For me, it can easily be disposed of as an non-necessary hypothesis. We don't know for a fact, and will never know, if everything in the universe is predetermined or not. It shouldn't bother people, therefore. — Olivier5

This seems to indicate that "determinism/indeterminism" isn't actually a property of the universe at all, but something else. When we think about "the universe", we presuppose determinism, because the "universe" we think of is actually a model of the universe, and a model relies on determinism to function.

a model relies on determinism to function. — Echarmion

That's a good point. We seem to project our need for predictability onto the word.

Indeed, determinism cannot be a scientific theory if it is not falsifiable. It is a matter of belief, and your belief in the matter seems quite strong. — Olivier5

Actually, we haven't touched on my beliefs yet which are a little more exotic. I'm merely criticising your belief atm.

It only takes one observation where the initial state is fully known, and where the assumption of determinism leads to a single expected outcome, and to not achieve that regularity of outcome, for that assumption of determinism to be ruled out. This could be, for instance, a new experiment that tests whether QM is Copenhagen-like or MWI-like. Or it could just be throwing a ball in the air and seeing it change trajectory mid-flight. That such non-deterministic outcomes are conceivable means that determinism can be falsified at any moment. And despite centuries of experimentation from Galileo to CERN, we've yet to see a single such event.

This does not prove determinism true, of course. Gun to my head, I'd wager on a deterministic resolution to the measurement problem, after which an assumption like Popper's that the natural universe is fundamentally non-deterministic would be quite mad.

only takes one observation where the initial state is fully known — Kenosha Kid

Fully known? You have any example of something that can be fully known?

Or it could just be throwing a ball in the air and seeing it change trajectory mid-flight — Kenosha Kid

Balls have been thrown up and down before, and we do seem to have difficulties predicting their trajectory. You've heard of the Galton box?

It only takes one observation where the initial state is fully known, and where the assumption of determinism leads to a single expected outcome, and to not achieve that regularity of outcome, for that assumption of determinism to be ruled out. — Kenosha Kid

This actually happens all the time. An outcome is predicted, the experiment is done, and the outcome is not what has been predicted. It's the basic scientific process.

However, the conclusion is never "determinism is false". It's always a specific "law" that is amended to account for the observation. And hence we get a more encompassing model of the universe.

So if we ever experienced a thrown ball exhibiting random movements, we'd come up with a system of physics that predicts those movements given the circumstances. In a way, that is exactly what happened with Quantum physics. The experimental results of setups like the "quantum eraser" are utterly bizzarre from the viewpoint of classical mechanics.

Fully known? You have any example of something that can be fully known? — Olivier5

Yes, lots of experiments have relied on initial states prepared to high precision.

Balls have been thrown up and down before, and we do seem to have difficulties predicting their trajectory. You've heard of the Galton box? — Olivier5

Yes, and that is a great example of unknowability in a chaotic system yielding unpredictable behaviour at a classical scale, but it is not shown to be an example of non-deterministic behaviour. A coin toss exemplifies the same principle. I do not know whether it will land heads or tails because I cannot personally model the flipping action. A robot would fare better. Similarly one could conceive of a version of the Galton box where the initial trajectory of the ball is precisely prepared, along with arbitrarily precise optical measurements to account for error, and one could determine the final resting place of the ball this way. The absence of such precision in preparation and measurement is why this is a problem of knowability, not determinism.

This actually happens all the time. An outcome is predicted, the experiment is done, and the outcome is not what has been predicted. It's the basic scientific process. — Echarmion

Note:

It only takes one observation where the initial state is fully known, and where the assumption of determinism leads to a single expected outcome, and to not achieve that regularity of outcome, for that assumption of determinism to be ruled out. — Kenosha Kid

high precision. — Kenosha Kid

High precision is different from full precision, and you know it. Or you should know it. Perfect precision of measurement of initial conditions is impossible. So you will have to think a bit more creatively than starting with "fully known initial conditions". That's never going to happen.

