What is the Measurement Problem and What Would Solve It- Tim Maudlin

This is a talk that was given in the Rutgers Graduate/Undergraduate Online Seminar in Mathematical Physics (GUOSIMP). We host talks every other week on Thursdays from 12:10pm to 1:10pm EST.

Speaker: Tim Maudlin (NYU)

Accessibility of Talk: Beginner

Date: July 16th, 2020

Title: What is the Measurement Problem and What Would Solve It

Abstract: There are many conceptual puzzles surrounding quantum theory. Perhaps the most widely known is called “the measurement problem” or sometimes “the problem of Schrödinger’s cat”. But a cursory look at the literature reveals substantial disagreement about exactly what the problem is and, consequently, what it would take to solve it. It will argue that there are several different problems one might have in mind, and for one of them there are only three possible routes for solution. Further, other nearby problems have uncontroversial solutions in two of these three approaches and no straightforward solution in the last. Yet another nearby problem has often been entirely overlooked. I will try to impose some conceptual order on this nest of problems.

[manuderezzo]

I don't think C is clearly stated. What do you mean, the cat 'is' alive or dead? Typically, one would think it means that you can measure it alive or dead, and you can certainly do that, even if the wavefunction is in superposition. How would you go about experimentally verifying this statement?

[alexskriabow]

again with the audio. do you folks not understand how to record audio?

[JoyceElroy-zw]

Hernandez Linda Harris Jessica Clark Anthony

[WarrenEden-uc]

White Scott Thompson Betty Gonzalez Nancy

[MiddletonEdgar-gr]

White Angela Young Jeffrey Young Joseph

[HubbardGavin-ex]

Lopez Jose Johnson Thomas Miller Elizabeth

[IsaiahUla-rw]

Anderson Charles Anderson Brian Lewis Barbara

[edwardlarson]

I believe that GRW is on the right track and identifies a real feature of nature. To my way of thinking, spontaneous collapse is at the heart of spontaneous emission, radioactive decay, and tunneling phenomena.

[walterbrownstone]

I bought a measurement device. Laser distance meter. It has a length of 150m and an accuracy of +-1.5mm. That is amazingly precise! That means it says 150m only when the actual length is between 149.9985 and 150.0015m! That also means that itreads 1mm for all objects between the length of -0.5mm and 2.5mm. The smaller you measure, the less precision you have. That is a problem of you, not the particle. Don't ever blame the particle for your shortcomings.

[LaboriousCretin]

⚫️ 🐈‍⬛️ 🐈 🐉 ☢️ . 📦 🐱 〰️ 🐲 🥷 🪤📊🎲🪩🌎🐇🧭🐰♾️💥🫧. Thank you for sharing the lecture. 😻 🪙

[whatsupdoc-fm]

34:26 You state that it is possible to prepare particles in such a way that they always have spin up (or down)? I thought it was impossible to predict for an individual particle whether it would end up UP or DOWN on the screen?

[johnmalik]

The thing that is measured does not exist independent of the measurement. They are not two independent things. It is that simple. Scientists assume that they are separate. Reality applies to life, not to theory.

[david_porthouse]

The collapse of the wave function is very likely to be
a nonlinear process, for which computer simulation is
needed. Any simulation needs to make use of a random
number generator. I believe I am stating the obvious.

I will propose in outline how to go about it. This is
likely to be a long posting, I am afraid, but I am
going to offer a tentative solution to the measurement
problem. Our first difficulty is that the Schrödinger
equation, or any equation like it, is very very
accurate at the ensemble level and we can assume that
the modification of it is forbidden. There is nothing
resembling the viscosity term to be found in the
Navier-Stokes equation. And yet we need to inject some
randomness. There are two ways to do it.

*The first way* is to hypothesise that some nonlocal
degree of freedom is involved, so even if we know
nothing about Bell's Theorem we could have guessed it
anyway. Just playing around with the Minkowski formalism,
we notice that there is more than one way to travel
faster than light. I suggest that the Schrödinger
equation describes an oscillation in one of the ways
which is capable of destructive interference with itself.
We can have an orthogonal tachyonic Wiener process in
the other way, which I will just call tachyonic Brownian
motion (TBM). This comes into action during the nonlinear
interaction between the wave function and the
electromagnetic field, and can then lead to an outcome
which does not have an issue with Schrödinger's cat. No
aetiology is proposed for this TBM and I am guessing that
it is quantified by having the Planck time as its
characteristic time.

Nitrogen tri-iodide has the unique property that it is
so unstable that it can be detonated by an alpha particle
from a substance like polonium-210. Nitrogen trifluoride
is stable by contrast. A computer simulation of tri-iodide
under bombardment needs to have an outcome which is
qualitatively different from the trifluoride. In any
well-written simulation the trifluoride behaviour will be
isentropic, but with the tri-iodide there will be a
destruction of unitarity and a substantial rise in
specific entropy. It is suggested that the missing
ingredient in the simulation is TBM, which being normally
orthogonal is quiescent in the trifluoride, but is
sufficient to detonate the tri-iodide once an
electromagnetic field is also present.

