# What is the difference between free-will and randomness and or non-determinism?

In relation to the question "What are the necessary conditions for an action to be regarded as a free choice?", it came up that one way to insure the possibility of free-will was to have more than one choice. But that doesn't separate free-will from non-determinism or from randomness.

In the context of a human making a choice, wanting to know 'Are my actions predestined?', 'free-will' means they can make the choice, 'random' means the choice is unpredictable (but by who), and 'non-determinism' means there is more than one choice. What is the relation of these three concepts? I think that non-determinism is a necessary condition for the other two, but that one could have free-will but still be predictable, and that randomness does not imply willful-ness.

Is that an acceptable reading of those concepts? If not, what are alternate workable definitions?

-

Let us imagine a simple case. We have three tiny people, each of whom is placed in a box with a pencil and a limitless supply of slips of paper. Every minute, they each write a number on a slip of paper, and slide it out through a slot in their box.

One of these people has "free will". By this we mean that the person can choose to write any number he or she feels like.

Another of these people has a book of predetermined numbers in the box, and is instructed to copy these numbers onto the slips of paper in sequence.

The third person has a source of entropy-- let us say a lava lamp-- and copies numbers based upon a suitable formula applied to an element of the entropic source.

Given this scenario, a few things become evident.

First of all: there is no way-- and I mean absolutely no way-- that we can rigorously determine which of the three boxes contains which of the three people based upon the numbers output. We don't know which of the number streams is the product of free will, which is the product of a determined process, and which is the product of chance.

Second: upon closer examination, we see that each of the three alternatives is actually problematic, even if we know which person is in which box. The predetermined list is following a deterministic process in that it is copying numbers from a book, but we don't know where the numbers in the book came from in the first place. Were they freely chosen, randomly chosen, or determined according to some other process? We can't tell (due to #1 above). Similarly, the allegedly random numbers are coming from a source of entropy (such as a lava lamp), but how can we tell if this is truly random, or if it is determinate? Can we be absolutely certain that there is no physico-mathematical model that will accurately predict the movement of the fluid? Finally, the alleged "free choice" was made by a person according to what he or she "feels like" choosing, but where do these feelings come from? Can we be certain that there is not a deterministic process that is controlling the brain chemistry and synapse-firing to cause certain numbers to be chosen?

So, where are we, then?

We are left with the sad fact that we have no means by which to provide a rigorous definition for any of the terms involved, and must rely purely on our intuitive sense of what they mean in context.

Fortunately, this is not really a problem, since the so-called problem of free will isn't really much of a problem, anyway. It certainly appears to us that we have free will, and our actions are predicated on that belief. If that belief is erroneous, precisely nothing changes in our actions-- it cannot, because if the belief is erroneous and our actions are determined, well, then they are determined and are what they would have been. So, at the end of the day, it is much ado about nothing.

-
Yes, that agrees with my preconceived notions of the three principles and their implications, e.g. no way to tell the difference externally. I was hoping though to get a more thorough explanation of the differences among the three: what does it mean that a 'person can choose', does a PRNG (pseudorandom umber generator) satisfy the same properties as the distribution of the output of a non-deterministic process? – Mitch Sep 19 '11 at 21:12
I never really read this answer before, but after reading it in full now it definitely gets my +1. :) You use a good thought experiment to highlight the futility of the situation. My only gripe is that you conclude by saying that at the end of the day, it doesn't really matter/amount to anything. This perhaps true, that indeed we still have to act as if we are free, but it doesn't mean the probability of us having free will and the probability of the universe being largely (if not all) deterministic are equal. (cont.) – stoicfury Mar 10 '12 at 6:00
We have massive evidence of a very causal universe, and essentially none for a non-causal universe. As a result, I've accepted that we live in a deterministic/deterministic-for-all-intents-and-purposes universe, and as such I do not hold people morally responsible for their actions. This removes blame and praise for me, is perhaps why I particularly enjoy certain philosophies (such as stoicism), why I don't support the death penalty, etc. So at the end of the day, while it doesn't necessary have to, it very well can make a big difference in someones life! :) – stoicfury Mar 10 '12 at 6:07
@KirkBoyer: You write: "Similarly, you can react to behavior in certain ways", which already assumes free will. If you have a choice whether or not to react, we're not dealing with a fully deterministic universe. – Michael Dorfman Sep 2 '12 at 10:14
Does it change how we think of and treat other people, when it comes to compatibilism vs. libertarianism? In other words, while the two might be identical in is-land, might they be different in ought-land? – labreuer Oct 20 '13 at 22:59
WORD               FORMAL DEFINITION        GENERAL USAGE

