Thursday, April 30, 2009

Suárez on Quantum Propensities

Interpreting dispositional or power properties as propensities seems to me to be a very promising avenue for ontology. This is because theories employing powers (the ones I’ve seen) don’t get the modal structure of the world correct: by assuming that powers entail their manifestations, they fail to provide truthmakers for possibilities. Taking powers to be probabilistically manifested propensities solves this problem. It also bolsters a realist account of causality: actualizing propensities into specific outcomes gives causation some “real work” to do.

Finally, propensities can serve as a link between a philosopher’s ontology and the interpretation of quantum mechanics.* I was happy to find recently (via Philpapers) that propensities have a champion in Mauricio Suárez, a philosopher of science at Complutense University of Madrid. In several papers he has explored and advocated the propensity approach to understanding the properties of quantum systems. He also has pointed out the value of propensities to the dispositional/power property approach to ontology.

Popper and propensities

Propensity theory seems to be a relatively neglected topic these days. The work of Karl Popper may be one reason for this. Given Popper’s stature, the fact that his propensity interpretation of probability is widely regarded as a failure is discouraging. But Suárez makes a strong case that the problems with Popper’s theory are the result of his emphasis on interpreting quantum probabilities specifically, and also due to some particular assumptions which can be set aside or modified.

Popper wanted to interpret quantum probabilities using propensities, and he also thought propensities could be used to interpret probability in an objective manner generally. This effort has been roundly criticized. An important criticism is that known as Humphreys' paradox (after Paul Humphreys, see this paper). Humphrey pointed out that the asymmetric causal nature of propensities made them inconsistent with the symmetric character of conditional probability. But this paradox is only a problem for propensity interpretations of probability. When it comes to quantum theory, Suárez makes a wise point when he says that what we want to do is interpret quantum mechanics, not quantum probabilities. The probabilities observed in experiments would be explained by our account of quantum mechanics. This “clicked” for me: in my own reading of papers on interpreting quantum probability, I had found that the arguments tended to point toward subjective or Bayesian interpretations (see posts here and here), but this work didn’t seem to help one progress toward a satisfactory ontological interpretation of the physics. Perhaps it is better to interpret the ontology first.

With regard to some of Popper’s other assumptions about the nature of quantum propensities, Suárez explains in the paper “On Quantum Propensities: Two Arguments Revisited” how two other criticisms of Popper’s view may be avoided by a revised account of propensities – specifically Suárez’ ”selective propensities” proposal.

Selective Propensities

In addition to the paper mentioned above, Suárez has two other papers which discuss his selective propensities approach. In “Quantum Propensities”, he looks at some other historical attempts to employ propensities to interpret QM, and then contrasts his own proposal. 2004’s “Quantum Selections, Propensities, and the Problem of Measurement” develops the approach in the most detail, showing how it builds on Arthur Fine’s “selective interactions” solution to the quantum measurement problem.

Suárez’ approach is new to me and I’m still trying to understand it (I had not been exposed to Fine’s work before either). It seems that in the selective propensity interpretation, a quantum system possesses a number of dispositional properties coinciding with the observables we measure in experiments involving particles. These properties manifest themselves consistent with the probability distributions we observe in QM. We assert that in a measurement one interacts only with the property of the system selected. The interpretation then says that to explain the result, we can employ a mixed state of that property’s eigenstates to describe the initial preparation, rather than plugging in the full quantum state of the system. (The full quantum state encompasses all of the system’s properties.) We can still interpret the interference effects which result in some experimental setups as due to the interplay among the system’s various properties consistent with the full state superposition of the system.

This seems to imply that the description of the initial state of the system (setting up either a mixed state over one observable or the full state) is altered by how we set up the experiment. This seems strange at first glance, but I guess there’s always going to be something strange when you’re working with QM. I also wonder how to think about generalizing this scheme to understand how interactions work beyond the laboratory setting.

