Hume’s Slip

I just noticed that Paul Russell’s contribution “Hume on Free Will” is published at the Stanford Encyclopedia of Philosophy (it was actually published 12/14/07, but I only read it now, perhaps because it was not free earlier).  As a big fan of Paul’s work, having both read his book on Hume and seen his wonderful lecture at Inland 2006, I had been waiting to read this article for years.  I was eager to read it because (i) I’m fascinated by the free will problem, (ii) Hume is my favorite historical philosopher, and (iii) his writings on the subject of free will always confused me, if only because he seemed to endorse compatibilism (gasp).

When I first started reading Paul’s work about Hume and free will, I couldn’t believe that Hume was a full blown compatibilist.  In particular, I wondered whether Hume’s true views about free will might have been kept somewhat hidden, like his views about religion, until after his death.  So I did some digging, and actually exchanged some lengthy emails with Paul about the topic, focusing on Hume’s later works.

To set the stage, let me say, I think Mele’s Zygote Argument is a great formulation of the classic dilemma facing any would-be compatibilist.  “You think you can have free will in a deterministic world?  Well, what if God designed your entire life, just like Shakespeare writing the life story of characters in his plays?  Surely *that’s* not free will?”  And, I genuinely believe, or at least suspect, that concerns like this motivate almost all incompatibilists.  So, in confronting Hume’s alleged compatibilism, I immediately focused in on this question: what does Hume think about God, predestination and responsibility for the story of the universe?  What would Hume say about the Zygote Argument?

Now, if my hunch about Hume was *entirely* wrong, Hume should have said, like most compatibilists (or quasi-compatibilists) today (e.g. Watson, Dennett, McKenna, Frankfurt---ignoring, for the moment, the subtleties of their views), “so what?  God designed my life?  Who cares?  I still satisfy the compatibilist requirements of your choice.  I still have free will.  I don’t see any tension here at all.”

But Hume, bless his heart, did *not* say that.  Here is Russell on Hume’s slip:

“It is evident that Hume is arguing two points. First, if God is the creator of the world and preordained and predetermined everything that happens in it, then the (obvious) existence of moral evil is attributable to him, and thus “we must retract the attribute of perfection” which we ascribe to him. Second, if God is indeed the ultimate author of moral evil, then no individual human being is accountable for the criminal actions he performs. The second claim does not follow from the first. Moreover, it is clearly inconsistent with Hume's general position on this subject. As has been noted, in this same context, Hume has also argued that no speculative philosophical theory can alter the natural workings of our moral sentiments. The supposition that God is the “ultimate author” of all that takes place in the world will not, on this view of things, change our natural disposition to praise or blame our fellow human beings. Whatever the ultimate causes of a person's character and conduct, it will (inevitably) arouse a sentiment of praise or blame in other humans who contemplate it. This remains the case even if we suppose that God also deserves blame for the “moral turpitude” we find in the world. In general, then, Hume's first formulation of the second alternative (i.e. that God must share the blame for those crimes that occur in the world) is more consistent with his naturalistic principles.”  (This quote actually first appeared in Russell’s original book Freedom and Moral Sentiment).

Now, I’ll be honest: my first reaction when I first found this quote is wanting to jump up and down and shout, “see, hard compatibilists, hard compatibilism is *too* hard!  Not even Hume can say it with a straight face!  Not even Hume can be consistent about his compatibilism, because it results in absurdities.”  If he *really* believed what he was saying about compatibilism, he could never make that sort of slip.  If he really believed what he said about compatibilism, that sort of slip would be as random, unnatural, and inappropriate as the Pope saying God is dead.  No, he made that slip because it revealed and betrayed part of Hume’s true feelings on the subject (whether he recognized them as such or not).

It would have been fascinating to confront Hume about this slip and let him reach some reflective conclusion.  Of course, that is impossible.

For example, Hume’s remark above leaves open two possibilities: (i) that he was a soft compatibilism, who distinguished between “God designing our entire life” and “mere” determinism, and (ii) that Hume was, at least partly, an incompatibilist.  A compatibilist unwilling to just outright ignore Hume’s slip might want to resort to (i).  However, soft compatibilism strikes me as very unpopular today, and for good reason.  Most everyone seems able to see that there is no relevant difference, with respect to a person’s free will and moral responsibility, whether that person was created through life design or “mere determinism” (I believe Mele calls this, or a similar principle, the “no-difference” principle).  Unfortunately, this only leaves (ii): Hume having (latent) incompatibilist sympathies.

