Friday, February 29, 2008

Hasker's critique of the case for determinsm

A redated post

Hasker consider the case for determinism

Two misconceptions about determinism

1) Determinism means that people do not make choices.

Determinists do not deny the existence of choices. They just maintain that actions are determined by past causes.

2) Determinism means that our choices make no difference.

My actions are causally effective even if determinism is true.

Arguments for determinism

1) Determinism is a necessary truth of reason. For every event that happens, there must be a sufficient cause. Otherwise the causes would be insufficient, and the event would not take place.
2) We always act on our strongest desire. Therefore, the strongest desire determines what we do. Our strongest desire is always the sufficient cause for our actions. Given our desires, we cannot do otherwise from what we do. (Psychological determinism).
3) Determinism is a presupposition of science. The scientists seeks to predict and control nature, this presuppose that nature is predictable and controllable. Science is the business of looking for and finding universal natural laws to explain everything. To deny determinism is therefore to oppose science.
4) Determinism is supported by the conclusions of the sciences.

Hasker, however, thinks that these arguments are not convincing.

1) Determinism is not self-evident. It can be asserted, but that doesn’t prove it.
2) There is no good way of deciding what one’s strongest desire is apart from seeing what motivates action. So “strongest desire” just means “the desire we act on.” Therefore the principle “We always act on our strongest desire,” just means “We always act on the desire we act on.” That’s trivially true.
3) Science seems to be getting along fine without determinism in physics. Quantum indeterminacy does not prove free will, but it does undercut the argument from presupposition for determinism.
4) The results of science also don’t support determinism. A good deal of behavioral science is statistical in nature. The existence of statistical laws is consistent with determinism, but also consistent with the rejection of determinism.

Therefore, Hasker concludes, the case for determinism is insufficient.

Another argument for determinism occurs to me. Suppose we assume a naturalistic or physicalistic world-view. If we do, then the physical world is a causally closed system. And everything else that exists, at least in space and time, is a necessary consequence of the state of the physical. Now it seems as if we don’t choose the state of the physical, since the physical is determined and determined only by other physical states. Nor are we responsible for the necessary consequences of the physical. But if our actions are the necessary consequences of the physical, then we are not responsible for our actions either.

Of course, one must be a physicalist in order to accept this argument, which Hasker is not.


w said...

I'm not sure that the 2nd "argument for determinism" is necessarily "trivially true."

It seems reasonable that there is a psychological explanation for every action, right?

There is a Twinkie in the pantry. I choose to eat it.


Well, because I desired it.


It tastes very good to me, and I associate pleasure with eating Twinkies.


(1) Because of biology--e.g. the way my taste-buds and brain work together. (2) Because I associate pleasant memories with eating Twinkies with my sister.

Why? . . .

We can continue, and at each stage, there seems to be some kind of psychological explanation. There is some reason why I made the choice I made. There is either a reason I did what I did or there is not.

Sure, that desire faced competing desires--e.g. the desire to remain thin and healthy--but there is a psychological reason why one desire outweighed the other at that particular time, right?

My psychological states are also caused. I have brain states caused by neurons firing, my upbringing, etc. all cause my psychological states at a particular time. There is a long chain of causes that lead to my choosing to eat the Twinkie.

Now, let's imagine that determinism is not the case. What explanation can you possibly give for eating the Twinkie? If eating the Twinkie was not the result of psychological states that were caused by biology and upbringing, then we are only left with chance, are we not?

And if we only have chance, then it seems we still do not have freedom. I see the Twinkie, a roulette wheel spins in my head and lands on "EAT." I eat. How is that freedom?

Now, move into the realm of morality. I see Tom. I feel a passionate hatred for Tom. I stab Tom in the eye with a fork.

If determinism is not the case, then upon seeing Tom, the roulette wheel in my head lands on "STAB WITH FORK," and I, obediently, stab Tom with a fork. I didn't make a "free choice"; it was governed by chance.

If determinism is the case, I saw Tom and my psychological state caused by my biology and upbringing caused me to stab Tom in the eye with a fork. But now you know my character. I am the kind of guy that stabs people I hate with a fork. I have a violent and cruel character. You know this because of my actions.

