Wednesday, September 29, 2010

Nomologicalism vs. Accidentalism

Reality is either governed by rules, or it isn’t.

If all events transpire according to some rule or law, this is nomologicalism.

If there is no reason behind why events unfold as they do, this is accidentalism.

There are two variants of nomologicalism: deterministic and probabilistic. The laws that govern the unfolding of events are either deterministic or probabilistic in nature.

Note, however, that deterministic nomologicalism could be seen as just a special case of probabilistic nomologicalism - the case where all conceivable outcomes of any particular event are assigned a probability of either 0% or 100%. This is analogous to a Turing Machine just being a special kind of Probabilistic Automaton, one with transition probabilities of 0% or 100%.

But if nomologicalism is true then the question is: why is it true? Why do these governing laws exist and how are they enforced?

If there is no reason that we have the laws that we do, or there is no reason that they continue to hold, then this itself amounts to a kind of accidentalism. If there is no reason that the laws are as they are, then they could have been and may yet be otherwise. And if there is no reason that they continue to hold, then they may very well cease to hold at any instant.

In this case, the current state of things is accidental...there’s no reason it couldn’t have been otherwise, there’s no reason it won’t become otherwise.

But, if there is a reason that nomologicalism is true, and thus a reason for why our particular governing laws exist and a reason for why they are consistently enforced, then what is the reason for that reason?

If there is no reason for the reason, then this again amounts to a kind of accidentalism.

The only way to avoid accidentalism is to posit an infinite hierarchy of reasons for reasons for reasons for reasons...and so on. An infinity of reasons.

17 comments:

Steve said...

Quantum Mechanics is an interesting case.

First, note that your account of (classical) probabilistic law in the post does not describe QM. QM's probability distributions cannot be seen as objectively set. It seems each measurement has a spontaneous aspect which cannot be absolutely ruled out, and the perceived distribution only exists relative to a particular observation. Indeed the non-fixed nature can be exploited in situations such as the quantum zeno set-ups.

Anyway, while I can imagine the particular formal structure involved in QM as contingent or accidental, the fact that QM seems to spontaneously draw out a concretely real outcome from a abstract cloud of possiblia fits with the overall spirit of accidentalism.

Anonymous said...

Well said Steve

Allen said...

First, note that your account of (classical) probabilistic law in the post does not describe QM.

Classical probabilities are not an intrinsic aspect of the system to which they are applied. Rather, they are a measure of how much information you have about that system.

Whether the system itself is deterministic or probabilistic doesn't matter. For example, whether you're using a "genuine" source of randomness when shuffling the cards or a psuedo-random source doesn't matter for the probabilities used in poker.

The probabilities used in poker reflect the fact that you have perfect knowledge about the rules of the game and perfect knowledge about the number of cards and the suits and the ranks, and *zero* knowledge about the ordering of the cards in the deck.

In contrast, according to the Copenhagen Interpretation of QM (as well as other indeterministic interpretations) the observed probability distributions are claimed to be irreducible. There is no reason the Born probabilities hold...they just do.

So, under Copenhagen, the probabilities aren't "classical", in that they don't reflect the amount of information that you have about the observed system. Instead, the probability distributions are a fundamental property of reality.

A real property, not an apparent (subjective) property.

QM's probability distributions cannot be seen as objectively set.

According to Copenhagen, the distributions *are* objectively set though...that's the Born Probabilities. Under Copenhagen, the "necessity" of the Born probabilities is a fundamental, irreducible, inexplicable, objective fact about reality.

It seems each measurement has a spontaneous aspect which cannot be absolutely ruled out, and the perceived distribution only exists relative to a particular observation. Indeed the non-fixed nature can be exploited in situations such as the quantum zeno set-ups.

The fact that it can apparently be exploited is evidence (though not proof) that it occurs according to some rule or law. It is "reliable". If it weren't reliable, it couldn't be exploited.

You can't exploit pure randomness. It's random.

THOUGH, you can sometimes be fooled into thinking that you're exploiting pure randomness. Roulette players who think that they're "on a roll" is an example of this.

And, if accidentalism is true, our belief that the order that we observe is caused by physical laws is another example of being fooled by randomness.

It seems each measurement has a spontaneous aspect which cannot be absolutely ruled out

I'm not sure what your use of the word "spontaneous" means here. Uncaused?

the fact that QM seems to spontaneously draw out a concretely real outcome from a abstract cloud of possiblia fits with the overall spirit of accidentalism.

It fits with the spirit of accidentalism, but it could still be a variant of nomologicalism if the Born probabilities are enforced by some law or are in some sense "necessary".

If there is any "ultimate" reason for how things are, no matter how vague, then nomologicalism is true.

If there is no ultimate reason for how things are, then accidentalism is true.

I personally lean towards accidentalism.

Allen said...

I think the question boils down to:

What is the Born Rule?

"Both Born and Heisenberg acknowledge the profound influence of Einstein on the probabilistic formulation of quantum mechanics. However, Born and Heisenberg as well as Bohr, Dirac, Jordan, Pauli and von Neumann differed with Einstein about the (allegedly) fundamental nature of the Born probabilities and hence on the issue of determinism. Indeed, whereas Born and the others just listed after him believed the outcome of any individual quantum measurement to be unpredictable in principle, Einstein felt this unpredictability was just caused by the incompleteness of quantum mechanics (as he saw it)."

Allen said...

How about this:

Classical probability reflects how much information you have about the system to which they are applied. When your information about the system changes, the probabilities must change also.

Quantum probabilities are intrinsic properties of *some* physical system.

Which physical system the probabilities apply to depends on which interpretation of QM you choose.

Either the physical system being observed (Copenhagen), OR -

The physical system consisting of both the observer and what is being observed (Relational), OR -

The entire universe (Many Worlds), OR -

Whatever holds for the other interpretations.

Steve said...

I think that's right. Whether copenhagen or relational, the probabilistic aspect of qm is irreducible, but whether the state is fixed or relational is a further question.

A relational or perspectivist interpretation is what I tend to think about. But whether or not you buy the whole ontological interpretation, I'm pretty sure it's clear you can't consistently interpret quantum probabilities as fixed properties of a system.

And since each interaction alters both systems (you can only step into the same river once), each event has a uniqueness, which makes it unlike our usual sense of a probabilistic outcome -- that of repeated trials of coin flips, say. This is what inspires me to see it as more a "spontaneous" or "free" event.

Allen said...

But whether or not you buy the whole ontological interpretation, I'm pretty sure it's clear you can't consistently interpret quantum probabilities as fixed properties of a system.

The probabilities aren't a *fixed* property of the system, but the Born Rule is. Or appears to be.

Right?

Steve said...

Right, there's no reason to think the Born rule formalism isn't fixed (in our local universe).

Allen said...

So if the Born Rule is "fixed", then in the same circumstances you should always see the same probability distribution. It is either a governing rule, or a consequence of some other governing rule (e.g. something to do with wave functions).

Though the question becomes, what does "fixed" mean? How is it fixed? What enforces the rule?

If nothing does...then it's not really fixed, it's just the way things currently are. And if there's no reason for them being this way, then there's no reason it won't change. There's no reason it won't come "unfixed".

BUT, again, I question the need to postulate causes for our observations.

Because then you just have to ask what causes the causes, and what causes the causes of the causes, and so on.

If you find it acceptable that the causes have no cause, then why not just say our observations have no cause? Why is that unacceptable?

Steve said...

It's OK. But here's the issue.

There is manifest order in our region of the universe. Understanding how this works is part of the task of science and metaphysics.

Now, it may be right that there is no reason for this local region to be this way. As in Meillassoux we could say that in the infinite reality of hyper-chaos there is no reason to be surprised that an ordered region popped up out the stew.

But it would be a strong statement to say that you're sure that there is nothing in our local region which gives a clue to the character of the larger reality of which we're a small part.

Allen said...

There is manifest order in our region of the universe.

What universe?

We have to start with our conscious experience. I think, therefore I am.

From our observations, we make inferences about the nature of reality.

Inferences such as: "there is an external world that exists independently of experience".

But this leads to two questions:

1) Why do we conclude this? What evidence leads us to to believe that conscious experience can't exist without a cause?

2) How do we conclude this? By what process is our evidence gathered and our inferences drawn? What is the nature of this process - deterministic, probabilistic, other? How likely is it that this process produces true beliefs?


Understanding how this works is part of the task of science and metaphysics.

I would instead say that the task of science is *describing* what we observe, primarily using mathematics.

The task of metaphysics is to draw conclusions about what exists.

The two tasks are only peripherally related. Science is about what we observe. Metaphysics is about what exists.

It requires a leap of faith to conclude that what we observe tells us *anything* about what exists.

Doesn't it? Why do you take that leap?

But it would be a strong statement to say that you're sure that there is nothing in our local region which gives a clue to the character of the larger reality of which we're a small part.

Well, no contingent feature of our local region would give us a clue to the character of the larger reality, would it? After all, it's contingent. It could have been otherwise. And thus should be otherwise *somewhere* in the larger reality.

So what feature of our local region *isn't* contingent? What necessary features are there?

Steve said...

We can think the order in our observable universe is contingent, but that's not the same as saying we can't know something about it.

Some of your comments are too skeptical for me. It's not a leap of faith to reject solipcism. All experience is of something. Experience takes place in a milieu which has order and structure. Our style of conscious experience intimately depends on this order (see how well you do without orderly brain function).

I don't know if there are parts of this local experienced reality which are necessary, but there are some candidates. Could there be a concrete reality without experience? Can there be things which offer no description in terms of mathematics? (You know the usual candidates.)

Allen said...

We can think the order in our observable universe is contingent, but that's not the same as saying we can't know something about it.

So I'm thinking that "know" is the problem here. I think that we can have true beliefs about reality, but that we can't actually justify those beliefs.

It's not a leap of faith to reject solipcism.

Okay. So, what are your reasons for believing that our experiences tell us something true about reality?

All experience is of something.

Dreams? Hallucinations? What are they experiences "of"?

Experience takes place in a milieu which has order and structure.

What causes you to believe this? What do you mean by "milieu"? What is the nature of this milieu?

Merriam-Webster definition of milieu: "the physical or social setting in which something occurs or develops."

Our style of conscious experience intimately depends on this order (see how well you do without orderly brain function).

Correlation is not causation. What is it about orderly brain function that causes conscious experience, do you think? Why would arranging quarks and electrons in certain patterns result in an experience of pain, or the experience of seeing colors?

Could there be a concrete reality without experience?

Again, I don't know what this would mean. I only know my experience. I don't know anything about what exists outside of it.

Can there be things which offer no description in terms of mathematics?

What do you mean by mathematics? The Russell/Whitehead axiomatic system, or the Zermelo-Fraenkel axiomatic system? If ZF, with or without the Axiom of Choice? And why?

If you start with the same assumptions as I do, and make the same inferences, then you will arrive at the same conclusions. That's logic. But so what? How do I know that my assumptions and inferential judgements were correct? My conclusion may be true, but is it relevant? How can I justify my belief that it is?

Classical logic, paraconsistent logic, quantum logic, infinitary logic, multi-valued logic, subjective logic, synthetic logic, etc. Which logic is the governing logic of reality?

In "The Evolution of Reason" William S. Cooper claims that logic is reducible to evolutionary theory. Which, of course, just raises the question of what evolutionary theory is reducible to...

With the correct choice of axioms and rules of inference, you can "logically" conclude anything you want.

These conclusions may not be consistent with what we observe, but who says that everywhere is the same as here? Why should it be?

Steve said...

So I'm thinking that "know" is the problem here. I think that we can have true beliefs about reality, but that we can't actually justify those beliefs.

Okay. So, what are your reasons for believing that our experiences tell us something true about reality?

Dreams? Hallucinations? What are they experiences "of"?


What would be a good answer to these skeptical questions? Without some degree of realism about the contents of experience, I'm not sure what is there to discuss

Re: "milieu" - I just meant environment or world.

We have investigated the structure of the world thru science. Thru metaphysics I arrive at my best guess that the content of the world being described is a network of micro-experiential events (as in Whitehead's actual occasions).

Correlation is not causation. What is it about orderly brain function that causes conscious experience, do you think? Why would arranging quarks and electrons in certain patterns result in an experience of pain, or the experience of seeing colors?

We've discussed this earlier. On the theory, the our minds are composed of the micro-experiential events.

Allen said...

What would be a good answer to these skeptical questions?

I think a good answer would be that there is no independently existing external world that explains our experiences.

The existence of an external world can only be a starting assumption...it's not something you can unequivocally derive from the available evidence. However, it's a starting assumption that leads to problematic conclusions...and to questions with no good answers.

Without some degree of realism about the contents of experience, I'm not sure what is there to discuss

All discussions are ultimately with yourself. We all live in a world of our own experience. The only question is what, if anything, lies behind that experience.

I'm not trying to explain *your* conscious experience, I'm trying to explain mine. And I'm not necessarily trying to convince you of my position. Rather, I'm trying to convince myself of it - or, failing that, to formulate a better theory.

So, I could write a random sentence generator and wait for it to produce interesting passages that sound like plausible explanations of my experiences...OR...I could go to this "Internets" thing and see what there is out there. This latter approach seems to work better...and is certainly easier. When I've exhausted the information available on the Interwebs, maybe I will then look into the random sentence generator.

Additionally, I could write a Eliza-bot program that parsed my essays and which then algorithmically generated variations in response ("I don't know, why do you think consciousness is fundamental?"). A sufficiently clever Eliza-bot program would give me something useful to bounce ideas off of...but, again, the Interweb seems to do the same job fine, and requires less work on my part.

We have investigated the structure of the world thru science.

I would say instead that we have investigated the structure of our experiences through science.

Thru metaphysics I arrive at my best guess that the content of the world being described is a network of micro-experiential events (as in Whitehead's actual occasions).

You still have to find some means of knitting the micro-experiences into my unified experience.

There is the network of micro-experiences - and then there is *my* experience. There must be something additional that connects the two, right? Surely it isn't admissable to just claim that my experience *is* the network of microexperiences?

If that's permissible then why not just say that my experience *is* the interaction of quantum wavefunctions and do away with all this micro-experience talk.

There must be something in addition to my experience and the network of micro-experiences...namely, that which connects them.

But if micro-experiences can "just exist", why can't macro-experiences like mine just exist?

Why *must* macro-experiences be ontologically, metaphysically reducible to micro-experiences?

Steve said...

OK. Please let me know if I start failing the Turing test!

You still have to find some means of knitting the micro-experiences into my unified experience.

This is a problem. The best I can see is that micro-experiences form up into a participatory pattern which is the macro-experience.

Why *must* macro-experiences be ontologically, metaphysically reducible to micro-experiences?

I'm not saying they must of necessity. But in our local reality I infer it from the fact that brain/body systems are composite structures and macro-experiences correlate with these structures.

Allen said...

OK. Please let me know if I start failing the Turing test!

The fact that you haven't been convinced by my arguments is beginning to make me suspicious...