Can someone explain compatibilism?

Real answer: No. No one can. Because it doesn’t make sense.

^ No, really. What am I missing? Is there really nothing here? I feel like there are serious, intelligent people who are making some kind of “if people feel that they have free will, well then free will exists, even if their thoughts and motivations are caused events” argument. But I am not informed on this issue, and I’m not a philosopher.

Compatiblism:
A concocted philosophy designed to sell meaningless books on the subject.

According to wikipedia compatibilism is just the belief that Free Will is compatible with Determinism. What’s so hard?

In what way does free will mean that thoughts and motivations aren’t “caused events”? Can’t they be part of a chain of causation?

What is free will, anyway? If the self is an illusion, and one’s memories are reconstructed when you recall them, then whose will is it? If you change your mind later about your intention, was your will still free?

@rexacoracofalipitorius: “According to wikipedia compatibilism is just the belief that Free Will is compatible with Determinism. What’s so hard?”

I don’t get it. I suppose if you define free will as “freedom to act according to one’s determined motives without arbitrary hindrance from other individuals or institutions”, then sure. But this isn’t really what we mean by “free will”? Is it?

@rexacoracofalipitorius: “In what way does free will mean that thoughts and motivations aren’t “caused events”? Can’t they be part of a chain of causation?”

…and still refer to it as “free will”? I just don’t get it. The reason for my question here is that I just don’t get it. Dumb it down for me.

@rexacoracofalipitorius: “What is free will, anyway? If the self is an illusion, and one’s memories are reconstructed when you recall them, then whose will is it? If you change your mind later about your intention, was your will still free?”

I can’t come up with a definition of “free will” that makes sense except for a very loose one in the context of that quote above from wiki. But that seems to be avoiding the issue. We’re not the author of our thoughts, so I can’t figure out how our will can be “free”. And yes, it’s becoming clear that the concept of the “self” falls apart fairly easily.

But where does compatiblism fit in here? Does it really just skip all of the self/consciousness concerns and go straight to freedom of action? As long as someone is not locked in a cage, then they are able to exercise some kind of “free will”? If this is the case, then does compatibilism even really address the type of “free will” that concerns most people?

This is my take on it. I am not claiming that this is right, but it is the only way I can make sense of it. I think it means that we are complex beings and that what we do is determined in large part by our inner state. Unlike a housefly, whose behavior is pretty much determined by its immediate surroundings, our actions are determined at least in part by past memories and reasoning. Although ultimately our actions are predetermined, we at least give the appearance of being self-determined, and in fact it is much easier to interact socially by making the assumption of free will.

It’s a shitty theory. Here’s a good resource: Stanford Encyclopedia of Philosophy.

Thanks everyone. I’m probably more confused then I was before. After combing through @gorillapaw’s and @rexacoracofalipitorius’ links, I am even more convinced that compatibilism is likely not addressing any of the relevant variables in what I would consider “free will”. It feels like an “intentional” change of the subject. I could be wrong, however.

Don’t go getting all “c.o.m.p.a.t.i…b.a.l” on us now @hominid.

@hominid I don’t think I can “dumb it down” for you, because I don’t actually have answers to any of the questions I asked.

I don’t think it matters at all that free will isn’t “real”, if that’s even true. Determinism isn’t real either: we don’t live in a deterministic universe designed by a watchmaker (blind or otherwise) but in a probablilistic one where the most basic events are governed by the laws of chance and the more complex ones based upon them are governed by chaotic dynamics.

It doesn’t matter if “free will” and “determinism” are compatible or not, because the philosophical usage of those terms still carries the baggage of a pre-scientific age. The questions of predestination and of free will belong more properly to theology or to folklore than to philosophy.

But even if “free will” isn’t “real”, it’s still an experience that all non-zombies have. We all speak and act as though we’re influenced by this “free will”. It seems to make a difference in our perceived quality of life; in as much as agency may be regarded as the exercise of one’s will, we’re happy when we have agency and we are unhappy when we’re deprived of it.
Similarly, even though “determinism” is not real, we still observe that effects follow causes, and we can to some extent use our knowledge of cause-and-effect to shape the world around us to our liking.

