Molinists and Occhamists on the Loose
This post is a response to Mark Linville's argument against the way Occhamists and Molinists reconcile God's foreknowledge with human freedom in his article "Occhamists and Molinists in Search of a Way out".
Linville’s Agument
Using Hasker’s arguments based on the combination of the necessity of the past and God’s essential omniscience, Linville concludes Occhamists cannot hold counterfactual power over the past (i.e. if I do X, the past would have been different). Rather Occhamists must hold to actual power over the past (i.e. I have the ability to move from the possible world I am in to a different one with a different past). Linville concludes this is the only valid way for Occhamists to reconcile God's foreknowledge with libertarian freewill.
However, 'actual power over the past' lets compatiblists off the hook on the consequence argument, since the consequence argument1 is based on the inalterability of the past.
But Molinists are committed to an independence thesis2, so they cannot assert actual power over the past.
So while Occhamists can escape Hasker’s arguments by asserting actual power over the past, and in doing so let compatiblists off the hook, Molinists cannot because of their commitment to the independence thesis.
Let's look at the various parts of the argument.
Hasker’s argument for the incompatiblity of divine foreknowledge and freedom based on the combination of the necessity of the past and God’s essential omniscience
William Hasker has offered a forceful version of this argument for the incompatibility of divine foreknowledge and libertarian freedom. Suppose that Clarence will have an omelet tomorrow morning, that God is essentially omniscient, and that omniscience entails foreknowledge. Suppose, further, that the past is unalterable in such a way that it is never within anyone's power to bring about any past states of affairs. Then the following seems to be true:
(1) It is not within Clarence's power to bring it about that God has never believed that Clarence would have a cheese omelet tomorrow.
(2) It is not possible that Clarence will refrain from having a cheese omelet and yet that God has always believed that Clarence would have a cheese omelet tomorrow.
(1) is thought to result from the premise of the unalterability of the past. (2) follows from God's essential omniscience. But then (1) and (2) seem to entail
(3) It is not within Clarence's power to refrain from having a cheese omelet tomorrow.
(Mark D. Linville. Ockhamists and Molinists in Search of a Way Out. Religious Studies, Vol. 31, No. 4 (Dec., 1995), p 501)
1 and 2 are true. The past is indeed unalterable. Further, God is essentially omniscient so the combination of God’s true belief that Clarence will eat an omelet and Clarence refraining from eating the omelet is logically impossible. However, all that follows from 1 and 2 is that Clarence will not refrain from eating the omelet; not that he cannot. The argument fails to distinguish between the causal ability to eat the omelet and the logical implications of actually doing so. Is Clarence able to eat the omelet is a different question than if he eats it, what are the logical implications? So Haskers argument subtly and invalidly slides from causal possibility and use of ability to logical possibility and the ability itself.
The Difference Between Counterfactual Power over the Past and Actual Power over the Past
(CPP) God has always believed that S does A at t, but it is within S's power to do something such that, where he to do it, God would always have had a different belief about what S does at t.
(CPPi) The actual world, W, includes God's always having believed that S does A at t. But there is a world W* such that W* includes S's refraining from doing A at t and God's always having believed that S refrains from doing A at t, and it is within S's power to bring it about that W* is the actual world.
(CPPi) is bolder than (CPP) in that it claims that, not only is there a world in which the agent does and God believes otherwise, but that at the time the relevant choice is made, it is within the agent's power to bring it about that that world is the actual world. This is a sense of counterfactual power with a causal element. (CPPi) commits one to the view that, for every autonomous agent, whenever that agent freely chooses from among possible alternatives it is partly up to that individual which world is the actual world. (P 509)
CPP is a resonable statement of conterfactual power over the past and CPPi is a good formulation of actual power over the past. I disagree that a ‘causal element’ need be imputed to the Occhamist and that they must accept CPPi. S is able to refrain from doing A and were S to do so, W* would be the actual world. So the connection between S and W* isn’t causal, it’s a counterfactual dependence. S’s actions can be the basis of truth of non-A and the truth of non-A is logically related to W*.
