Skip to content

Instantly share code, notes, and snippets.

@Nikolaj-K
Last active November 21, 2021 17:32
Show Gist options
  • Save Nikolaj-K/37faa8a7a7afb5fa376ee09ebba0a545 to your computer and use it in GitHub Desktop.
Save Nikolaj-K/37faa8a7a7afb5fa376ee09ebba0a545 to your computer and use it in GitHub Desktop.
Sabine Hossenfelder's Superdeterminism

Sabine Hossenfelders Superdeterminism

Introduction

This weeks paper is

Superdeterminism: A Guide for the Perplexed by Sabine Hossenfelder with some notable reference to other papers.

Exposition paper. In 10 section, 20 pages. Read

https://en.wikipedia.org/wiki/Superdeterminism https://en.wikipedia.org/wiki/Interpretations_of_quantum_mechanics

  • QUOTE Outlook

Read

https://en.wikipedia.org/wiki/Lindbladian

PLEASE PROPOSE PAPER IN THE READING STREAM.

Person

https://en.wikipedia.org/wiki/Sabine_Hossenfelder

https://www.amazon.com/s?k=Walter+Greiner&i=digital-text&__mk_de_DE=%C3%85M%C3%85%C5%BD%C3%95%C3%91&ref=nb_sb_noss

Her youtube channel (200k subscribers): https://www.youtube.com/results?search_query=Sabine+hossenfelder+Superdeterminism+Bohm

Blog (since 2008): http://backreaction.blogspot.com/

Paper skim:

  • 80 papers going back to 2002 (over 4 papers per year)

Her arxiv papers https://arxiv.org/search/?searchtype=author&query=Hossenfelder%2C+S

  • QM Foundations
  • Otherwise dark matter / cosmology
  • Minimal Length

Search "determinist"

Relevant:

  • "A Superdeterministic Toy Model" - https://arxiv.org/pdf/2010.01327.pdf

  • "Superdeterminism: A Guide for the Perplexed" - https://arxiv.org/pdf/2010.01324.pdf

  • "Rethinking Superdeterminism" The present paper has a lot of vocabulary used and possibly discussed in that paper. That one is similar but also coauthored and touches on other (counter-)arguments. Advertises Palmers "Invariant Set Theory" => Talk "Rethinking Superdeterminism" (May, 2020) + Discussion https://youtu.be/YglT09Korr0

  • "Testing Superdeterministic Conspiracy"

  • "Testing super-deterministic hidden variables theories" => Talk "Sabine Hossenfelder: Testing Superdeterministic Conspiracy (EmQM13)" https://youtu.be/cbSc-PLGU8o Proposes an experiment to distinguish QM and e.g. superdeterministic theories See 2014 paper (which itself relates to a 2011 paper) https://arxiv.org/pdf/1401.0286.pdf

Other people

################################# Toy model paper, brief intro

5

  • QUOTE "The model, when averaged..."

################################# Questions

Toy model with 2 non-commuting observables A and B - does it still depend on 1 of the observables (or both)?

Why not consider a theory with a plurality of dynamical equations (if you measure A, you gotta compute the expectation with the dynamical equations for A.)

################################################################################

Abstract

Only local, deterministic description of QM. (Says it's a 'description', not an 'interpretation')

1 Introduction

Wants to leave aside the "free-will debate", which she discussed in a previous paper

2 How to think about superdeterminism without tying your brainin knots

Vocabulary

"Psi-epistemic": Wave function is not ontic ("not an object in the real world") "Statistical Independence": Degrees of freedom of spatially separated systems can be considered uncorrelated, even if they have a common past.

The "Hidden variables" in the language of e.g. Bell-type experiments are not necessarily unmeasurable, but "hidden" only means they are not part of the QM theory.

Superdeterministic theory: The evolution equation of the measured system s is correlated with the parameters l that are part of the measurement (the values "choosen" by the measuring actor, where this "choice" is then also not necessarily uncorreleated/independent of said evolution.)

Note that here this does not mean a live dependency of the former on the latter: We merely demand the possibility of depencency of the evolution of s on l-at-measurement-time (not before, although if the theory is also deterministic, then we can also say s depends on l at setup).

(Sidenote: So for "superdeterminism"+"not necessarily deterministic", it is not as general as one mathematically would have the freedom to model.)

In this discussion, no assumption or (explanation, if you will) is made about where this correlation comes from

  • it might e.g. be a result of past vicinity of aspects that are part both of the measured system as well as of the measuring agent.

3 The role of non-linearity for superdeterminism

(2 pages)

"All the trouble with interpreting quantum mechanics comes from taking this probabilistic prediction to be fundamental, rather than a consequence of missing information. This is because if one takes the probabilistic prediction to be fundamental (ie takes a Psi-ontic perspective) then [...]"

"we do not want to make the Schr ̈odinger equation non-linear, we are looking for an underlying non-linear theory that gives rise to the ordinarily linear evolution law of quantum mechanics"

?Call for non-first-order-in-d/dt equation of motion

See toy model

4 Superdeterminism is neither classical, nor realist, nor an inter-pretation of quantum mechanics

(S. Hossenfelder) Writes that she's not a Realist.

