LHC - the last chance for all theories of everything?

[Dmitry67]

From a computational-complexity point of view there is no difference between the two statements
a) one universe exists
or
b) all possible universes exist - except one
Of course specifying the one in both cases should require the same amount of information.

Similarly computationally the statements
a') no universe exists
and
b') all possible universes exist
have the same complexity.

If you argue strictly based on computational complexity and with Ockams razor, then you cannot distinguish between a') and b') nor can you distinguish between a) or b). So in order to do that you have to refer to some experimental result, i.e you have to enlarge your reasoning beyond computational complexity.

You do that by taking into account "known facts" and by going from a mathematical to an ontological reasoning. Of course you can rule out a') immediately. You can trust in a least one universe which means that ontologically and based on Ockams razor you should not add more assumptions than strictly necessary. So you add the assumption that one universe exists and - wow - it agrees with the known facts. If you would instead add an uncountable set of universes it would still agree with the known facts, but you have to explain why "they are there but invisible". Therefore you add assumptions which you cannot prove.

And don't forget: you cannot use your "mathematical multiverse hypothesis" as starting point for your proof as your quest is just to prove exactly this "mathematical multiverse hypothesis".

We have almost agreed :)

Yes, empty set and universum contain the same amount of information - ZERO
So our multiverse contains NO information (of course not for "frogs") as it is Universum and it always exists in one and only one possible state.

And yes, my reasoning is:
2 possible options: NOTHING or EVERYTHING, we observe SOMETHING, hence the second option is valid. Note that "assumption that one universe exists" is not an assumption - it is an experimental fact :)

Your last point is valid: I use MUH as an axiom, because I don't see any choice.

I have 2 questions if you have time:
1. Do the inner regions of black holes (inside their horizons) exist? Are any claims of what is going on there falsifable?
2. Do you think that TOE would use only formulas or it would need to use some words? In another words, do you expect TOE to be interpretation-less? What do you expect TOE to do with the interpretations? (give a new one, tell us what is a right one, making the whole notion of interpretation a nonsense)?

 

[tom.stoer]

1. Do the inner regions of black holes (inside their horizons) exist? Are any claims of what is going on there falsifable?

Yes, inner regions do exist and in principle they are accessible for an experiment.

Think about a brave colleague (an expert in theory, experiments and results regarding GR and QG). His last project (together with his students) is to travel into a gigantic black hole and to spend the rest of his life in order to complete the knowledge regarding black holes. Within this large black hole he will - besides his experimental activity - publish research papers, create an own arxiv repository, organize conferences etc. Of course he has to hurry up, but in principle nothing will be in conflict with his objective.

2. Do you think that TOE would use only formulas or it would need to use some words? In another words, do you expect TOE to be interpretation-less? What do you expect TOE to do with the interpretations? (give a new one, tell us what is a right one, making the whole notion of interpretation a nonsense)?

I am not so sure.

A short-term candidate will certainly not be free of interpretation and non-mathematical rules ("if X ... then apply rule Y"). Even the basics of that theory may be easier to formulate in english language ("the speed of light is constant in all reference frames").

A long-term candidate may exist on the "next stage"; it could be some kind of meta-theory that addresses these formalization-issues; but perhaps it will aim for something totally different - today we do not know.

I expect that a candidate ToE will be based on few (hopefully :-) mathematical and non-mathematical "axioms". It will be able to address questkions like:
- what is the structure of space-time
- how does space-time, local gauge-symmetry, ... arise from X (where X is a more fundamental structure)
- why do we live in 3+1 dimensions
- why do we observe U(1)*SU(2)*SU(3)
- why do we observe three fermion generations, ...
- what is the origin of the cosmological constant
- what replaces the black hole and big banhg singularity, ...
- ...

After we are able to study one candidate ToE, of course new questions will arise, especially regarding the mathematicsl structure(s) X. Perhaps there will be competing structures X, X', ... (strings, loops, non-commutative geometry, ...). I do not know if this will be an iterative process (just as it was the last centuries) or if there will eventually be a paradigm shift which allows us to go the meta-theory level.

I agree that we need some deeper insights regarding development of physical theories, especially as we will have less new experimental facts - just because of the required energy scale. That means we must focus more on abstract ideas, mathematical justifications, completeness, consistency etc.

As you say, something like MUH will be an axiom (you could even say dogma, doctrine) and will therefore belong to / be related to the philosophical realm. I am pretty sure that physics and philosophy will again come closer together has they are today (there was an intensive dialog in the early twenties when QM was developed; in the following decades "shut up and calculate" became the major guideline.

 

[Dmitry67]

1 I expect that a candidate ToE will be based on few (hopefully :-) mathematical and non-mathematical "axioms".

2
- why do we live in 3+1 dimensions
- why do we observe U(1)*SU(2)*SU(3)
- why do we observe three fermion generations, ...

1 I always asked - but never got an answer - for an example of a "non-mathematical" axiom. An example of an axiom which can not be expressed in mathematical terms. Could you provide any examples (even not realistic, not about our universe)?

2 What kind of "why" do you mean?
a. anthropological principle
b. cosmic darwinism
c. proof of mathematical inconsistency of other possible universes

 

[tom.stoer]

Regarding a non-mathematical axiom: take constant c in GR as an example: as long as you do not have the full developed GR based on manifolds it's hard to write this as a mathematical axiom; of course you can start with a local description which will eventually correspond to the tangent space, but you don't know this in advance. So take this as a an example.

Regarding "why": the basic problem is that it's easy to ask these why-questions, but it's hard to answer them; therefore I don't think that it's a good idea to exclude ideas at this early stage. So right now I am trying to be open-minded to all three directions. My problem with
a) is that it appears to me as self-immunization against falsification
b) is that it sounds nice but nobody can tell me how evolution, mutation and evolutionary pressure could work; what is the DNA of the physical laws?

Of course I cannot prove that other universes do not exist; nevertheless I would take c) as starting point and search for
i) a theory that describes this specific universe and
ii) a selection principle (could be consistency or something else)
[this would be MY research program - I don't want to force anybody to abandon other lines of research]

 

[tom.stoer]

@friend & fra: I think you are trying to do something like "finding the DNA for physical law".

I have no idea how I would start and I am not familiar with the ideas and research programs you are mentioning. However, it becomes clear that even if you want to avoid axioms and rules but try to let physical laws emerge from something "deeper", you need a set of rules. Logic is perhaps the simplest mathematical structure, but I am afraid that logic alone will do the job. And negotiation must follow some logic, too. So it's at least logic that serves as a basis for you.

One remark: I am still not sure if we all understand what Tegmark wants to say. What does the small word "IS" mean? Is it absolute identity, not only isomorphism? Are the U(1) and the SUO(2) identical - or just isomorphic? Do they become strictly identical if I remove all "human baggage"? Is it correct, sufficient and reasonable to assume that there IS NOTHING ELSE but a relation between mathematical entities? He is not very explicit when it comes to relations to different philosophical schools ...

 

[Fra]

So it's at least logic that serves as a basis for you.

Sure, of course I kind of rely on some kind of logic in the general sense, but I thought that from the context the association to logic here was to ideas that mathematical consistency and deductive logic. Ie. that you can get to KNOW the certain laws by though along - no interaction. I beg to differ with that view.

The kind of logic I do rely on is loosely speaking inductive logic, not deductive logic. Clearly the traditional quantification of inductive reasoning is probability theory. Ariel and Jaynes makes this point strong. However, even the rules of inference themselves are not unique. Here I differ with them.

They formalise inductive reasoning, into probability theory and then use various bayesian or max entropy methods as the RULES of inference. But of course, the rule of inference is chosen and the isntant you choose the specific entropy measure. Similarly there are objections to bayes rule.

I am trying to generalise inductive reasoning, by suggesting that by taking the proper intrinsic view, other rules of inference other than bayesian and max ent method are possible, and bayesian and max end methods with a fixed entropy measure are NOT the optimal inferences. Sometimes they are of course, but it's not a general case.

I have hopes that quantum logic to mention one think should b4e satisfactory explain as a unique choice of optimal inference RULE in particular situations! But again, but understanding hte general case, I also expect to understand the generalisation of Quantum logic, which will help solve QG problems and unification.

So in a certain sense, I am looking for a mathematical reconstruction, but it is not possible to understand the motivation from a pure mathematical perspective. Also the ides suggested does contain self-referential elements, and the this self-interaction should amount to a kind of self-inference, a kind of self evolution.

I think a correspondence here in simple case would be that the schrödinger equation is really just the expected self-evolution, or the self-inference. The optimal inference when external feedback is taken into account is the collapse thing.

In this view there is as I see it no mystery with the collapse at all.

Given that I want to take this further than Ariel and Jayes, who basically reconstructs the same old continuum probability theory and use that as a basis for inference, one of my basic conjecture, is like theirs that the laws of physics ARE more or less the rules of rational inference. And the point is then tht the optimal inference is a matter of point of view, since the instrinsic view allows no external measures of optimality.