Free will          strictly undefined       some property or essence which
grants unrestrained action

randomness         strictly undefined       the state or condition of having an
unpredictable outcome such that any
outcome is equally probable

indeterminism /    [1]                      See [1]
non-determinism

1. (a) the doctrine that human actions, though influenced
somewhat by preexisting psychological and other conditions,
are not entirely governed by them but retain a certain
freedom and spontaneity.
(b) the theory that the will is to some extent independent
of the strength of motives, or may itself modify their
strength in choice

Sources: Dictionary.com / my brain

-
@DuckMaestro: I would argue that randomness as defined above with "such that any outcome is equally probable" is necessary for the definition to remain valid. Any event in which the potential outcome is biased in favor of a particular event or set of events—even if not wholly fixed—is not truly random (it becomes pseudo-random). – stoicfury Oct 25 '11 at 5:35
thanks for the comment back with your reasoning. I suppose relevant to this context the definition could swing either way. Coming from a formal mathematics background (its formalizations not always useful or applicable in philosophy, granted), I felt as written "equally probable" could be distracting. To counter your last argument, I'd say, well, an unfair coin will have a non-uniform distribution yet the outcome is still random; but then it occurred to me that an unfair coin is perhaps characterizable as "deterministic (to some degree)". Perhaps this is your point. :) – DuckMaestro Oct 25 '11 at 7:01
Yeah it's a fairly convoluted topic, for sure. I suppose you are correct though, that such a coin toss as you describe would be considered "random" in a general sense (after all, my definition is listed under "General Usage"). But in a strict sense I think it's needed, although I'm not sure it would yet be enough to strictly define "random" (which is why I wrote "strictly undefined"). The very concept of a "strict" definition of "random" is paradoxical to me... :P – stoicfury Oct 28 '11 at 1:18
Mostly agree. I would drop the "equally probable" part. As an example, a weighted coin would be random but not where each output is equally probable. Another example from quantum mechanics is that a qubit in a mixed state could have a 75% chance of resolving to one of two states when measured. In no useful sense are these things "non-random." – James Kingsbery Apr 24 '14 at 17:09
@JamesKingsbery - I am sure we agree with each other; few things in this universe can be shown to be exactly equal to something else. I just mean that in the general sense of randomness ("General Usage" column) people treat coin flips as if they are totally fair, that the chance of either outcome is equal, that the chance for heads or tails is 50/50. Strictly speaking (as you describe) they are not, but functionally they are for us. – stoicfury Apr 24 '14 at 19:23

The first problem is that you cannot tell what free will is without giving a definition, but it is hard to give an appropriate operational definition without already implicitly answering your question, viz. taking a stance.

The second problem is that randomness can be characterized in many different ways and these definitions are not fully equivalent. Unpredictability is an important criterion in cryptography, but does not work as a general criterion for itself, for it is too unclear. For example, if you record a sequence of random numbers (generated from radioactive decay) for later use, you may "predict" them before they are used, because you have already recorded them, yet they would still be random numbers. Vice versa, many deterministic processes can be unpredictable in practise. A slightly better practical definition of randomness is that a sequence of numbers is random if you cannot compress it and it satisfies some additional statistical criteria. (That practical measure is related to Kolmogorov complexity.) The problem is that a sequence of numbers might pass all these tests and still not be random, i.e. you cannot really tell from a sequence whether it is random or not by looking at the numbers alone. A random generator outputting characters may spit out all works of Shakespeare at any time, although this is highly unlikely.

Third, non-determinism involves a random choice between alternatives.

So what does this have to do with free will? Since a random choice would not be a choice of your free will but rather a choice made by the respective random source, a choice made of free will does not seem to be the same as a non-deterministic choice. Since on the other hand we do not consider a deterministic choice to be a choice of free will either - perhaps not even a choice at all - the concept of free will might be inconsistent after all. In any case, without a proper definition of free will we cannot give an answer to your question, and if you had this definition you would likely have answered the question already for yourself.

There are two easy ways out of this dilemma: (1) Discard free will in the sense of accepting determinism. (2) Discard free will in the sense of accepting that free will consists of highly complex, non-deterministic, yet rule-based choices.

I personally opt for (2), because it is more in line with modern physics than (1) and supported by the view that human brains are open information processing systems. (Notice that a computational system with the ability of making non-deterministic choices is not a Turing machine, although it might be an extension of one.)

-

Peter van Inwagen in his "Metaphysics" (2009) explains point of view that both determinism and indeterminism (which he understands as randomness) are incompatible with free will.

He concludes that as such, free will cannot exist or merely an illusion.