Actual Power over the Past Lets Compatiblists off the hook on the Consequence Argument
Shall we say, then, that agents may have causal power over their actions, but only counterfactual power over the causal antecedents of those actions? If the Plantingean is permitted such a distinction, I fail to see the relevant difference between the cases that would justify our denying such a move to the compatibilist with regard to determinism. (p 506)
Linville makes two mistakes. First, the causal antecedents in the case of Occhamism (God’s past belief) does not determine or necessitate our actions. But in the case of determinism, the causal antecedents do determine our actions. So in the case of Occhamism, the ability is actual and the hypothetical use of that ability corresponds to a hypothetical past; but in determinism the ability itself is only hypothetical and not real.
Second, in Occhamism, we have counterfactual power over logical antecedents, not causal antecedents. I am able to choose chocolate or vanilla. Let’s say ‘I will choose chocolate’ is true. I am able to choose vanilla and have the counterfactual power to bring it about that “I choose vanilla” is true (i.e. if I choose vanilla, then my action is the basis of truth of the proposition ‘I choose vanilla”, but I don’t cause a truth only a basis of truth).
This counterfactual power entails the counterfactual power to bring it about that God’s past belief that ‘God believed I will eat chocolate’ is false. It however, does not entail that I have the ability to retroactively cause God not to have believed “Dan will eat chocolate” (the power is counterfactual not factual), nor does it entail that God believed ‘Dan will eat chocolate’ and ‘Dan will eat chocolate’ is false. Rather, in entails that I have the causal ability to choose vanilla such that if I were to do so, God would have believed “Dan will eat vanilla”.
Conclusion
Linville's arguments are thought provoking, but small subtle mistakes lead in the steps lead to large mistakes in the conclusion; that Molinism is reducable to a form of compatiblism. Molinism reconciles God's providential governence of the world and libertarian freedom and it stands up against Linville's argument.
-------------------------
1 The Consequence argument has many specific and nuanced forms, but the basic picture is: if determinism is true the laws of nature and the past determine everything, including my acts. But I cannot control the past or laws of nature, so I cannot control their consequences, including my acts.
2 The truth of conterfactuals is logically prior to anything anyone actually does.
Linville’s Agument
Using Hasker’s arguments based on the combination of the necessity of the past and God’s essential omniscience, Linville concludes Occhamists cannot hold counterfactual power over the past (i.e. if I do X, the past would have been different). Rather Occhamists must hold to actual power over the past (i.e. I have the ability to move from the possible world I am in to a different one with a different past). Linville concludes this is the only valid way for Occhamists to reconcile God's foreknowledge with libertarian freewill.
However, 'actual power over the past' lets compatiblists off the hook on the consequence argument, since the consequence argument1 is based on the inalterability of the past.
But Molinists are committed to an independence thesis2, so they cannot assert actual power over the past.
So while Occhamists can escape Hasker’s arguments by asserting actual power over the past, and in doing so let compatiblists off the hook, Molinists cannot because of their commitment to the independence thesis.
Let's look at the various parts of the argument.
Hasker’s argument for the incompatiblity of divine foreknowledge and freedom based on the combination of the necessity of the past and God’s essential omniscience
William Hasker has offered a forceful version of this argument for the incompatibility of divine foreknowledge and libertarian freedom. Suppose that Clarence will have an omelet tomorrow morning, that God is essentially omniscient, and that omniscience entails foreknowledge. Suppose, further, that the past is unalterable in such a way that it is never within anyone's power to bring about any past states of affairs. Then the following seems to be true:
(1) It is not within Clarence's power to bring it about that God has never believed that Clarence would have a cheese omelet tomorrow.
(2) It is not possible that Clarence will refrain from having a cheese omelet and yet that God has always believed that Clarence would have a cheese omelet tomorrow.
(1) is thought to result from the premise of the unalterability of the past. (2) follows from God's essential omniscience. But then (1) and (2) seem to entail
(3) It is not within Clarence's power to refrain from having a cheese omelet tomorrow.
(Mark D. Linville. Ockhamists and Molinists in Search of a Way Out. Religious Studies, Vol. 31, No. 4 (Dec., 1995), p 501)
1 and 2 are true. The past is indeed unalterable. Further, God is essentially omniscient so the combination of God’s true belief that Clarence will eat an omelet and Clarence refraining from eating the omelet is logically impossible. However, all that follows from 1 and 2 is that Clarence will not refrain from eating the omelet; not that he cannot. The argument fails to distinguish between the causal ability to eat the omelet and the logical implications of actually doing so. Is Clarence able to eat the omelet is a different question than if he eats it, what are the logical implications? So Haskers argument subtly and invalidly slides from causal possibility and use of ability to logical possibility and the ability itself.
The Difference Between Counterfactual Power over the Past and Actual Power over the Past
(CPP) God has always believed that S does A at t, but it is within S's power to do something such that, where he to do it, God would always have had a different belief about what S does at t.
(CPPi) The actual world, W, includes God's always having believed that S does A at t. But there is a world W* such that W* includes S's refraining from doing A at t and God's always having believed that S refrains from doing A at t, and it is within S's power to bring it about that W* is the actual world.
(CPPi) is bolder than (CPP) in that it claims that, not only is there a world in which the agent does and God believes otherwise, but that at the time the relevant choice is made, it is within the agent's power to bring it about that that world is the actual world. This is a sense of counterfactual power with a causal element. (CPPi) commits one to the view that, for every autonomous agent, whenever that agent freely chooses from among possible alternatives it is partly up to that individual which world is the actual world. (P 509)
CPP is a resonable statement of conterfactual power over the past and CPPi is a good formulation of actual power over the past. I disagree that a ‘causal element’ need be imputed to the Occhamist and that they must accept CPPi. S is able to refrain from doing A and were S to do so, W* would be the actual world. So the connection between S and W* isn’t causal, it’s a counterfactual dependence. S’s actions can be the basis of truth of non-A and the truth of non-A is logically related to W*.
Actual Power over the Past Lets Compatiblists off the hook on the Consequence Argument
Shall we say, then, that agents may have causal power over their actions, but only counterfactual power over the causal antecedents of those actions? If the Plantingean is permitted such a distinction, I fail to see the relevant difference between the cases that would justify our denying such a move to the compatibilist with regard to determinism. (p 506)
Linville makes two mistakes. First, the causal antecedents in the case of Occhamism (God’s past belief) does not determine or necessitate our actions. But in the case of determinism, the causal antecedents do determine our actions. So in the case of Occhamism, the ability is actual and the hypothetical use of that ability corresponds to a hypothetical past; but in determinism the ability itself is only hypothetical and not real.
Second, in Occhamism, we have counterfactual power over logical antecedents, not causal antecedents. I am able to choose chocolate or vanilla. Let’s say ‘I will choose chocolate’ is true. I am able to choose vanilla and have the counterfactual power to bring it about that “I choose vanilla” is true (i.e. if I choose vanilla, then my action is the basis of truth of the proposition ‘I choose vanilla”, but I don’t cause a truth only a basis of truth).
This counterfactual power entails the counterfactual power to bring it about that God’s past belief that ‘God believed I will eat chocolate’ is false. It however, does not entail that I have the ability to retroactively cause God not to have believed “Dan will eat chocolate” (the power is counterfactual not factual), nor does it entail that God believed ‘Dan will eat chocolate’ and ‘Dan will eat chocolate’ is false. Rather, in entails that I have the causal ability to choose vanilla such that if I were to do so, God would have believed “Dan will eat vanilla”.
Conclusion
Linville's arguments are thought provoking, but small subtle mistakes lead in the steps lead to large mistakes in the conclusion; that Molinism is reducable to a form of compatiblism. Molinism reconciles God's providential governence of the world and libertarian freedom and it stands up against Linville's argument.