There are also the symmetry principles hidden here, symmetries are emergent as a result of interactions, and are not fundamental. Understanding the interactions here should in my expecation help explain why certain symmetries in the rules of rational inference are selected. And thus the symmetries of physical law.

I have understood that this is hard to convey. Having thought of this now for a new years I think the conceptual part is becoming pretty clear, but still I see that not many seem to connect, with a few exceptions. Probably because I do not know of any current papers that does exactly this. The related ideas are from smolins, evolving law, ariel caticha and ET jaynes, as well as some other. Olaf Dreyer has partly acknowledge the inside view.

I suppsed this will remain foggy until substantial progress is made.

Edit: I also associate time evolution with the inference processes. The relativity of time in relativity should be reproduced from the relativity of inference, as in the emergent symmetries. Note that both Jaynes and Olaf Dreyer belives that GR could be DERIVED from the proper reconstruction. I fully share this view, although I have a different view of the starting points. Instead of derive, I prefer to say emergent, and this emergence is a physical equilibration process.

/Fredrik

 

[Fra]

Obvious this talk about optimal inference and selection for inference rules, what is the assumption on the seleciton?

My conjecture is that the selection is the obvious one - rules that are destructive are disfavoured and those that are self-preserving and constructive (that grow more competitive) are selected for.

So the inference rule is the DNA of physical law IMO. But the DNa can not be fudnamental, clearly the DNA itself must have evolved as well. Thus I more think of different levels of this code.

So the rules of inference that are selcted in our universe and those that are optimally self-preserving. Ie. the game preserves itself.

I'm trying to model this by pondering how to construct optimal measures, that are basis for actions, given limited resources.

My starting point is the low end of the complexity scale, becuase here the options are finite, and then ponder what structure emergenes as the resources are scaled up. During this journey spacetime and it's symmetries should follow - I hope.

/Fredrik

 

[tom.stoer]

You have to define how a rule looks (or a law or whatever) like, and you have to define how the negotiation between rules (which are mathematical entities) looks like.

Can you explain how this approach could look like mathematically?

What are your symbols, relations, axioms etc.? How does a rule look like?

You have to define how "rules act on rules": you have a negotiation process for which you need rules; "rules acting on rules" can therefore be negotiation between physical laws, but it can also be the evolution of some physical entity.

How and when do which rules interact? How are two (or three? four?) interacting rules selected? How do they "come together"? How does the DNA look like? How does mutuation, crossing-over and spawning of new rules look like?

How do you count rules or members of classes of rules in order to decide which rules are successfull = dominant?

What will our universum be? One master rule or a colletction of the most successfull rules?

Don't you need a meta-rule which "initiates" this whole process - which then becomes part of it and is subject to negotiation as well? what are your initial conditions?

 

[Dmitry67]

Regarding a non-mathematical axiom: take constant c in GR as an example: as long as you do not have the full developed GR based on manifolds it's hard to write this as a mathematical axiom; of course you can start with a local description which will eventually correspond to the tangent space, but you don't know this in advance. So take this as a an example.

Could you clarify?

c=1

so I don't see any problems with it. However, it reminds me about the question I wanted to ask about the parameters of the Standard Model. How do you expect the axctual values be explained by the TOE?

Say, Mass of Up Quark / Mass of Down Quark?

1. Some analytical expression (even very complicated) derived from TOE, say Mu/Md = sin(ln(2*pi) / sqrt(e))

2. It is just a parameter. It is an axiom. (possible justification using AP)

3. It is a parameter but it can vary (cosmic darwinisim, or alternatively superstiring theory with bulk, colliding branes giving birth to the universes with all possible combination of parameters, then AP)

 

[Dmitry67]

One remark: I am still not sure if we all understand what Tegmark wants to say. What does the small word "IS" mean? Is it absolute identity, not only isomorphism? Are the U(1) and the SUO(2) identical - or just isomorphic? Do they become strictly identical if I remove all "human baggage"? Is it correct, sufficient and reasonable to assume that there IS NOTHING ELSE but a relation between mathematical entities? He is not very explicit when it comes to relations to different philosophical schools ...

Let me quote him (chapter Description versus equivalence)

Whereas the customary terminology in physics textbooks is that the external reality is described by mathematics, the MUH states that it is mathematics (more specifically, a mathematical structure). This corresponds to the “ontic” version of universal structural realism in the philosophical terminology of [14, 22]. If a future physics textbook contains
the TOE, then its equations are the complete description of the mathematical structure that is the external physical reality. We write is rather than corresponds to here, because if two structures are isomorphic, then there is no meaningful sense in which they are not one and the same [19]. From the definition of a mathematical structure (see Appendix A), it follows that if there is an isomorphism between a mathematical structure and another structure (a one-to-one correspondence between the two that respects the relations), then they are one and the same. If our external physical reality is isomorphic to a mathematical structure, it therefore fits the definition of being a mathematical structure.

If one rejects the ERH, one could argue that our universe is somehow made of stuff perfectly described by a mathematical structure, but which also has other properties
that are not described by it, and cannot be described in an abstract baggage-free way. This viewpoint, corresponding to the “epistemic” version of universal structural
realism in the philosophical terminology of [14, 22], would make Karl Popper turn in his grave, since those additional bells and whistles that make the universe nonmathematical by definition have no observable effects whatsoever.

I 100% agree and have nothing to add.

 

[tom.stoer]

Could you clarify?

c=1

you have to embed this statement in a context which you call "baggage". w/o context c=1 is meaningless; therefore you have to define the context mathematically as well. But no you have the problem that you
either have to specify (via an axiom) that spacetime is a 4-dim pseudo-Riemann manifold (but this is certainly not a nice and easy-to-beliece axiom)
or you have to find simpler axioms for which a formal definition makes sense.

Einstein had the c=1 "axiom" in mind but derived the context (pseudo-Riemann manifold) later. That's why I still think that it's a good example, at least for GR (possibly not for the ultimate theory).

How do you expect the axctual values be explained by the TOE?

Say, Mass of Up Quark / Mass of Down Quark?

1. Some analytical expression (even very complicated) derived from TOE, say Mu/Md = sin(ln(2*pi) / sqrt(e))

2. It is just a parameter. It is an axiom. (possible justification using AP)

3. It is a parameter but it can vary (cosmic darwinisim, or alternatively superstiring theory with bulk, colliding branes giving birth to the universes with all possible combination of parameters, then AP)

I don't know; the ToE must deliver both the value and the way how and why this value is produced:-)

1. I guess it will not be an analytical expression, but of course it could be the (implicit) solution of an explicit equation.

2. no!

3. Could be, but then I would prefer an answer why it's this value or at least some range. If somebody claims a kind of evolutionary process then she/he must specify the rules regarding evolutionary pressure, selection, spawning of baby-universes, cosmic DNA and all that.

Compare it to evolution in biology: Darwin had a couple of ideas (e.g. "survival of the fittest") and some mechanisms for selection (population, predators, ...). In the meantime we were able to figure out the rules for the DNA (at least partially).

I would expect something similar for a ToE claiming that a specific theory (or parameter set) emerges from an underlying structure (multiverse or whatever). What I have seen is that nobody was able to answer these questions so far. In string theory nobody is able to construct M-Theory, nor has anybody developed a clear idea what the mathematical structure of the landscape is, nor is there some kind of measure on the multiverse (you need a measure to count the population ...), nor do I see a clear prediction regarding spawning of baby universes etc.. There are some nice ideas, but sooner or later people start to wave their hands and cry for the anthropic principle.

The whole discussion did not start because of a clear fundamental principle but only because people where not able to do the calculations!

 

[tom.stoer]

Let me quote him (chapter Description versus equivalence) ... I 100% agree and have nothing to add.

OK, I got the point (missed it when I was reading the paper).

So he says that iff two entities are isomorphic to 100% and in all their aspects and properties, then they are identical. Therefore iff the universe can be described in pure mathematical language w/o any baggage, then the universe IS this mathematical structure - and the mathematical structure IS the universe.

I agree that from a purely mathematical point of view this is sound!

Of course it's misleading to discuss the simulation approach, as this approach explicitly introduces a meta-level to the simulation (the program plus its output), namely the simulation engine (HW, operating system, ...). The paper could very well live w/o these remarks.

 

[Dmitry67]

Of course it's misleading to discuss the simulation approach, as this approach explicitly introduces a meta-level to the simulation (the program plus its output), namely the simulation engine (HW, operating system, ...). The paper could very well live w/o these remarks.

Well, he actually denies that it makes sense to talk about the simulation:
check pages 18-21

 

[Dmitry67]

you have to embed this statement in a context which you call "baggage". w/o context c=1 is meaningless; therefore you have to define the context mathematically as well. But no you have the problem that you
either have to specify (via an axiom) that spacetime is a 4-dim pseudo-Riemann manifold (but this is certainly not a nice and easy-to-beliece axiom)
or you have to find simpler axioms for which a formal definition makes sense.