Perhaps the explanation of the fact that both compatibilism and incompatibilism seem to lead to mysteries is simply that the concept of free will is self-contradictory. Perhaps free will is, as the incompatibilists say, incompatible with determinism. But perhaps it is also incompatible with indeterminism, owing to the impossibility of its being up to an agent what the outcome of an indeterministic process will be. If free will is incompatible with both determinism and indeterminism, then, since either determinism or indeterminism has to be true, free will is impossible. And, of course, what is impossible does not exist.

It should be noted that there is a chance that can save the free will.

A research by Thomas Breuser concluded that neither deterministic, nor indeterministic universally-valid theories are possible. That means that no theory can predict (even probabilistically) the future of a system which contains the observer himself due to self-reference problem.

As such it seems that the free will of at least the observer can be saved while all other people will appear to him as following the laws of deterministic or random theory and as such, not possessing free will.

Particularly, regarding quantum mechanics Breuer proves that a system which includes the observer has states in phase space which are in principle cannot be distinguished by observer himself however good measurement devices he would employ. Yes these states affect the future evolution of the system.

This can be understood as that there is hidden information which real but unreadable by any physical device which affects the future behavior of the observer.

-
This (also) doesn't actually answer the question... – stoicfury Mar 18 '13 at 1:01
@stoicfury: doesn't it partially answer the difference between non-determinism and free will? – Mitch Mar 19 '13 at 0:49
@Mitch: I only see it talking about the compatibility of in/determinism with free will, not their definitions or what makes them distinct from each other. – stoicfury Mar 19 '13 at 8:11

Short Answer: I think the necessary conditions for calling something a 'free-will' should be that 'you can take an action unhindered and uninfluenced by current and past events. Because this is what 99% of people all around the world try to defend when they defend 'free-will'. I think other types of (scholarly defined) necessary conditions have to be thrown out since they don't correspond with what most people mean when they claim people have a 'free-will'.

Long Answer: I think the key in sorting out the free-will confusion is to separate 'choice' from 'free-choice' ('will' from 'free-will'). If you claim that you have free-choice you have to define what makes it different from choice. If you claim that we live in a non-deterministic universe you have to describe what you mean. Then you have to provide at least an example (an evidence) for this.

When it comes to randomness and non-deterministic view points you will always be able to present things that seem random or non-deterministic until you understand all of the workings of the universe, which we will never do, and so you will always find evidence for this e.g. before Newton objects seem to move at random, before we found the atom things on that level seemed random, now quarks seem random, next something else will be proven to be random. Determinism (cause and effect) is like math; we just have to agree that all things (abstract things are not things in their own right since thoughts do not exist outside of our scull) can be explained through cause and effect until we find too many things not happening according to cause and effect, just like we just have to agree that 1 + 1 = 2 until too many people start adding up 1 pile of sand with another pile of sand and get confused over why they don't have two piles. Unless we agreed on how knowledge can exist (cause and effect and math being 2 of the most vital) are agreed upon we will get nowhere and have to believe that knowledge is impossible.

When it comes to 'free-will' as oppose to 'will' I don't see how anyone has ever defined 'free-will' in a useful way as something different from 'will', and then shown an example of when this happened. To make it different you would have to make the necessary conditions of 'free-will' something similar to 'you can take an action unhindered and uninfluenced by current and past events' and try to present examples of when this happened. Because no one is arguing that we don't make choices; whether they are free or not they are still choices; choices which are either caused or not caused by something else.

I think the reason for why you find so many people defining 'free-will' basically as 'sometimes we make choices' is cultural pressures. You don't become popular anywhere by talking away this long lasting concept that we base all our blame and credit ideas on; our legal system; our meritocracy. So it is easier for people to define it in a way that lets them say that they believe in 'free-will' so that they don't have to deal with the real consequences of the fact that every effect has one or more causes which ultimately can be understood (even if it is complex esp. all social sciences).

I have been having the free-will - determinism debate at least once every week for over 10 years now and I can tell you that 99% of the people I have been talking to (I am well traveled; have constantly been moving around the whole globe over this time) defend 'free-will' by claiming that you can 'take an action unhindered and uninfluenced by current and past events'

-
So, are you saying that free-will is the antonym of determinism? And are you saying that non-determinism, which I stipulated as having more than one choice, is not the right way to define it? – Mitch Mar 19 '13 at 0:52
Yes, but not necessarily because that is the only way to see it but because that seem to be the way most people view it before having read 2000 years worth of literature trying to defend free will at any cost. When it comes to 'non-determinism' I have a hard time even comprehending what it means. Because I can't imagine how a non-deterministic universe would work. How would it function if it did not function according to cause and effect – Kriss Mar 19 '13 at 2:09
(unless by non-deterministic you mean that we cannot determine the outcome because 50/50 type scenarios exists? This would make it impossible to pre-determine what will happen but it would not be an argument against determinism) – Kriss Mar 19 '13 at 2:11
Making choices is something we all do 1000s of time each day. This mean that we also have 1000s of choices each day. But choosing is a mechanism used to describe how humans function, not a claim about how the world works (whether it is deterministic). It is not a physical explanation but a design explanation (see the intentional stance by Daniel Dennett en.wikipedia.org/wiki/Intentional_stance) which is a useful shortcut for talking on a macro level about what is happening. Psychology often use design level explanation while neuroscience are better at mapping up the physical evidence. – Kriss Mar 19 '13 at 2:29
sorry for leaking :) but I hope it will help :) – Kriss Mar 19 '13 at 2:30

While Michael's answer contains an awesome thought experiment, I disagree with its "fortunate" conclusion so much I think it leaves margin to a better answer. Also I'm not glad with basically all the definitions I've read here, except one - stoicfury's (read below). Also Karl's input (about the brain) is very insightful.

Contrary to Michael, I think definitions are very, very important. I'd argue it's fundamental for one's well being and it can lead a society to glory or doom.

I once had too loose definitions of words and no religion. So I was often confused with what I should think, with what I should do. Hell, I still have this problem today, and I will probably have to live with it for ever. As I began to give proper definitions in my mind, I also enabled myself to move forward, set a direction to move in my life. It's not much different accepting a definition I created myself or accept one from religion, but I think having it defined is very important if you want to build anything in your life.

As for a society, giving words to concrete objects is easy. And it facilitates communication to do practical things. "Pass me the hammer, please". Similarly, giving words to abstract concepts can help we orient each other to better position each one of us in society and build it up. "Should I get married?". "How many kids?". Yes, "kid" is quite an abstract concept. And I think the facility we have to communicate about it today is in part responsible for how big our Earth society is, despite the babylon of different languages.

So, allow me to give you my own personal definitions in hope to bring some new perspective here. I'm pretty sure they go against some common definitions, such as ones we can see in wikipedia, but those are how I see the concepts they're associated with, based on my social intuitive sense. Of course you can choose to use different words to try and add different definitions, give different meanings. But I think the current meaning people give to those words are quite broken...

Universe = every single thing we can even imagine

Some may perceive an idea / information / data as being something outside our reality. Dreams, for instance, can easily break rules of physics. Or can they? This is no simple concept and I won't digress about it here... Please, just take this definition and see if you can cope with it: there is nothing we can even think of that would be outside the universe. Not just the common concept associated with the "observable universe". There is no "multiverse". Everything is included. What exists, what doesn't exist and what might exist. If there is anything beyond the big bang that might affect us, it is within the universe. This is an important common ground definition I'll use for the other 3 you asked.

Free will = ability to think

I think free will has nothing to do with actions. I would say a restrained prisoner still have free will, even though he won't be able to do anything with it. A coma patient have no free will. Sure, thinking in itself can be considered work (i.e. an action), in a physical sense, as there are universe-embedded-neurons acting in our brains. And, as such, even our thoughts are restricted by the universe rules we live in. But, as long as we are able to rationalize anything, I call that free will. Maybe "free will" is just another word for "intelligence", after all. I love a definition I once saw for intelligence as "capacity to predict the future".

Randomness = events which were unpredictable

stoicfury's definition is spot on to me, just there's no need to say "any outcome is equally probable" as it is implied in the "unpredictable". To me "random" means "unpredictable". The moment some future event becomes predictable it is no longer random. If we live indeed in a deterministic universe (I think we do, but that's beyond the point), nothing is random by definition. Even then we can still perceive things as random while we can't predict them.

Non-determinism = events with unpredictable outcomes

What? So it's the same as random? Yeah, basically. A deterministic universe is a connected universe. If it originated from a single point in space time, the initial variables there and then have determined every single thing that will ever happen within it. Nothing in it is truly random. So a non-deterministic event is an event that will bring at least 1 random result. And indeterminism is just a prettier name for this same thing.

Conclusion . . .

Free will has nothing to do with randomness or determinism. You say random means the choice is unpredictable, but how could it ever be unpredictable if there was just 1 option to begin with? You're confusing definitions there. Unpredictable implies more than 1 option, more than one possible choice. Randomness and determinism are different ways to look at the exact same thing.

