possible "more physical". I still one can still understand the enormous succes of calculus as introduced by newton and leibnitz in physics, but the reason for that success does not contradict the possibility that this framework isn't of uniersal utility. But maybe we have a bit lost ourselves in the contiuum formalism due to it's success. There is a book by ET Jaynes "probability theory - the logic of science" where he by following an idea that he is to argue on counting evidence and rating degrees of belief, and he shows that by independent reasoning he is naturally lead to the koglomorov axioms of probability theoyr. But very early he just makes an assumption that "degree of belief" is represented by real numbers, so he puts on that the set of possible beliefs forms a continuum! That's something I object to, and I think it's even wrong if you consider the set of possible beliefs as "distinguishable possibilities" to a physical observer/system. It doesn't make sense to consider a "mathematical observer". /Fredrik

The perspective I insinuated above is not cured by this example. If you consider the state space or allowable states. It's infinite regardless of wether you picture angles or slopes. So the deeper issue of "infinity" is deeper than there mere extension of all finite real numers to include also [math]\infty[/math]. It has to do with how to measure sets of distinguishable possibilities. Does it make sense to consider an infinite of physically distinguishable states? This is IMO the more interesting side of the issue, that goes beyond the angle vs slope example you raise. Now, if you combine this question with the meaning of probability and in particular subjective bayesian probability. Then the question seems to get also physical interpretations that is not just human concepts. Does the action of two interacting systems, really reflect a continuum of distinguishable possibilities or not? /Fredrik

Maybe one can interpret the question wider, and ask wether the real numbers exist in nature? This questions can be motivated by the observation that if we do not talk about symbolic mathematics, but actual numerical computations in the context of predictions etc, pretty much without exception all our representations of any real number is truncated, and the set of rela numbers is also truncated to a finite set, in a computer for example. That makes for an interesting reflection also if nature itself, say one proton responding to an electron, if each subsystem can only code finite amount of information, then that may suggest that somehow the continuum models we often use contains a redundancy. Continuum models, tend to contain an infinite and uncountable set of possibilities. And it seems ambigous to see howto compare two infinite set of possibilities. Usually the only sensible way to do this is with limits, but the choice of limiting procedure rarely connect to physics - maybe they should? I for one think it's doubtful wether contiuum models has a justified place in a fundamental reconstruction of physics and a theory of QG. If we admit that real numbers are just a mathematical tool, could it be that, there exists a better mathematical tool with LESS redundancy that we chould use? Then maybe the infinities that we know to appear in some calculations just wouldn't appear in the first place. /Fredrik

My favourite way to try make sense out of all possibilities beeing realised at the same time is a game theoretic analogy. First acknowledge, that a system is characterized in the way it behaves. In other world, in the way it interacts with it's environment. This means for superposition that the system actually behaves sort of as if it realized all possibilities at once. Where have we seen that before? Interaction is something mutual. If you are in an environment, a rational player will ideally act consistently with his own expectations of the environment he is in. Similarly, the environment will backreact upon him as per the expectations(read information and subjective relative information state) it has of this system. So a rational player does not choose what to do, from randomly choosing one possibility of the unknown premises. He acts upong all of them, calculating the risks based on all available info. The action based on such total risk analysis can intuively be seen to be different, than a randomly chose action, based on random premises. Roughly speaking one can conceptualize this from an intuitive point. This means that the information interpretation of QM, makes sense. A system interacts with it's environment as if the mutual information is lacking, which is exactly the case. This will persist until there is a mutual information exchange; an interaction; a measurement. Then the behaviour changes, just like we would expect from a rational player. I personally think that the single best abstraction to try to understand the logic of QM is this. All of the mechanical or visual analogies are inadeqaute. Instead, try to think in terms of players actions, based upon incomplete information. And how a set of such players behave in interaction. Each player is an obsever, observing the others. In the particles, one observer is not on par with the others. One observer constitutes an earth based lab, which is massive compared to the particles beeing studied. In this sene, the state of the observer is largely that of an outside obserer, since it's not much influence by the game going on. /Fredirk