Einstein had the c=1 "axiom" in mind but derived the context (pseudo-Riemann manifold) later. That's why I still think that it's a good example, at least for GR (possibly not for the ultimate theory).

In SR and GR there is no 'c' if you work in Planks units.
"c" is a thing invented by humans, it does not have any fundamental meaning.

Like people used degrees to measure angles. But in mathematics it is more natural to use radians. So conversion constant for conversion from grads into radians does not have any fundamental meaning.

But again, any claims that c=1, x=3, space has 4 (10,11,26) dimensions can be encoded in a pure mathematical language. That is why the idea of Max Tegmark is so solid: I had never seen anything that could qualify as "physical" axiom. Something when can not be expressed - in principle ! - in mathematical language.

The only candidate is Smolin's "evolving law". It is pure handwaving without a single formula :) Sorry Fra

 

[tom.stoer]

In SR and GR there is no 'c' if you work in Planks units ... "c" is a thing invented by humans, it does not have any fundamental meaning.

Really?

"c" has two meanings:
- it is a fundamental constant with the unit "m/sec"; this is somehow invented by us humans
- it is the speed of light (or better: propagation of signals); this is a result of the theory

It is interesting that even in GR locally the light cone = propagation of signals respects "c=1". That's not trivial but emerges from the theory.

Compare it to mass. You can argue that with "Planckmass=1" the meaning of mass dissapears. Still, GR is based on the assumption - and it reproduces this assumption - that inertial and gravitational mass are equal. This is non-trivial and emerges from the theory (I think it is not understood to 100% as GR does not always allow for an unabmiguous definition of mass).

Therefore I think that these two examples serve as non-mathematical axioms.

 

[Dmitry67]

Really?
1
- it is the speed of light (or better: propagation of signals); this is a result of the theory

2
It is interesting that even in GR locally the light cone = propagation of signals respects "c=1". That's not trivial but emerges from the theory.

3
Compare it to mass. You can argue that with "Planckmass=1" the meaning of mass dissapears. Still, GR is based on the assumption - and it reproduces this assumption - that inertial and gravitational mass are equal.

1 if it is a RESULT of a theory then it is NOT an axiom!
2 same
3 Well, GR is not a final theory. In TOE mass is a tricky thing (HUP, virtual particles, Unruh effects-accelerated frames etc). So that equivalence should emerge from TOE as a result. Note that in QM equivalence principle does not work on short timescales (because time and mass do not commute) - another proof that that principle is not fundamental.
But even if we forget it we can still write Mg=Mi :)

Sorry, I don't see any physical axioms.

 

[Fra]

You have to define how a rule looks (or a law or whatever) like, and you have to define how the negotiation between rules (which are mathematical entities) looks like.

Can you explain how this approach could look like mathematically?

What are your symbols, relations, axioms etc.? How does a rule look like?

You have to define how "rules act on rules": you have a negotiation process for which you need rules; "rules acting on rules" can therefore be negotiation between physical laws, but it can also be the evolution of some physical entity.

How and when do which rules interact? How are two (or three? four?) interacting rules selected? How do they "come together"? How does the DNA look like? How does mutuation, crossing-over and spawning of new rules look like?

How do you count rules or members of classes of rules in order to decide which rules are successfull = dominant?

What will our universum be? One master rule or a colletction of the most successfull rules?

Don't you need a meta-rule which "initiates" this whole process - which then becomes part of it and is subject to negotiation as well? what are your initial conditions?

I think I've possibly got the swine flue or something and I feel a bit lame and fevery today. I'm supposed to go on a business trip on thursday so I'm hoping to rest and get rid of the fever before that.

I'l try to comment more later.

I've never tried to seriously present much of the specifics here for several reasons.

- As I understand you're not allowed to post personal research, except possibly in the indepdendnt research section.

- There is soo much left to do, that I consider it my own problem to sort it out. This is work in progress but I'm an amateur and progress is slow on hobby basis.

This is why I mainly try to *discuss* things at just an intellectually sound and conceptual level, which is within the guidelines as I understand. Another reason for this is that this is the motivation also for my CHOICE of mathematics.

I'll try to give you a some more hints later.

/Fredrik

 

[tom.stoer]

Sorry, I don't see any physical axioms.

They are there - directly in front of your eyes.

The weak equivalence principle = the universality of free fall = the equality of inertial and gravitational mass is a physical, non-mathematical axiom of GR. It has been formulated w/o a mathematical framework. Einstein then derived this framework = GR from which this principle (as a law) did emerge. The equation Mg=Mi as a starting point is mathematically nice but physically meaningless w/o specification of the framework.

 

[tom.stoer]

... so I'm hoping to rest and get rid of the fever before that.

Get well soon!

There is soo much left to do, that I consider it my own problem to sort it out. This is work in progress but I'm an amateur and progress is slow on hobby basis.

I fully understand your situation as mine is similar :-)

Thomas

 

[Dmitry67]

They are there - directly in front of your eyes.

The weak equivalence principle = the universality of free fall = the equality of inertial and gravitational mass is a physical, non-mathematical axiom of GR. It has been formulated w/o a mathematical framework. Einstein then derived this framework = GR from which this principle (as a law) did emerge. The equation Mg=Mi as a starting point is mathematically nice but physically meaningless w/o specification of the framework.

no and no.

Equivalence principle, mach's principles etc were just a MOTIVATION to create a theory. It is cristally clear if we take SR:

2 axioms (in fact, observational fact) -> einsteins version of SR -> deeper understanding of spacetime (Minkovsky) -> 2 'axioms' now are derived from the formalism.

Lets return to Mg=Mi. Yes, it was an oversimplification. But if you insist... What EP says? That gravitational force is proportional to the mass? So, we get Mg/Mi and mass dissapears from the formula for the acceleration? Hence, in weak g fields objects fall with the same acceleration, correct?

Lets make it more formal. Fall=move. So 2 objects with different mass starting from the same point of 4D spacetime move by the same worldline trajectory, correct? But this result is rather trivial if we look at gravity as curved spacetime. How else it could be?

But even if was not trivial it is just a statement regarding the form of a worldline. So we get rid of the baggage about 'object', 'fall', etc etc. It is just mathematics and nothing more.

 

[tom.stoer]

Sorry, but it's hairsplitting to ask for non-mathematical axioms and to explain later that what I gave you are not axioms but principles ...

You do not have to explain how the mathematical formulation based on the world line of an object emerge from the theory:
1) I know how it works
2) it's exactly my statement:
- you start with an "axiom" (you call it principle)
- derive the full theory in mathematical language
- and check how your axiom or principle is represented in the full theory

I agree that you can get rid of the baggage in the full theory, but I don't see how to manage it in terms of axioms. Can you formalize the Peano axioms?

 

[Fra]

Some things to get the right background.

- Like I said before, I share a good deal of the sentiment of Jaynes and Ariel. Their "mathematics" is essentially that of probability theory. I am reconstructing another formalism, that is IMO a generalistion of probability theory as the answer to inductive inference.

- I share the vision of Ariel and laws of physics can be seen as rules of inference, however I differ in that I complicate their picture by suggesting he the inference system is subject to evolution. unless it's already obvious the analogy goes like this...

Standard physics is usually formulated as an initial value problem, where you have a initial state in a given configuration space, from which the future state follows from the laws of physics (equations of motion, or schrödinger equation etc). Not that this is a deductive logic.

premise ~ initial conditions, or the current state
deductive system ~ laws of physics

In the inductive inference view, the idea is that there exists no certain deductive system, only an inductive system. But you can still have different views on induction.

For example unpredictable phenomena like QM, where you can not deduce the outcome of an experiment, can then be said to be an form of induction, but one can easily describe this as a form of probabilistic deduction, or almost equivalently a deductive probability.

So QM still fits the deductive inference model if we accept that state of matter are only statistical.

But we have implicitly used probabiltiy theory and statistics here quite uncritical. But as I've argued several times, and like Smolin argued in his argument "against timeless laws" - which in my opinion is also an arrgument against deductive inference - this model is not universally sensible. It makse sense when we study a small subsystem, but not when a small system studys it's own environment.

So, the idea is that the laws of physics would be identified as the laws of inference. But since I reject the validity of standard probability, what do I suggest instead?

I suggest to replace flat usage of probability theory with a reconstruction of an intrinsic measure that is really a way to count evidence. Actually this is the original way Jaynes argued, he pondered over how to count evidence, and represent degrees of belief. Philosophical arguments was translated into axioms, when ultimately lead to a formalism that is standard probabiltiy theory. This is in his book "probabiltiy theory - the logic of science".

IMO, he makes some mistakes that makes me reject the reconstruction. He introduced the continuum too lightly, that's the first mistake, and it proves to be a key one.