To all that, Michael's 3 tiny people experiment still applies, and the conclusion is still the same. Today (and maybe for ever) we can't tell with certainty if our universe is deterministic or not. Even if we define the universe as a non-cyclical infinite space-time continuum it still could be either random or intrinsically predictable. And the foreseeable future indeed make it seem like it is practically irrelevant to have a precise definition for those concepts.

And who's to say it really is irrelevant? So far, even if the universe is not random, we can't predict the future. I can only say, from past events, many definitions some philosophers did in the past are quite valuable today. I don't know why we like to argue about future concepts so much, but the value in defining all this may arise much sooner than we can expect. There is the very small chance we may live much longer to even see it happening. I don't think getting to a good and solid common definition is unimportant at all. Just the personal enjoyment of digressing about all those subjects are of quite enough value to me. Specially if I find someone who share the same view!

If our so called capacity of choice (often associated with "free will") is indeed an illusion, I find it very enjoyable how this deterministic universe can still uphold such a magnificent existential living experience. And finding a person who can agree, who got the same idea coming from a completely different point of view is kind of magical. It's probably the main reason why I go through so much trouble to write questions and answers here.

-

It is perfectly reasonable to suppose that someones choices can be derived from previous causal factors. I agree with the derterministic model. In every scenerio the initial conditions and attitudes (like neural conditions) are there, and if it were possible to perfectly identify these conditions with accuarcy and account for ALL variables one could rule out the illusion of randomness. But I would like to introduce the concept that these factors, and their interactions, are literally INFINITE. Chaos theory being what it is, the interactions of all these variables in a system such as the human mind, cannot be quantified or predicted, simply because they number in infinite. And because of this, choice might indeed be free because out of an infinite series of happenings- there is one result. Only AFTER a choice is made, can it be traced to its preceding deterministic factors. But that choice was not predestined. Whether or not this adds up, I believe that the concept of deterministic factors being infinite, means there is something more to this problem of free will then we can really comprehend.

-

Okay, hold up.

All of your actions is a result of things happening in your brain. Inside your brain is a lot of neurons and other things. Keep that in mind.

Kahnemann has a lot of publications in cognitive psychology that show that brains are not in fact as unpredictable as we think. In fact, most modern neuroscientists agree that all human actions can be explained by way of cognitive algorithms.

So, why do we feel like we have unrestrained action?

Eliezer Yudkowsky gives the answer that we as humans have to take into account how the aforementioned cognitive algorithms feel from the inside. That is, when asking about free will, you have to ask "why does it fell this way" in stead of "does the world look like this?" (given that neuroscience has a negative answer to that question).

Putting it simply, it feels like we have unrestrained action when our planning-algorithm runs. We subconsciously plan everything in terms of sensory-operations and motor-functions. To get down to the corner store, you first have to stand up from the chair.

It feels like we have unrestrained action because all of our immediately available motor-functions and sensory-operations are labeled "free" inside our brain. How much does it cost you in calories to raise your left hand? Virtually nothing compared to running a mile.

This also explains how you can talk about unrestrained action in past-tense. You simply remember (or hindsight-evaluate) that it was a free action in the given situation.

-
You do understand though that "All of your actions is a result of things happening in your brain" is a super positive way of thinking and that everything you say could be simply dismissed with that single premise, which you didn't justify at all? – iphigenie Jan 27 '13 at 21:22
Karl's answer challenges the hidden presuppositions of the question, I see nothing wrong with that. In fact, he has a very good point. There are not many viable alternatives to "All of your actions is a result of things happening in your brain" that do not fall prey to the interaction problem or a variant of the Homunculus problem. These problems have never been answered in a satisfying way, whereas his position remains compatible with current scientific knowledge. Perhaps Popper/Eccles and Penrose/Hameroff should have been mentioned as alternatives, though. – Eric '3ToedSloth' Mar 18 '13 at 11:50
I also think that 'All of your actions is a result of things happening in your brain'. But that's not a particularly satisfying thing (too dependent on local anatomic knowledge). Is that what you mean by it's a 'premise', @iphigenie? – Mitch Mar 19 '13 at 0:56
@Mitch No, the premise I find more than just a little problematic is that every intellectual activity turns into a property of your brain. I think it was Dennett who actually claimed that that's the case, or maybe it was Searle. There's a whole book on this conversation, but I'm having trouble recommending it: this one. I don't favour the old dualism, but I think it's way too easy to say it's our brain that makes our decisions. I make – iphigenie Mar 19 '13 at 1:41
decisions, and while it's perfectly clear that my brain has something to do with it, it's not the one and only part of me that does that, because brains don't think, we do, and brains don't choose, and brains don't make wedding plans, and neither do they go on elections. – iphigenie Mar 19 '13 at 1:44