All we can do is speculate.  Gardeners, do you think Hume, free from the religious and social pressures of his time, if confronted about his slip, would:

1.    retract it;
2.    embrace soft compatibilism;
3.    embrace “selective compatibilism” (Paul Russell’s own recommendation, which is that other people can hold the designed agent responsible, but not the cosmic designer); or
4.    repent and embrace incompatibilism (perhaps distinguishing between “free will” and liberty and responsibility)?

Furthermore, why do you think Hume slipped?  Here are two possibilities:

1.    Hume was so against religion that he got carried away and was willing to say anything that tarnished religion, even if it was inconsistent with the rest of his view about free will.
2.    Compatibilism is false.  Even Hume’s vigorous defense of it can’t avoid such telling slips, and these slips betray the falsity (or perhaps, as Smilansky would say, the “shallowness”) of the view.

Comments

[Neil]

Werking's Slip....

what if God designed your entire life, just like Shakespeare writing the life story of characters in his plays? Surely *that’s* not free will?”

But of course that's not (even) analogous to the zygote argument: Shakespeare didn't design his characters to murder and so forth, and then let them loose. He wrote each act. And it is uncontroversial that if my mental states are irrelevant regarding what I do, I am not responsible.

Second, the zygote argument isn't a good analogy to causal determinism, because it is intentionally created. The obvious response is to ask "what difference does intentional creation make? Now suppose that the zygote comes into existence by random forces, what difference does that make with regard to the MR of the agent (who, after all, acts in just the same way, and may have exactly the same mental states)". Now that seems right to me. But there are powerful order effects here: if I approach the question from the other end, first asking about the MR of an agent who is moderately reasons-responsive, and only then turning to the zygote argument, I can find myself sympathetic to a hard compatibilist line on the zygote argument.

[Brandon Watson]

I don't actually think Hume has committed a slip here. He does argue two points, and the first is much as Russell suggests. But the second is explicitly on the supposition that God has "infinite perfection," is "infinitely wise," and "who can intend nothing but what is good and laudable," and thus it is not surprising that it doesn't follow from the first point. Neither is actually intended to represent Hume's naturalistic principles; it's explicitly a dilemma that Hume thinks his religious opponents will raise against him ("I can foresee other objections" and following) . He responds to the second point on Russell's list (actually the one Hume lists first) by giving the account of responsibility Russell identifies; and he responds to Russell's first (Hume's second) by shrugging and saying nobody knows how to handle it (thus implying that he has no particular need to do so to answer his opponents, given that they can't answer it themselves). So I think Hume is entirely consistent here; his account of responsibility throughout is in terms of approbation arising from our disposition to praise and blame. He is considering an objection by opponents, however.

But despite my disagreement here I strongly second the approval of Paul's work; far and away the best that has been published on the subject.

[Kip Werking]

Neil,

I don't understand your objection. You say that "But of course that's not (even) analogous to the zygote argument: Shakespeare didn't design his characters to murder and so forth, and then let them loose." I was only comparing the design part to Shakespeare's writing, not the "letting loose." I agree that Shakespeare never instantiated any of his characters so that they could carry out his story in real life. But that doesn't seem relevant to me, their fates seem just as circumscribed in either case.

Brandon:

The apparent inconsistency is this: Hume's account of responsibility, in "terms of approbation arising from our disposition to praise and blame" (as you say, for example), never mentions a requirement that the agent's life was not cosmically designed. Everything about Hume's account asserts or suggests that the cosmic designer is irrelevant, because even cosmically designed agents can satisfy all of the conditions for liberty and responsibility Hume discusses. And yet (here's the inconsistency), he says that such agents are not accountable for their actions, only the deity is.

[Neil]

Kip,

I don't mean that Shakespeare didn't *actualise* his characters. I mean that Shakespeare didn't give them a set of internal dispositions and then out them in an environment in which they would then act as he predicted. That's what the zygote thought experiment involves. For each act that Shakespeare's characters perform, Shakespeare is the proximate cause of that act. When Hamlet stabs Polonius, he does so because Shakespeare writes that Hamlet stabs Polonius. But no compatibilist thinks that if perform an action, and that I perform that action because some other agent is the proximate cause of that action (for instance, they move me like a puppet) I am responsible for that action.