You have to define how a rule looks (or a law or whatever) like, and you have to define how the negotiation between rules (which are mathematical entities) looks like.

Can you explain how this approach could look like mathematically?

What are your symbols, relations, axioms etc.? How does a rule look like?

To stick with close terminology to show that it's close to probability theory, I replace probability distrubiotn with microstate, and the space of distributions with microstructure.

I work with finite sets of natural numbers, to represent microstates.
Also, every microstructure has a complexity number, I call M.
Every microstructure has a eventspace volume k, this is the number of disntighuishable types of evidence/events.
The numbers of each set sum up to this complexity number.
The microstructure is the set of these sets, with complexity M.

From this point what I have is a discrete probability, where not only the event space is finite but - more importantly - the probability itself is discrete/quantized simply because there is no finite representation of the continuum.

similarly, by combinatorical considerations, you can define a natural way to measure one microstate relative to another one. If you calculate the probability for one microstate taking another microstate to be the prior, then in the expression you can separate information divergence and the complexity of the microstate in an interesting way.

You get P = w e^{-M{S_{KL}}} where S_KL is the information divergecen between the states, which is independent of complexity. w is a factor that scales with complexity, but w -> 1 asymptotically approaches as complexity -> infinity.

Thisi s very simple, but this is the seed to what later will generate physical actions. The least action principle is in this view, simply the principle of maximum probability, except that in my view, the "probability" is an INSIDE counting of evidence. And in the case of high complexity this also conicinced with the principle of minimum information divergence (minimum speculation), but this is not generally the case for LOW complexity.

Some postulates will be that for example the expected action of an inference system, is that which maximised the transition probability - this is the correspondence of the least action principle.

But this is just the starter, next I consider that an observer actually consists of sets of microsturctures (sets of sets) that have relations by means of transformations, which is really a form of recodign of information. Thus information (and evidence counts) can flow between the "spaces", and that an observer complex is infact a system of communicating spaces constructed in a special way.

THIS is where interesting things will happen. Becuase new logic appears here that does not comply to standard probability. For example quantum logic should be explainable as a result of inforamtion equilibrating between the dual spaces and their relation would explain how to make sense of out negotiations of X and P, when they belong to different spaces.

But there are a lot of opne things here for sure.

But so far we haven't even gotten to the evolution part yet. By my idea is to start descrive how ANY inside view should look like, this is possible since the options are finite if you start at the low complexity limit.

Then one can see which combinations of observer complexes that can coexist so to speak.

Note that I start with a distinguishability index - not a spacetime. space should also emerge as a preferred index structure, certainly I expect the dimensionality to be related to stability of hte complexes. But I am not there yet :) the problems is that several problem are related so it really doens't quite work to solve one problem at a time; this is why even my very approach to this is evolutionary, I work on all things at once and make broad but slower progress.

You have to define how "rules act on rules": you have a negotiation process for which you need rules; "rules acting on rules" can therefore be negotiation between physical laws, but it can also be the evolution of some physical entity.

How and when do which rules interact? How are two (or three? four?) interacting rules selected? How do they "come together"? How does the DNA look like? How does mutuation, crossing-over and spawning of new rules look like?

How do you count rules or members of classes of rules in order to decide which rules are successfull = dominant?

What will our universum be? One master rule or a colletction of the most successfull rules?

process - which then becomes part of it and is subject to negotiation as well? what are your initial conditions?

Alot of questions, all justified. but each comment would be be long, and still incomplete due to the nature of the incomplete progress... need to goto sleep now. maybe the previous comments helped also on some of the latter questions?

---
some final comments on how the reconstructed inference system would differ from the standard probabilistic and entropic on of jaynes and ariel:

Each distribution would come with a natural kind of "mass", thus information has mass.
This serves as inertia when these two distrubions are forced into negotiation.
This intertia also fills a purpose in the process of increasing the complexity of the observer - this will I think relate to gravity (certainty attracts more certainty)

More importantly the reconstruction comes with a natural information mesure, no need to postulate a specific choices of entropy. The natural information measure comes in the form of probability of probability in the framwork I reconstruct.

But the major point is that the meaure I reconstruct, is not to be interpreted as a frequentist thing, or rely on ensembles of systems (say ensenmbles of universes) it is simply an inside COUNT of evidence. You need no ensemble or repatable experiments. The new concept is a proper evidence count.

/Fredrik

 

[Fra]

Does anyone know what the latex formulas nowadays shows up as black unredable? I looked at old threads where they look fine?

Has the latex plugin to this site changed?

/FRedrik

 

[Fra]

Note that I suggest a closer match between the scientific process and physical processes.

About science, it's undoubtedly so that the deductive system we call physical law is inferred or abduced from our experiecen with nature. In a sense we can say that the inference system itself, is inferred, and constantly evolving.

This is why my program is not conceptually consistent with the realist view of timeless physical law. Or put differently, my view of inference is that there is not objective timeless reality to the inference system itself! so the inference system itlsef is a result of an inference, this is the circularity that is the evolution in my view.

/Fredrik

 

[Fra]

So the somewhat killing of realism that to a certain extent was initiated by the relational thinkers like Einstein and mach, and then QM, is not take take to it's full implication. I want to take this another level. And then away goes the realism of physical law.

I know a lot of people think this would crash all we know, and without fixed deductive systems we are toast. But this is not so. We can just remind ourselfs that human science has not crasched in despite of it's deductive system is beeing inferred, and sometimes renegotiated. The human brain certainly doesn't crash.

On the contrary would I say that the opposite - a fixed deductive system - would risk going into a halt! I consider this conceptually related to someone Christine Dantas wrote about in some old posts and on her blog (don't remember) - that nature avoids deadlocks. IMO, she expresses something that is very close to this issue, but perhaps put in different words.

/Fredrik

 

[tom.stoer]

I think I understand your point. Fact is that you have to specify some physical laws that apply here and now. Then there is the question how they can change in time (what is time?)

We must not mix the following approaches:
1) the believe in the existence of the moon even when nobody is looking at it
1') the believe in Kepler ellipses even when there are no planets
2) the description of (some of the aspects of) the moon by a mathematical framework

1) is the starting point for any kind of realism
1') is the starting point for something like structural realism
2) comes a a later stage and tries to map some aspects of 1) or 1') into a mathematical framework

What Dmitry67 is discusses is to identify 2) itself with something like 1') and 1) As I tried to explain I am not convinced.

As I understand

... And then away goes the realism of physical law.

it is an objection against 2) and some interpretations, not against 1) itself.

Now you can say that its useless or even nonsense to believe in 1) if you can neither be sure that 2) applies nor be sure that 1') applies NOW. My believe is that at least 1) will always survive as long as you do not go for solipsism.

So at which level does you inference "happen", 1), 1') or 2)

 

[Fra]

I'll comment more later. I've better from the flu and just got back from a business trip to paris. The internet at the hotel was a disaster, it was working on and off intermittently and I got fed up with trying to use it. Ontop of that they had multiplexed the phone on the ethernet wires in the building so when I plugged in the laptop the phone didn't work. No wonder the connection was unstable.

/Fredrik

 

[Fra]

I think I understand your point. Fact is that you have to specify some physical laws that apply here and now. Then there is the question how they can change in time (what is time?)

I certainly acknowledge these questions. Here are some incomplete further comments.

> Howto specifiy current law

This can be interpreted in severa ways. I guess the most obvious and honest answer is that our best state of physical law is simply the standard model. Because after all, that is extremely successful.

One would have to distingusih acknowledging the standard model to be the best of our current knowledge and the quest for candidates of improved models, that solves some of the open questions, that have both possible experimental preductions but also refer to conceptual and consistency issues in the current model.

The way I am reasoning would also suggest a change in the framwork in which the standard models if phrased. This in practice means that the suggestion I have would imply a minor revolution of thinking. It's not just that I suggest shifting a couple of parameters in the standard model, I'm suggesting a completely different framwork for posing the scientific questions. Thus do I expect a revolution type of evolution.

But until then, there is no question that the standard model (which it's flaws) is the best de facto model we have.

So howto translate the "standard model" into the new framework? The idea is to first of all work on the basis I referred to previously, and then try to find the emergent observer complexes and emergent stractures, and hopefully match those structures with the standard model. so in short, the new ideas would suggest a new framwork that constrains the models, and then the standard model and the body of experimental data can be used to "train" or tune the new model. This "training" would not be part of the proper inference I refer it, it's simply a practical thing needed to translate the "current state" between two frameworks when the framework is revolutionized.

What is time is a very big equestion in itself, I think your questions are well motivated and firing so many of them into one thread leaves me no choice but to be brief. Time in my view is somewhat analogue to the thermodynamic(TD) arrow of time, but the difference is that remember that ordinary TD is founded on plain probability theory. I was previoulsy suggesting a reconstruction of that, which important differences, and in that sense similarly will the arrow of time we more complex. In particular will the usually issues ot heath death not necessarily apply to such evolutions.

