Philosophy Now Forum

“Incompatibilism” is the claim that free will is incompatible with a deterministic world.

Take four similar Meccano strips (or equivalent) and bolt them loosely together at their ends to form a rhombus. Bolt a fifth strip between two of the opposite corners. The resulting assembly has no degrees of freedom: it is determinate. Note, however, that different lengths of the diagonal strip result in different shapes of the assembly.

Consider this as a model for a world containing at least one mind. The mind is modelled by the diagonal strip, and the length of the diagonal represents its will.

Then we conclude that this world is determinate, but ALSO that the will of the included mind is able to determine the state of the world. This apparent contradiction arises because the determinism of the world equates to an absence of externally accessible freedom, whereas the freedom needed for a will to determine what happens is internally accessible freedom – accessible to a will whose state is itself a variable of the world state. The model illustrates that it is a fallacy that a lack of externally accessible freedom necessitates a lack of internally accessible freedom. This is the first Incompatibilist fallacy.

The standard reply to this kind of argument is that in addition to the ordinary kind of freedom – the freedom of a free-range chicken, for example – “freedom of will” requires the ability of the mind to choose its own state at the time of choice. This is impossible, it is claimed, because prior causes in the world provide necessary and sufficient conditions to determine that state. Can anyone guess why I claim that this too is a fallacy?

RogerSH wrote: ↑Thu May 13, 2021 1:30 pm Take four similar Meccano strips (or equivalent) and bolt them loosely together at their ends to form a rhombus. Bolt a fifth strip between two of the opposite corners. The resulting assembly has no degrees of freedom: it is determinate. Note, however, that different lengths of the diagonal strip result in different shapes of the assembly.

This makes no sense to me on a few ends:

One, it's unclear what "determinate" would refer to above--that is, it's unclear what sense of "determinate" you'd be using.

In free will discussions, and normally in general, "determinism" refers to the fact that antecedent states causally dictate a single subsequent state. But that doesn't seem to be what you'd have in mind above.

Two, "freedom" in the context of free will discussions refers to whether it's possible to make a decision, in the sense of there being at least two incompatible subsequent states possible immediately following the same antecedent state, and then the issue is whether we can consciously cause a particular one of those subsequent states to obtain rather than being forced (whether we realize it or not) to follow a single possible subsequent state.

(And thus we have the reason that many people, including me, say that compatibilism is incoherent: determinism means that only one immediately subsequent state is possible; freedom means that at least two immediately subsequent states are possible; those two notions are incompatible.)

Additionally, when you say that what you describe "has no degrees of freedom," it's not clear what you'd have in mind in general, as the Meccano strips, the bolts, etc. are all in motion on a microscopic level (as are all things that are more than a single elementary particle), and even on a macro level, what you're describing isn't going to be "absolutely rigid," "absolutely nonpliable," etc. Nothing could withstand the pressure that would require--which would probably be an infinite amount of pressure.

So it seems that "This set-up has no freedom"/"This set-up is deterministic" is not the case.

RogerSH wrote: ↑Thu May 13, 2021 1:30 pm Then we conclude that this world is determinate, but ALSO that the will of the included mind is able to determine the state of the world. This apparent contradiction arises because the determinism of the world equates to an absence of externally accessible freedom, whereas the freedom needed for a will to determine what happens is internally accessible freedom – accessible to a will whose state is itself a variable of the world state. The model illustrates that it is a fallacy that a lack of externally accessible freedom necessitates a lack of internally accessible freedom. This is the first Incompatibilist fallacy.

By the way, what's required for the world to be deterministic is that NO phenomena (that is, no occurrences), period, have any ontological freedom.

If any phenomena are ontologically free, then the world isn't deterministic. In other words, a single ontologically free phenomenon falsifies "the world is deterministic."

So if there is such a thing as "internally accessible freedom," that means that there's at least one thing with ontological freedom, and thus the world isn't deterministic and compatibilism isn't the case (at least not in general, in a nonqualified way).

That the world has some ontologically free phenomena doesn't imply that free will is the case, of course. Free will is only the case if there are ontologically free phenomena that will has some influence on. But even if there would be no free will, if there are some free phenomena somewhere, "The world is deterministic" is not the case. You could say something like "Most phenomena are deterministic," of course, but that wouldn't help compatibilism make sense, as compatibilism claims that the SAME phenomena can somehow be both deterministic and free.

Yes, I did worry that my use of a static model for a dynamic world might pose a problem! The idea was simply to make the distinction between internally accessible freedoms and externally accessible freedoms as vivid as possible by using the simplest possible model. In fact, a static model is just a special case: it is still the case that

antecedent states causally dictate a single subsequent state

since the subsequent state in this case is identical to the antecedent case.

the Meccano strips, the bolts, etc. are all in motion on a microscopic level (as are all things that are more than a single elementary particle), and even on a macro level, what you're describing isn't going to be "absolutely rigid," "absolutely nonpliable," etc.

It's only a model, and I'm intending that we consider its behaviour as though it were made of idealised rigid components with no internal structure.

... in the sense of there being at least two incompatible subsequent states possible immediately following the same antecedent state...

By this definition, the entire assembly is clearly determinate, as there is only one possible subsequent state: the state identical to the initial state.

...whether it's possible to make a decision...

Where is the mind where this decision is made? If it is external to the world - if the decision is represented by the length of a separate strip still in the box - then you are right, there are no externally accessible freedoms. But if the diagonal represents a mind that is part of the world, then a "decision" is a particular length of the diagonal, and the question facing that mind is, would a different length have different outcomes? And the answer is yes. In other words, it is not the adjustability of the entire world that is in question from the point of view of a mind that is part of the world, but the adjustability of the part of the world external to the mind.

It seems to me that incompatibilists think of the mind as external to the world when conceiving of a decision, but as a part of the world when conceiving of the world as determinate. That is what I see as incoherent.

determinism means that only one immediately subsequent state is possible; freedom means that at least two immediately subsequent states are possible

Of course, only one immediately subsequent state of the entire world is possible because that means that the mind, being part of the world, is in the same state. Freedom of the mind to choose requires that that there are at least two immediately subsequent states of the world external to the mind whose possibility is only excluded, in all except one case, by the fact of not being chosen.

To ask for more than that as a requirement for freedom is to ask that a mind can choose, as well as the state of the external world, its own state at the moment that it is doing the choosing - related to what I call the second fallacy, which I will come to soon when I've cleared up any confusions about my first post - which I hope to have done here!

By the way, what's required for the world to be deterministic is that NO phenomena (that is, no occurrences), period, have any ontological freedom.

Of course, I don't think the world is perfectly deterministic, but I am assuming it is for the purposes of the current argument, since I think the conclusion is unaffected and it makes the point at issue clearer. (On the other hand, it is possible that at least some varieties of perfectly deterministic world could never generate the complexity necessary for consciousness, so there couldn't be a will as I understand the term.)

So if there is such a thing as "internally accessible freedom," that means that there's at least one thing with ontological freedom

I don't think that calling a freedom "ontological" provides justification for failing to distinguish external from internal freedom. Any system, however deterministic, has internally accessible freedom. It just means that there is a causal relationship between the state of part of the system and the remainder, so the state of the part is able to have a determining influence on its surroundings.

RogerSH wrote: ↑Sat May 15, 2021 6:58 pm since the subsequent state in this case is identical to the antecedent case.

No. That's incoherent. It wouldn't be an antecedent and a subsequent state in that event.

It's only a model, and I'm intending that we consider its behaviour as though it were made of idealised rigid components with no internal structure.

Then we're not actually talking about ontology, but some fantasy.

By this definition, the entire assembly is clearly determinate, as there is only one possible subsequent state: the state identical to the initial state.