[Brandon Watson]

Kip, you say,

Everything that Hume's account asserts or suggests about the cosmic designer is irrelevant, because even cosmically designed agents can satisfy all of the conditions for liberty and responsibility Hume discusses.

On the last part of this, I would agree, and I think, actually, Hume explicitly says this in this very context (Selby-Bigge 102; i.e., the penultimate paragraph of the section). But this is in response to what is asserted or suggested about the cosmic designer, which is not part of Hume's account, but an objection he foresees arising in response to certain aspects of his account of liberty and necessity. The whole discussion begins with the following sentences:

"I pretend not to have obviated or removed all objections to this theory, with regard to necessity and liberty. I can foresee other objections, derived from topics which have not here been treated of. It may be said, for instance..."

and then it begins talking about the the "Author of all our volitions" and so forth. Thus all the talk about the cosmic designer is talk being attributed to his critics who reject Hume's account because they think it leads to impiety on the one hand and absurdity on the other. Thus Hume sees the claim that 'cosmically designed agents are not accountable' not as a result of his account but as something that his critics might well try to argue is a result of his account.

[Kip Werking]

Brandon,

You're right. My understanding was based on my reading of Paul Russell's quote, above, which gave (at least) me the impression that Hume making an odd statement, inconsistent with the rest of his view. Reviewing the source materials, I see that he was just paraphrasing hypothetical arguments against his view. Paul has confirmed this understanding in an email.

So I guess (much to my chagrin) his slip wasn't much of a slip at all. But, in the interest of salvaging this thread, I still ask the Gardeners which is preferable:

1. soft compatibilism
2. hard compatibilism
3. selective compatibilism (Russell's own recommendation)?

[Joe Campbell]

I’m not sure that the Hume passages in question – at the end of section 8 of the Enquiry (46 ff.)* -- are talking about the manipulation argument. I don't think there is a slip either.

Hume considers “other objections” to his views on liberty and necessity that are “derived from lines of thought that I haven’t here discussed.” [That is, they are based on principles which Hume would not endorse.] The main objection involves the hypothesis that God is “the original cause of everything” and has two parts (46).

(1) If human actions can be traced up by a necessary chain to God, they can never be criminal; on account of the infinite perfection of the being from whom they are derived, and who can intend only what is altogether good and praiseworthy (46).

(2) If they are criminal, we must conclude that God isn’t perfect after all, and must accept that he is the ultimate author of guilt and moral wickedness in all his creatures (46).

Notice that in part (1) Hume claims that human actions that are traced back to God “can never be criminal” BECAUSE God has “infinite perfection” and can only do what is “good and praiseworthy.” The point is not that if God were the cause of everything, then human actions would not be free. The point is more that if God is the cause of everything, then why is there evil? One explanation is that all evil is apparent. This strikes me as a discussion of a version of the causal problem of evil.

* Page numbers refer to Jonathan Bennett’s Early Modern Philosophy, http://www.earlymoderntexts.com/pdfbits/he2.pdf

[Kip Werking]

Joe:

My remarks were based on the following quote, which Paul Russell noted is inconsistent with Hume's entire view:

"So we have to conclude either that those actions are not criminal, or that
God and not man is accountable for them." Let's call the big block quote containing this sentence A.

Hume goes on to paraphrase these arguments:

Paragraph B: "This objection consists of two parts, which I shall examine separately. (1) If human actions can be traced up by a necessary chain to God, they can never be criminal; on account of the infinite perfection of the being from whom they are derived, and who can intend only what is altogether good and praiseworthy. (2) If they are criminal, we must conclude that God isn’t perfect after all, and must accept that he is the ultimate author of guilt and moral wickedness in all his creatures."

He paraphrases them again, in the respective discussions beginning with the paragraphs starting "(1)" and "(2)", on p. 46 and 47 of your linked text, respectively (let's call these C1 and C2).

What's very important, and quite subtle (at least to me), is that although A says "God and not man is accountable for them", Hume's later paraphrases in B and C2 only concern themselves with the problem of God's possible guilt, and not humanity's possible innocence.

That is, the quotation in A clearly contemplates that if God is accountable, we're not. But B and C leave it quite open that God is guilty, and so our we. Hume seems to avoid the issue of humanity's innocence, perhaps because he recognized the possible inconsistency with his larger view.