Also, since my construction is based on evidnce count, it is a subjective inference system and not objective. This also naturally renders my reconstructed "thermodynamic arrow of time" subjective, or observer dependent. It's my conjecture that this can be entirely consistent with relativity in the appropriate approximation.

Since a basis trait of the interacting inference systems is that the "inconsistencies" between two inference systems does not halt the system - instead it DEFINES a new interaction, and a new symmetry between inference systems. In my view symmetries of laws are emergent in this way. This would also suggest a completely new way to seeing at symmetry breaking etc. Again it's a different way of thinking than the standard model framework.

We must not mix the following approaches:
1) the believe in the existence of the moon even when nobody is looking at it
1') the believe in Kepler ellipses even when there are no planets
2) the description of (some of the aspects of) the moon by a mathematical framework

1) is the starting point for any kind of realism
1') is the starting point for something like structural realism
2) comes a a later stage and tries to map some aspects of 1) or 1') into a mathematical framework

What Dmitry67 is discusses is to identify 2) itself with something like 1') and 1) As I tried to explain I am not convinced.

As I understand it is an objection against 2) and some interpretations, not against 1) itself.

Now you can say that its useless or even nonsense to believe in 1) if you can neither be sure that 2) applies nor be sure that 1') applies NOW. My believe is that at least 1) will always survive as long as you do not go for solipsism.

So at which level does you inference "happen", 1), 1') or 2)

I will respond more to this later. But in short, I certainly see the difference between some of the old style realism, and the more modern less objectionally "structural realism" which is when you have a realist view of physical laws.

I reject both as fundamental.

however, one very important distinction I would like to add is that there is a difference between

1. The rationality in a belief in a fixed basis of your own actions - this fixed basis can be thought of as realistic.

2. the idea that realist elements are timless, eternal and not subject to inference.

Even in my view there is a defendable rationality between action upon certain beliefs AS IF they were elements of reality in the realist sense. This is because due to the limited complexity of the inference systems, some uncertainties can not be quantified/measured. This is the lowest level in the information hierarchy, and IMO consititues the most stable physical law.

It's the idea of rational action, the rational action acts upon it's premises regardless of wether this will later be changed.

IMO, the only influence on the action that the believe in realism has, is that your actions are so constructed as to NOT QUESTION the validity of this. This exists also in my way of thinking ,but I do not mix up this rational action based on non-deductive inference and imperfect initial conditions with realism in hte sense it's usually meant.

I consider Dmitry's reasoning to have strong realist traits. About your view I'm not I formed an opinon yet, but it seems that I'm probably closer to your reasoning than to Dmitrys. Buit I could be wrong.

/Fredrik

 

[ConradDJ]

There is definately a physical reality, but if it is independent from any perception is not clear to me. Especially in quantum mechanics it's no longer clear how to make a cut and define exactly this external physical reality.

Compare this to a rainbow: a rainbow somehow exists independent from our perception of the rainbow, but the independent existence reduces the rainbow to a collection of raindrops, light rays and geometrical concepts. This collection could be called its reality / existence, but it can hardly be called rainbow ...

Unfortunately for the physical reality of quantum objects we are not able to say what corresponds to the raindrops, light rays etc. Atoms? quarks? fields? quanta? operators and Hilbert spaces?

"Does the moon exist even if nobody looking at it?"

Now the problem for me is that even if I deny that q.m. can tell us anything regarding "reality", I have the feeling that there is something behind this "empirical film" of the world...

So my conclusion is a) that there is a mind-independent, external reality but b) that we cannot even know about the very concept / meaning / notion of its existence. We do not know to what the "rainbow" reduces - but we are definately sure that its reduction causes it to lose its "rainbowness".

Tom -- I think you're pointing toward something fundamental here, that doesn't become quite clear.

The rainbow is not "mind-dependent" -- in that a camera located at the same place would "see" the rainbow too. But its "rainbowness" -- i.e. the aspect of its structure that's not captured in a description of raindrops and light-rays per se -- depends on the point of view from which it's seen.

Classical physics was based on the assumption that a complete description of the world can be developed without reference to "points of view" -- or to put it more generally, without reference to any specific physical conditions under which measurements are possible.

Relativity then required that spacetime "points of view" be taken into account. And QM generalized this requirement to include any physical conditions of measurement.

So 20th-century physics raises a problem that I don't think can be solved so long as we think in Cartesian / Kantian terms -- i.e. "external reality" vs "mind-dependent" appearances. Your rainbow represents the missing third category -- i.e. the aspects of the world's structure that belong to the real, "external" physical world, just as much as the moon does -- but aren't describable as objects "in themselves" (light rays, raindrops).

The third category is what Rovelli points to (in his Relational QM paper) when he says physics is about "the information things have about other things". This also describes Fra's project, though his premises and starting-point differ from Rovelli's.

As for me, I would frame the issue a bit differently. I'm trying to understand the relational structure of the "external world" in terms of the conditions under which physical interactions can communicate information, in the context of other kinds of physical interaction.

But the basic point is that physics includes not only structures that can be described "in themselves", but also structures of relationships between things, like the rainbow. The classical assumption that the latter structures are reducible to the former seems now clearly untenable... so the deep unresolved issue in contemporary physics (as Smolin and many others have argued) is understanding how to describe what's going on in physical relationships.

Unfortunately, we're all so used to the dichotomy of "mind-dependent" vs. "mind-independent" that it's hard to grasp this challenge with any clarity. What you call the "empirical film" confuses what appears to us, subjectively, with the structures in the world itself that make physical "appearances" -- rainbows -- possible.

 

[Dmitry67]

THere is another interesting option as I mentioned in another thread.

Very likely the complicated things like consciousness/mind can't be explained based on the properties of the parts they consist of. Even more, if the world is deterministic, mind can be... well... I can't say 'non-deterministic', but rather 'non-applicable'.

 

[tom.stoer]

Hi Frederik,

I think I somehow start to understand your (excellent) ideas

Howto specifiy current law ... our best state of physical law is simply the standard model.

Currently - yes.

... would also suggest a change in the framwork in which the standard models if phrased. ... It's not just that I suggest shifting a couple of parameters in the standard model, I'm suggesting a completely different framwork for posing the scientific questions.

I agree - at some time or stage this is required.

... I certainly see the difference between some of the old style realism, and the more modern less objectionally "structural realism" which is when you have a realist view of physical laws.

I reject both as fundamental.

Why? Can you explain in more detail, please?

... there is a difference between

1. The rationality in a belief in a fixed basis of your own actions - this fixed basis can be thought of as realistic.

2. the idea that realist elements are timless, eternal and not subject to inference.

I never mentioned that explicitly as it was clear for me from the very beginning. Whatever this reality is, there is no requirement that it's timeless - whatever timeless means :-)

... About your view I'm not I formed an opinon yet, but it seems that I'm probably closer to your reasoning than to Dmitrys.

let's wait and see ...

 

[Fra]

Hello Tom,

... I certainly see the difference between some of the old style realism, and the more modern less objectionally "structural realism" which is when you have a realist view of physical laws.

I reject both as fundamental.

Why? Can you explain in more detail, please?

My way of abstraction of physics is a reconstructed information theoretic view that follows from the reconstructed "evidence count theory". But in short in this view, both types of realist views are a result of a physical inference. The only difference is that the structural realism is more sophisticated since it's somehow recodes the observational data to find the invariants (the laws) - in this sense, if I have to choose I would prefer structural realism, but it still suffers from the same principal disease ;)

This is pretty much what relativity did as well, when the reality of absolute space and motion was rejected, the rescue was to find the transformations that connects the relational views. And of course this structure (symmetry of spacetime) is more stable than the previous views.

A prime examply of this reasoning is Rovelli's reasoning, in this various papers, partial/complete observables etc etc. It's one of the issues I have with this reasoning, because he doesn't seem to take the observers inside view seriously, since he keeps referring to the transformations between observers, while not acknowledging that hte inference of these transformations/symmetries that he considers to be the more physical thing, can only take place relative to the inference machinery defined by an inside observers. In this sense, I don't think Rovelli's reasoning is self-consistent - from the point of view of physical interaction as inference. The paradoxal part is that in this RQM paper he at least partially aims to this, but in the paper there is a turn wheere he avoids some important keys. (For example the meaning of "probability" - that's exactly the question I acknowledge and the reason why I want to reconstruct it)

This is also how QM saved determinism. By bundling the unpredictable outcomes into statistical distributions, the determinism was recovered at statistical level.

It's a repeating pattern, and it's certainly a natural one, and a rational one. I do not object to that itself, on the contrary. But the symmetries or laws, found from a previous diversion of observations made by a different observers, the relationa between the observers, recovered a connection between their observations, by symmetry transformations.