Again, the very idea of that is incoherent. A subsequent state has to be different than the antecedent state. antecedent/subsequent (or consequent) are temporal modalities. Time is change. If the states are identical, there has been no change, no time, and there aren't two states, there's just one.

Where is the mind where this decision is made?

Minds are subsets of brain function (at least--we don't know if other materials can have mental properties or not).

If it is external to the world -

First off, nothing can be external to the world in that sense. "The world" in that sense is basically the same as "the universe," which should be everything that exists, whatever its nature.

But brains aren't in any significant ontological way separate from the rest of the world. They're in different locations and made of different materials than other things, but in the same way that, say, a shoe is in a different location and made of different materials than a star.

if the decision is represented by the length of a separate strip still in the box - then you are right, there are no externally accessible freedoms. But if the diagonal represents a mind that is part of the world, then a "decision" is a particular length of the diagonal, and the question facing that mind is, would a different length have different outcomes?

I'm really not able to follow this. It would have helped if you'd answered the questions I asked regarding how you're using the word "determinate" etc.

It seems to me that incompatibilists think of the mind as external to the world when conceiving of a decision,

I'm an incompatibilist. I don't believe there's any way to make compatibilism coherent without simply changing the discussion and using common terms to refer to something very different, something relatively unusual (to common usage). I explained why I'm a incompatibilist in the post you're quoting. It has nothing to do with mind being "external to the world." It has to do with determinism being a situation where only one consequent state can follow an antecedent state (and where of course the two are not the same, because the very idea of that is incoherent), whereas freedom obtains when at least two possible states can immediately follow an antecedent state.

Of course, only one immediately subsequent state of the entire world is possible because that means that the mind, being part of the world, is in the same state.

Maybe you don't understand what I'm saying? The antecedent state could be a particular particle, A, in a particular motion, the consequent state would be A "striking" particle B, say. And if it's an immediately consequent state, we're talking about the very next temporal moment (the very next change in A, so we could say A moving the smallest possible distance just prior to "striking" B). If there's only one possible immediately consequent state, then B can only move with a particular amount of force, in only one direction, etc. If there are at least two possible immediately consequent states, then, for example, B might be able to move in direction x or slightly different direction y, with no intervening forces, changes, etc. that are different. The second state there (B's motion) can't be the same as the first state (A's motion immediately prior to striking B).

So I don't know what you'd even be thinking re saying that subsequent states would be the same/would be identical to consequent states (aside from that being incoherent).

I don't think that calling a freedom "ontological" . . .

The whole point there is to clarify that we're not talking about freedom in, say, the political sense, and we're not talking about free will, necessarily. We have ontological freedom (not political freedom, obviously, and not free will), when particle B can move in two at least slightly different directions after being struck by particle A (where for each option, particle A is striking particle B just the same way).

By the way, the state identical to A moving at velocity v at time T1 is A moving at velocity v at time T1. How is A moving at velocity v at time T1 antecedent or consequent to A moving at velocity v at time T1? That makes zero sense, and suggests unfamiliarity with the terms "antecedent" and "consequent" (or "subsequent").

Terrapin Station wrote: ↑Sat May 15, 2021 8:05 pm No. That's incoherent. It wouldn't be an antecedent and a subsequent state in that event.

I am using “state” in the sense of mechanical science, which evidently differs from ontological usage. Typically, the state is represented by a state vector comprising positions and orientations, and their rates of change, of all the components of the mechanical system in question. Time is not included, which is what differs from your sense, but the state vector S varies with time so that S(t) represents the vectorial value of S at time t. If t2 is later than t1, then the system is fully determined if there is a vectorial function F[ ] such that S(t2) = F[S(t1)]. In the example case I am considering, F[ ] is the identity function, so that S(t2)=S(t1), although t2 differs from t1. I hope that clarifiers matters and corresponds well enough to your sense of determinate/fully determined.