Thanks for another good highlight Martin! I'll add my partly critical view to Rovelli's ideas. I see the picture Rovelli paints by noting this trends, is part of the motivation for his own expectation of the future of physics. I think his idea is to argue that he takes the relational ideal, to it's extreme, but IMHO he still maintains a great deal of realism (making the revolution still incomplete) Rovelli suggests that the root of some confusion is not beeing clear on what an observable means. In his paper http://arxiv.org/abs/gr-qc/0110035, he tries to classify observables as partial and complete. "Partial observable: a physical quantity to which we can associate a (measuring) procedure leading to a number." "Complete observable: a quantity whose value can be predicted by the theory (in classical theory); or whose probability distribution can be predicted by the theory (in quantum theory)." He notes that classical (ie. non.quantum) mechanics is all about relations between partial observables. Ie. the theory does not determine partial observables, it only determines relations between them, and these relations are envisioned as timeless. In the QM case, he says that the theory should predict conditional probabilities, (or probability amplitudes) between partial observables. Again these conditional probabilities are timeless. He says that that partial observables has a physical correspondence in that that correspond to actual measurements, such as clock readings, rod readings etc, detector readings etc. The complete observables OTOH, are what the theory predicts. We have learned from SR and GR that coordinate measurements are ambigous because the coordinate lacks physical basis. This is what we get if we think in terms of geometry. The parameterization of a section of a curved surface has no relevance for it's geometry. The geometrical properties OTOH, are those that are diffeomorphism invariant, and thus observer invariant. So what's the possible objection to this picture which seems to be the basic lesson from the relational tradition from Mach to Einsten? My objection is that it's not relational enough, in particular does it ignore the relativity of the observers internal structure. In SR and GR, an "observer" is reduced to a choice of coordinate gague. A reference frame equipped with clocks and rods. No attention is ever payed to how the physical manifestation of these clocks and rods are actually to be realized, in sharp situations, and wether there are constraints on the freedom to build accurate clocks and rods, and what happens to the observational data, once observed? IS there any hidden memory storages somewhere? In QM and observer is reduce to a choice of measurement operator. Then the geometrical picture means, that what a single observer is completely arbitrary since in the geometrical picture the observer is simply an arbitrary gauge fix. It's only the relations between all possible observers that we can predict from a theory. But how do we know when we have distilled the full symmetry transformations between all observers? And more important, how can an single observer acquire and hold this information? There is a birds view here, that breaks the relational ideal. It implicitly makes use of a birds view that IMO is principally no more sophisticated that newtons idea of absolute space. I don't think such an abstraction acknowledges the meaning of the observer. It is still true that an observer can only be characterized by it's interactions with other observers. And thus we expect that our theory should predict how observers interact. In a sense this is what gauge theory does, two different observers, corresponding to different gague choices, are related by the gauge transformation. So the gauge symmetry, predicts how any two arbitrary observers relate. But this isn't good enough. What we need is one observers view, of how two OTHER observers relate. Not Gods view, of how two observers relate? The first point is that we can not so easily dimiss the concept of observer, a fundamental object in any theory of scientific inquiry, to the status of an arbitrary gauge choice. There is something deeply wrong with that, that seems to imply a hidden birds view, where you picture "all possible observers" and that the collective somehow describes a geometry. But the only one that can be at a single moment, be informed about this is an omnipresent observer that can see everything, but without interfering. But somehow, the very notion of theory, belongs in the context of scientific inquiry itself. Thus the idea of observerindependent laws, are an almost religious realist idea I can't accept. I would expect that laws are what an observer infers from measurments, as per some scientific method, and not ONLY the other way around. I think there IS a infinite regress here, that is unavoidable. And this regression IS evolution. The question I have at my hand, is not how the unknown laws of physics can explain my observations, but rather how I can infer from them, knowledge that allows me to LEARN about the laws. It's here the internal structure of the observer comes into play. To make a comparasion between partial observables, some sort of memory structure is needed. Also the amount of memory limits what is possible to see. These things are not even mentioned in Rovelli's reasoning. Which I find surprising. Therefore, i think that an alternative reasoning than that of rovelli, is to consider evolving and emergent laws. And these laws, an in particular, different observers differing VIEW of law, explains their relation, and mutual impact. This could be an alternative logic, still effectively consistent with all we know of SR, GR and QM. It might however, suggest a different way forward, that is even more relational that the geometric view rovelli advocates. In that it totally does away with the realism of geometry as well. And not only in the probabilistic sense, that we only know nthe probabiltiy for the geometry, but in a even deeper sense of evolving law. This paradox here between the fundamental importance of the observer, and on the other hand the geometrical view that hte observer is ambigous and that only the relationas to other observers, can justify it needs to be resolved. I can't see that rovelli tries to. The gods view may seem like a technical possibility, but I think it's conceptually inconsistent with a fit method of scientific inquiry. The person I think is at least starting to acknowledge some of these questions is Smolin. He seems to have a chapter in the later part of the book, but it doesn't seem to focus on this. I think the relational ideas, and relations between partial observables, is a good perspective. But I think something is missing.Rovellis seems to want to tie up that sack in a realist manner. I think the picture, in which the actual complete observables emerge, must be an inside view, not a birds view, which in turn imples that the basis of this view is evolving. Rovelli on one hand, seems to try to very much acknoqledge the deep lessions from relativity, but when he tries to merge it with measurement theory, I can't see that his synthesis there is satisfactory. I think another way to do it, would instead place the typical symmetry views, in a dynamical context, and thus the emergent symmetries would correspond to emergent observable law. The concept of an observer independent law should not be needed, no more than an external embedding space is needed to describe differential geometry. So the early part of Rovellis abstraction of mechanics, as well as the early part of his relational QM(http://arxiv.org/abs/quant-ph/9609002), is I think good, but when this is to be cast into the form suitable for a theory of scnetific inquirry, it does not quite do to reduce the observer to a arbitrary gauge choice and be done with it. I'll end with noting that I think that Rovelli's RQM paper is one of not so many favourite papers of mine. Not because I like all of it, but because it contains part that speaks for themselves. /Fredrik