But as this is described, this symmetry is inferred from experience in one observers view, and to this particular observer this reasoning seen in isolation is on par with any other inference. An in general an inference, may it be about the state of an small system in the environment, or a statemeny about the symmetry of actions of all systems observed in your record, is nothing more than a basis for his actions. It's not possible to turn this inference into a deduction and arrive at a certain "symmetry". What can happen, is that the uncertainty of hte symmetry is not distinguishable, and in this case it's IMO rational to act as if it was an element of reality, but that doesn't make it universally, timless and objectively real.

I never mentioned that explicitly as it was clear for me from the very beginning. Whatever this reality is, there is no requirement that it's timeless - whatever timeless means :-)

Yes, that's true. I like the point Smolin made in one talk on this, where he pointed out the ambigous notion of eternal law when the universe is only 14 billion years old - and still growing older.

The reason I used the word is that I have a feeling that a lot of people that perform realist type of reasonining (also structural realists), doesn't seem to be bothered by the ambigous notion of what realism is, since if time is on one hand realtive, if the realist elements aren't eternal then exactly what does it mean? This realism seems to me to be a poorly defined guide.

OTOH, the practical difference to the immediate action of someone, between the old style realismm and just the rational action view I hold is probably non-existent. But somehow most realists doesn't seem to think of this of an opinion as as something that just influences the action but rather something more. I find that guide to be very unclear, therefore I reject it.

/Fredrik

 

[Fra]

Some additional comments...

We must not mix the following approaches:
1) the believe in the existence of the moon even when nobody is looking at it
1') the believe in Kepler ellipses even when there are no planets
2) the description of (some of the aspects of) the moon by a mathematical framework

1) is the starting point for any kind of realism
1') is the starting point for something like structural realism
2) comes a a later stage and tries to map some aspects of 1) or 1') into a mathematical framework

What Dmitry67 is discusses is to identify 2) itself with something like 1') and 1) As I tried to explain I am not convinced.

I probably lost track of that part to be able to comment. I recall from past threads that Dmitry is more of realist than me at least. But if he suggest that these are somehow versions of each other than I sort of agree. They all fit into my inference view.

Now you can say that its useless or even nonsense to believe in 1) if you can neither be sure that 2) applies nor be sure that 1') applies NOW. My believe is that at least 1) will always survive as long as you do not go for solipsism.

Like I tried to say, I distinguish between the *rationality of belief* that does impact your action, and an illusion of some external reality to this belief.

To make another association/analogy here, there are some models on how the human brain works that considers a form of feed-forward state machine where the action of the machine contains expectatins of the future based on the past. this view is basically that the brain is an inference system to predict the future, and learning could then be feed-back which is driven by the reaction from the environment depending on the expected future vs that actual future.

I don't suggest we should turn this into a discussion of the brain or conscioussness but there are analogies here that might help grasp the idea, and if you think about it there are some interesting connections to how actions are formulated by means of feynmann style sume over paths, that is in effect a kind of action based on expected evolutions based on an initial condition.

This QM logic, apparently works, but if we could find a deeper understanding of it, in a larger context, merging it with relativity would I think be easier.

So at which level does you inference "happen", 1), 1') or 2)

Mmm I'd say probably all three, unless I misunderstood your question?

/Fredrik

 

[tom.stoer]

... I think you're pointing toward something fundamental here, that doesn't become quite clear ...

The rainbow is not "mind-dependent" -- in that a camera located at the same place would "see" the rainbow too. But its "rainbowness" -- i.e. the aspect of its structure that's not captured in a description of raindrops and light-rays per se -- depends on the point of view from which it's seen.

... raises a problem that I don't think can be solved so long as we think in Cartesian / Kantian terms -- i.e. "external reality" vs "mind-dependent" appearances. Your rainbow represents the missing third category ...

ConradDJ,

you are definately right. My feeling is that our discussion (and possible science) scratches the surfaces of something deeper than just new equations or just another discussion about Cartesian/ Kantian / ... reasoning.

Regarding the rainbow: it is both "external" and "internal" - and it is "relative".
It's "external" as it "consists" of fundamental "things" like raindrops, light rays etc. I don't want to say that these external things or their existence in itself is understood to 100%, I only want to stress that a rainbow has "aspects" that are somehow independent from any observer.
It is "internal" as its representation in our brain or mind (whatever this is) depends on us. The color "red" is not a piece of the rainbow's existence but it's the mind-internal representation of some specific wave length of a specific light ray between a raindrop and my retina.
It is "relative", as e.g. it's position depends on the position of the sun, the position of the raindrops and my position.

The "external existience" is justified by the miracle argument: It would be a miracle if there were nothing else but "mind-internal" existence along with coinciding phenoma perceived by different individual observers. At least some aspects of the rainbow are perceived identically (similarly) by different observers (including the camera). This is a strong hint (not a proof) that there is something that "exists even if nobody looks at it".

The problem with this "external existence" is that we cannot say what it really IS, because our perception is always filtered by our senses, certain devices etc. So it seems that there IS some regular structure behind the phenomenological level which has an own, independent, external existence.

The structural realism deals with this regular structure but (at least partially) ignores the problem that the "physical laws" are not only relations between "external entities" but also relations between an "external entity" and "me". The position of the rainbow is a perfect example. Therefore the clear cut suggested by realism does not work - neither for old-fashioned "naive realism" nor for any other school of though.

It becomes clear when we start discussing the terms "external" or "independent". They point towards something that exists w/o "me", but as soon as we try to pinpoint them, theyescapes from any sound definition.

Now the problem is that we are used to think in two categories - objective and subjective. The objective, realistic position is problematic as explained above, but the subjective one is problematic as well. If you look at idealism, phenomenology or positivism you always have the feeling that they miss an important point. If they (as positivists do) insist on "predicting experimental phenomena only" they neglect "the existence of the moon if nobody looks". They claim that the whole discussion is meaningless, but if you look at a positivist getting up, going to the kitchen and using his coffee machine you cannot avoid the impression that the coffee machine existed all over the night and has not been created by opening the kitchen door (other positivists create their coffee machine by switching on the light; fortunately one coffee machine can be re-created eiter way :-) This is OK as long as you are clear and aware about this blind spot. If you build an subjective ontology that does not even mention this blind spot you are on our way towards solipsism.

What I am trying to say is that I like Kant's idea of the "thing-in-itself" that is unreachable by our senses, but that serves as an "x" around which we construct our reasoning. This "x" is like the rainbow itself (or a your shadow) - if you try to catch it, it slips away, but nevertheless it's clear that it is "there".

I think that every ontology either explicitly using the cartesian cut or explicitly denying the cut is doomed to fail. The cut exists in some sense, namely as a structure within our categories of thinking, so we must not deny its existence.

Perhaps the situation compares somehow to the early years of quantum mechanics. Neither waves nor particles are the correct description of nature. It is neither exactly one of these pictures, nor is it a naive combination of them. It is something deeper that only in some rare cases (experiments) presents itself in such a simplified way.

In a similar way I expect that a deeper level of the (description of) physical reality will reveal a harmonization of these different and mutually exclusive points of view - idealism (subjectivism) and realism (objectivism) ...

 

[Fra]

Tom I'm starting to see your reasoning, but let me ask, do you consider yourself somewhat of a structural realist?

I only want to stress that a rainbow has "aspects" that are somehow independent from any observer.

To me the questioning of what the rainbow is or isn't wouldn't take place without a context to encode it.

I'd rather like to say that there are aspects of the rainbow that are apparently independent of a choice of observers among a defined choice, but then certainly begs the question why are we constrained to the choice of observers that "happens to be in agreement"

That is basically questions is the one you raise here

The "external existience" is justified by the miracle argument: It would be a miracle if there were nothing else but "mind-internal" existence along with coinciding phenoma perceived by different individual observers. At least some aspects of the rainbow are perceived identically (similarly) by different observers (including the camera). This is a strong hint (not a proof) that there is something that "exists even if nobody looks at it".

In my view, it's not a conincidence at all. It's a result of evolution, and only those observers (read physical matter systems) that implements consistent inference systems are able to coexist in equilibrium. The disagreement between the systems implies interactions that forcefully deforms the inference machinerys.

Thus the "consistency" we see around us, is I think no conincidence at all. Just imagine that opposite, it would be wild and I presume destabilise in fractions of seconds.

So when did this evolution take place? I guess my personal idea is that this would (informally) have taken place very very early in what be today consider to be the big bang, at the point where the forces was indistiniguisahble and there was no localised matter. Probably at some point, the equilibration/laws of some of the basic laws was complete.

/Fredrik

 

[tom.stoer]

Tom I'm starting to see your reasoning, but let me ask, do you consider yourself somewhat of a structural realist?

It's too early to answer this question; but I am definately MORE a structural realist than a naive realist.

... It's a result of evolution, and only those observers ... that implements consistent inference systems are able to coexist in equilibrium.

I don't agree. If you insist on evolution of the observes' rules, you have to answer what the driving factors of this evolution are. Every evolution takes place in an environment which decides about the "survival of the fittest". So again you are referring to something that is objective = external to the observers.

 

[Fra]

If you insist on evolution of the observes' rules, you have to answer what the driving factors of this evolution are. Every evolution takes place in an environment which decides about the "survival of the fittest". So again you are referring to something that is objective = external to the observers.

Yes, of course. But i think this is a bit tricky, because in my view EVERY inference is made fomr the point of view of an inference system (inside observer) - this includes the decription of the evolution.

Some things are simply undecidable, or unpredicable, and this is in particular hte observers OWN FULL evolution. I suggested that an observer is a predictive system of the future events whose actions follows from this. But certainly no observer can KNOW the future. In particular not it's own future.

To explain the meaning of my idea of the subjective basis for the "environment of an observer", then the trick is to consider a second observer B. You have one observer B, observing it's environment. And this observer might distinguish more or less coherent subsystems in his environment (say one subsystem A), and this observer can thus try to understand the ACTION of these subsystems(A), and B can then describe the evolution of A relative to the observed B-subjective embedding of A.

This is why it's like smolin pointed out a large different between
1. observing an isolated subsystems, (which is effectively the case in particle lab experiments) since the observing environment is enclosing the observed system effectively.
2. a small observer observing it's own overwhealming environmnet.

So your point is good, but I have thought of that and I do not talk about an objective evolution, I suggest that there are only inside views even of the evolution itself. BEcause ultimately the evolution of law, and the evolution of states - as per law, are really principally the same in my view, the different is that the evolution of time and normal timevolution takes place at different levels in the hierarcy of the inference system.

Also on top of this, it is also admittedly unavoidable that every single thing I say here, are implicitly conditional on my current evolved brain. There is nothing whatsoever I can do about this. But it's good enough for me.

So I hope I tried to make the point. I am not envisioning this evolution as an objective evolution. Rather there are several views of the evolution as well. Clearly the view an observer has on it's own evolution, is simply the more simplified normal timeevolution, which is defined only differentially. As the actual time progreses, the feedback of the environment might generally induce a revision of the inference system in ways that is of course by construction completely and fully undecidable from the pooint of view of the observer itself.

However, that does not mean that this evolving inference system would be totally unpredictable for another (usually much larger observer).

It's exactly in this sense I distinguish between the inside view of the players in say a particle experiments. This particle I think simply have no clue, they can not predict their own future very well - which is revelaed in the way they act - their actions encode their "simple" expectations.

OTOH, the massively complex laboratory frame rules by reasonably intelligent human beeings can predict the future of these particle more than they can themselves do, because
1. we have much more information
2. we have a muhc more developed inference system

Does that make sense?

/Fredrik

 

[Fra]

The idea that there are only inside views of the evolution as well, is also why I think that the environment and observer evolve together. Both are needed. It's like spacetime and matter. One without the other makes no sense. So I think the inflation of space is a process that must by consistency of this idea, go hand in hand with the creation of matter. They somehow both drive each other.

Ie. the evolution doesn't take place in an external fixed environment, because the environment is itself evolving, because the environment is simply another inference system. It's like you have two inference systems fighting each other, and the one that predicts the other one better wins control.

/Fredrik

 

[ConradDJ]

The problem with this "external existence" is that we cannot say what it really IS, because our perception is always filtered by our senses, certain devices etc...

The objective, realistic position is problematic ... but the subjective one is problematic as well. If you look at idealism, phenomenology or positivism you always have the feeling that they miss an important point.

Tom – Your post makes a lot of sense. In the background of all this thinking is that Cartesian sense of being “inside our heads” and looking “out” at an external world through some filtering screen that may or may not allow actual information to get through. As radicalized by Kant, it’s the sense of a mysterious unknowable reality “out there” that our mysterious consciousness “in here” can never get to.

From this viewpoint, the fact that information has to be communicated through physical interaction is seen as something negative. Ideally the mind (as something existing “in itself”) should have direct, unmediated access to the “external reality” (as it is “in itself”). This notion of what it would mean to have genuine knowledge is at the root of our tradition, from Plato to Hegel.

But of course we humans are physically “out there” in the world interacting with things, and with each other. To the extent we develop an “internal” consciousness, it grows out of this communicative environment. And though our senses and devices certainly have their limits, what’s really remarkable is how much information is there in this interactional environment, and how thoroughly we’ve been able to explore it.

In fact, QM seems to be telling us that all the determinate information there is in the world gets communicated through interaction. Ultimately,“behind” the interaction there isn’t any definite “thing in itself” to be known, but only – as described by the wave-function – some kind of structured potential for communicative interaction.

So it seems as though the kind of being your rainbow has, may be what’s fundamental. Both the “external existence” of things and our own “internal existence” seem to depend on the very special character of the interactional environment between us all, that can actually define all this information.

In any case, I don’t see the problem in present-day physics as one of struggling to grasp a mysterious reality we can’t experience. I think the hard problem is one of understanding the structure of what we do experience and measure – not as something going on “in our heads”, but as what goes on “out there” in the physical world that constitutes our shared informational existence.

I tried to point out in another thread why this is a hard problem. It’s not that we lack access, but that we have so many ways of access, that all depend on each other.
https://www.physicsforums.com/showpost.php?p=2319688&postcount=3"

I think what we lack is an adequate analytic approach to this kind of information-system. And the more basic problem is that in these debates over objective vs subjective thinking, what’s so remarkable about our physical communications environment is just taken for granted -- or treated as though it were an obstacle to true knowledge.

 

[tom.stoer]

... the evolution doesn't take place in an external fixed environment, because the environment is itself evolving, because the environment is simply another inference system. It's like you have two inference systems fighting each other, and the one that predicts the other one better wins control.

I have no problem with the idea of an evolving environment. Whar I miss is the expolanation for "survival of the fittest", population, "... wins control". What are the rules to decide what fits to what? what are the rules for producing predictions and comparing them with "reality"? what does it mean that one systemwins over the other? what happens to the winner (loser)? What sets the "timescale" or clock for the evolution (it's of course not the physical time).

 

[tom.stoer]

... is at the root of our tradition, from Plato to Hegel.

I agree.

Ultimately,“behind” the interaction there isn’t any definite “thing in itself” to be known, but only – as described by the wave-function – some kind of structured potential for communicative interaction.

Yes, something like that.

..., I don’t see the problem in present-day physics as one of struggling to grasp a mysterious reality we can’t experience. I think the hard problem is one of understanding the structure of what we do experience and measure – not as something going on “in our heads”, but as what goes on “out there” in the physical world that constitutes our shared informational existence.

You are right, it's not a problem directly related to physics, it's more are problem of the ontological position of present day physics. The question is critical to physics because currently I do not see that philosophy has managed to understand the most important results from the last century (QM, GR). You can see that simply be the terminology we are using - it' out-dated! Philosophy stopped to develop a fundamental ontology as they stopped to understand the physical progress = with the beginning of the 20th century ...

The more I think about all these ideas I come to the following conclusion:
1) either we accept that the physical reality (including us as observers, physical laws and their evolution, ...) indeed IS a purely mathematical framework w/o any additional baggage (like interpretation, recipe for application, ...) that is embedded in a mathematical multiverse consisting of all (sound) frameworks
2) or it is impossible to find a consistent ToE at all because by Goedels theorem every candidate-ToE would allow to ask questions like "is THIS ToE consistent" - and of course we know that the answer can be given iff (if and only if) the ToE is inconsistent - and if it cannot be given, the ToE could be consistent but this is not provable.

I do not like this conclusion, but I do not see a way out.

 

[John86]

I have no problem with the idea of an evolving environment. Whar I miss is the expolanation for "survival of the fittest", population, "... wins control". What are the rules to decide what fits to what? what are the rules for producing predictions and comparing them with "reality"? what does it mean that one systemwins over the other? what happens to the winner (loser)? What sets the "timescale" or clock for the evolution (it's of course not the physical time).

"There are no winner and losers" it's just quantum mechenical evolution at work call it uncertainty. the consequence of this is that it disturbs last strong holds of realism. this will although ultimately have no certain consequences for our daily classical world though.

 

[Fra]

I have no problem with the idea of an evolving environment. Whar I miss is the expolanation for "survival of the fittest", population, "... wins control". What are the rules to decide what fits to what? what are the rules for producing predictions and comparing them with "reality"? what does it mean that one systemwins over the other? what happens to the winner (loser)? What sets the "timescale" or clock for the evolution (it's of course not the physical time).

Ok, I guess you demand something very explicit here, and of course so do I.

I do have some partial answer to your questions but they aren't very mature or sufficiently developed to knock some someones head quite yet. Every question here is very well justified. I'm not hiding from them :)

But given the work in progress instead of trying to convince you by more fuzzy or incomplete arguments I will settle with that I am trying to find explicit answers. The ultimate argument will certainly have to be developed and verified predictons.

