3

Suppose you have two intelligent agents interacting, with each one one's behavior based on their prediction of what the other will do. For example, suppose you have two mind readers playing rock, paper, scissors. It is impossible to predict what the other will do. Alternatively, suppose someone attaches a machine to my brain that reads my mind and tells what I will be doing for the next 10 minutes. I can always improvise by making a deviation from the predicted behavior.

It would seem that there are limits to how well a person's behavior can be predicted. Can there truly be determinism without predictability? Without being able to set up experiments to successfully test the prediction of a person's behavior, doesn't determinism become scientifically meaningless?

user1153980
  • 141
  • 2
  • 6
    Chaos theory looks at dynamical systems that exhibit/assume determinism and yet also exhibit unpredictability under any finite precision. – Galen Apr 23 '22 at 21:49
  • Consider that predictability is relative, depending on who makes the prediction. Determinism implies full predictability under 2 conditions: perfect knowledge of the situation, and a perfect theory of reality. Only both together allow for perfect predictions. Alas, we will never have any of those but for very trivial cases. As an exemple, think about drawing the top card from a shuffled deck: the card you are going to draw is already fully determined, but because you have only partial information the outcome is unpredictable to you. – armand Apr 24 '22 at 08:27
  • Does this answer your question? Is the debate on free will over? – tkruse Apr 24 '22 at 10:55
  • You may be interested in determinism vs prediction – Nikos M. Jul 05 '23 at 11:22
  • @armand uncertainty is a result of having more than one possible outcome and not a result of imperfect knowledge of the underlying mechanism of the only possible outcome (as would be the case if determinism is true). This is true both in science and in everyday life. Science rarely has an exact mechanism and infinite precision, yet some events are known to be certain (when this happens) because they are the only possible outcomes to expect and there is no uncertainty in these cases. In determinism for every given state there is only one possible outcome thus there can be no uncertainty ever. – Nikos M. Jul 07 '23 at 14:10
  • @Galen in chaotic systems for prediction in a finite interval of time, finite precision is adequate. Finite precision is inadequate only for prediction over a roughly infinite interval of time. – Nikos M. Jul 07 '23 at 20:09
  • @NikosM. Thank you for the correction. Lyapunov stability considers infinite intervals of time, for example. I was conflating "chaos theory" with a somewhat more specific discussion about the butterfly effect. See Palmer et al 2014 for more information on that discussion. – Galen Jul 07 '23 at 20:49
  • @NikosM. You're placing yourself from a perspective of perfect information and perfect model. But we don't have those. Even in a (imaginary) deterministic world if I throw a dice but don't know the strength and angle of the throw, or I don't know how to model it's trajectory, or I don't have the time to actually compute before it lands in a face, I will be uncertain about the outcome although only one is possible. When we read a novel the end is already written, but since we don't know it we can still be surprised. I'm not saying the world is deterministic, I'm saying this argument is poor. – armand Jul 08 '23 at 00:37
  • @armand don't you ever have certainty about something (as the only outcome) and yet not know how exacrtly it will happen? It is the same. In determinism one cannot separate the ontology from the epistemology. In a deterministic world the physical and mental state of a subject is the way it is exactly because everything else is the way they are or will be, else it would not be determined by others being fixed. The subject's state reflects the state of any other thing, imperfect knowledge cannot even arise in a deterministic universe. The ontology fixes the epistemology in a certain sense. – Nikos M. Jul 08 '23 at 07:33
  • @armand that means the subjects state S is related to the state of any other thing, by fixed one-to-one deterministic mappings F_i(S). There can be no uncertainty ever. This is the point. – Nikos M. Jul 08 '23 at 07:44
  • @NikosM. I see word salad is at the menu today, but no addressing of my points. That's not hard though. All you have to do is demonstrate how the premise "the world is fully deterministic" implies that cavemen without any knowledge of physics or adequate measure tools can predict with 100% accuracy the result of any dice roll (no uncertainty). Big words and mockery of mathematical formalism won't help you to do that. In fact nothing will because the idea that determinism implies predictability is completely absurd. Have a good day. – armand Jul 08 '23 at 11:34
  • @armand if you had read carefully my reply you would have found the answer to your original comment. Being rude and dogmatic is not a refutation of any position. Cavemen don;t have to calculate anything in a deterministic universe because what happens is what is by default expected anyway, there is no uncertainty. – Nikos M. Jul 08 '23 at 12:28
  • @armand there are various notions of predictability implied by determinism see for example On Separating Predictability and Determinism by Robert Bishop for some of them. Cheers. – Nikos M. Jul 09 '23 at 07:32

4 Answers4

1

In general determinism also leads to predictability, but in special cases this is not the case.

This was discussed also here Is the debate on free will over?

As i have said there, it is logically impossible to predict a deterministic system from within when that prediction is used by a so called Counter that will act the opposite of a prediction. This is known as Scrivens paradox.

In the linked answer that is demonstrated with a very simple setup, your example is more complex but still similar.