Balls have been thrown up and down before, and we do seem to have difficulties predicting their trajectory. You've heard of the Galton box?
— Olivier5

Yes, and that is a great example of unknowability in a chaotic system yielding unpredictable behaviour at a classical scale, but it is not shown to be an example of non-deterministic behaviour. — Kenosha Kid

So, 1) you propose throwing a ball as an attempt to falsify determinism, 2) I show you some balls being thrown and landing on the ground haphazardly, and 3) you say: "no no no this doesn't count".

Ergo, determinism is not falsifiable. It's metaphysical.

I get what you're saying, a single, fixed deviation isn't the same as constantly changing results.

But effects on the micro scale do exhibit constantly changing results. And it is important to remember, when doing these kinds of thought experiments, that we're not talking about the world suddenly changing from one set of rules to the other. We would of course notice if tennis balls suddenly started exhibiting erratic flight patterns. But this is because we have already extensively catalogued the behaviour of classical objects. The question is whether what we are and have been seeing is determinism or merely sufficiently deterministic behaviour.

High precision is different from full precision, and you know it. Or you should know it. Perfect precision of measurement of initial conditions is impossible. So you will have to think a bit more creatively than starting with "fully known initial conditions". That's never going to happen. — Olivier5

It only has to be known to the requisite precision of the experiment. In the Galton box, this would be extremely high, but not necessarily perfect. In the case of throwing balls in the air to catch them, little precision is required.

So, 1) you propose throwing a ball as an attempt to falsify determinism, 2) I show you some balls being thrown and landing on the ground haphazardly, and 3) you say: "no no no this doesn't count". — Olivier5

Do you want to refer back to my criteria for falsification? Throwing balls haphazardly is a tractable problem, but unless you know how you're throwing them, it is not a knowable one. It is not beyond out current technology to fire a tennis ball at a precise point on a wall, for instance. Throwing them like a maniac disproves nothing.

But effects on the micro scale do exhibit constantly changing results. — Echarmion

Agreed, but it's important to be aware of why, and not lump unknowns, intractability, and genuine non-determinism into one catch-all. Otherwise you get Olivier5's claim that randomly chucking balls about demonstrates non-determinism.

Taking a moment of perspective, everything you're doing on this site relies on your expectation of regularity. You do not seriously expect the E key on your keyboard to spit out a random character on the screen. We are all comfortable with the idea that science-driven technology manifests regularity with utmost reliability. One must measure this against claims that the Universe is fundamentally random and ask: which seems to explain my experience?

It only has to be known to the requisite precision of the experiment. — Kenosha Kid

If you don't know exactly what the original conditions are, how can you expect your predictions to be exact?

Even if you say (as actually done in real science) "we can predict a result within a certain margin of error commensurate with (or calculated based on) the margin of error in initial conditions", what you are testing in such an experiment is just whatever law or principle you're testing, the deterministic law you use to derive final conditions from initial conditions. But you're not testing all the other possible deterministic theories that could be invented, are you? Therefore, if your prediction is not exact, if it is off by a greater margin than expected, it won't disprove determinism at all. It will just disprove that particular law you tested.

Do you want to refer back to my criteria for falsification? — Kenosha Kid

Yes. Let's do a little bit of analytical work here on your criteria for falsification. Especially since I've never come across those criteria before, and yet have read 90% of Popper's opus, and he's the guy who came up with the criterion of falsifiability in the first place... which means that your criteria appear very much to be your criteria. So let's see some definitions.

I'm not convinced you've understood me at all. The phrasing for knowability I gave above was "all relevant information". For an experiment that is insensitive to the exact position and momentum of a ball, the exact position and momentum of the ball is not relevant information. For an experiment that is sensitive to these, the initial state is not knowable, and thus the experiment is not a good test for the theory's presumed determinism, or its null hypothesis. So this is really not a useful non sequitur.

Again, in order to falsify the presumed determinism, it is necessary to consider experiments that are knowable (we have all

relevant

information), tractable (we can make sufficiently accurate predictions of expected deterministic behaviour), and unpredictable (regular behaviour is not manifest). That is required in order to discern deterministic behaviour from non-deterministic behaviour. Anything else is irrelevant.

Agreed, but it's important to be aware of why, and not lump unknowns, intractability, and genuine non-determinism into one catch-all. Otherwise you get Olivier5's claim that randomly chucking balls about demonstrates non-determinism. — Kenosha Kid

I am not trying to demonstrate non-determinism. In fact my argument is that it's essentially impossible to do so. All absudities we can come up with rely for their absurdity on the contrast with the real world as we experience it. But if that world is not actually deterministic then of course a non-deterministic world looks exactly like our world.

Anyways, I think your first sentence illustrates that. Why, or rather how, do we know what the things we observe on the micro scale actually signify?

One must measure this against claims that the Universe is fundamentally random and ask: which seems to explain my experience? — Kenosha Kid

But who are we asking? Ourselves. Are we an unbiased observer? There doesn't seem to be a reason to think that our brains are somehow designed to answer the question.

Especially since I've never come across those criteria before, and yet have read 90% of Popper's opus, and he's the guy who came up with the criterion of falsifiability in the first place... which means that your criteria appear very much to be your criteria. So let's see some definitions. — Olivier5

I think you’re misreading Kenosha. He isn’t proposing alternative criteria for falsification generally. He’s saying what criteria would need to be met specifically to falsify determinism. Every hypothesis will have specific criteria by which to falsify it. He's suggesting the criteria by which determinism could be falsified. In other words, what observations would show that determinism was false.

These criteria seem to come from nowhere and go nowhere. I didn't ask about any of that. I asked you to imagine an experiment that could disprove determinism. And you have failed to do so.

Let's see if he can explain.

I can explain for him.

Determinism is true if and only if:
- given a particular state of affairs,
- and the laws of nature,
- a particular future state of affairs is guaranteed.

So to prove determinism false, you need to:
- know a particular state of affairs
- know the laws of nature
- observe something contrary to what those ought to guarantee

Those are basically his three criteria.

You mean, if only we knew the true and exact laws of nature, we could tell whether they are determinist or not. Well, that's true, but we will probably never know the real and exact laws of the universe. We certainly don't know them now. So your proposal doesn't work.

That just means we are unable to be sure whether determinism is true or not.

More precisely, we will never know for sure. Like we will probably never know for sure if gods exist or not.

Maybe I missed part of the earlier conversation, because I thought you were on the “we can falsify determinism” (as in “it is actually false”) side of things, and Kenosha was saying that nobody has yet done what would be necessary to falsify it.

The discussion is now on whether determinism is falsifiable in Popper's sense, and hence whether it is a scientific theory. Ken contends it is (but has failed to show how); I contend that determism is not falsifiable, and hence a non-scientific question.

How would you disprove a probabilistic theory based on Popper? At no point in time you can know if the 90% chance will be met at the end of all times.

A probabilistic theory still makes predictions, testable with statistics.

Maths say if you throw a coin it is a valid result to get 1bln heads in a row. This does not contradict the 50-50 probability.

In practice it does. Because that result is so improbable as to be next to impossible.

But you were talking about scientific theories.

And how to falsify them. By experience.

Your chances to draw 10 times 'head' in a row are 1/1024.

Your chances to draw 100 times 'head' are 1 / 10^30 or thereabouts.

Your chances to draw 1000 times 'head' are approximately 1 / 10^301. That's

0.0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001 if I'm not mistaken.

You think you can draw heads a billion times? You want to calculate the odds of that?

So what? IF that happened, would it prove the 50-50 chance wrong? Given it is wrong as the coin may land on its edge - is that what you are trying to say?