/Fredrik

 

[Fra]

Just one more fuzzy comment if it helps, but I think I already said this previously.

I have no problem with the idea of an evolving environment. Whar I miss is the expolanation for "survival of the fittest", population, "... wins control". What are the rules to decide what fits to what? what are the rules for producing predictions and comparing them with "reality"? what does it mean that one systemwins over the other? what happens to the winner (loser)? What sets the "timescale" or clock for the evolution (it's of course not the physical time).

"the population" ~ matter or "confined energy" systems

"intrinsic inference system" ~ encodes the physical action of above matter

the looser ~ gets destabilized (looses mass/energy) and are not longer part of the population, and thus neither is the inference system / action the looser encoded

the winner ~ either maintains status quo, or better grows more degrees of freedom (~mass)

This is why as in high energy experiment you break up the matter systems into smaller parts, the smaller the parts the less encoding capacity do we have to encode complex interactions, and therefore would one expect unification at the zero complexity limit.

The complexity ~ mass, of a system is the natural regulator cuttoff, that constrains the space of possible inferences. This would suggest a feynmann kind of paht integral but with a built in renormalisation, the natural cutoff is simply the complexity of the context, the observer, than encodes the integral.

IE. a "simple observer" can not possibly "observe" (infere/abduce) a sufficiently complex law in the environment, therefore the immediate actions of this observer must also be invariant with respect to this indistinguishable complexity.

A successful observer, who are able to predict the environment is effectively conquering degrees of freedom of the environment.

/Fredrik

 

[John86]

Yes, that's true. I like the point Smolin made in one talk on this, where he pointed out the ambigous notion of eternal law when the universe is only 14 billion years old - and still growing old

what role does time play before that phase

 

[Fra]

what role does time play before that phase

My personal view is simply that at that supposed point, it is not possible to encode anything that implements any kind of arrow of time, therefore I think time is simply undefined, or a question that isn't raisable, because the context neede to encode an arrow of time isn't there.

/Fredrik

 

[ConradDJ]

I'd rather like to say that there are aspects of the rainbow that are apparently independent of a choice of observers... but then certainly begs the question why are we constrained to the choice of observers that "happens to be in agreement"

In my view, it's not a coincidence at all. It's a result of evolution, and only those observers (read physical matter systems) that implements consistent inference systems are able to coexist in equilibrium. The disagreement between the systems implies interactions that forcefully deforms the inference machinerys.

Thus the "consistency" we see around us, is I think no coincidence at all. Just imagine that opposite, it would be wild and I presume destabilise in fractions of seconds.

I have no problem with the idea of an evolving environment. What I miss is the explanation for "survival of the fittest" population, "... wins control". What are the rules to decide what fits to what?

Fredrik – The issue with this language about “equilibrium” and “destabilzing” – as I’m sure you recognize – is that there’s an implication that time is running on in the background, and some systems survive longer in time than others. (That of course makes sense for evolution in biology, where the issue is indeed the survival of complex systems over time, against the threat to their stability / integrity posed by the physical environment.)

I don’t think this is what you have in mind, so the question is how to frame the issue differently. In physics, it seems that the fundamental issue is how to define (determine) information in the first place – including information about the time-sequence of events, etc. The “threat” posed by disagreement among observers is not that some equilibrium state will break down, but simply that the conflicting information provides no definite basis for further information to be defined.

It’s only to the extent that all present “observers” agree on information, that it gets passed on as a basis for future agreements. “Future” here does not refer to a background time-continuum so much as to “what’s next” in the evolutionary process itself. I’m thinking that distinction will make sense to you. Here it’s not a matter of “survival” over a period of time – the issue rather is what gets defined / agreed on “here and now” that can be “passed on” as a basis for other moments of observation, both here and elsewhere. There is something like "time" only insofar as this progressive determining-and-agreeing keeps happening.

And there are various levels of agreement – e.g. agreement about specific facts, and agreement about the general “laws” or patterns into which all the facts must fit. The one refers to the ongoing evolution of information, moment by moment, as we observe things, while the other refers to patterns that evolved long ago, which are the preconditions for our coming to agreement about specific facts. (If nothing were lawful and predictable, clearly there could be nothing like observation or communication.)

And of these general patterns, some are presumably more "primitive" and others evolved "later" based on those "earlier" patterns.

But to me, the crucial point is that we’re not talking about some mysterious process that takes place only at the Planck scale or at the beginning of time – this is something that still happens in every “observation” and every communication, all the time. So the problem is one of sorting out the tangle of many different kinds of physical interaction that make up our present-day “communications network”, to try to see which are the more “primitive” and which the more “highly evolved” interaction-patterns.

By the way, I'm glad you're being pushed to get more of your unfinished work out in the open!... though some of this is way over my head.

-- Conrad

 

[BigFairy]

It certainly has a lot to do with the ideas ever so proposed

 

[ConradDJ]

You are right, it's not a problem directly related to physics, it's more a problem of the ontological position of present day physics. The question is critical to physics because currently I do not see that philosophy has managed to understand the most important results from the last century (QM, GR). You can see that simply be the terminology we are using - it's out-dated! Philosophy stopped to develop a fundamental ontology as they stopped to understand the physical progress = with the beginning of the 20th century ...

Yes, I very much agree. There was actually a very important breakthrough in Heidegger's Being and Time, published just as the QM formalism was getting established in the late '20s. (Heidegger was certainly aware of the turmoil in physics, though I'm sure he made no attempt to understand it.) But Being and Time was published as the first half of a project that he never completed, and his later work did not succeed in making any connection with the sciences.

I spent years (in long-ago grad-school days) deciphering his work, and learned a lot from it -- but I was tremendously disappointed by his "followers" (not to mention his detractors). I couldn't find anyone who was actually doing anything useful with his work, and for the most part it was understood merely as another contribution to phenomenology, relevant only to understanding human existence. His task of fundamental ontology was never taken up.

The more I think about all these ideas I come to the following conclusion:
1) either we accept that the physical reality (including us as observers, physical laws and their evolution, ...) indeed IS a purely mathematical framework... in a mathematical multiverse consisting of all (sound) frameworks
2) or it is impossible to find a consistent ToE at all because by Goedels theorem every candidate-ToE would allow to ask questions like "is THIS ToE consistent"...

I do not like this conclusion, but I do not see a way out.

But I don't see mathematical consistency as the issue. For one thing, mathematics and logic only have meaning because the physical world gives us distinct, identifiable objects that persist through time. "A=A" may seem incontrovertible, but in the very early stages of the universe there would have been nothing to which this statement corresponds.

I think we fall back on Platonic notions of a world made of mathematics, only as a default -- because it's so difficult not to take for granted how well the physical world works, to define things and relate them to each other in so many ways. We treat physical existence as if it were a logical "on/off" switch -- either something exists or it doesn't. But to "exist" means to participate in a very complicated web of ongoing physical relationships with other things that exist. (This was Heidegger's basic point about "the meaning of being" -- that it has to do with "being-there" or "being-in-the-world"... as opposed to treating it merely as a logical "is"/"is not".)

And though we can model distinct aspects of this web in mathematics, with great success, it doesn't follow that it is essentially a mathematical structure. What Fra is trying to develop is a functional model -- in that only a functional system (e.g. a self-replicating one) can be imagined as evolving, in the sense he intends. Likewise Smolin in The Life of the Cosmos... where the functionality of the universe is taken to be a weird kind of self-replication.

Smolin's idea seems to me far-fetched, physically, and uninteresting from a philosophical standpoint. But I think his point is well-taken that the physical world looks in many ways like a highly-functional and indeed highly-evolved system. So the issue is how to define its basic functionality -- what exactly it means that the world "works" to support physical existence -- not just human existence, of course.

 

[tom.stoer]

"the population" ~ matter or "confined energy" systems

"intrinsic inference system" ~ encodes the physical action of above matter

the looser ~ gets destabilized (looses mass/energy) and are not longer part of the population, and thus neither is the inference system / action the looser encoded

the winner ~ either maintains status quo, or better grows more degrees of freedom (~mass)

This is why as in high energy experiment you break up the matter systems into smaller parts, the smaller the parts the less encoding capacity do we have to encode complex interactions, and therefore would one expect unification at the zero complexity limit.

The complexity ~ mass, of a system is the natural regulator cuttoff, that constrains the space of possible inferences. This would suggest a feynmann kind of paht integral but with a built in renormalisation, the natural cutoff is simply the complexity of the context, the observer, than encodes the integral.

IE. a "simple observer" can not possibly "observe" (infere/abduce) a sufficiently complex law in the environment, therefore the immediate actions of this observer must also be invariant with respect to this indistinguishable complexity.

A successful observer, who are able to predict the environment is effectively conquering degrees of freedom of the environment.

wow - interesting idea - I have to think about that for a while:-)