In general, apart from practical issues in gathering perfect knowledge and computing fast enough ahead of time, there are other known issues that can cause even simple deterministic systems to not be predictable by computation. Such systems do not become indeterministic though, and can be theoretically perfectly predictable from the outside.

However, even a system that is not deterministic can allow to make perfect predictions in many cases. More importantly, within a non-deterministic system, elements can still remain perfectly determined by past events, and sometimes elements can even be perfectly predictable, though not the system as a whole.

So while a perfectly deterministic universe would settle the question of free will, a universe with indeterminism still allows deterministic models of the mind to be true. In simpler terms, even if quantum mechanics generated a lot of random noise, this might not render neuronal activity less deterministic, the same way small waves on a lake do not prevent to steer a boat.

tkruse
  • 3,787
  • 7
  • 21
  • No system could be partially deterministic. In a deterministic system there is nothing indeterministic and vice versa. – Pertti Ruismäki Apr 24 '22 at 11:16
  • A rock is not a system. – Pertti Ruismäki Apr 24 '22 at 20:44
  • There are no deterministic systems in reality. Nothing to sit on. In an imaginary deterministic system there could be no humans to sit on anything. Sitting is an act of free will, you have to decide where to sit and when. – Pertti Ruismäki Apr 25 '22 at 02:51
  • Do you have any reference for those claims? – tkruse Apr 25 '22 at 02:58
  • Look up the definition of determinism. See, if you can find anything like that in reality. Experiment, try to sit somewhere without deciding where and when. – Pertti Ruismäki Apr 25 '22 at 03:51
  • So you have no such reference, and you are pushing your personal views. – tkruse Apr 25 '22 at 04:45
  • My reference is the definition of determinism. I have no personal opinions about that. – Pertti Ruismäki Apr 25 '22 at 11:58
  • @PerttiRuismäki: A two-body system can be considered detetministic, in a way a three-body system is not. We tend to call the latter unstable, & expect it over time to reach a stable state, like say a binary pair with a third in orbit so effectively two two-body systems. There will be gravity waves, tidal forces, stellar evolution, quantum indeterminacy etc etc but if those impacts are too small to impact the system we can treat it as deterministic, in the timespan of order the current age of the universe. Causal-closure is more useful an idea than determinism (nb: conservation of information) – CriglCragl Feb 03 '24 at 23:31
  • @CriglCragl Determinism is not dependent on the number of objects in the system. The idea behind classical physics is that we can treat physical systems as deterministic, even if they aren't. Classical physics of large objects is sufficiently accurate for most practical purposes. – Pertti Ruismäki Feb 04 '24 at 06:23
1

Let's suppose an atom is deterministic in the sense that you can logically/mathematically predict it's behavior. Using calculus you can approximate a continuum of a gazillion atoms, though it won't be an exact prediction. Still, the model works because of the deterministic behavior of a single atom. If you require exact predictability, then I imagine you'd need a computer the size of the universe to simulate/predict the entire universe. Does that mean that the universe is not deterministic?

Zinklestoff
  • 13
  • 3
-3

No and all the comments/answers suggesting this to not be the case are wrong or are confusing predictability in principle with predictability in practicality.

If it is deterministic, then one can in principle predict X. Otherwise, it is not deterministic. However, can it be practically impossible to determine something despite determinism being true? Yes.

Baby_philosopher
  • 1
  • 1
  • 4
  • 22
  • Where's the in principle possibility of predicting light-speed particles coming in from distant space (so called space invaders)? – J Kusin Feb 03 '24 at 19:51
-3

Determinism is scientifically and logically meaningless when studying human behaviour. Determinism specifically assumes that there is no such thing as human behaviour (=free will). There is only a causal mechanism with a fixed but unpredictable future. A deterministic system is constantly "predicting" its own future as fast as physically possible.

Determinism is meaningful only in classical physics, as a simplified model of reality. A very practical tool that makes understanding physics much easier and calculations simpler.

Besides, your example of two mind-readers is meaningless speculation on something that has never happened and probably will never happen.

Pertti Ruismäki
  • 2,520
  • 4
  • 15
  • Determinism assumes no such thing. Determinism is open to compatibilism and incompatibilism. Compatibilism explicitly contradicts what you say. So your claims directly contradict sources like https://en.wikipedia.org/wiki/Compatibilism : "Compatibilism is the belief that free will and determinism are mutually compatible", and you are obviously pushing your own philosophy, in violation of the rules of this site. See : https://philosophy.stackexchange.com/help/on-topic: "Please note that this site is not a personal blog or a pulpit for you to express your own personal philosophical beliefs." – tkruse Apr 26 '22 at 00:58
  • Compatibilism is both an illogical and a completely useless idea. Free will and determinism cannot be reconciled without redefining both beyond recognition. Besides, there is no determinism in reality to be compatible with. – Pertti Ruismäki Apr 26 '22 at 03:20
  • Again, pushing your opinion. – tkruse Apr 26 '22 at 04:13
  • Not an opinion. The definition of determinism does not allow any kind of free will by any definition. – Pertti Ruismäki Apr 26 '22 at 05:26