Determinism

Great Debates
121

No. I’m asking what ‘follows non-causally’ means.

But is there a factual one? If both A and B are possible at some point, and only A happens, then there must have been some choice – I don’t mean in an active agent sense, but merely in the sense that a ball running down a hill chooses whether to roll to the left or to the right, at some point. Otherwise, it ought to be the case that still A and B are possible! You tell yourself that things can ‘just happen’. To me, that makes as much sense as an underdetermined equation just having one specific solution.

If A and B are possible, and A happens, a choice must have been made somewhere – a ball rolled down the left path rather than the right one, anything like that. In a deterministic universe, such choices are trivial – they’re enforced by causality. But in an indeterministic one, there does not seem to be a way for them to be made at all; and asserting ‘it just happens’ is at best (if it is logically coherent at all, which I don’t think it is) intellectual surrender of the ‘god did it’ type – if one were to accept this, then debating about the fundamentals of the universe is useless, anyhow.

How? There are multiple possible paths, right? And in the end, only one gets actualized, right? So how doesn’t that mean that a single path was selected?

No, I’m removing the randomness. In any case, this part of the discussion has probably run its course. There’s just no inherent stochasticity in physical modelling; all the stochastic effects emerge from models with perfectly deterministic microdynamics, and all randomness is just accounting for our ignorance of the precise microstate. Of course, one can arbitrarily introduce stochasiticity, just as well as one can introduce faeries, but then that’s not scientific modelling anymore.

It’s just accounting for the fact that ultimately, classical physics fails. Nothing very shocking, really.

No. The hyperplane of simultaneity is well defined for any inertial observer in any single instant; it doesn’t depend on causality in any way.

Einstein wasn’t a fan of it, but relativity only forbids information transfer >c, which entanglement doesn’t provide; entanglement is perfectly coherent with special relativity (which is a good thing, otherwise quantum field theory would be impossible).

Superdeterminism is usually regarded as a loophole that opens up the possibility to have local realistic theories despite Bell inequality violation (basically, the argument is that the experimenters had to set their measurement devices according to some predetermined scheme – it means ‘an exaggerated form of determinism’ in the sense that there must be a predefined plan according to which the actions of the experimenters have to occur, which is certainly not what I’m arguing for), but other than that, it doesn’t have anything to do with local realism.

Uh, I’m not sure I want to go into all this as well. Basically, I follow Dan Dennett’s thinking – he defines a concept of ‘evitability’, which quantifies how well an agent can escape certain consequences in a variety of similar situations. The analogy is that of a pro golf player, who will make a given putt in more cases than an amateur will – his evitability, i.e. the capacity of evading the bad consequences of missing, is higher than that of the amateur; even though in any given situation, whether or not the putt is successful is determined completely by the situation, i.e. there is no ‘could have done otherwise’ in the same situation. But in a given set of similar situation – putts in different weather conditions, on different days, on different courses, from different distances --, the pro will outperform the amateur.

The same concept of evitability can be applied to moral judgements. ‘Committing a crime’ describes a set of similar situations, particularly when you consider one specific crime, in which a ‘moral’ agent will commit a crime less often than an ‘immoral’ one – the moral agent’s evitability is higher. Now, the evitability of any agent is probably not a fixed quantity – and that’s where punitive or other corrective measures come into play. Like the amateur can practice to improve his putting, the immoral agent can ‘practice’ to enhance his ability to not commit a crime in a given situation – or if he will not practice on his own, he can be trained.

That’s it in a nutshell, I think. Overall, I’m not sure it’s of any use – or even possible – to try to codify morality into a rigid set of rules. In the end, moral judgements are a bit like grammatical ones – you judge a set of words to be grammatical similarly to how you judge a set of actions to be moral. The rules for this are essentially heuristic, to a certain extent malleable, sometimes inconsistent, they contain certain exceptions etc., which is only natural if you consider their origin – by and large, they have evolved alongside human society, the selection acting on them being the survival of the society itself, similarly to how grammatical rules have evolved, with those being selected that allow for efficient communication within certain limits. So trying to find a single guiding principle for moral judgements is a bit like trying to find a single – descriptivist – principle according to which grammar works.

122

MrDibble, in an entirely deterministic universe, do you think that there would not be something like Brownian motion? More broadly, what are the observable differences between a deterministic and an indeterministic universe/the observations that suggest that this is an indeterministic universe?

124

one thing happens, and then another thing happens. things follow each other non-causally (at least apparently) all the time. maybe i’m missing something, because i can’t understand how this is a difficult concept.

125

well, of course debating about the fundamentals of the universe is useless, if they’re hidden. we only care about the universe insofar as it can be known.

