Comment by jostylr | Hacker News Reader

Comment by jostylr | Hacker News Reader

Parent Full thread

Bohmian mechanics is based on the idea that we perceive stuff to be in a certain position in a single reality because there is a correspondence to stuff being actually there. That's nice. If the particles are surfing a wave and not impacting it, so be it.

It is also rather nice to think of the particles as just being points in space with nothing else associated with them; an electron is just an electron because the portion of the wave function that is relevant and guiding it is the electron portion; see a paper from 2004 entitled "Are all particles identical?" [1] (I am a coauthor on that). If one thinks about it, we only know about particles through their motion so having things like mass and charge linked to the object guiding the particle seems perfectly reasonable. Points are not only not labelled by numbers (particle 1, 2, etc) but also not labelled by mass and charge.

The nondeterminism of not knowing the initial conditions is fine; the point was to have a theory with well-defined objects that give some plausible story and connection to our experiences, such as stuff existing and being somewhere. The fact that non-relativistic Bohmian mechanics happens to be deterministic is just happenstance for many of its supporters. In some QFT versions, the dynamics of creation is not deterministic and there is no reason for that to be a problem. But it is well-specified without having to invoke some special magic action called "observation".

As for QFT, the biggest problem for Bohmian mecahnics is the need to have an actually well-defined evolution of the wave function. The idea of particles being created and annihilated is not particularly hard. And, in fact, recent work has shown that if one takes that seriously and respects probability leaking from n particle space to n+1 and n-1, then at least some of the divergence problems go away. See [2]

1: https://arxiv.org/abs/quant-ph/0405039 2: https://arxiv.org/abs/1809.10235

> Bohmian mechanics is based on the idea that we perceive stuff to be in a certain position in a single reality because there is a correspondence to stuff being actually there. That's nice. If the particles are surfing a wave and not impacting it, so be it.

At that point it's very obviously a violation of Occam's razor though. It's like positing that the content of my field of vision is an objectively real thing, that the reason the universe looks like a video projection is that there really is a video projection going on, even though that video projection has no physical effect.

> If one thinks about it, we only know about particles through their motion so having things like mass and charge linked to the object guiding the particle seems perfectly reasonable.

Indeed. But if one thinks a little more, what's the point of positing a particle at all, if all of the physics is in the pilot wave?

The definite positions of your brain states evolution is correlated to the other positions of all the stuff. The other particles do have an effect on your evolution and there is a "you" set of particles one can talk about. Remember the wave function is a function on configuration space so evaluating the guiding effect on the particles is to have to know what point in configuration space it is at; this is actually the troubling bit and leads to the nonlocality concerns, but that problem is common to any quantum theory with definite results happening.

The physics, therefore, is not all in the pilot wave. If you take as the point of a particle theory that there should be particles with positions changing in time, then that is what is being given in Bohmian mechanics.

Also, ask yourself, if the wave function is on configuration space, what constitutes a configuration? In Bohmian mechanics, it is clear, but if the wave function is all there is, then why are we talking about configuration space at all? It is just this abstract vector in Hilbert space evolving and many different representations can happen. Why do we not perceive reality in terms of these other representations?

If it helps, you can think of the wave function a bit like a dynamic law. In [1], the authors suggest thinking of log( psi) analogously to the Hamlitonian H on phase space in classical mechanics. There is no back action on H and most of it is irrelevant to the evolution of a particular particle system in that framework and yet everyone recognizes it as just a convenient way of describing the dynamics.

The difference is that psi evolves but even that may only be true on a subsystem point of view. It is theoretically possible to have a stateless universal wave function which, when particular particle positions of the environment are plugged in, nonetheless gives evolving subsystem wave functions.

Occam's razor is difficult to apply here without a prejudice. If you want to minimize the number of equations, then sure, "the wave function is everything" works, but it comes at the cost of there being what could be considered an infinite number of "you"s and everything else, all slightly different and whole existing other expressions of the universe with no connection to us. If you want collapse somewhere, then you have to posit that mechanism.

On the other hand, by adding in particles and the guiding equation, one gets a singular "you" and everything that we experience is, more or less, definite and singular. So the "existing" stuff is dramatically reduced.

Which one of this is truly simpler is a matter of taste, I would say. I think in terms of communicating with people, the Bohmian version of "there is this universal wave and the positions of stuff are guided by it" is pretty simple. The law itself is so trivially a part of the Schrodinger equation that it could easily be derived before the Schrodinger equation itself. Contrast this with other versions which is "reality collapses to a definite state when we look at it" or "there are infinitely many different universes". None of those seem as simple.

1: https://arxiv.org/abs/1101.4575

> there is a "you" set of particles one can talk about

We know that particles don't have identity though - exchange of identical particles is a symmetry and physics would be very different if it wasn't. I won't claim it's compelling, but to me that suggests that a particle is more like a pattern or a field excitation than a thing with its own concrete existence.

> Why do we not perceive reality in terms of these other representations?

What would be different if we did? I mean obviously at a macroscopic level particles moving through space is a model that gives a good approximation and is easy to think in, but that doesn't mean they're any more physically real than e.g. temperature.

In physics, particles not being labelled by anything other than their trajectories is a very natural starting point. When one uses the natural configuration space, one without labels in which a configuration is a set of n points in physical space rather than an ordered n-tuple, then the complex-valued wave functions on that space are exactly those of boson type. To get the fermions, one replaces the value space with a 1 dimensional complex bundle over the configuration space, one which twists in the right way. A paper I coauthored explores this in a general context: [1]

The "you" is then a rough set of particles whose trajectories roughly coincide with your macroscopic trajectory. Their identity is just given by where they are.

As for representations, I feel like I can easily understand how to get momentum or temperature from particles with their time evolution (trajectories), but I do not see how, say, to get positions of particle just from knowing what their momentums were and their time evolution.

1: https://arxiv.org/abs/quant-ph/0601076

But we don't have even a set of definite trajectories. If we see e.g. an electron coming towards a hydrogen atom and then an electron moving away from it, not only do we not know whether the incoming electron "bounced off" or whether it settled into the orbital and "kicked" the electron that was already there out, but in a fundamental physics sense what occurred is some weighted average of both (in the same way that we don't merely "not know" which of the two slits an electron went through but in an important physical sense it partially went through both).

It depends on the theory. The Bohmian theory, which is what I have been using, is one in which electrons have actual positions that change over time along trajectories. We may not have access to that data, but that is fine. Certainly in simulations one would be able to see which scenario happened. For some, it might be the same electron moving away, for others it would be kicking one out. One could definitively say which one is happening in the simulation. In experiments, we cannot say that because our access to the knowledge is limited by quantum equilibrium. The quantum formalism is very much like thermodynamics in that regard; the individual details are missing, but the larger picture can be computed. Nevertheless, in a Bohmian world, the electrons have their distinct identities as distinguished by, and only by, their trajectories.

Reply on news.ycombinator.com