In sum, I feel that not only is “compatibilism”, so far as I understand it, a pointless waste of time, but I feel that way also about the entire “free will vs. determinism” dilemma, and about much of philosophy. The tree of knowledge is healthy, but could benefit from some pruning.

@rexacoracofalipitorius , It may be a matter of semantics, but I would say that the world is deterministic even if laws have to be interpreted probabilistically. It is not necessary to be able to predict what will happen in order to say things are predetermined. I am not suggesting that there is some God who knows for certain what is going to happen, just that laws of nature are still being followed even when the laws must be stated in terms of probability.

@LostInParadise Even if the world is non-deterministic, we still perceive cause and effect. Some of these cause / effect relationships are repeatable. This allows us to reliably predict certain events. In that sense, those events are “predestined” by us, or by whomever or whatever set up the preconditions we observed.

If it’s not necessary to be able to predict events in order to call those events “predetermined”, then what is the defining criterion?

The defining criterion is to have mathematical equations to define natural laws, which we do. It is just that the equations are in the form of probabilities.

Then, if an event does not follow a natural law defined by a mathematical equation, it is not a “predetermined” event?

If the event itself can’t be predicted by the equations of the relevant natural law (for example, the equations of fluid dynamics can’t predict whether a storm cell will form at a particular place at a particular time) do we still call that event “predetermined”? If so, why?

For weather we do not need to get to the quantum level. We can make all the necessary measurements. The difficulties in prediction relate to chaos theory. We are limited in how accurately we can make our measurements. At some point, the inaccuracies will mount sufficiently so as to make prediction impossible. That does not mean that weather is not predetermined.

I’m confused by the statement that you can have a deterministic world that depends on probabilities. This doesn’t make sense. In a deterministic model, the same outcome occurs every single time it is run. There is no room for randomness.

We have two different views of what should be meant by deterministic. My view is that deterministic means laws following mathematical equations, even if the equations involve probabilities. Otherwise, you would have to say that neither determinism nor free will applies.

There are those who say that quantum uncertainty provides a possibility for free will, but they are mistaken. Consciousness cannot resolve uncertainties at the quantum level. Resolutions at the quantum level either can or cannot directly impact consciousness, but in either case there is no room for free will.

@LostInParadise “We have two different views of what should be meant by deterministic. My view is that deterministic means laws following mathematical equations, even if the equations involve probabilities.”

This may be your view, but it is incorrect. In mathematical determinism, as in philosophical determinism, there is only one possible outcome, given the same initial conditions. The outcome cannot be affected by “probabilities”.

We are going to have to agree to disagree. Take a look at this entry, section 4.4 on quantum mechanics and determinism. There is also a many worlds interpretation of quantum mechanics that does not require probability

@LostInParadise In your link, I see nothing more than a confirmation that probabilism (Jesus, is that even a word?) and determinism are opposites. The author presents a few “what if?” statements, but never continues his/her own thought.

If you want to offer some words of your own on this topic, I will read them, but I’m not doing any further “homework” on this topic.

I concede that my interpretation of determinism is not the standard one. All I wanted to point out in the articles is that there are ways of interpreting quantum mechanics that do not relate to probability. In the many worlds interpretation, all possibilities occur in different universes. This strikes me as a bit bizarre, but there are at least a few respectable physicists who take it seriously.

@LostInParadise The possible worlds thing is crap. If there are possible worlds that really do exist, they exist in a way that’s completely disconnected from our universe. That being the case, it’s impossible to know anything about them or learn from them in any way since information cannot be passed from one world to the next (that would produce cause/effect across possible worlds which can’t happen). It’s functionally identical to a hypothetical thought experiment.

What is the alternative? You have a perfectly determined wave function that suddenly “collapses” when observed. There is something very unsatisfying about that.

@LostInParadise Leaving aside quantum mechanics for the moment, you stated:

“At some point, the inaccuracies will mount sufficiently so as to make prediction impossible. That does not mean that weather is not predetermined.”

I maintain that if prediction is impossible then that does in fact mean that the phenomenon is not predetermined.