126

How is it that one thing happens, rather than another, which would have equal reason to do so? Yes, I know, it ‘just does’, but to me, that’s the same as saying x = 3, y = 2 ‘just is’ the solution to x + y = 5 (I think that’s actually a fairly close analogy, as any mathematical modelling of indeterministic chains of events would include just such underdeterminations). It’s certainly a possible solution, but how can it be ‘the’ solution? It just is – I don’t know what to make of that.

127

Half Man Half Wit, thanks for a thoughtful response. For the benefit of any lurkers, I will mention that Dennett is a traditional compatabilist, and a very highly regarded one at that. Although, of course, not everyone in professional philosophy circles agrees with him. His is but one of several compatibilist models.

FWIW, my problem with Dennett’s model is two-fold. First, I don’t think determinism is true, although it’s patently obvious there are influences on behavior. Volition-subject-to-influences seems to me a better fit with the evidence. And a better fit with how most people (except professional philosophers) look at the problem. Second, if evitability is to serve the function of supplying a basis for assigning responsibility, it seems to me to be volition in sheep’s clothing. After all, the whole point of choosing the label was to distinguish his model from inevitability.

When push comes to shove, I don’t think traditional compatibalism goes through. What persuaded me of this was a careful reading (and rereading) of the Stanford Encyclopedia of Philosophy article on Compatibilism. Many models have been proposed, of which as I said Dennett’s is only one. Yet, it appears, none will work if one takes determinism seriously. The solution, I think, is to let that go. The mind isn’t like an ideal gas or a ball rolling down a hill. It’s a different kind of thing. Swallow that pill and volition-subject-to-influences follows rather easily.

128

No, I don’t. Superficially similar, sure, but “like” it, no.

The unpredictability of nuclear decay, Heisenberg and the non-uniform nature of the broader Universe. If the Universe were entirely deterministic (no randomness whatsoever) then it wouldn’t look as it does.

129

I am generally loathe to get involved in Great Debates, but because a number of people have been so very helpful in answering my own questions, I feel obliged to respond since I have a studied determinism for quite some time. My responses will generally be in a GQ sort of perspective. That said . . .

This is a doctrine known as illusionism. Its primary relevance is in assigning moral responsibility for various human actions, in particular crimes. This perspective eliminates the argument that one was not unable to not do what one had done because of the laws of cause and effect. This is generally tricky to justify, since it’s the falsity of the assertion of free will is in the name, but it is also generally acknowledged that this kind of thing isn’t exactly rare.

This is at the heart of the free will debate. Generally speaking, on one side you have people saying that the actions of the will essentially are random, in that they have no cause. That is, the will determines itself. On the other hand, you have those who say that finding a motivation for most human action is not difficult, and if you understand one’s thinking process, that is avoid projection, everything pretty much makes sense as following cause and effect. So, this leads to the conundrum that if the will is not acting randomly, but is following a predictable course, then how can it not be that the will isn’t being determined by the motivations at the moment? That there are so few truly crazy people with utterly inexplicable behavior is generally considered among the strongest evidence for determinism, and explaining this in a free will context is the chief problem of adherents of free will.

It gets a bit trickier than this, though, which is why materialism today is more commonly called physicalism. Mostly because there is more than matter in the universe today, which makes it harder to say just what counts as physical. If there is some distinct substance that affects thought, then whatever the hell it is, it would be physical under most conceptions of the doctrine of physicalism even though the entire point of the doctrine is that this doesn’t and can’t happen. I suspect you know this already though, but I thought I’d point it out.

This does not necessarily have to be true. Even if humans are completely determined by the chemical and electrical and so on structure of their bodies and outside influences, people are widely acknowledged to be able to do things that no other extant collection of matter can do. It does not necessarily follow that human value does not exist because of determinism.

Although, you do bring up a side concern of the free will debate, which is theodicy, or why God allows evil to exist. This question is not so important today, but it caused huge disputes in the past. Free will is one of the more common answers to the question of theodicy, since by placing moral responsibility into human hands, it absolves God of all wrongdoing, that is of moral responsibility for his actions.

This is one of the most common ways to avoid the random/determined thing Der Trihs was talking about. The idea, especially common in the seventeenth century, was that our motives influence our actions without determining them. So, for example if you’re hungry, you want food, and this leads the will to lean toward the direction of looking for food without making that happen. And this idea has some credence. We don’t always eat when we’re hungry because of many, many reasons. But what happens (and I am condensing a lot here) is that this argument is very slippery because if motives can influence the will to some significant degree, it becomes difficult to see how exactly the will is distinct form other physical processes, which on the whole, are considered to be deterministic. That is, if the will is beholden to the physical universe to any significant degree, how can it be made of different stuff.

This is a general shorthand. It is generally considered that the universe is determined or at the least mostly so. The question is how the mind fits into it. So, the stimulus, whatever it is, is usually considered to be determined, and the question is how does this affect someone’s mind, where the determinism issue isn’t quite so clear.

The problem is that it’s not always clear whether knowing the answer to a question is going to be helpful or not. Often times applications of ideas and whatnot are not apparent for quite some time.

This is another common way to adhere moral responsibility to an individual’s actions. When an action (or less often an omission), however it comes to pass (or doesn’t), can be reasonably attributed to an individual, then you can adhere moral responsibility to it. And this is how a lot of things work. For example, duress: Person A is doing something under threat of Person B, and clearly we attribute the fault to Person B since that is the “origin” of the action. Why determinism makes that tricky is that this sort of thing relies on an idea of agency, an idea which tends to disappear under determinism. And if there is no agent, there is no moral responsibility. So, here, to resolve the dilemma, it is necessary to explain agency and why exactly people can act in given ways, or the argument presupposes its conclusion.

Largely, because if outside actions have such an effect, then the will is not acting of its own accord, which, according to a free will doctrine it must. That negative reinforcement is effective implies that the will is at the least not completely free.
Okay, those are my comments on the first page. Now it is time for me to eat dinner. If these comments prove helpful, I will do the other pages when I can. Let me also say that I am not a physicist, and most of my knowledge of physics comes form this very message board, so I can’t say for sure whether determinism is true or not.

ETA: Please ignore all grammatical errors. I see a whole bunch, and five minutes will not be enough to fix them. I’m sorry about that.

130

Are there, though?

131

Well, then I should mention that I don’t regard myself as a compatibilist – I don’t think what Dennett calls ‘free will’ has enough in common with the everyday notion of the term to be called such.

I think the important distinction is that his model doesn’t apply to singular situations, but rather to classes of broadly similar ones – in any given situation, what happens is indeed inevitable.

I, on the other hand, don’t think that any ‘different kind of thing’ (dualist) model can work – the problem of how two fundamentally different kinds of thing should interface appears insurmountable. This was already realized by Descartes, who postulated a special sort of ‘gateway’ to be located within the pineal gland, without ever expounding on how that gateway supposedly worked; other people (I forgot who) came up later with the metaphor of synchronized clocks, where the only interaction between the two substances is apparent. In truth, both substances just share a common evolution, like a clock on your desk and a clock in the hall – that you hear the chimes of the clock in the hall whenever your desk clock shows the full hour then isn’t evidence of an interaction between the two, but merely brought on by the synchronization. Of course, the problem here (apart from the question of how the two substances could possibly become synchronized) is that since whatever we see, feel, or experience is ultimately mental, we could completely leave out the physical, without any noticeable difference.

But what exactly would be the difference? If I am handed a system exhibiting Brownian motion, how can I tell whether it is deterministic or indeterministic at bottom? It seems to me both would lead to indistinguishable behaviour on the accessible level – a grain of pollen (say), whose motion is best modelled by a random walk.

is only indeterministic if one interprets quantum mechanics in such a way.

only puts limits on how exact certain observations can be made.

can easily come about through purely deterministic means – take a simple structure, such as the Rule 30 cellular automaton: starting from very definite, very special (i.e. non-random) initial conditions, it develops to a very non-uniform, random (in the sense that it passes all tests of randomness) state by means of the application of a simple, deterministic rule.

What, concretely, would look different? How does one distinguish true from apparent randomness? If I were to initialise the momentum vectors of particles producing a Brownian process using the middle column of Rule 30 as initial data, how would you tell it’s not a ‘true’ Brownian process, but only ‘superficially like’ one?

132

Do you read your own links: “Rule 30 is a class 3 rule** from random initial conditions**” - run rule 30 with an initial 0 rather than a 1 in the initial cell (simple coin flip) and tell me how chaotic it runs. Or run it with a more complex starting condition…

There are an infinite number of initial conditions for Rule 30 that yield repeating results. And apparentlyit doesn’t do so well on chi-squared testing, compared to some other GAs. It remains pseudorandom, anyway, because knowing the rule and the initial conditions, you can run it again and get the same result. Definitely chaotic, yes, but that’s not what we’re talking about here, as you know. And I can’t help but start to think that you are confusing the appearance of randomness with the fact of it.

133

Because it would be the same every time you run it.

134

The point was that from a decidedly non-random initial condition – one single cell set – using a deterministic rule, a pattern with an apparent complexity emerges that rivals, in the long run, any complexity you find in the universe, so that this complexity is no indication of fundamental randomness.