Thus, when you write:

"The point is not that if God were the cause of everything, then human actions would not be free. The point is more that if God is the cause of everything, then why is there evil?"

This only seems to address the first prong of Hume's critics' argument, as discussed in A, and the bullet point (1) in B and C1. The second prong, as described in A, leaves open the (impious) interpretation that when humans commit wrongs, God and not the human beings are accountable for them.

What's also very interesting to note is that, although Hume believes that the objection in C1 is mistaken, as he explains at length, he believes that he, and philosophers in general, are powerless to answer the objection discussed in C2 (leaving his critics with his impious conclusion) ("The second objection can’t be answered so easily or satisfactorily: it isn’t possible to explain clearly how God can be the ultimate cause of all the actions of men without being the author of sin and moral wickedness.")

Thus, my current position on Hume's alleged slip is this: although the statement regarding God's guilt and humanity's innocence in A is attributable to Hume's critics, and not himself, Hume regards at least the conclusion about God's guilt as being unanswerable, and therefore one can fairly impute it to him.

However, Hume paraphrase's his critics as objecting that "God, and not men, are accountable", when he discusses this same objection in C2, he only considers God's guilt, and does not explicitly endorse humanity's innocence. Further, doing so would have been inconsistent with his larger view on agency (as Paul Russell notes), thus one can only fairly attribute the second claim, about humanity's innocence, to Hume, to the extent that his discussion of argument (ii) in B and C2 is meant to mirror the argument given in A, which explicitly says "God and not man is accountable."

Personally, I think this extent is significant. I don't know whether Hume's silence regarding humanity's innocence in B and C was (i) inadvertent or (ii) carefully done, to avoid raising trouble for his larger project. But Hume's possible slip is a possibility worth considering.

I'll end this oversized comment with two more remarks:

1. Actually, on first blush, the two pronged statement in B above seems to be a false dichotomy (if they are intended to be mutually exclusive): I don't any reason why all of the acts couldn't be good, in the Liebnizian sense, while it is *also* the case that only God, and not man, is accountable for these acts. That is, the question of whether God, and not man, is accountable seems orthogonal to the question of whether the acts are "criminal".

2. The Zygote Argument seems to be unduly narrow: it only contemplates Diana ensuring that Ernie makes a single decision. But, if she could design one choice, she could surely design all of them, and design his entire life story. This is more along the lines of what Watson considers in his Soft Libertarianism and Hard Compatibilism. Perhaps this should be called the "Super Zygote Argument"?

[Joe Campbell]

Kip,

Thanks for your comments.

I should start by saying that this section on Hume is difficult to understand and differing minds will disagree. As you note Paul Russell seems to understand the passages in the same way that you do. And if Hume had intended it to be understood that way, it would raise the kinds of questions that you ask: is Hume a soft compatibilist or a hard compatibilist? What would Hume say about the manipulation argument?

I'm not going to answer those questions. I'm just going to repeat my claims that you are reading more into the passages than is strictly speaking there. Note that Paul does so in part to motivate an interesting puzzle about manipulation, and to shed light on the differences between soft and hard compatibilist responses to the manipulation argument. This is all good and I don't have a problem with Paul's passages on this subject.

But as you understand (1), it suggests that "if God is accountable, we're not." But note first that Hume does not say that we are not accountable for our actions in this case. Rather he says if God is the cause of everything, our actions "can never be criminal." His comments apply only to wrong actions.

The reason for this -- for the fact that if God is the cause of everything, our actions can never be criminal -- is then given by Hume: "on account of the infinite perfection of the being from whom they are derived, and who can intend only what is altogether good and praiseworthy."

God can only do what's good. If he is the cause of everything, then nothing can be criminal. Considerations of manipulation do not come into play.

What about passage (2)?

"(2) If they are criminal, we must conclude that God isn’t perfect after all, and must accept that he is the ultimate author of guilt and moral wickedness in all his creatures."

This problem cannot be dismissed so quickly, as you note. Hume adds: "it isn’t possible to explain clearly how God can be the ultimate cause of all the actions of men without being the author of sin and moral wickedness."

Note that what can't be explained is how God can be the ultimate cause without being the author of sin.

You think that the conclusion to draw from these passages is a "statement regarding God's guilt and humanity's innocence" but so far Hume has said nothing about our innocence. He is only talking about God's guilt.

I do not see, in any of these passages, a worry about whether human beings are responsible for their actions if God is the cause of everything. