How Do We Determine the Most Fundamental Laws of Nature?

[vectorcube]

Some laws of nature are not fundamental. These laws are explained why more fundamental laws. These more fundamental laws would be explained by more fundamental, deeper laws. Assumpting that this process do not continue indefinitly. We will eventually arrived at a set of fundamental laws, S. Why do S obtain? Since there is no fact of the matter that would explain the obtaining of S, it then follows that S is a brute fact.


Claim: We start with S that is not fundamental. We can always find a set that is fundamental using the following algorithm:

While( S is not fundamental is true)
{
Construct set S* from S, such that the laws generated by S* produce all the laws in S.
Replace S with S*
}

Claim: If algorithm returns S, then the S is most fundamental.

Proof: Suppose S is not fundamental, then the loop would not end, and thus, S cannot be returned. Contradiction Therefore, the algorithm returns S is fundamental.


Analysis:
In the special case, the algorithm would not return S, then the above claim( conditional) would not apply. This could happen if there is not fundamental set of laws that would generate S.

What does this all mean? There is either 1) a infinite regress of laws where no laws are more fundamental, or 2) there is a set of laws that is most fundamental, and are just brute facts.


Note: If you are going to reply. Please, explain yourself in easy to understand terms. Please, Do not try to show off by using "big words", or being "vague, and profound". It never works. Imagine yourself writing a actual philosophy paper in order to get a grade. Please, no new new age crap.

 

[apeiron]

Alternatively the flaw is in adopting a linear notion of logic. If everything must have a crisp prior, or equally, be nested in a crisp superset, then the infinite regress is built into the logic (and is not necessarily a truth of the world).

To escape such regress, others would explore bootstrapping approaches to worlds and therefore to their logics that model worlds.

Peirce for example considered how laws would develop from vagueness and so would take their "fundamental" shape only in the infinite future. Or rather, there would be no limit on their continued development, even though actual change in the laws might become asymptotic.

 

[vectorcube]

Alternatively the flaw is in adopting a linear notion of logic. If everything must have a crisp prior, or equally, be nested in a crisp superset, then the infinite regress is built into the logic (and is not necessarily a truth of the world).

What is "crisp prior", "linear notion of logic", and "crisp superset".

can you give me some references?

To escape such regress, others would explore bootstrapping approaches to worlds and therefore to their logics that model worlds.

What is "bootstrapping approaches to worlds"?

Peirce for example considered how laws would develop from vagueness and so would take their "fundamental" shape only in the infinite future. Or rather, there would be no limit on their continued development, even though actual change in the laws might become asymptotic"

don ` t know

 

[apeiron]

don ` t know

And ignorance gives you a free pass?

 

[vectorcube]

And ignorance gives you a free pass?

It is fine if you think you are smarter, and more knowledgeable than me. What interest me are real, clear, and unpretentious philosophy. I am not afraid of what other people think of me. I don` t care if you think of me as stupid, or ignorant.

 

[Fra]

Ok here is my personal opinon.

I like the example and I opt for infinite regress, but during the infinite regress there is always the "current set" of "best" laws.

While( S is not fundamental is true)
{
Construct set S* from S, such that the laws generated by S* produce all the laws in S.
Replace S with S*
}

If you are serious about this, then this algorithm must be computed by a real computer in real time, leading to a number of complications:

The set S is all the apparent non-fundamental non-regularities, thrown into the input of your computer. So by the time your algorithm has concluded the successor S*, the input has changed - this can go on forever as inifinte regress, since the input keeps changing.

One can also imagine that your algorithm is too fast, and prematurely concludes it has found the fundamental laws, and then secons later new data of non-fundamental regularites arrives that is inconsistent with your previously inferred laws. Again you are lead to further regressions.

Instead, I think this regression is how nature works, and to understand nature and it's effective - but generally evolving - laws is to understand these cycling inferences that take place in nature, this means to look inside that algorithm and find out exactly HOW the set S* is constructed from S. What is the logic here? And if we consider two interacting algorithms, can be make an inference about how they are likely to interact given these insights? Maybe!

Infinite regress, is often used as a bad thing. But if we just open our eyes, does reality look anywhere static or non-evolving? Not to me.

In this framework, effective laws are simply the "best inference", given the constraints of complexity and computational power. Ie. the best inference possible before the conditions for the infernece (ie the set S beeing the input) has changed!

/Fredrik

 

[vectorcube]

I like the example and I opt for infinite regress, but during the infinite regress there is always the "current set" of "best" laws.

In a infinite regress, no law is really fundamental, because there is always a more fundamental law behind it.

If you are serious about this, then this algorithm must be computed by a real computer in real time, leading to a number of complications:

No. Algorithm do not need a computer. Algorithm is just a set of percedure aim to complete a certain task. I write it that way because it is more fun. I am not saying there is a giant computer popping out our universe, or popping out laws.

 

[arithmetix]

Then could there be a prime cause? (pet subject of mine)

 

[Fra]

In a infinite regress, no law is really fundamental, because there is always a more fundamental law behind it.

They way I picture this, there is no fixed "fundamentality scale" where you are certain of making objective progress. Yes, there is always a "better law", but since the only way to judge the law, is relative to the prior law, you are only defining a differential progress, not a global one.

No. Algorithm do not need a computer. Algorithm is just a set of percedure aim to complete a certain task. I write it that way because it is more fun.

Ok but then your argument is a lot weaker.

Anyway, the idea of the algorithm is pretty good, and if you take it more seriously, your algorithm is the "inference system" I am talking about. Ie, the improvement of laws are a result of an infernece process (your algorithm here). And the current state of the laws, determined the action of the computer itself, and even the algorithm.

This is getting closer to my thinking. But then then algorithem itself is also evolving, so is the computer. To understand the action of a physical system - say a proton - amounts to understand the inside view of the inference the proton makes if you take the reaction from its' environment to be the input into the inference. Ie, the action of the proton can hopefully be understood by assuming tha the proton is acting in consistentcy to it's own inferred laws. But to understand the proton, I think we need to understand how this "inference machinery" scales down to a low complexity level such as the proton. And then also how this "inference machinery" itself is subject to evolution, in the sense that the stability of the proton reflects a kind of "stability of the inference system".

The preferred inference system I'm referring to here - to correspond to say a proton, is to compare with a STABLE law. In therms of your algorithm it would correspond to solutions where your algorithm enters a steady state where S is not changing - it's stable. But to jumpt from such a local equilibrium, to assuming a global or "fundamental nature" of these laws are unwarranted in my view. OTOH, to the inside observer, the local equilibrium is at least transiently! indistinguishable from a global one.

/Fredrik

 

[seycyrus]

Claim: We start with S that is not fundamental. We can always find a set that is fundamental using the following algorithm:

Your algorithm is just as crucial as your set of laws.

We call this algorithm science and are continually trying to improve it.

Your assumption of a perfect algorithm is invalid.

 

[apeiron]

The set S is all the apparent non-fundamental non-regularities, thrown into the input of your computer. So by the time your algorithm has concluded the successor S*, the input has changed - this can go on forever as inifinte regress, since the input keeps changing.
/Fredrik

We probably need to distinguish ontic and epistemic versions of the argument here.

This would be the epistemic angle - the better the model, the better the measurements. An infinite loop of possible improvement is implied.

But then the ontic angle would be that we are really talking about actual observers inside the systems they model. And measurements can't procede to infinity. Event horizons exist in either direction of scale.

So what is possible (infinite refinement) for an imaginary perfect observer is not practical for any real one (as worlds without event horizons don't make sense - a separate argument ensues here...)

More needs to be said to unpack the notion of fundamental law as well. Does the OP actually have a definition in mind?

Clearly the usual cannonical idea of a fundamental law is a symmetry expressed as a dichotomise asymmetry - so this one kind of thing equals that other kind of thing. It is also taken to be a universal constraint. So that locally anything is permitted that is not globally forbidden.

So once you begin to unpack what is actually meant by "fundamental law", we can see that it is just the global half of the story. We can see that constraints must interact with local freedoms in good old systems fashion. And so there aways remains that room for local surprises. An irreducible indeterminancy which allows systems to keep developing.