Really, you’re shifting the goalposts here a little. At first, you argued that the apparent stochasticity of Brownian motion (or similar effects) came from the stochasticity of the microdynamical interactions; when I showed that even with deterministic interactions, Brownian motion emerges, you claimed that the stochasticity resides in the random initial conditions; now I show you a system where highly (apparently) random behaviour derives deterministically from non-random initial conditions, and still you quibble. The point is, any complexity, any apparent randomness of the universe can be reproduced deterministically.

Only finitely many of which, of course, one would recognise as being repetitive, as the rest are far too complex in themselves to be surveyed as a whole. But that’s a minor point.

You can do the same with reality, at least classically.

That’s why I’ve asked a couple of times how to tell ‘the appearance of randomness from the fact of it’.

You seem to believe that there is some ‘fundamental layer’ of reality, in which is held the matter of fact of how things really are, as opposed to how they appear; hence, you concede that there may be deterministic models of Brownian motion that seem like real Brownian motion, but aren’t really. And while that may be so, appearances are all we have access to – and to all appearances, deterministic models work at least as well as indeterministic ones, and in most cases, much better, as determinism seems to be at least an approximate property of our everyday experience, and strategies built under the assumption that effect follows cause have typically a larger chance of success. There is only the necessity of a fundamental indeterministic layer if you assume that there must be; without that assumption, there are deterministic models in perfect accord with observable reality, and that’s it.

Determinism is embodied in the general predictability of large numbers of natural phenomena; nothing points to indeterminism in such a way, that can’t be equally well described deterministically.

So would the universe. Go ahead, try it!

135

That’s not non-random. You have you choice of 2 initial settings (for the Universe, remember). You choose black. Why? Why not white?

Funny man.

136

How is ‘making a definite choice for the initial condition’ random? Of course, that choice is arbitrary, but why not? And choosing all white would be like choosing all momenta = 0 in a Brownian motion example. Sure, you don’t get Brownian motion then, nor evolution in rule 30, but so what? The point is that choices of initial conditions and rules exist such that observed complexities are reproduced, deterministically.

The thing is, if you have a physical system, you simply generally can’t re-set it with the necessary precision and watch it evolve again. In principle, if you could set up the system – say, a box of gas particles with some larger object in it – exactly as it was in the beginning of its evolution, then yes, we would get the same motion again. Problem is, you can’t; so you effectively only get to observe each system once.

So if I showed you a video of the ‘natural’ system’s evolution, and a video of the rule 30-seeded one, could you tell the difference?

137

Are you claiming the *Universe *made a definite choice?

138

Half Man Half Wit, if you’re not a compatibilist, you’re not really following Dennett. And you’re not furnishing a basis for assigning responsiblity. See the Stanford article. Mind, this isn’t why I think determinism isn’t true. The reasons for that are explained in the links supplied in Post #107. But, if determinism is true, I think we have to change how we do things and how we think about responsibility. Indeed, because I think determinism is partly true, in the sense that some people have impaired volition, I think we should be taking a different view in those cases. In real life, that’s the side of this debate I generally find myself having to argue, since most folks operate under LFW assumptions.

BTW, I’m not sure where you got the idea that I’m arguing for a dualist position, but I’m not. No ghost in the machine for me. When I say the mind is a different kind of thing, I mean it’s not like an ideal gas or a ball rolling down a hill, i.e., not something which is necessarily deterministic.

139

No, there presumably was only one possibility – that’s how determinism usually shakes out. Are you claiming it rolled dice?

I disagree with Dennett about his being a compatibilist, too. :wink: (Well, it’s actually more that I’m not sure that I think compatibilism is really a coherent stance; to me, the question of whether or not there is determinism is a bit of a red herring concerning the question of whether or not there is free will. Even in an indeterministic universe, I still see the (usual) concept of free will as logically problematic. Dennett sees these problems to, and resolves them through coming up with a new concept of free will; I disagree that this concept really exemplifies what is usually meant by the term.) But his evitability concept doesn’t mind what one calls oneself.

140

But there are two possibilities for Rule 30 (at its simplest case, that is)

Yes, effectively (not that I think there’s only one starting parameter for the actual Universe). That we are living in the Universe that resulted, doesn’t mean we *necessarily must * assume this is the only possible one.

Think of it as a parallel to the Strong Anthropic Principle.

141

Well, one might imagine a multiverse, one universe corresponding to every possible set of initial conditions. Or anything else; as I said before, I don’t think the ‘how did it all come about’ question is terribly important here.

It’s the conservative assumption to make, though. And I still can’t see how this dice rolling is supposed to work, but it seems I’ll have to live with that.