-------------------------
1 The Consequence argument has many specific and nuanced forms, but the basic picture is: if determinism is true the laws of nature and the past determine everything, including my acts. But I cannot control the past or laws of nature, so I cannot control their consequences, including my acts.
2 The truth of conterfactuals is logically prior to anything anyone actually does.
Comments
That which an agent can do. Another way to say it would be the range of effects and agent can produce or the range of actions an agent can perform.
God be with you,
Dan
I suppose we can ask--what does "can" mean, to you?
If the agent's X-ing is incompatible with the laws of nature and the prior states of the universe in conjunction, then presumably he cannot X. Yet you'd like to say that he can indeed X. How so?
That's not entirely helpful, because it isn't clear to me that, given God's foreknowledge, the agent can produce any effect except that which is foreknown by God.
Well, in everyday conversation we certainly talk about people being able to to this or that so it's unusally you find such an idea objectionable or suspicious.
Hum... Do you think this based on some other argument besides the Hasker one I addressed above, or perhaps do you not like the way I addressed it?
I suppose we can ask--what does "can" mean, to you?
Can just means possible or ability or power or something of the sort. For more I would need a context.
If the agent's X-ing is incompatible with the laws of nature and the prior states of the universe in conjunction, then presumably he cannot X. Yet you'd like to say that he can indeed X. How so?
The agent x-ing is hypothetical. Why wouldn't a hypothetical future be associated with a hypothetical past?
By 'prior states', I assume you mean God's foreknowledge or prior true future tense propositions (i.e. yesterday 'Bob will X today' was true).
So if (hypothetically) Bob does Y, it's unsupprizing that in such a hypothetical future, there's a hypothetical past in which 'Bob will do Y' was true.
Now of course Bob will not do Y, but that doesn't me he cannot, just that he will not. So what Bob will do is compossible with 'prior states'. Thus the abilty to do otherwise is compossible with 'prior states' even if the actual use of that ability is not.
God be with you,
Dan
Maybe our ordinary conversation is misleading. I don't think our use of language is a true and infallible guide to proper understanding of the nature of reality.
I'm not clear on how hypothetical X-ing is anything like an actual ability to X.
There may be worlds where I X, because X-ing is not inconsistent with the pasts states of the universe in conjunction with the laws of nature in that world. But that is all uninteresting to the question of what it is that I can do in the actual world.
It is plain as day that I cannot actualize worlds where the past is different from the past of the actual world. I can't change the past like that. But then it seems like what I can do is actualize a world that is an "extension", as Fischer puts it, of the actual past in conjunction with the laws of nature. But that's only one thing, if God exists. So I'm only able to do one thing.
Hypothetical worlds where I do otherwise are not helpful to me, because at a moment of deliberation, I can't make actual a world where the past is different from the actual past.
There may be worlds where I X, because X-ing is not inconsistent with the pasts states of the universe in conjunction with the laws of nature in that world. But that is all uninteresting to the question of what it is that I can do in the actual world.
I agree. But you asked me about hypothetical X-ing, so I addressed it. In other words, the problems you developed were based on X-ing, not the ability to X. The ability to do otherwise is quite consistent with ‘past states’.
It is plain as day that I cannot actualize worlds where the past is different from the past of the actual world. I can't change the past like that.
The term ‘actualize worlds’ is usually reserved to God’s decision to create one of many possible worlds. In this sense, of course, we cannot actualize worlds. Likewise if by ‘actualize worlds’, you mean cause such and such a world to be the actual world, we cannot ‘actualize worlds’. But if you mean if by actualize worlds you mean, hypothetically, if we were to X, a world consistent with X would be the actual world, then yes, we can (but do not) ‘actualize worlds’.
However, this does not entail the ability to change the past. Arguing that it does overlooks the distinction between causal ability and counterfactual dependence (please carefully consider the difference between CPP and CPPi above and why I might accept CPP but reject CPPi). So my view does not violate the intuitive ‘plain as day’ principle you seem to think it does.
God be with you,
Dan