I’ll try follow up with more after re-reading and digesting this material further.
----------------------------------------------------------------------------------
* I’ve often thought about the issues involved when philosophers try to make sure their metaphysical ideas comport with physical theories. On the one hand, philosophers very much want to avoid proposals which seem to conflict with science. On the other hand, since our physical theories are provisional (and likely to be replaced in time by improved theories), perhaps philosophers shouldn’t worry if well-motivated ideas imply revision to current scientific understanding. I’ve seen relativity theory invoked to criticize philosophical positions (e.g. presentism in the discussion of time – see an abstract of what looks like an interesting paper here), but many recent research programs in quantum gravity explore the idea that relativity is an effective (low-energy regime) theory rather than something fundamental. The search for a theory of quantum gravity implies relativity, quantum mechanics or both will need to be revised.

So, while I personally want my metaphysical theory to accommodate quantum mechanics (and worry less about conflicts with relativity), I realize that this is tricky territory. It seems best to just be explicit about one’s presumptions.

8 comments:

Allen said...

Hey, what do you think about this:

http://www.higgo.com/quantum/fourreasons.htm

Steve said...

I'll check it out. thanks.

Steve said...

My feeling about this is similar to that other paper which extolled the simplicity of the information-theoretic idea that the simplest answer to the question "What exists?" is "everything".

I say not that this is wrong per se, but that it is only half the story.

There's is a tension throughout: he can't avoid repeatingly discussing "views", "observers" and "perspectives", but wants to devalue these as compared to the objective truth that all information exists. But the existence of the myriad subjective points of view goes hand in hand with the existence of the totality of "objective" information.

I've made this point when discussing the interpretations of quantum mechanics:
Copenhagen goes wrong in trying to deny the reality of the wavefunction. But MWI goes equally wrong when trying to deny the existence of wave function collapse. Both are unavoidably part of reality.

I agree that our view is subjective, not objective. But the collection of all subjective views=the collection of all objective information. That's why consciousness is an integral part of reality. It is the processor of information, and all information exists to be processed.

Allen said...

Hmmmmmm. I have something to say on this, but it's still percolating.

Also of interest:

http://www.overcomingbias.com/2008/05/if-many-worlds.html

Steve said...

Thanks. I actually read most of those lengthy qm posts at overcoming bias and found them 100%wrongheaded. In fact the irony given the name of the blog is that it is a preconceived *bias* that a collapse process is "unphysical" that drives the argument for many-worlds. As if nature has to conform to one's (classically formed) notion of what counts as physical.

(In other posts he argues decoherence theory supports many-worlds. That is also dead wrong. It does not support any particular interpretation.)

Unlike many critics of many-worlds, I have no problem with proposing a multiverse per se. But trying to interpret QM without measurement makes no sense to me. QM theory exists to make sense of experimental results, i.e. measurements. And you can see this in the tortuous arguments many-worlds folks try to make to recover the role of probability in our world.

Allen said...

Possibly of interest:

Study Rules Out Fröhlich Condensates in Quantum Consciousness Model

http://www.physorg.com/news155904395.html

Steve said...

Thanks for that link. I think Hameroff probably does tries too hard to invoke quantum processes which are extended across the brain. I think the non-trivial quantum effects will be restricted to shorter distances.

Thoughts said...

Your link to the article by Eagle is interesting. I am always depressed by how many philosophers are closet presentists. If they came out of the closet it would be easy to say "hey, so you think the world is just a succession of frozen instants, no wonder you find philosophy intractable!" In the absence of presentism philosophy is challenging but not intractable.

See Materialists should read this first

Suarez' paper is intriguing, he summarises his view as "I defend the view that a selection is an interaction of the pointer position observable of a measurement device with a dispositional property of a quantum object."

The issue then becomes how the measuring device got a defined state. This reduces the measurement problem to the preferred basis problem. The interesting feature of the preferred basis is that it ends up with anthropic interpretations - ie: the observer must exist with a particular state so the state of the universe must be selected to support the existence of the observer. My own guess is slightly different, I would postulate that observation depends upon a particular form of space-time and this means that observers only exist in the part of the multiverse that has this form.

See Time and conscious experience for a discussion of how a particular form may be needed for observation