Time and space as assymmetries in stochastic evolutions

[fredrik]

Background: Currently the most sensible procedure to me is to consider spacetime basically as defined through stochastic evolution (relating to a kind of information entropy), and that the "classical spacetime" will be reached as the "best guesses" or expectation values. Time would be considered naturally as the "most probable disturbance" given our conditional information (which could be anything btw - as long as it's correct and not just an assumption).

 

In between measurements the information entropy either increases or is held constant, but during interactions the entropy is decreases as new information is gained. So interactions (communication) can be considered to maintain information.

 

I have an idea that out of this it might be possible to derive some very general equations of dynamics, that would sort of correspond to the "a priori" dynamics, given no further info. Any further info, is simply added ontop of the former as a constraint to the equations of dynamics.

 

The idea is not to start out with a spacetime notion - rather to let the spacetime dimensions emerge as assymetries in the stochastic evolution. So there will be no presupposition of spacetime. If data then requires spacetime, it will come to us as the "most natural" interpretation. It will be more complex, but I think also more correct. And it would be extremely general, and easily extend beyond physics to general intelligence applications.

 

This is very fuzzy I know. And my intent of this post is definitely *not* to present a new theory here: I simply want to ask if anyone on here that has been working more on this than I have lately, is aware of what the current work is on this? And if anyone has any links?

 

I have some ideas but before I start fiddling with equations from scratch I wonder if anyone knows if there are any good papers on this already that I better read before I get to it?

 

Also this post could I guess as well have been posted in some of the other sections, but I had to pick one

 

Ideas or pointers anyone?

 

/Fredrik

[fredrik]

Time would be considered naturally as the "most probable disturbance" given our conditional information (which could be anything btw - as long as it's correct and not just an assumption).

 

This came out wrong. What I meant was that time would be a "parametrization" of the most probable disturbance, and the actual "rate of time" would simply be related to changes in some reference subsystem, for example a clock device taken as definition of the rate of time. So while the flow of time might be independent of the clock as such, the definition of a "second" would obviously be a matter of definition (arbitrarily chosen reference).

 

/Fredrik

[fredrik]

I'll fiddle on and see what happens. I've found a number of interesting papers but IMO they don't address the main point, they rather elaborate the existing ad hoc theories from the point of view of probability and information theory which is illumanating itself but I rather have in mind using the full power of information theory together with some natural philosophical arguments to see what the consequences of these things are when required as constraints. I see that along the way many many assumptions are made to simplify, and I am not convinced that this is acceptable, and it may itself be the reason to several current problems.

 

The obivous main problems I can predict so far are of technical nature, such as hard to manipulate formalisms and one has not choice but to rely on numerical methods in the general case. I think it should ultimately be formulated as an optimation problem under a number or different constraints.

 

The main task is to define the ultimate measure with a minimum of a priori assumptions. This is a similar procedure I had in mind when I was thinking on howto properly simulate a cell. The main problem was to find the ultimate measure of success. Once this is found, it's merely a technical problem to find the best path.

 

While I am not religious, I think that if there was a god who created this, all he had to do was to defined the magic universal measure of life, and leave the rest to itself, and there would be sponatenous evolution.

 

/Fredrik

[Norman Albers]

Physics shows me more and more the unity in things: surely the most disunified thing around here is us. Also, there is fundamentally even in the grain of sand or in the vacuum, desire. This I see in the sense of the formation of a dipole fluctuation, or balancing regions of divergence. This is the "one stick" of Yin and Yang. The fact that there is such possibility of form at higher levels and such an unending dance of creation and destruction of form, this gives me the Blues for Allah.

[fredrik]

I agree breaking of vacuum symmetries and zero point energies etc is a key topic which is interesting/fun too. This would be the easiest test case where I hope to try out some new ideas eventually. I don't want to go back and contaminate my brain with the standard procedures yet... I am trying to take advantage of that fact that I am reset now... and I try to rethink critically without falling back into the standard lines of thinking. I am trying to rework all of it in a purely sort of probabilistic/learning model approach without overly silly prior assumptions. I want to view physics as a distinguished application in a general learning theory or information theory, or if possible formalize the scientific method. My instinct is that there is no doubt this is possible. It's just a matter of exactly how complex it will be and what the preferred formalism is.

 

In parallell I'm briefly scanning existing approaches to see if there are parallell projects going on that I like. In a certain sense alot of the current theories already tangent to all this... it's just that I don't see why it's not abstracted harder because I think there is alot of power in the approach. What bothers me is that there is a mix of old "classical mechanics" stuff with good new ideas and this mix makes the logic blurred as far as I can see. This is why I want to clean out anything that's out of faishon and start from square one.

 

I think that correspondance principle might follow if the new theory is trained in the classical domain so to speak. There should not need to be a priori assumptions hard coded in it to fake classical mechanics in the limit. I think it will be a matter of training.

 

/Fredrik

[Norman Albers]

Too bad there is no Institute nor any support for those of us who dare to try to think further. In the words of Mother Theresa, it is between you and God. Observe the response to my "open question" on QM. I wish you good meditations.

[fredrik]

> it is between you and God

 

Even though I am not one bit religious myself, I like this statement in the general sense.

 

At this point, your QM questions seems to be formulated - from my point of view - somehow in the "application layer". And to not blurr my own thinking up, I will try to not enter these things until I reworked the basics (in my own head that is

 

Right now I am on a very abstract general level, and trying to define things like the event space, and how to treat the conjugate event spaces properly.

 

I am not sure if it's because I am stupid and everyone thinks it's too obvious to mention, or if they are admittedly ignored but to take and example I tried to extract the philosophy of the CDT approach in one of the links posted by Martin, and already in the foundations there are some things that make me leery, and I need to verify if it's valid.

 

In a very general sense, the of feynmann pathintgerals (conditional or transitional amplitudes) can be considered as nothing but a summation or a complete partition of the event space - this partition is by definition arbitrary and could be exploited to simplify the formalism, which I supposed could be the exploit in their philosophy, however in the treatise I sense some assymmetry in the treatise when time comes into play... there is also extensive use of the classical mechanics "action" as per the feynmann postulates. These jumps need to be motivated because they are not obvious to me at least. I also need to analyse their incorporation of the causality constraint. My first simple impression on reading was that it was used in choosing a partition that "a priori" was incomplete. But I am not sure if that was the right interpretation. The way they present the stuff in the papers are in my mind lacking philosophical argumentation, so I feel that I have to recrify everything I read into my preferred thinking.

 

The definition of event space, and treatise of time is crucial... and already at this level we make assumptions.

 

Since I am rethinking at this basic level I haven't participated in any higher level discussions yet. The whole "concept" of photon localization and the concept of energy is (in my rethinking here), not yet formalized.

 

My motivation is that I think alot of the headache are founded at this low level. This is why I consider it well worth the time to rethink this.

 

When this is in place, and only then, will I try to reconsider the applications and particle definitions.

 

/Fredrik

[Norman Albers]

I am just approaching this level and quite enjoy what you say. I do not yet know what a "momentum loop" is, for instance. I know Planck's constant is in units of angular momentum and its quantization seems basically loopy to me.

[fredrik]

My main concern in what things are, isn't what are they are mathematically, it's more what is it, defined in terms of reality and knowledge. Although this connection is bound to be fuzzy, it is still important. Attaching the formalism to reality does require some philosophical argumentation, some sloppy half hearted analogies isn't good enough.

 

I do not like when mathematical concepts are pulled out of nowhere, with from a philosophical point often very vauge/weak connections to the real issue, and then a machinery of mathematical excercises are initiated and all further work is made with blindfolds. I find such treatments both unreadable and objectionable. It is more like something I would expect from matematicians, that find physical models "fun to play with" After all there are plenty of researchers at mathematics departments where they don't do physics, but rather simlpy "study the mathematical models suggested by physicists". Which is of course excellent and necessary, but that is sort of only part of the issue. I remember one of my old math teachers who if I remember exclaim complaints about the quality of the math seen in physics papers. His expressio was that "someone has to do it properly", and indeed when it comes to evaluating the mathematical properties of certain objects, that is indeed a job more a mathematician. He was interested in the models, not the purpose or usability of the models. And I think everyone should follow his instinct.

 

There has seemingly been a mutual synergy between mathematics and physics, which is important.

 

I like mathematics, but I am mainly a philosopher to mind, and has always been. Stringent formalisms are important, but the joy of playing with formulas must not take over the original purpose of someone that studies reality.

 

/Fredrik

[Farsight]

I agree with the sentiment, fredrik. Mathematics is a vital tool, but sometimes it seems like 'the only tool in the box and it's got no handle. So we can't get a grasp of what we're dealing with. I like to think ontologically rather than mathematically myself.

 

Ditto, Albers. I have problems with quantization of anything to do with rotation. And when I look at action and think about energy multiplied by time, or momentum multiplied by distance, I think of energy = stress multiplied by volume, of stress = force times area, of c = a distance/time conversion factor, of kinetic-energy stopping distance, and momentum stopping time. It seems to be telling me I'm looking at the same thing from different viewpoints. I want to grasp that thing via the ontology of what's there, not get sidetracked into mathematical quanta and mind games telling me that things jigger when they rotate.

