I surely can't be the first person to think that the free will/determinism debate is rife with scope ambiguity. Consider the natural language claim that: if the world was in state S1 in 1900, and the laws of nature imply that if the world was in state S1 in 1900, then John will do act A at midnight, 1st January 2010, then John must do act A at midnight, 1st January 2010.
Evidently, this claim should be understood so that the must has wide scope. That is to say, this claim should be read as having the form N([[P&Q]->R]) i.e. It is a necessary truth that, if the world was in state S1 in 1900, and the laws of nature imply that if the world was in state S1 in 1900, then John will do act A at midnight, 1st January 2010, then John will do act A at midnight, 1st January 2010.
And yet, everyday I read both students and professional philosophers who misread this claim as having the logical form [[P & Q]-> N(R)] i.e. if the world was in state S1 in 1900, and the laws of nature imply that if the world was in state S1 in 1900, then John will do act A at midnight, 1st January 2010, then it is a necessary truth that John does act A at midnight, 1st January 2010, and hence that it is impossible that John should do otherwise than A at midnight, 1st January 2010.
Anyway, as I was saying, I'm certain I'm not the first the note this scope ambiguity, but is there a particular paper you know that is clear about the need to avoid it, and succeeds in setting up the argument from incompatibilism without it. Answers on a postcard please!
- Geoffrey Ferrari's blog
- Login or register to post comments
Dear Geoffrey, Sorry for
Ryo Chonabayashi on Thu, 06/11/2009 - 12:17pmDear Geoffrey,
Sorry for this late comment, I just noticed your interesting post this lunch time!
This is just for clarification:
Although my knowledge of the debate is really limited, I am surprised by your presentation of the debate. You wrote: if the world was in state S1 in 1900, and the laws of nature imply that if the world was in state S1 in 1900, then John will do act A at midnight, 1st January 2010, then John must do act A at midnight, 1st January 2010.
I am surprised by this because you wrote ‘John will do’. Don’t many writers characterize the debate in terms of ‘causal necessity of one’s actions’? Isn’t this the reason why many writers use the phrase ‘John’s doing A at midnight is a necessary truth’ in a sense that one’s action in question is causally necessitated?
About the paper which clarify the distinction between N((P&Q)→R) and (P&Q)→N(R): I don't really know. Sorry!!
Best,
Ryo
Necessity, Laws and Scope
Geoffrey Ferrari on Sat, 06/13/2009 - 9:08pmHi Ryo,
You're quite right of course that one might include a statement of necessity within the statement of the law, but this really only pushes the question one stage further back. Compare these two views:
I take it that any necessity within a statement of law must take the wide-scope form of the first of these formulations. Thus understood, the premisses of the determinism argument must be:
From these premisses, the best one can hope to show is that the world will be in state s2 at t2, not that the world must necessarily be in state s2 at t2.
'must' and 'will'
Ryo Chonabayashi on Sun, 06/14/2009 - 3:15pmHi Geoffrey,
Thank you for clarification.
I still have got two questions though these questions are again for clarification. I must be missing something...
1. What exactly do you mean when you use the following two expressions, ‘John will do A’ and ‘John must do A’ [also, 'the world will be in S2' and 'the world must be in S2']? I suppose that you mean ‘John’s action A at that time is causally necessitated or determined’ by the expression ‘John must do A’. But I’m not clear what you mean by the expression ‘John will do A’. I thought you wouldn’t mean that John had a will to do A at that time, rather than John is causally necessitated or determined to do A.
2. Given the formula N((P&Q)→R), why we cannot say that it is a necessary truth that if the world is in S1 and the laws of nature imply that if the world is in S1, John must do A at that midnight, then, John must do A at that midnight? Well, your original claim says, ‘if the world was in state S1 in 1900, and the laws of nature imply that if the world was in state S1 in 1900, then John will do act A at midnight’. So, if we follow this claim, obviously, we should say that R means ‘John will do’, not ‘John must do’. But, if we take Q as, ‘if the world is in S1, then John’s act A at that midnight is causally determined’, R would be ‘John’s act A at that midnight is causally determined’. In a similar way, if we take Q as ‘if the world in S1, then John must do A’, R would be ‘John must do A’. I thought the real issue at stake is not where we should put 'N', rather how we characterize Q. I suppose how we take Q is depending on our understanding of causal necessity of actions.
Ryo
'must' and 'will' reply
Geoffrey Ferrari on Tue, 06/16/2009 - 8:31amHi Ryo,
I apologise for any misunderstanding. By 'John will do A' I mean the future tense of the verb 'to do' i.e John will do A iff John does A at some time in the future. I wasn't talking about will in the sense of 'volition' at all.
By 'John must do A' I mean that it is necessary, in some sense of 'necessary', that John does A.
Insofar as I understand your point of view, I think that you are making the very mistake that I wanted to highlight. Unless I misunderstand you, you seem to want to say that given the past and the laws of nature, it is causally necessary or determined that certain events happen in the future. But strictly speaking, that is false! In the past, let us suppose that the world was contingently in state S1 in the past. What is causally necessary is only the claim that if the world was in state S1 in the past, then the world will be in state S2 in the future. And the only conclusion one can derive from this is that the world will be in state S2 in the future, not that it is causally necessary that the world will be in state S2 in the future.
So far, I may simply be repeating myself. So let me put things another way. Let us say that the 'causally accessible' possible worlds are all and only those which have the same laws of nature as the actual world. In all causally accessible worlds, it is true that if the world is in state S1 at t1, then the world is in state S2 at t2. Now, when we say that the world is in fact in state S1 at t1, we are not saying that the world is in state S1 in all causally accessible worlds. We are saying that the world is in state S1 in the actual world. In other causally accessible worlds, the world may not be in state S1, because, there are possible worlds which have the same causal laws as the actual world but in which there is a different quantity and arrangement of matter and energy. And that is why this argument will only show that the world will be in state S2 at t2, not that it must be. Because if we say that the world must be in state S2 at t2, we are saying that the world is in state S2 at t2 in all causally accessible worlds. But our argument does not show that. Indeed, it is consistent with everything that we have said that there is a causally accessible world, one in which if the world is in state S1 at t1 then it is in state S2 at t2, but in which it is not the case that the world is in state S2 at t2. How is that possible? Simply because that world was also not in state S1 at t1. In such a world, the conditional claim that if the world is in state S1 at t1 then it is in state S2 at t2 is still true, because its antecedent is false.
Apologies for spelling this out in excruciating detail, but it seemed like the only way to prevent misunderstandings.
Hi Geoffrey, Thank you for
Ryo Chonabayashi on Tue, 06/16/2009 - 9:34amHi Geoffrey,
Thank you for correcting my misunderstanding. Sorry for bothering you again and again!
Ryo