It seems that determinism, not its opposite leads to moral responsibility. You know my character BECAUSE you can explain my actions and trace it back to my character. If determinism is not the case, you cannot explain my actions, and you do not know my character.

The problem comes, however, if you think in terms of punishment rather than restraint and reformation. The rapist could not have done otherwise, therefore, punishment and vengeance are unmerited. Still, though, the rapist can be restrained so that he does not hurt anyone else. He may also be reformed so that he does not do further harm.

We condone or condemn people based on the character they possess. We judge their character. Someone's character may be violent and cruel or kind and tender.

Determinism allows us to judge someone's character by offering explanations for their behavior. If not determinism, then chance rules (i.e. actions are not caused by previous occurrences) and there can be no moral responsibility.

If one seeks to punish or seek revenge, though, (as many Christians believe God does) something seems very wrong about this.

stunney said...

The most obvious problem in physics for the last seventy-plus years is that of interpreting the wavefunction 'collapse'. Does it really 'collapse', and if it does in some sense, what causes it to do so? In thinking about that issue in relation to human minds, it's always worth remembering something which is quite easy to forget, which is that the actual wavefunctions of all the particles of all actual human brains are actually not all, er, known.

So, it may be the case, for all we know, that there is a significant number of occasions in which the probability of the particle trajectories occurring which are required for a particular brain (mine, let's say) to be in state A ('choosing vanilla ice-cream soon'), and the probability of the particle trajectories occurring which are required for it to be in state B ('choosing mint chocolate chip ice-cream soon') are each significanty less than, but jointly equal to 1. If I then choose vanilla ice cream, no known law of physics will have been 'violated', even though we cannot identify a physical cause of the wavefunction collapsing.

So why not assume what appears obvious to consciousness–—that on a significant number of occasions, what collapses the wavefunction is an irreducibly mental state, such as 'freely deciding to eat vanilla ice cream soon'?

It might be different if we actually knew that the probability of state A occurring was 1 and the probability of state B occurring was 0, such as might be the case if I hated mint chocolate chip inordinately and loved vanilla inordinately. But let's say I like both and don't have any strong preference for one over the other, and the probabilities for A to B are roughly equal. Consciously I experience this as feeling quite open to choosing one or the other. No prior mental state seems to be determining my choice, and no prior material state seems to be determining my choice. And ex hypothesi, nothing is determining my choice, given a relevant wavefunction like the one governing the probabilities of A and B respectively. It is reasonable to conclude that nothing is determining my choice, except my will at that moment.

If I choose vanilla, it would be silly to say 'chance' chose vanilla. No, chance didn't choose vanilla. I did. And I did so without violating the wavefunction that describes all the physical facts about my bodily state at the relevant times. That wavefunction gives, ex hypothesi, roughly equal probabilities at t for brain-states A and B to be observed when 'measured' at t+1. To all appearances I caused the wavefunction to collapse by choosing vanilla. Asking for a specification of an underlying mechanism that 'produced' or ’caused' this choice in fact is to ask for something that's not physically consistent with the physical facts specified by the wavefunction itself, for it does not describe any ’cause' if that term is taken to mean a determining, physical necessitation of what the actually 'measured outcome' will be.

Moreover, as I said previously, even exhaustive investigation of a purely material process will eventually yield partless ultimate parts of the physical world that act not in virtue of some other, underlying mechanism, but rather, immediately, directly, and non-mechanically. Such ultimate partless parts of the physical world just do whatever they do, and there is no sense to the question of how they do it.

If ultimate material realities can act without some further underlying process but directly (and they must, given that they're ultimate), it seems inconsistent to demand that mind must require an underlying material mechanism in order to be causally potent in the physical world, for the ultimate parts of brain-matter itself don't need any such mechanism.

To be sure, doing things to brains will normally affect experience. But this just serves to demonstrate that causal relations exist between material states and mental states, which we knew anyway. It doesn't show that all mental states just are material states, since causes are normally defined to be non-identical with their effects. And is it even possible coherently to doubt that many mental states have physical effects, such as ordering US aircraft to bomb targets in Iraq?

stunney said...

What is it for one material something to cause another material something?

If the right answer is along the general lines of, 'There's a lawlike observed correlation between A things happening and their soon being followed by B things happening', how would that account of causation rule out mental states causing physical states?

If instead the right answer is along the lines of 'A things have an intrinsic causal power, ICP, to make B things happen', how would that account of causation rule out mental states having an intrinsic causal power, MICP, to make certain physical states happen?

It strikes me as silly to ask for an explanation of how an immaterial thing can cause a physical state to occur, if by 'explanation' you mean a physical explanation, since such explanations are all explanations of physical things and immaterial things aren't physical things. Hence immaterial things don't cause their activity's effects the same way physical things, like rocks, cause their effects.

For example, a set's mathematical properties are not generally regarded as being physical properties. How then can they be in causal contact with brains? And if you think you can shortcut that question by identifying sets and their properties with human brain states, be aware of the fact that a mathematical proof can be given for there being an infinite number of such sets and properties. So, one would have to do an awful lot of work to show why there are not really an infinite number of such sets and properties. The Quine-Putnam indispensability thesis is that we need to posit sets because our best scientific theories require us to posit them in rather like the way we posit other strictly non-observable theoretical entities (such as certain fields, superpositions of particles, etc), especially given how much reliance science places on mathematics and how important set theory is to mathematics.

The problem then lies in thinking that mind or personhood reduces to successive more basic configurations of matter. The motivation for thinking that minds/persons do reduce to such configurations stems from the idea that minds/persons can only be understood by figuring out the workings of sub-mental, sub-personal material parts and processes. But why think such a thing?

Asking for a mechanistic process that ’causes' or 'produces', say, a particular conscious thought is like asking what are the inner mechanistic workings of a particular fundamental particle. But there just aren't any inner mechanisms or parts that combine to ’cause' or 'produce', say, the property of electrons being negatively charged. (I'm assuming for the sake of the argument that electrons are fundamental particles, or fundamental quanta of the electromagnetic field. One can use strings instead if strings are the truly fundamental physical things. Or one can use quanta of any ultimate energy field, whatever those may be.)

In other words, regardless of whether materialism is true or not, some things must have their nature, or essential properties, and must engage in the activity that is specified by that nature or essential properties, not in virtue of some underlying parts and processes that 'enable' or ’cause' that nature or activity, but immediately, directly, and hence non-mechanically. Otherwise one must posit an infinite regress of underlying parts, processes, causes, and mechanisms.

An objection to the anti-materialist position usually takes the form of posing a question: how can an irreducibly mental state cause physical behavior? In the absence of a satisfactory account of the wavefunction 'collapse' problem, it is not unreasonable to suppose that irreducibly mental phenomena play a role.

One possibility, which I think is more Ockhamist than the many-worlds interpretation of QM, is that the wavefunction is collapsed by just one mind—--God's. That is how interpret what Dummett says in the final chapter of his latest book, Thought and Reality. The idea that God does the collapsing raises the question of whether the universe is a 'closed system'. Some recent scientific theorizing about brane worlds and gravitational 'leakage'(the Randall-Sundrum models) suggests that it might well not be. And another theory that has a similar relevance is occasionalism.

But leaving the details of speculative quantum, string, braneworld, and occasionalist theorizing to one side for the moment, I think the answer to the 'how do minds cause bodies to move' issue will depend on what your theory of causation in general is; and it's well known that the concept of causation is the subject of very longstanding, very vigorous, and ongoing debate.

But suppose you believe, as most people do, that qualia are real. Well here's what I regard as a compelling example of a quale causing some physical behavior, in some reasonable sense of ’cause':

A comedian tells a joke, which results in the following composite mental state: you understand the joke and find it very funny. Call that mental state M. M then causes you to laugh, which is an instance of bodily behavior. Call that behavioral state L. A couple of plausible claims are that 1) a necessary property of M is that it be a conscious state; and that 2) states essentially involving consciousness, in particular qualia, are irreducible to states of matter. (An epiphenomenalist would standardly accept both 1 and 2.) One can in many contexts make a distinction between what something is––the 'real nature' of a thing–—and how it contingently appears to us, such as how the moon appears to wax and wane. Of course's the moon itself isn't really waxing and waning as much as it appears to. But in the case of states whose nature, unlike lunar states, essentially involves consciousness, such as qualia states, there logically can't be a distinction between the 'real' nature of such states and how such states contingently appear to us. In these cases, the reality of the state and its appearance to conscious minds are, in effect, one and the same thing. To take Kripke's famous example, if one leaves out of a list of pain's essential, constitutive properties (the properties that go to make something actually be pain) the phenomenal property of how pain feels, one would be leaving out the crucial, most essential property pain has. If some state S doesn't feel painful, S just isn't a state of pain. Or, as Kripke famously put it: "For a sensation to be felt as pain is for it to be pain."

Now, one may adapt this reasoning to my example: if one leaves out of a list of M's essential, constitutive properties, its amusing feel—-its 'funny' quale—-then one would have left out a property that is essential to M's being the mental state of 'understanding a joke and finding it very funny'.

And here's the punchline: what would have happened if that quale had not been part of M? Would you still have laughed—--would the behavioral state L still have occurred? By far the most plausible answer to that is, no. But then it seems very plausible to say that the state of M including the 'finding it funny' quale—which, like qualia in general, is not reducible to states of matter---—caused L; or, more accurately, was one of the causes of L (with brain signals, throat muscles, etc also being involved).

Now, there is a typical type of materialist answer to this type of example, and that is to say that in fact what happened was that you laughed first and only then experienced the funny quale. Obviously, that is prima facie implausible. But I wouldn't say it's out of the question, because it is often the case that we react first behaviorally to something, and only consciously process it after a moment or so. However, it's important to distinguish the occurrence of qualia on the one hand, and an ability to think about, and accurately report on, one's own qualia. The former is far fom being equivalent to latter. Which is why the oft-cited Libet data are far from being decisive.

Consider also that sometimes you hear a joke and don't 'get' it right away. You think for a few moments, then you 'get' it, then you think, boy, that is funny, and then you feel a pleasant mental awareness of rising amusement, and then you decide to refrain from laughing because you're in church for your mother's funeral, and the joke was told by the priest, and so you sit rigidly out of fear that bursting out laughing at your mum's funeral will be considered to be in bad taste, and then you hear others trying to stifle some sniggers, and only then do you roar with laughter, as does the rest of the congregation, the priest, your mum (she's in heaven), and Jesus (he's there too with your mum).

Oh, the joke? The joke was about these two materialists who walked into a bar. It was an iron bar, you see. And in between squealing because of their pain qualia while in the ER where your mother was attending to them in her capacity as a nurse, they congratulated each other on 'once again' confirming materialism.

Steve Lovell said...


There is another important argument for determinism which you've missed. I'm not pursuaded myself, but it's a hard argument to fault.

(1) For any future event e, either e will occur or e will not occur.
(2) If e will occur, then it was the case yesterday that e will occur.
(3) If e will not occur, then it was the case yesterday that e will not occur.
(4) Facts which obtained yesterday are not under our control today.
(5) If it was the case yesterday that e will occur, then it is not under our control today whether or not e will occur. (From 4)
(6) If it was the case yesterday that e will not occur, then it is not under our control today whether or not e will occur. (From 4)
(7) Therefore, it is not under our control whether or not e will occur. (From 1,2,3,5,6)

Obviously e can be any event at all, and there is nothing special about "yesterday" and "today" other than that yesterday is in the past that some of today is not in the past (a tiny part being present and the rest future).

There are a couple of obvious responses:

Deny (1): There are no fixed facts about the future (apart from those of logic), so neither disjunct here is true.
Deny (4): Some facts which obtained yesterday are under our control.

On the face of it these are the only options. Neither looks particularly attractive: Denying (1) looks like denying a version of the law of excluded middle, while denying (4) looks like endorsing backwards causation.

Personally I think either could be denied and that the consequences of these denials are benign.

I'd be interested in hearing other views though.

Perhaps the argument isn't valid (though it looks it), or perhaps one of the other premises could be denied (though I doubt it).