Terrapin Station wrote: ↑Sat May 15, 2021 8:05 pm Then we're not actually talking about ontology, but some fantasy.

I’m surprised that you don’t think there is any value in models which isolate a particular feature of ontology, but if you think that, my original post is obviously not going to be very helpful! That is a shame as it is the easiest way to explain the difference between external and internal freedoms, (which are routinely distinguished in mechanical science but seem to pose a problem for Incompatibilist metaphysicians).

Terrapin Station wrote: ↑Sat May 15, 2021 8:05 pm First off, nothing can be external to the world in that sense. "The world" in that sense is basically the same as "the universe," which should be everything that exists, whatever its nature.

In that sense of the world, I agree, nothing can be external. My sense of world is intended to include everything that is causally connected to the part of the world we are in. In a thought experiment, one can imagine a demon that is not part of that world. Other strands of the multiverse at the time in question, if there be such, are also excluded by my definition.

Terrapin Station wrote: ↑Sat May 15, 2021 8:05 pm The whole point there is to clarify that we're not talking about freedom in, say, the political sense

OK. So I am definitely talking about ontological freedom, which I think of as externally accessible formal freedom, as well as about an internally accessible formal freedom.

and we're not talking about free will, necessarily. We have ontological freedom (... not free will), when particle B can move in two at least slightly different directions...

To clarify, I think you are saying that ontological freedom is a necessary but not sufficient condition for Free Will, because, for example, particles don't have Wills.

I'm an incompatibilist [which] has to do with determinism being a situation where only one consequent state can follow an antecedent state, whereas freedom obtains when at least two possible states can immediately follow an antecedent state.

Of course, this definition of ontological freedom is coherent on its own, and the concept is a useful one; but because it is an external, “god’s eye view” freedom, I am arguing that it is not intelligible as the freedom of a Will that is internal to the world. So it is necessary to get clear what a "Will" or "choice" or "decision" is before establishing how it can be free.

I don't believe there's any way to make compatibilism coherent without simply changing the discussion..

I know some compatibilists focus on a psychological sense of freedom, but I am referring here to freedom in a formal sense, but defined in such a way that it can meaningfully be exercised by a mind that is part of the world. (Hilary Bok, for example, is a compatibilist in this sense). What exactly that sense is can be discussed more easily after I have explained what I mean by the “second fallacy".

...using common terms to refer to something very different, something relatively unusual (to common usage).

That is a separate question that I do not intend to pursue under this topic, though I cannot hep smiling in passing at the thought that a free-range chicken means it (or its range) has ontological freedom! Perhaps you mean "common usage among ontologists"?

If we're going to keep going back and forth I don't want to do a bunch of different points/issues each time. Let's see if we can settle anything by doing one point/issue at a time:

RogerSH wrote: ↑Tue May 18, 2021 2:52 pm I am using “state” in the sense of mechanical science, which evidently differs from ontological usage. Typically, the state is represented by a state vector comprising positions and orientations, and their rates of change, of all the components of the mechanical system in question. Time is not included,

So a big problem with this is that time is identical to change. If we include change in something, then we have time.

At any rate, I'm talking about states themselves, not representations of states, which wouldn't be identical to the states themselves.

I was hoping that my first few points would just be understood as clarifications of terminology.

Terrapin Station wrote: ↑Tue May 18, 2021 2:57 pm So a big problem with this is that time is identical to change. If we include change in something, then we have time.

Yes, we have time, and the physical state is defined at each time, which is change. That's just a different sense of "state" surely? I don't see how it can be a big problem when it works perfectly well in (pre-Einstein) physics.

RogerSH wrote: ↑Tue May 18, 2021 4:19 pm I was hoping that my first few points would just be understood as clarifications of terminology.

Terrapin Station wrote: ↑Tue May 18, 2021 2:57 pm So a big problem with this is that time is identical to change. If we include change in something, then we have time.