And yet as I say, it is also true in the systems view that because the constraints DO keep developing, the space for local surprises keeps shrinking. There is an asymptotic approach to the limit. And that is what "fundamental laws" attempt to do - take the limit, extrapolate the breaking of symmetry to its most crisply, asymmetrically, broken state.

Which is why it is better to think of fundamental laws as states of affairs infinitely far in the future. The laws exist as potentials at the vague beginning, for sure. But invisible. They emerge into view as a world forms and stand as its final expressed shape. Yet never actually achieved.

 

[Pattonias]

Doesn't this system only form a contradiction when you try to form it into a set algorithm.

Wouldn't the solution be to look for the factor that makes the final result true. The answer is that today we set up laws, backed by precedence, that allow for better laws to be formulated in the future. The question should be posed where were the did the original laws come from. I don't think at the time of their conception there was a standing preference to use laws with precedence. I think that concept evolved over time.

The original were most likely laid out and decided upon in answer to a particular event and evolved to what they are today. Unless of course you want to argue that they were written on stone. In which case the thread will most likely end up banned.

 

[Fra]

This would be the epistemic angle - the better the model, the better the measurements. An infinite loop of possible improvement is implied.

But then the ontic angle would be that we are really talking about actual observers inside the systems they model. And measurements can't procede to infinity. Event horizons exist in either direction of scale.

We can never sensibly release ourselves from the observation/interaction angle, unless you take on some realist view, or structural realism views - but this is not my view.

So, to translate what you say, without going into realism, it sound like you introduce a third observer, that is observing his environment where other observers are interacting? Then you are sort of giving these inside observers an excelt context, but this context is still merely just an "image" encoded in the third observer. Sure, this is a perfectly valid angle, but this doesn't change anything as I see it. The observable parts relevant to the modelling in this case, relates to the third observer. The expected causality horizons of the othre observes interacting, are just part of the "model" the third observer is a manifestation of.

The main point I want to add here is that the notion of causal horizons is not objective, it's always relative to another system/observer. I don't think there is anything objectively ontological about horizons.

This would be the epistemic angle - the better the model, the better the measurements. An infinite loop of possible improvement is implied.
...
So what is possible (infinite refinement) for an imaginary perfect observer is not practical for any real one (as worlds without event horizons don't make sense - a separate argument ensues here...)

I'm not sure what you mean here, but of course with my opt for "infinite regression" doesn't mean infinity is REACHED (in finite time), that's one point of the argument. I just mean more like "indefinite ongoing differential regression", and the actually regression takes place in time, in the observing systems internal degrees of freedom, because this is the home of the "model", where the best laws are encoded.

/Fredrk

 

[tauon]

let's start properly, from the beginning:
what's a "law of nature"?

o.0

how did you conclude that the things you call "laws of nature" exist?

you must describe the background first, you can't throw metaphors at us like this. your thought exercise about laws of nature and the supposed algorithm that finds fundamental "laws of nature" is meaningless since you did not even bother to define "law of nature" let alone attempt to somehow show that these things exist...

 

[vectorcube]

let's start properly, from the beginning:
what's a "law of nature"?

o.0

how did you conclude that the things you call "laws of nature" exist?

you must describe the background first, you can't throw metaphors at us like this. your thought exercise about laws of nature and the supposed algorithm that finds fundamental "laws of nature" is meaningless since you did not even bother to define "law of nature" let alone attempt to somehow show that these things exist...

No. Car companies do not invent the wheel everytime they want to build a new car.

If you want the background, then you can read philosophy of laws of nature some where. Tho, it is a lot of fun to distinquish contingently true generalizations from fundamental laws of nature, it is not really my concern. Well, you need to know that i do believe the laws of nature are real. They are contingent, but necessary relations between universals. That is if all F are G, then it is necessary that every instance of F is G. That is all.

 

[tauon]

No. Car companies do not invent the wheel everytime they want to build a new car.

If you want the background, then you can read philosophy of laws of nature some where. Tho, it is a lot of fun to distinquish contingently true generalizations from fundamental laws of nature, it is not really my concern. Well, you need to know that i do believe the laws of nature are real. They are contingent, but necessary relations between universals. That is if all F are G, then it is necessary that every instance of F is G. That is all.

I know very well what you meant by "laws of nature", I was however alluding to something else...
I for one, stand by the contrary: there are no laws of nature. what you're referring to as "law of nature" is merely an artificial description.

this whole rant is only interesting if the laws of nature are assumed to actually exist outside a mere artificial conceptual framework that we created (be their existences as parts of a platonic realm, or a meta-reality or the universe itself).

I think this would be a whole lot more interesting if you'd try to convince me of the existence of the laws of nature. otherwise it's just a boring hypothetical exercise that I can't really "get into"...

but what the heck, I'll byte:

Proof: Suppose S is not fundamental, then the loop would not end, and thus, S cannot be returned. Contradiction Therefore, the algorithm returns S is fundamental.

incorrect phrasing. the correct wording here is:
if an S is returned than S is fundamental. however, the problem with this algorithm is that even if S exists, the algorithm will never return anything. why? because your algorithm does not find fundamental laws, it always verifies if laws are not fundamental.

Analysis:
In the special case, the algorithm would not return S, then the above claim( conditional) would not apply. This could happen if there is not fundamental set of laws that would generate S.

no, if the algorithm never stops that means that the set of fundamental laws may or may not exist but if it exists than it is not recursively enumerable, hence not computable by the algorithm (which is exactly the case).

What does this all mean?

that your algorithm will never stop regardless if there is a set of fundamental laws of nature or not (even if we consider ideal conditions), because "S is not fundamental" can never become false- the only way it can become false is if the algorithm exhausts all the possible cases of S* and verifies that no S* can give S, something which is impossible because the set of potential S*'s that the algorithm must check is infinite, you can always construct new sets of supposed "laws".

the only thing that algorithm can compute (if altered a little) is if S is not fundamental, you only need add "return S is not fundamental" if there is S*, instead of S* -> S. by the method you described, you can eventually find -given sufficient time- whether S is not fundamental.

 

[vectorcube]

I know very well what you meant by "laws of nature", I was however alluding to something else...
I for one, stand by the contrary: there are no laws of nature. what you're referring to as "law of nature" is merely an artificial description.

QUOTE]


Your life.

incorrect phrasing. the correct wording here is:
if an S is returned than S is fundamental.

No. To prove p. You assume -p, and thus have to lead it to a contradiction.

What are you doing here?

because your algorithm does not find fundamental laws, it always verifies if laws are not fundamental.

Yes it does. See the word "find"? The loop test if the fundamental law is fundamental.
I leave the implementation details to the engineers.

recursively enumerable, hence not computable by the algorithm (which is exactly the case).

why is it not computatable? Curious. turtles all the way down.

because the set of potential S*'s that the algorithm must check is infinite,

Why? Suppose there are originally 30 laws, then there are 10, then there are 4... all the way to 1. The algorithm stops. It can` t get below 1.


For example:
There are 4 maxwell ` s equations. If you can express it in terms of an action, there would be only one equation. The algorithm stops.

 

[tauon]

Your life.

what is that supposed to mean?

No. To prove p. You assume -p, and thus have to lead it to a contradiction.

What are you doing here?

the question is: what are YOU doing here?
your algorithm never employs any contradiction.

it assumes the starting condition that S is not fundamental (meaning that there is S* from which S can be derived) and to prove that, searches for an S*. this is the very core of your (broken) algorithm.

did you not write it like this?

Yes it does. See the word "find"? The loop test if the fundamental law is fundamental.
I leave the implementation details to the engineers.

discussing anything with you is proving to be quite frustrating!
you leave the implementation to the engineers? what?

than you defined no algorithm. it seems you're clueless even about what you yourself wrote!

your algorithm does nothing then.
certainly it does not what you claim it does.

the only thing you wrote for your algorithm is a search mechanism: that's it! a mechanism that searches for systems which can yield S... try and edit it, maybe you forgot to write some parts so we're left hanging with an incomplete algorithm.

why is it not computatable? Curious. turtles all the way down.

because it never terminates! the program/algorithm never stops!
also, from what I found out now, it's not even an algorithm! you have no idea what it's supposed to do! you just wrote pretentious stuff using fancy symbols that hint towards a pseudo-algorithm. GOOD JOB!

:D

Why? Suppose there are originally 30 laws, then there are 10, then there are 4... all the way to 1. The algorithm stops. It can` t get below 1. For example:
There are 4 maxwell ` s equations. If you can express it in terms of an action, there would be only one equation. The algorithm stops.

hmmm, please enlighten me on how exactly a "more fundamental" system implies that there are fewer rules?
your inconsistent blabberings make no sense!

so there can absolutely be only 1 fundamental rule? hmm, okay...but how in the world do you justify this proposition?

 

[vectorcube]

what is that supposed to mean?

It means very slow.

the question is: what are YOU doing here?
your algorithm never employs any contradiction.

No, i never employs any contradiction in the algorithm.

I used it in my proof (which you quotes on the previous post). Read.

it assumes the starting condition that S is not fundamental (meaning that there is S* from which S can be derived) and to prove that, searches for an S*.

The assumption is "S is not fundamental when the algorithm ends" in the prove. I am talking about the proof that came after the algorithm. What are you talking about?

discussing anything with you is proving to be quite frustrating!
you leave the implementation to the engineers? what?

than you defined no algorithm. it seems you're clueless even about what you yourself wrote!

your algorithm does nothing then.

So when is it the norm that algorithm deals with implementation details?
The word "find" in this context would just be a module that i don` t necessary need to know, or define. I put the module there so that i can use it. Well, I guess you don` t know much about that as well.

because it never terminates! the program/algorithm never stops!
also, from what I found out now, it's not even an algorithm! you have no idea what it's supposed to do! you just wrote pretentious stuff using fancy symbols that hint towards a pseudo-algorithm. GOOD JOB!

It is pseudo-algorthm, and good job.

 

[vectorcube]

hmmm, please enlighten me on how exactly a "more fundamental" system implies that there are fewer rules?
your inconsistent blabberings make no sense!

Are you sure i am the one that is inconsistent?

so there can absolutely be only 1 fundamental rule? hmm, okay...but how in the world do you justify this proposition?

The point you need to know is that process from specific laws to more general laws cannot continue indefinitly. In each iteration of the loop, you get a much smaller set of laws, or equations. There are two ways to end. It either ends with some finite set of laws, or just one law. You say the iteration will not end shows that you don` t really understand the proof, and the algorithm Too bad that you don ` t even know what you actually don` t know to make a case. Does it make sense?

 

[tauon]

I used it in my proof (which you quotes on the previous post). Read.

oh that?

I already addressed that. it's a broken "proof". sure if the algorithm ever returns an S than that S is fundamental (I never claimed otherwise) the problem that I pointed towards is that the algorithm can never return such a result!

can you not read?

The assumption is "S is not fundamental when the algorithm ends" in the prove. I am talking about the proof that came after the algorithm. What are you talking about?

precisely about your opening post, and your defective argumentation!

So when is it the norm that algorithm deals with implementation details?

what?
I didn't ask you to write code in X programming language, I asked you to consistently and completely define your algorithm.

start> find fundamental system >if found> end

is not at all a well-formed algorithm. actually it's no algorithm.

The word "find" in this context would just be a module that i don` t necessary need to know, or define. I put the module there so that i can use it. Well, I guess you don` t know much about that as well.

you need to define what finding such a thing means, and when it needs to stop! learn a thing or two about algorithms and then come back so we can have a proper discussion.

It is pseudo-algorthm, and good job.

hmmm, never before was my spelling corrected, the correction resulting in a misspelling. very nice. :D

 

[vectorcube]

oh that?

I already addressed that. it's a broken "proof". sure if the algorithm ever returns an S than that S is fundamental (I never claimed otherwise) the problem that I pointed towards is that the algorithm can never return such a result!

But the iteration process is finite, and the loop does end. How is it "broken"? lol

can you not read?

Weak.

precisely about your opening post, and your defective argumentation!

defective to you? Well, i am not "broken".

hmmm, never before was my spelling corrected, the correction resulting in a misspelling. very nice

I am offend!

 

[Evo]

Locked pending moderation.