We can predict weather up to a week in advance. Are you saying that the weather is predetermined up to a week from now but then becomes undetermined? That seems a peculiar use of language.

@LostInParadise No… weather is not predetermined. Being able to predict weather with some degree of reliability on the scale of “sunny/not sunny” over a small region does not mean that the exact outcome of the model is predictable. There is a lot of randomness built into weather models. If you run the model more than once, you will get a different outcome. That immediately shows that it is not deterministic.

If you run the model with the exact same data, you get the same results. That is how computers work. If the data is slightly modified, the model will eventually give very different results.

This is a matter of semantics. We are going to have to agree to disagree. It just seems odd to me to say that next week’s weather was not predetermined last week but is partially predetermined this week.

@LostInParadise “If you run the model with the exact same data, you get the same results. That is how computers work”

No, no, no. Absolutely no. I have run meteorological models as part of my work. They have stochasticity built into the models. Typically, it is the average of many results that is presented for eventual use. The absolutely do not produce the same result every time. This is not about semantics.

@dappled_leaves If you run a pseudo random number generator with the same seed, you’ll get the same outcome every time. Computers cannot, by themselves generate truly random numbers without accessing some external source of randomness.

@gorillapaws For the moment, I will only say that these models are extremely complex, and use a much more sophisticated approach than simply inserting the occasional RAND() function.

@dappled_leaves Sure they’re complex, but if you input the exact same set of initial conditions (including seeds for the random number generator) then they should output the exact same results every time. That makes them deterministic. The variability you’re seeing is from having different seeds for each run.

@gorillapaws I think you are missing your own point. The same set of initial parameters are subjected to many types of random fluctuation, based on probabilities set in the model. You would have to remove the random elements in order to achieve the same results every time. With the random elements included, no, the model is not deterministic. This is not difficult to understand.

I mean, what you are basically saying is that if you make the random elements non-random, the model becomes deterministic. Well, duh. That supports what I have been saying.

This conversation is way over my head. But I’m curious – would someone be able to tie this back to the concept of free will (and possibly compatibilism)? The addition of randomness doesn’t get us any closer to free will, right?

@dappled_leaves , Varying data, one way or another, is a necessary part of running simulations. What would be the point of building a complex simulation program if you got the same result each time? The computers that do weather forecasting get their data from sampling current atmospheric conditions and then run their model to make predictions. Over a short time range, the results are consistent over different runs. Otherwise the forecasts would be worthless.

@hominid , Randomness does not affect the free will argument. If what you do depends on some random state of a subatomic particle, that does not imply free will on your part. It could conceivably mean that your behavior is in theory unpredictable, but that is not the same as having free will.

@LostInParadise I will not respond any further on this thread. It is like beating my head against a brick wall. There is a reason that we distinguish between deterministic and stochastic models. The names have meaning. Repeatedly denying that it is so does nothing to affect this. Good luck in your future endeavours.

@hominid Frankly, this discussion has convinced me that those arguing for compatibilism have not the faintest clue what they are talking about. So perhaps there was a point to this tangent after all.

@dappled_leaves I think the issue is different definitions of determinism. As I understand it, determinism (in philosophy) means that given an initial set of conditions there can be exactly one possible outcome. Whereas the deterministic vs stochastic types of modeling is using a more colloquial use of determinism. I’m not arguing in favor of compatibilism by the way. This discussion is related to compatibilism because in order for free will and determinism to be compatible, it requires a looser interpretation of determinism.

Stochastic modeling makes things complicated. It fits somewhere between complete knowledge and complete ignorance of the future. In the case of weather, it is interesting that it is possible to make statements about large scale climate predictions over a far longer period of time than it is possible for local weather forecasts. I remember years ago scientists were predicting that global warming would cause more extreme weather patterns. They have been proven correct, though the specific locations and nature of the extreme weather changes could not be predicted.

@LostInParadise stochastic modeling doesn’t violate the laws of the universe. It’s still behaves deterministicly even though it’s not labeled deterministic. It’s a convention of language usage and not a formal description of behavior. Just as rolling a die is random, but if rolled in identically the same manner it should yield the same result every time.