Yes, we have time, and the physical state is defined at each time, which is change. That's just a different sense of "state" surely? I don't see how it can be a big problem when it works perfectly well in (pre-Einstein) physics.

Alright, so for the next part:

but the state vector S varies with time so that S(t) represents the vectorial value of S at time t. If t2 is later than t1, then the system is fully determined if there is a vectorial function F[ ] such that S(t2) = F[S(t1)]. In the example case I am considering, F[ ] is the identity function, so that S(t2)=S(t1), although t2 differs from t1. I hope that clarifiers matters and corresponds well enough to your sense of determinate/fully determined.

We don't actually have a different time (insofar as we're talking about just these states--we might have time relative to something else) if t1 is identical to t2. Again, since time is identical to change, if there's no change (in state t1 to t2), there's no time (insofar as t1 & t2 go), and t2 isn't a subsequent or consequent state to t1.

I didn't say t1 is identical to t2, I said t2 was later. It is all the values in the state vector at t2 that are identical to the values at t1. (The identity function makes the variables in S identical, not its argument.) The universe has gone to sleep, so all the positions of every particle are unchanging and all the velocities are zero. Time is ticking along in a Newtonian way, though you wouldn't know it.

RogerSH wrote: ↑Tue May 18, 2021 4:34 pm I didn't say t1 is identical to t2, I said t2 was later. It is all the values in the state vector at t2 that are identical to the values at t1. (The identity function makes the variables in S identical, not its argument.) The universe has gone to sleep, so all the positions of every particle are unchanging and all the velocities are zero. Time is ticking along in a Newtonian way, though you wouldn't know it.

Again, if there are no changes (from state 1 to state 2), there is no time, so it is NOT later. You can't have it so that there are no changes yet time passes. That's incoherent.

Terrapin Station wrote: ↑Tue May 18, 2021 4:56 pm Again, if there are no changes (from state 1 to state 2), there is no time, so it is NOT later. You can't have it so that there are no changes yet time passes. That's incoherent.

No changes except the passage of time! That's surely quite a familiar concept?

In any case it isn't essential to my argument, it just makes it easier to visualise. You can have all the bars rhythmically lengthening and shortening if you like!

Anyway, sorry but I have to sign out now for a couple of days probably.

Before I go I'll post part 2 of the argument, which no doubt you will like even less, but I'd still like to know why....

Terrapin Station wrote: ↑Tue May 18, 2021 4:56 pm Again, if there are no changes (from state 1 to state 2), there is no time, so it is NOT later. You can't have it so that there are no changes yet time passes. That's incoherent.

No changes except the passage of time! That's surely quite a familiar concept?

In any case it isn't essential to my argument, it just makes it easier to visualise. You can have all the bars rhythmically lengthening and shortening if you like!

Anyway, sorry but I have to sign out now for a couple of days probably.

Before I go I'll post part 2 of the argument, which no doubt you will like even less, but I'd still like to know why....

[My original post...]

The standard reply to this kind of argument is that in addition to the ordinary kind of freedom – the freedom of a free-range chicken, for example – “freedom of will” requires the ability of the mind to choose its own state at the time of choice. This is impossible, it is claimed, because prior causes in the world provide necessary and sufficient conditions to determine that state. Can anyone guess why I claim that this too is a fallacy?

Divide the second Incompatibilist claim into two parts. (1) It is impossible for the mind to choose its own state at the time of choice. (2) The impossibility arises because prior causes in the world provide necessary and sufficient conditions to determine that state. I entirely accept (1): for the mind to choose its own contemporaneous state is certainly impossible, since whatever state it is in, it cannot be in any other state at that moment, so there is no alternative possibility to choose. In terms of the Meccano model, the length of the diagonal strut tautologically selects that length from any other length. However, I claim that (2) is false because the first conclusion does not require any assumption of determinism. Suppose the diagonal of the model is an elastic band rather than a rigid strut. This introduces an externally accessible freedom but not an internally accessible one: although the length can vary, at any instant the length that is willed is by definition the length it is. Whether the length/state of mind at that instant arises from external influences, internal processes, a quantum leap, agent causality, transcendent divine intention or whatever makes no difference: it still can’t choose to be other than it is at that instant.