Thanks for your comments ajb! I definitely share both these "quests" with you. I guess I my personal guess is that even normal "first quantiztion" QM will be understood just as what you say in the latter note - as a limit. I do not think the structure of QM as it is known to be successful can be explained within itself. Maybe we need to think up different abstractions, where hilbert spaces and unitary transformations are understood and special, or limiting cases. The major problem seems to be how to control the madness that presents itself when you allow non-unitary transformations and fuzzy hilbert spaces as you loose the solid reference. I think however that goes well with nature, there are not fixed references. But hopefully once we understand the origin of inertia in these terms, we can also see how stability and effective hilbert spaces can emerge and persist even while embedded in chaos. /Fredrik

Aside from the fact that it hasn't been proven that it is not possible, what are the motivation for asking this specific question? As I understand it, that abstraction of quantizátion, and classical vs quantum worlds, seems to be the standard view. Do you really think that abstraction, will do also for the future generation of theories of TOE unification and new understanding on QG? I personally have doubts in that, and only the lack of proof that it is not impossible seems insufficient motivation to support the question. I guess I have a more philosophical angle than you, but I'm curious where your motivation to the more stronlgy mathematically abstracted quests comes from. As I see it, the physical and conceptual support is becomer weaker these days? do you disagree? What I mean is that, perhaps an ultimate hilbert space with unitary evolution isn't the way our ultimate understanding will end up? It sure is still a possibility, but my confidence in that idea is dropping. /Fredrik