I agree, but from a pragmatic of view the behavior is indistinguishable from a pure probabilistic model. In that sense, I can see I can see what @dappled_leaves is saying.

In short.. if you know all the factors involved, prediction is easy.
If you don’t know all the factors involved, prediction is hard.

Still though, in 100% of the unpredictable (to observers) cases involving weather or whatever, hindsight is always 20/20. An indeterminate world would be one where hindsight was not so and things tend to come about without any specific causes or contributing factors.

Computer models don’t enter into it. The actual phenomenon can be, in principle, unpredictable. There are limits to precision in measurement, limits imposed by our tools and limits which are imposed by nature. Without infinite precision, the “same” measurement could in fact be two distinct values for the purposes of a chaotically-dynamic system, which exhibits nonlinear sensitivity to initial conditions (that is, the decimal places furthest to the right don’t necessarily affect the outcome to a smaller extent than do the digits left of them.)
Some things can’t be measured at all, some things can’t be measured with sufficient precision to make good predictions. These are fundamental aspects of the universe, not a failure of our technology. Some things are actually and in principle impossible to predict.
If it’s not predictable, then it can’t be said to be predetermined (because if it were predeterminable, then it would be predictable by the process of determination.) So much for predestination.

“Free will” is an attractive idea, but there’s reason to think that it doesn’t exist either. Even if it does exist, it could be a localized phenomenon. There’s no reason to think that one’s “free will” extends beyond a small subset of the conscious thoughts one has. Your “free will” doesn’t control your autonomic responses or your reflexes, and it controls fewer of your “conscious” actions than you perceive. The philosophical idea of “free will” is bunk in as much as it implies agency that doesn’t exist.

I therefore maintain that in as much as both concepts are bunk, they are quite compatible.

@gorillapaws Modern computers generally contain a weak hardware random generator measuring thermal fluctuation or clock drift. They run out of random fairly quickly, but can provide good seeds for a seeded PRNG, making it sufficiently entropic as to not render @dappled_leaves’ models deterministic. My /dev/random on this computer is good for about 96 bytes of “good” entropy, for example.

@rexacoracofalipitorius There’s an important distinction between something being non-deterministic in practice and the thing being deterministic in actuality. Just because we lack the ability/tools to measure the initial conditions with perfect accuracy and precision doesn’t mean it’s not behaving deterministically. Just as before we had microscopes, that didn’t mean microscopic organisms didn’t exist.

So yes, a computer model with good hardware seeding of a pseudo random number generator may yield results that to us appear to be non-deterministic in practice, that doesn’t mean they’re not acting deterministically in actuality.

@gorillapaws It’s not that we lack the ability or tools to measure with perfect accuracy. It’s that perfect accuracy is not attainable in principle. Perfect accuracy implies zero measurement error, which implies infinite precision. Infinite precision is not attainable.

If non-predictability isn’t enough to call something adeterministic, then what is your defining criterion?

@rexacoracofalipitorius It’s not whether a person can realistically predict the outcome; it’s if a hypothetical person with perfect knowledge of all relevant inputs could. Free will necessitates that even this hypothetical person with perfect knowledge couldn’t do so. Another example would be things like behavior of certain subatomic particles. As I understand it (and I could be wrong here) according to quantum mechanics the behavior of certain subatomic particles are inherently unpredictable. This is not due to a limitation of our ability to measure them, but a property of the particles themselves, and so even a person with perfect knowledge couldn’t predict the outcome.

You didn’t answer the question I asked, and I don’t understand the point of the rest of your post.

Is it your position that fundamental unpredictability is a sufficient condition for free will, or just a necessary one?

In quantum physics, complementarity means that certain properties of particles can’t be co-measured: for example, the precision in measurement of momentum can’t be measured to greater than (reciprocal times bar-h)* of position. I’m not sure whether this implies that complementary properties can’t coexist, or if there are other properties which conditionally don’t exist.