Two molecules of nitrogen tri-iodide are in fact a
detector in the classical sense, and constitute the
smallest detector that I can think of. What is called
the Heisenberg cut comes between one and two molecules
of tri-iodide. Maybe in the future somebody will think
of a smaller detector, but it won't really affect the
argument to be given here. The computer simulation of two
molecules of tri-iodide will need to run in at least
twenty four dimensions of configuration space, just
counting atoms. This is impossible in practice, and gives
us a hint of what we are up against. All detectors are
just too complicated to model by the formal method using
TBM. We really do need a detector to get the collapse of
the wave function in our simulation, but we must adopt
other ideas.

*The second way* to reconcile the immutability of the
Schrödinger equation to the need to use a random
number generator involves a bit of handwaving. We just
throw away the Schrödinger equation and replace it by
a classical system with some ordinary Brownian motion
for any object heavier than the Planck mass. The
Heisenberg Uncertainty Principle is replaced by the
Fürth Uncertainty Principle on the same scale, so hardly
anyone will notice. Yes, it is indeed handwaving, but we
have TBM as an aetiology and no known practical
alternative.

Classical BM will be much more disruptive than TBM and
of course we are reinventing decoherence. We shouldn't
have much trouble collapsing wave functions using it
in our simulations. The usual objection to decoherence
is the lack of any means of destroying unitarity in any
closed system, which has been answered by proposing TBM
as the aetiology.

If we have an electron in a potential well, then the
electron is modelled by the Dirac equation plus TBM.
The electromagnetic field is modelled by correlated
TBM so the wave function and the electromagnetic
field working together can act like a nonlocal Vernam
cipher. The potential well is considered to be a dimple
in a heavy object so it is modelled with a bit of
classical BM. I have already written a little computer
simulation of the Dirac wave packet, and I am guessing
that the propensity of the monochromatic wave packet to
be a tachyon is also going to be significant.

I intend to put a series of computer simulations in the
public domain which anyone will be able to modify. If
they wish, they can rip out TBM and install some other
way of doing things. What has been described here is
the projected solution to the measurement problem as
we start our programming. We will just have to see how
we get on.

[HoyleBarret-pe]

Miller Angela Thomas Donald Perez Barbara

[Nathillien]

Generally; "Wave function collapse" puts the quantum system into a eigen-state. But is the Dead and Alive an eigen-state of a Cat???
I think it is NOT - and that is why this whole quantum mechanical experiment is not applicable to a system like a Cat.
Also:
Can a Cat be described through a wave-function that is some kind of a combination of all the Cat particles wave functions?
I don't think so. All those particle wave-function are collapsing all the time when interacting with other parts of the cat and the environment.
So constantly, parts of the Cat wave-function are collapsing while some other parts are not.
The Cat wave-function cannot be linear. So it cannot be described with Schrodinger Equation.
Not applicable.

[zeroonetime]

We cannot measure a measurement. Or, can weee??

[BessieOscar-eb]

Clark Matthew Williams Scott Young Donna

[sonarbangla]

Recently t' Hooft published a video explaining how probabilistic nature of reality turns deterministic and classical. Tim Maudlin was recently publishing theories of consciousness, biology and genetics, when did he become a physicist?

[chrimony]

A potentially good talk ruined by an invite for interruptions and an incredibly loud mic on the other end. Can't watch. Oh well.

[davidhand]

GRW is not a solution. Whatever actually does happen during a collapse, no matter what causes it, can be linear or local. The Schrodinger equation either describes time evolution or it doesn't. Copenhagen has the Born rule, which is not the Schrodinger equation and cannot be linear, and refuses to give any idea what determines which time evolution equation nature uses. A random collapse theory just disposes with the pretense of defining what moments of time have a non-local, non-linear transformation instead of following the laws of physics. It sweeps the problem under the rug even farther than Copenhagen does, but it stinks all the same.

Given the additional constraints of satisfying GR and the experimental realization of the Bell inequalities, the only reasonable way to solve the measurement problem is the Everett interpretation. The additional "hidden" variable is _your_ state, or equivalently, that of the measurement device. You are looking at a live cat _and_ you are looking at a dead cat, but those are two states of _you_ that can't causally affect each other in order to communicate.

I _hate_ the other term for the Everett interpretation; it's highly misleading. There's only one wavefunction. There are no new copies of anything being spawned every time something is measured. You didn't start with one cat and end with two. You and the cat both started in two indistinguishable states and ended up in two distinguishable states, separated by a wall of noise. Once you get to the actual physics and math, Everett is merely the elimination of the Born rule axiom, which is itself in contradiction with the Schrodinger equation in the first place.

Why can we reliably calculate probabilities, then? If a system has exactly N outcomes and there is no reason for one to be any more likely than another, they can only be equally likely, i.e. p = 1/N. No matter how complicated the system is, the states can always be subdivided into equally likely parts. Mapping those parts to experimental outcomes and adding them up gives you a probability. It's really not that complicated. There's an episode of PBS SpaceTime that went into this recently and explained it in detail, and my comments there address all of the reservations they raise.