Can slow change solve the metalaw problem?

[Martin Sallberg]

The standard argument for the claim that time "must" be an illusion is the metalaw problem, that asymmetry such as fundamentally irreversible time in which the future literally does not exist would allow the laws of physics to change and we have not detected any change of the laws of physics over time. Can the problem be solved by setting the laws of physics from just after the Big Bang to slowing down any change of the laws of physics in the extreme? So extreme that the 13,8 billion years that have elapsed since could not produce a sufficient change in the laws of physics to be detectable even by the best instruments we have today? Does anyone know if this is mathematically possible?

 

[Fra]

If the context of your question is Smolins evolving law and corresponding metalaw dilemma then yes i think the only sensible way of thinking is that for all practical purposes laws do not evolve anymore.

1) smolins cns says that the laws mutate only at the big bang. But what is insatisfactory about that is that the state space pf physical law is assumed and i see no satisfactory mechanism for how "genes" mutate in big bangs. Its more a concept to me than final.

2) another option is to consider a kind of principle of prescedence smolin sniffs on where some self organisation randomly takes place and laws formed/selected. Then the important changes would be tiny fractions of the first second. After that the "laws" as we know them are fixed.

But like many other hypothesis none is a coherent mathematical reality yet. It seems too few people are working on this. I am myself fully convinced that any rational notion of law must evolve in response to interactions. But this idea is very deep and can itself only be grasped in a learning perspective. Ie its the one asking about the distinction between if laws are created or "discovered" that is really confused. In the learning perspective the two scenarios are indistinguishable.

/Fredrik

 

[Dale]

slowing down any change of the laws of physics in the extreme? So extreme that the 13,8 billion years that have elapsed since could not produce a sufficient change in the laws of physics to be detectable even by the best instruments we have today?

Such a proposal would introduce some additional uncertainty without explaining any additional data.

 

[Fra]

Such a proposal would introduce some additional uncertainty without explaining any additional data.

Yes but any idea of evolving law must be used together with selection principles and other constraints without running into the metalaw dilemma. Otherwise you are right that all we would do is make the landscape problem even worse.

But HOW todo this is still open and to my knowledge while smolin raises and elaborates the idea i see why is seems strange to ppl that don't see how a missing ingredient can help.

In ST for example, i think the landscape considered seems larger or more unmanagable simply due to some missing selection constraints.

/Fredrik

 

[Dale]

Yes but any idea of evolving law must be used together with selection principles and other constraints

That doesn't resolve the issue I mentioned above. It just moves the additional uncertainty from the laws to the constraints, but since both are needed to make an experimental prediction you are still left with additional uncertainty without explaining additional data.

 

[Fra]

That doesn't resolve the issue I mentioned above. It just moves the additional uncertainty from the laws to the constraints, but since both are needed to make an experimental prediction you are still left with additional uncertainty without explaining additional data.

I suppose that depends om how you implement this, i do not see it as necessarily so.

In the way i have in mind the conjecture is that there is a natural balance between the constraints and the mutation of law - just like there is in biological evolution. The motivation is obvious - as any other way would not yield a steady state. If done right it will show self organizing behaviour and this should yield predictions and if none of the predicted steady states correspond to standard model i would consider the program falsified. Its the relation between constraint and mutating law that must be right or we will see what tou say.

/Fredrik

 

[Dale]

there is a natural balance between the constraints and the mutation of law - just like there is in biological evolution.

[sarcasm] Right, because being similar to biological evolution would definitely eliminate additional uncertainty since the result of evolution is completely certain and uniquely predictable [/sarcasm]

There is no way around this as far as I know. If you want laws that vary then the predictions have additional uncertainty. That is acceptable if it explains additional data, but if not then many scientists prefer the less uncertain laws either using Bayesian reasoning or Occham's razor

 

[Fra]

If you want laws that vary then the predictions have additional uncertainty.

You mean that the prediction is more uncertain if the rules of the prediction may change in unknown ways relative to if the rules where fixed.

This is of course true, but the point is that we do _not known_ which these fixed rules are - there is rather a landscape of them and no mechanism for selection. We need to describe the transitions in the landscape. That is also rhe key to Explain the standard model from firar principles by reducing the number of fundamental parameters.

So we still have the uncertainty.

Evolving law rather improves the situation by considering the physical process whereby law might change. If we can do this and solve the metalaw dilemma by postulating that the fundamental mechanisms here are random walks then the fitness of our paradigm has been improved.

if it explains additional data, but if not then many scientists prefer the less uncertain laws either using Bayesian reasoning or Occham's razor

I agree here. It should explain additional data, that the idea otherwise we gain nothing. Wether someone will succeeds in this remains to seem. But i insist that its by no means unreasonable.

/Fredrik