Archive for June, 2010

Getting Real

28 June 2010

Getting Real

Chapter five of Bernard d’Espagnat’s On Physics and Philosophy is entitled “Quantum Physics and Realism.”

D’Espagnat attempts to demolish various arguments for conventional realism even as he pokes holes in anti-realist arguments, finally settling on a kind of unknowable realism — unknowable except indirectly through the patterns predicted by the laws of quantum physics.

The last few sections of the chapter feel somewhat disjointed as he discusses related issues but kind of runs out of steam.

Here in more detail are some of the dichotomies (and similarities) he raises.

Physical Realism: Instinct vs Argument
D’Espagnat says scientists and “laymen” generally support physical realism, not just because it’s “instinctive” but because of some explicit arguments.

Practical vs Counterfactual Definitions
Some philosophers consider what is real to be what we can act on. But we can’t act on stars. So other philosophers speak of what would be the results if one performed an action. This is a counterfactual.

Classical vs Modal Logic
When we bring in the conditional we leave classical logic and move into the realm of modal logic.

Actual vs Counterfactual Measurements
Quantum formalism says little about counterfactuality. We can anticipate what information we’d gain if we actually performed a measurement. But this expectation doesn’t guarantee anything about the system if we perform a different measurement instead.

Disturbing vs Not Disturbing the System
The uncertainty over a system’s state persists even if we perform measurements that couldn’t possibly “disturb” the system. If we perform one measurement we can’t say it has the state that some other measurement would reveal — unless we actually make that measurement.

Conceivable vs Actual Tests
A realist who uses so much counterfactuality faces a strict litmus test for any strongly objective statements: no consequence of such a statement can be false if a test is actually performed.

Intuitive vs Rigorous Realism
It might seem self-evident that at least on the macroscopic level realism works. However, these arguments end up failing.

Predictions vs Proof
The “no-miracle argument” (or “inference toward the best explanation”) says a theory that makes lots of successful predictions is likely to be correct.

Realist vs Quantum Predictions
The problem is that quantum theory makes macroscopic predictions that match those made by realism of the accidents, so there’s no proof that objects exist with attributes the way our common sense tells us they should.

Realism vs the No-Miracle Argument
If the “no-miracle argument” fails to prove something as “obvious” as the existence of objects, can it be rescued or does it fail entirely? D’Espagnat considers two counterarguments.

“Equivalent” vs “No Equivalent” Option
You can try eliminating the quantum option by removing the “equivalent theory” option from the “no-miracle argument.” But some philosophers say the “no-miracle argument” still fails, because realism of the accidents doesn’t explain enough.

Minimal vs Generous Explanations
These philosophers say a scientific theory has to prove more than the problem at hand. Newton explained planetary orbits but in the process also explained gravitation, the Moon’s motion, and the return of Halley’s comet.

Realism of the accidents “explains” how we make predictions in our daily lives, but appears to offer no corroboration beyond this domain. D’Espagnat sympathizes with this argument but notes it doesn’t offer an alternative to the realist position.

Raw Observations vs Constructed Entities
A second anti-realist argument is that observations have to be interpreted: the sun just doesn’t sink into the western sea. But scientific revolutions can replace some entities with totally different ones (Newton vs. Einstein’s theories of gravitation, for instance).

If a theory can junk old entities, or offer two equivalent but very different mathematical formalisms, how is a realist to know which interpretation to trust?

Again, d’Espagnat says this argument should be taken seriously, but doesn’t undermine quantum theory as a possible replacement for realism of the accidents.

Realism’s Flaws vs Disproving It
Acknowledging these flaws in the case for realism of the accidents, d’Espagnat says these problems don’t prove realism is entirely wrong.

Descriptions vs Open Realism
D’Espagnat advocates an “open realism.” He says the counterarguments attack realism’s “power to describe,” but it might still be “a miracle” if there didn’t exist a mind-independent reality beyond words.

Laws of Physics vs Whimsy
The “no-miracle argument” comes in handy when considering the laws of physics. We can’t just decide the electromagnetic field is a scalar. Something constrains our imagination as we discover such laws.

D’Espagnat says the no-miracle “postulate” can’t prove conventional realism, but it justifies a kind of “open realism” with its mind-independent reality.

No-Miracle vs Intersubjective Agreement
We’ve looked at the “no-miracle” argument based on successful predictions. Now we look at an argument based on agreement between observers.

Contingent vs Non-contingent Facts
The intersubjective argument looks at agreement between observers about “contingent” facts. These are statements about how things are in reality rather than as a logical necessity.

Reality vs Mental Organization
A contingent fact might be that there’s a teapot on the table. If two people agree that’s the case then the simplest explanation is there’s really a teapot there.

One could also argue that the concept of “teapot” just mentally organizes our sensations, but then (d’Espagnat says) it would be hard to see how two people could agree on what they’re seeing.

Phenomena vs Noumena
Some anti-realist philosophers object that the concept of causality applies to phenomena, not noumena (such as Kant’s). They refuse to assume a relationship between a person’s mental images and the real world.

Phenomena vs Ad Hoc Objection
We’ve just seen the anti-realist objection about phenomena. There’s another objection that the realist argument based on intersubjectivity is too ad hoc.

Minimal vs Generous Explanations (Encore)
The realist’s explanation for the intersubjective agreement (so the objection goes) only explains the agreement, nothing more. The claim is that you should be able to apply a good theory to more than just the initial problem.

Noumena vs the Objectivist Realist
D’Espagnat says both the phenomena and ad hoc counterarguments rely on the concept of noumena (some reality not evident in the phenomena).

He adds that the objectivist realist would reject the idea of a noumena.

Objections vs Alternatives
Also, neither counterargument offers a better explanation of intersubjective agreement.

Objections vs Disproof
So neither objection delivers a knockout punch against our intuition that objects exist because we mutually agree they exist.

Realist Expectations vs Verification
A more detailed scenario is this: Alice predicts that whenever she writes in her notebook that she sees a teapot, Bob will write a similar prediction if he’s in the same room.

If Alice didn’t believe in objects’ existing independently then she’d be surprised to learn Bob agrees with her so consistently.

Conventional Realism vs Quantum Non-realism
But if Alice knew about quantum mechanics then she’d know you can believe in non-realism yet still make predictions that both she and Bob can agree on.

The quantum formalism predicts probabilities of observations that all observers will make.

But it doesn’t claim a pointer or teapot is “really” there, at least not before the measurement.

Disproving Realists’ Proofs vs Any Explanation
D’Espagnat says quantum mechanics shows philosophers’ objections to realists’ proofs are valid.

But quantum formalism provides an alternative “explanation” despite philosophers’ doubts than any explanation is possible.

Open Realism vs Radical Idealism
D’Espagnat reiterates that physical laws don’t exclusively depend on us, so radical idealism doesn’t work.

When you combine intersubjective agreement and quantum mechanics, he says, you end up with a reality beyond what the human mind creates, but this reality is also beyond description.

Classical vs Quantum Broglie-Bohm
D’Espagnat looks at a Broglie-Bohm model that is conceptually classical but makes quantum predictions.

The Broglie-Bohm model imagines “real” physical particles guided by a wave function, but this function ends up having to be non-local.

Classical vs Non-local correlations
It turns out you can’t load up the two particles at the source with supplementary (commonly called “hidden”) variables to predict the correlations.

“Bell’s calculation” (named after John Bell) shows that Bob’s measurement of one particle depends on Alice’s earlier measurement of its twin.

Classical vs Non-local Correlations
So even when you assume classically physical particles, if you want to make predictions compatible with standard quantum theory then you need to accept non-locality.

Fact vs Law
The correlation between the particle measurements depends on quantum law rather than any facts (such as the additional variables) that you add on.

Contingent Features vs Deep Structures
So the predictions depend not on “contingent” aspects of reality but rather its “deep structures.”

The deep structures of reality are mind-independent, and cannot be described except through the laws that predict our observations.

Experimental Data vs Contextuality
In order to match the experimental data, any theory you want to interpret ontologically will need to incorporate “contextuality.”

This just means that the measurement of one quantity depends on whether another quantity is simultaneously observed, and what that quantity is.

Contextuality vs Nonseparability
Besides contextuality, an ontologically interpretable theory must take into account the nonseparability of one part of a quantum system from any other part.

These two considerations derail objectivist realism’s attempts to interpret quantum phenomena.

Personal vs Impersonal Probabilities
The Born rule requires that anyone — and everyone — viewing a particular measurement will get the same “impression.”

D’Espagnat points out his way of adding a “personal” rule.

Physical Observer vs State of Mind
The measurement is finally “registered” not by the observer as a physical system but by the observer’s state of mind.

His model, later revived by others, suggests Alice and Bob could measure entangled particles and end up with different mental states.

Measuring State of Mind vs Neurons
But quantum mechanics demands a strict correlation between these measurements.

D’Espagnat says that when Alice asks Bob for his measurement she is measuring Bob’s physical state of neurons, vocal mechanisms, etc., not his state of mind.

The quantum formalism will apply to this kind of physical measurement and therefore will guarantee a correlation, although d’Espagnat does ask if quantum physics could really be so peculiar.

Relativity of Knowledge in Theory vs Practice
Starting a long time ago various philosophers have acknowledged they might have to give up on objectivist (or “transcendental”) realism.

Science would be allowed to examine our experience rather than what is “really” out there. But later philosophers blurred the language and the “empirical” distinction got lost when they talked of “reality.”

Scientists in turn felt justified in taking the intersubjective agreement of shared observations proves that what’s seen is really there.

Macroscopic Reality vs Quantum Superposition
But once you go beyond the macroscopic world to quantum states in superposition “reality-per-se” breaks down and no longer matches empirical reality.

Pure vs Quantum Philosophy
D’Espagnat then speaks of how in the twentieth century various philosophers developed theories without paying attention to quantum theory, yet their conclusions show some parallels to the more scientifically aware.

Wittgenstein vs Carnap
Wittgenstein spoke of the world as a set of facts, not things, but d’Espagnat finds Wittgenstein’s language ambiguous, with “fact” used either in a realist or mind-centered fashion.

D’Espagnat finds Carnap much clearer with the notion of “linguistic framework,” or Quine’s similar “relative ontology” (or just “ontology”).

World of Things vs Sense Data
Carnap said that in ordinary life we might use the “world of things” as our linguistic framework, but philosophers might use the framework of “sense data.”

Carnap vs Quine
Quine said the question of whether an object or attribute exists is answerable only in the right linguistic framework.

Carnap similarly spoke of “the ontology to which one’s use of a language commits him.”

Relative vs Classical Ontology
Carnap used “relative ontology” to describe a linguistic approach without meaning ontology’s classical meaning of “Reality as it really is.”

Big vs Small Range of Linguistic Frameworks
Carnap (and maybe other philosophers) could believe in a free choice of linguistic framework, but a quantum physicist has to take nonlocality into account.

In the basic version of Broglie-Bohm a pilot wave depends on the coordinates of all particles in the Universe. This “thing” is nonseparable and therefore is nothing like our ordinary concept of a thing.

Knowledge Through vs Beyond Language
D’Espagnat says a philosopher may say the lack of any (strongly) objective knowledge about “reality-per-se” means the concept is meaningless.

Scientists generally believe there is some real “outside stuff” so they in turn think there’s more to the world than language.

Abstractions vs Ontic Systems
Despite the holistic nature of quantum systems we tend to look at just part of it through “abstractions.” The partial systems are called “ontic” by physicist Hans Primas.

Intersubjective agreement exists because people who use the same abstractions come up with the same ontic approximations.

Exophysical Ontologizations vs Endophysics
When we get more ambitious than just simple statistical interpretations we develop versions of reality called “exophysical ontologizations” or “contextual ontologies.”

Primas also conjectures a reality-per-se he calls “endophysics,” but this cannot be described directly.

Reality vs Its Forms
D’Espagnat concludes the chapter by saying that “unquestionably” some reality exists on its own, but what form it takes depends a lot on ourselves and the abstractions we perform.