This impossibility of choice arises from what-is-being-chosen being made identical to what-is-doing-the-choosing. To show this more formally: from the very concept of “choice”, what-is-being-chosen must include two or more possible states right until the act of choice makes one state correspond to the state of what-is-doing-the-choosing, whereas what-is-doing-the choosing must arrive at one single state immediately prior to the act of choice, so the contradiction is unavoidable: two or more states cannot equal one state. So the second incompatibilist fallacy, as I see it, is blaming determinism for the impossibility of “choosing one’s contemporaneous state of mind”, when it is logically impossible in any case.

Of course, some people say that their own state of mind is up to them, but on cross-examination they mean only that their state of mind at a given moment in time is in part the consequence of mental choice made earlier. If I am feeling angry, I can decide to take a deep breath and with luck will feel calmer – so that calmness is indeed, in part “up to me”. That doesn’t mean I am choosing my own state of choice at the moment of choosing. It couldn’t mean that, because there is nothing that could mean.

The implication is that a concept of “freedom of will” that includes this self-referring requirement is completely useless, since it does not refer to any state of affairs that could pertain in any possible way that the world could be. The compatibilist concept of formal free will is logically speaking the only concept that makes sense. This concept, to recap, refers to the freedom of a will whose state is an internal variable of the world to access internal freedoms of that world (even if that world were perfectly deterministic – which, of course, it isn’t) to make choices that, among other restrictions, exclude its own state at that instant.

For readers who don’t like mechanical analogies, another way of defining internal freedom is that it is a counterfactual freedom. (There is nothing radical about this: one definition of causality is that A has caused B, where A and B are both the case, IF it is true that if A were not the case, B would not have occurred. This criterion is not invalidated just because A being false is counterfactual.)

Thus the simplest expression of freedom of will is “I did C but I could have done D if I had wanted to, so I was free to choose between C & D”. The clause “if I had wanted to” refers to a counterfactual because in the event I didn’t want to. There is nothing in this expression of free will that assumes externally accessible ontological freedom, because the state of mind doing the choosing was not implied to be part of what was being chosen, so whether I had been “able” to want to doesn’t enter into it. That is the point of the “second fallacy”.

To combine the first and second arguments, even if the entire world were fully determined, the world external to my mind would not be, because there are alternatives which depend on the state of my mind. Thus if my mind has logically to be excluded from what-is-being-chosen, what-is-being-chosen is already underdetermined. So full determinism would not be a problem. (I am assuming, of course, that consciousness would be possible in a fully determined world, which is not beyond question.)

Ironically, the argument in this post really only came clear to me after reading Sam Harris’ polemic against the assumption of Free Will. He argued, essentially, that in a fully determinate world choices are necessarily unfree, but in a quantum-indeterminate world, choices are necessarily unchosen. (The alternatives that exist are not correlated to different mind-states, so no conscious choosing is involved.) Catch-23? Nowhere did he describe what the world would have to be like for Free Will to be possible. Neither did he seem to reflect on the fact that if there is no such conceivable world, this shows his concept of Free Will to be without meaning, or that this makes his claim that it doesn’t exist as meaningless as the claim that it does…

RogerSH wrote: ↑Tue May 18, 2021 6:52 pm No changes except the passage of time! That's surely quite a familiar concept?

It might be something people say and think makes sense, but it's incoherent. Again, time is IDENTICAL to change. If there are no changes, there is no time. Something has to be changing in order for there to be time. Time doesn't exist "on its own." It's simply the dynamic relations of matter.

So the whole point is that state A can't be identical to state B where state B is subsequent to state A. The importance of this is in understanding the definitions of freedom and determinism I gave earlier, under which compatibilism doesn't work.