[Norman Albers]

It is not a mathematical game that charges always give the same eigenvalue regardless of what sort of "particle" you speak of. I do not appreciate your viewpoint because I do not play idle mathematical games. I feel quite nicely how things are on the mental monkeybars of which I speak. There may not be much for us to share here.

[fredrik]

I don't want to obstruct my rethinking by advancing myself but my view of the various quantization approaches is that it simply depends on what partitioning of the event space, and prior you make.

 

The simple canonical way "[math]p = -i\hbar\frac{\partial}{\partial q}[/math]" is just one way, inspired by the familiar fourier theory, and historically I guess a natural way. Logically I think it is by no means distinguished and the only way.

 

We define a logical prior which is information about p, to evaluate the conditional probability of q. I think of the quantization in this case is a plain conseqeunce of our choice of partitioning of the event space and "prior event space".

 

There is nothing that stops us from defining a mixed partitioning that could be anything we want, as and alternative to the normal (q,p) phase, if we think that makes our treatment easier. As long as we do it properly they should in a certain sense be equivalent, so changing partition IMO I think should not need to involve "assumptions" as such. It's just a change of representation, we are free to make.

 

I didn't explain everything in detail here, but IMO these things can grasped on a principal level, and this I find important for internal guidance. The formalism itself is just a language, but it's meaning can be seen without the full formalism.

 

So far I think it's not so overly weird, but I think that this alone, is not sufficient. We need to go yet deeper, or alternative iterate the same logic used a few times.

 

I will have no choice but to analyse this in more detail eventually.

 

/Fredrik

[Norman Albers]

This is nice. I have been questioning the underpinning of Fourier analysis. It all seems hunky-dory, I am mathematically comfortable up to two points: are we violating causal implications by involving infinite wave composition elements, and, as always, GIGO, referring to the event space of the model. I am developing an expression of the vacuum response as superconducting, which is equivalent to QM vacuum polarization. Black bodies are quantized but I'm not convinced enerything else is.

[fredrik]

In my thinking, causality is treated on a logic or probabilistic level, not "physical level". I think of the "physical" level as built ontop of the abstract information level. I see different interpretations of causality. Locality OTOH is yet another story. These things will in my thinking at least enter the model higher up in the development. At the most basic low level treatise, there is no such thing as locality or the cloesest thing to causality I can think of is the obvious fact that the prior state influences the evoltion into the posterior state. This would generate a causal like behaviour, at least locally. But the details remains to be elaborated. I believe strongly in the suggested route here and I intend to try and walk it and see where it takes me.

 

/Fredrik

[fredrik]

> at least locally

 

bad word... what I meant instead of locally is "with respect to a particular observer", but OTOH the nature of the observer is sort of incorporated into the priors.

 

/Fredrik

[fredrik]

GIGO, referring to the event space of the model.

 

I'm sure if I understood you right here but I agree that the *probabilistic event space* must be defined better. This is one major point indeed. To simple take some classical physical spacetime concept as part of eventspace of particle positioning is wild and I think is plain wrong, at least in the general case, and at this basic level I find it unacceptable to step aside from the so called "general case".

 

That was the focus on the first post in this thread. I am not sure yet what the proper abstraction is, but I am working on it.

 

/Fredrik

[Norman Albers]

Garbage In, Garbage Out "I want to grasp that thing via the ontology of what's there, not get sidetracked into mathematical quanta and mind games telling me that things jigger when they rotate." Farsight, here your I/O rate exceeds your CPU process.

[fredrik]

are we violating causal implications by involving infinite wave composition elements

 

While I'd like to avoid the word "causal" at this point, when I read your comments again I think this is a good question and maybe I misunderstood your comment... there is something to it, one way or the other. No doubt the boundaries and discreteness are related too... the boundaries and general properties of the samples space will definitely have an impact on the relations too. This is why the whole issue of the event and samples spaces should be carefully considered, and the properties of the basic set will have a big impact on the rest (and probably even also dictate which method of quantization to use, who says we have to stick to only one scheme?), and also on convergence properties of any expansions.

 

There are several important issues at this point that isn't clear to me and need work. One complicating factors is that my instinct suggest that the event spaces themselves are somehow not fixed, they should be "dynamic" or changing and evolve in response to changing information. Now, we only have to get this to fit together.

 

/Fredrik

[Farsight]

It is not a mathematical game that charges always give the same eigenvalue regardless of what sort of "particle" you speak of. I do not appreciate your viewpoint because I do not play idle mathematical games. I feel quite nicely how things are on the mental monkeybars of which I speak. There may not be much for us to share here.

 

Garbage In' date=' Garbage Out "I want to grasp that thing via the ontology of what's there, not get sidetracked into mathematical quanta and mind games telling me that things jigger when they rotate." Farsight, here your I/O rate exceeds your CPU process.[/quote']

 

Your comments duly noted.