*from memory, probably inaccurate, couldn’t be bothered

@rexacoracofalipitorius It’s clearly a necessary condition, but not sufficient. Free will also necessitates the person being able to control their decisions. The point is that if the universe is deterministic, then it’s a property of the universe, and not dependent on someone actually being capable of determining behaviors. e.g. before mankind had telescopes to see planets orbiting the sun, their orbits were still deterministic even though nobody at the time was able to predict their motion.

I just met Dan Dennett while getting my eyeglass lenses replaced. I have to admit that I was caught off guard and felt uncharacteristically nervous/star-struck. I was with my boys and the 4 of us were the only people there. I fumbled through some “I’m a fan of your work” garbage before leaving and trying to explain to my 6 and 9 year-old boys what I think compatibilism means.

I just stumbled across this blog post from Daniel Miessler. It’s relevant, and the author seems to sum up the suspicions and problems I had with (my understanding of) compatibilism.

I don’t want to prolong this too much, so I will briefly state my objection to the blog post. Just because a person has options and is aware of them does not mean that the option that is actually chosen is done so through free will. Being aware of the options is what distinguishes us from houseflies. The actual choice that is made is based on who we are as a result of our genetics and history and, although we fully feel that we are choosing freely, that feeling is an illusion.

@LostInParadise – I don’t think you meant to say that you objected to the blog post. You seem to agree with Miessler (and me) 100% in that free will is an illusion.

Yes, but I think that there is still significance to the fact that we have a sense of self and that we are aware of there being different options. It is definitely not free will, but neither is exactly like being an automaton the way insects are. Maybe it is a matter of complexity. It seems like there is some middle ground between acting like a simple automaton and having free will.

I think the fact that you’re compelled to believe there is a difference indicates that there is exactly no such thing.

@LostInParadise: “Maybe it is a matter of complexity. It seems like there is some middle ground between acting like a simple automaton and having free will.”

How does adding complexity allow for free will? Or should I say, what is this “middle ground”? If you could not have chosen differently, complexity doesn’t really play into it, right?

In any decision that I am aware of my choices, I’m inevitably unaware of why I chose what I chose.

I am not saying that complexity allows for free will. I agree that there is no free will, but complexity can create a situation that is functionally equivalent. If you are unable to predict what a person is going to do then from a purely pragmatic point of view the person can be treated as if there is free will. The distinction between determinism and free will would have no practical value. We all visualize ourselves and others as if we have free will. Society works perfectly fine that way. Unless you can point out an advantage to seeing things differently, we might as well treat our illusions as reality.

Here is another way of looking at complexity. Suppose we want to create a robot that simulates human behavior. Suppose further that the only way that we could do this is to create a machine whose behavior is indistinguishable from that of a human. In that case the simulation is useless, because the only way of simulating a conscious being would be to create a conscious being. In other words, the complexity would be irreducible.

@LostInParadise: “We all visualize ourselves and others as if we have free will. Society works perfectly fine that way. Unless you can point out an advantage to seeing things differently, we might as well treat our illusions as reality.”

I think there are practical and moral advantages to doing away with the free will delusion. I don’t think this is purely an academic discussion. But I think that we’re probably in agreement regarding the question of whether or not the concept of free will can withstand scrutiny. Any further discussion would be centered around the implications of this lack of free will – and the effects of dispelling that delusion on the public.

I’m not sure I follow your second paragraph. If you implied in your first paragraph that deterministic behavior manifests in a way that appears to allow for libertarian free will, then wouldn’t a project to create a robot that simulates human behavior only need to meet this criteria?

@LostInParadise said: “there is no free will, but complexity can create a situation that is functionally equivalent.”

Just playing with the ideas here: Since we have no idea what free will would be like in real life, complexity can only ever give us this thing that SEEMS like free will. If we create a robot that is functionally as conscious as a human, it only means we’ve created an advanced robot… like us. But I guess it doesn’t mean we’ve created free will.

An interesting observation from A Course in Conscionsness,:

It might be determined through an unconventional causality that operates in a time-reversed direction so that the future rather than the past determines the present.

The discussion goes on to state that this is not precluded by either quantum physics or conventional physics because “because microscopic physical laws are equally valid in the time-reversed direction and in the forward direction”.

This model would preclude the existence of free will, but not necessarily the appearance of free will.