Steve Lovell said...

I see we have no-one offering a response to the argument I put up from determinism.

Is this because the argument is so implausible that it doesn't deserve a response or because it's difficult to see what can be said in response.

Or perhaps it's both!

For what it's worth, I think the argument goes wrong at (4). This is not because I accept backwards causation.

My objection can be put by either saying that (4) is false or better saying that that (4) is ambiguous depending on facts being both states of affairs (truth-makers) and truths.

The truth makers of past truths are not necessarily past facts. The facts which make these pasts truths true, are future facts which therefore do not fit under (4).

How does this seem to other readers? I've had colleagues who found the argument very convincing and didn't accept my response, though I could never figure out why.


Timmo said...


We can deny that there are facts about the future and let statements of the form 'it will be the case that p' be neither true nor false (be gappy) without dropping the law of the excluded middle if we employ van Fraassen's technique of supervaluations.

Let L be the language of classical propositional logic. A valuation v for L is a total function mapping atomic sentences to the set {0, 1}, and mapping molecular formulae as follows:

V(~A) = 1 - v(A)
v(A V B) = V(A) + v(B)

Since the other connectives can be defined in terms of these two, that's enough clauses.

A partial function s:L --> {0,1} is a supervaluation of L just in case there is a set V of valuations of L such that for every formula A, s(A) = x iff for every v in V, V(A) = x, and s is undefined at A if there is no such value.

Let S be the language which shares the syntax of L, except valuations for L are exactly those supervaluations of L. You can prove that, for any A, |=L A iff |=S A. While formulae of S can be neither true nor false, all classical tautologies are valid in S.

So, you can setup a system of tense logic in which every proposition of the form 'Gp' (G is the operator with the intuitive meaning that 'it will always be the case that...' and p is an atomic formula) is neither true nor false, but still have |= Gp V ~Gp.

I don't know if this is the right way to go philosophically, but it does show that you can set up systems of logic which allow you to keep all classical tautologies while admitting truth-value gaps.

Timmo said...

Uh, I did not intend for the weird caps in my definition of valuations for L. It should just read:

v(~A) = 1 - v(A)
v(A V B) = v(A) + v(B)

Timmo said...

Whoops! I noticed another typo. This sentence "Let S be the language which shares the syntax of L, except valuations for L are exactly those supervaluations of L." should read:

Let S be the language which shares the syntax of L, except valuations for S are exactly those supervaluations of L.

Sorry. :-(

steve lovell said...


Thanks for your comments.

I agree, we can also go that route, although the extra sophistication of supervaluations.

If I were to deny the relevant premise, I'd rather say that all statements about contingent future events are false.

This is analogous to Russell's discussion of statements like:

(1) The king of France is bald.
(2) The king of France is not bald.

While these two statements are contradictory, (2) is not the straightforward negation of (1). Therefore we are not denying the law of exlcuded middle when we say both are false. Rather (1) and (2) should be interpreted as

(1') There is a king of France and he is bald.
(2') There is a king of France and he is not bald.

The negation of (1') is therefore not (2') but rather (3')

(3') Either there is no king of France or there is a king of France and he is not bald.

or more simply

(3'') It's not the case that (There is a king of France and he is bald)

Along these lines statements about contingent future events are to be interpreted thus:

(Fa) It will be the case that P.
(Fb) It will be the case that not-P


(Fa') There is a fact of the matter of whether P, and it will be the case that P.
(Fb') There is a fact of the matter of whether P, and it will be the case that not P.

Where P is some future contingent state of affairs, we can then consistently deny both (Fa') and (Fb').

Still, I don't think this is the correct way out of the argument. I'd rather go the route I suggested a few comments above.


Darek Barefoot said...


For what it's worth (which I freely confess is not a lot, given my limited experience with these kinds of questions) I think your final analysis and solution is correct. It is not incoherent to assert that the fact that I will willingly perform or refrain from performing an action tomorrow may make it true today that tomorrow I will perform or refrain. But it is tricky terrain. William Lycan, in his essay on dualism recently cited by Victor, asserts that knowledge need not always entail causation, based on the fact that we have at least some knowledge of the future.