Superdeterminism is emphasized as above-QM theory framework, as opposed to a mere interpretation of QM. Draws relatively close comparison (spontanous) collapse theories https://plato.stanford.edu/entries/qm-collapse/ https://en.wikipedia.org/wiki/Objective-collapse_theory in this regard

Makes quite the strong commitment (or at least preference for) superdeterminism.

5 Superdeterminism isn’t as weird as they say

Fending off philosphical objections (also done in the prev. paper with palmer)

6 But someone told me superdeterminism would kill science

Fending off philosphical objections (also done in the prev. paper with palmer)

7 The relevance of future input for superdeterminism

"As previously stressed, the superdeterministic evolution law depends on the measurement settings at the time of detection. There is no non-local interaction here. The prepared state need not interact with the detector before measurement, it’s merely that the setting appears in the evolution law. For a simple example of such an evolution law, please refer to the accompanying paper [5]." (Sometimes called "retrocausal", but she doesn't like that term.)

See next section of use of this for potential superluminal signaling.

8 Superdeterminism may or may not allow superluminal signaling

See image. Considering measurement of P at D1 and D2, which may thus depend on both.

Clarifies form of information gain.

Discusses/Comments on a certain paper [16] that speaks about the potential capability of superluminal messaging.

9 Superdeterminism is not necessarily finetuned

"A “finetuned” theory or a “conspiracy theory” is a theory which lacks explanatory power. Alack of explanatory power means that the theory requires more information as input than justcollecting the data. This notion of finetuning makes sense because a theory that is finetuned inthis way is unscientific. We should prefer just collecting the data over such a theory." "To summarize this section: A model is finetuned if it lacks explanatory power. To evaluatewhether a model is finetuned, one therefore needs to look at a model."

Discusses two notions of fine tuning (pen example) and how scientifically based/useful they are.

10 Outlook

"The lack of experiment could be remedied with existing technological capacities even in absence of a fully-fledged theory if the quantum foundations community could overcome their obsession with Bell-type tests. Bell-type tests will never allow us to tell superdeterminism apart from quantum mechanics. Instead, we need to look for deviations from Born’s rule in small andcold systems [1, 23]"

? attractors

"What prevents us from making progress here is simply that we haven’t made enough effort. Because superdeterminism has historically been stigmatized as “unscientific”, pretty much nowork went into formulating a useful theory for it. So, at present all we have are toy models. Isincerely hope that these notes will serve to encourage some readers to explore the benefits ofgiving up Statistical Independence and make headway on developing a theory for superdeter-minism."

Style

Overall clear.

Too casual for me, especially section 7. This chapter has a particular conversational tone (in the paragraph talking about muffins.) and beyond. Next page: "Thing is though, .."

Often the first paragraph gives clear definitions which are then used to argue for or against something.

This article likes to take apart missconceptions.

Thoughts

I'm not inclined to reject consistent possibilities by virture of non-falisbility, so the superdeterministic scenario - which allows for more theories - should be welcomed.

However, I also don't demand that any possibly QM-underlying theory look like a classical one either.

References

This paper "Superdeterminism: A Guide for the Perplexed": https://arxiv.org/pdf/2010.01324.pdf

Her other papers: https://arxiv.org/search/?searchtype=author&query=Hossenfelder%2C+S

From the present paper (from section "10, Outlook"): "The toy model put forward in [5] should be understood as a effective limit of a superde-terministic theory. In this effective limit, the detector settings are hard-coded into a modifiedHamiltonian evolution. The theory we are looking for would explain how one obtains such aneffective limit."

Has an evolution equation (d/dt)rho = -i [H, rho] + K_{l_hidden, born_rho} rho where K depends on hidden variables and, though some transition probabilities, on the observable (its eigenstates) as well as the latest state (depending on H).

And this is then generalized to models of more complicated states. ? Does this QM-reproduction work when you have two non-commuting observables, i.e. so that one averaging produces reproduces QM-claims of two non-commuting observables at once.

From 5, the conclusion: "The model, when averaged over the uniformly distributed hidden variables,λ, reproducesthe quantum mechanical predictions. If the variables were not uniformly distributed, for exam-ple because the sample space is not large enough, this model would give predictions that differfrom quantum mechanics. That is, even in an experiment where one expects to have a goodenough sampling to reproduce Born’s rule with only small statistical deviations, the observeddistribution could still be heavily skewed in the space of hidden variables, hence not approxi-mate Born’s rule as expected. This makes clear that superdeterminism is not an interpretation ofquantum mechanics; it is a concrete possibility for replacing quantum mechanics with a bettertheory and more effort should be made to experimentally test it." "Exactly when the deviations from quantum mechanics become non-negligible depends onwhat the hidden variables are; the above introduced toy model cannot answer this question.The toy model is therefore, strictly speaking, untestable, because it does not specify where thedistribution of hidden variables comes from."

Keywords

Superdeterminism, Statistical Independence, Pointer state, Decoherence,

############################################

Questions

? Regarding messages (e.g. in the context of superliminal messaging), how does (super)-determinism go together with intent of agents (messenging around)

? "on-linear evolution law of the type discussed above, where the locations of the attractors depend on the detector setting"

? "pointer states of the detector"

? What are "spontaneous collapse" versions of quantum mechanics?

? What's this whole fine-tuning discussion in this context?

? Is there one unique evolution equation for a physical system or - if there are various observables - are there several at once. (I guess there's only one and Statistical Independence makes it so that no other "can" be chosen?)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment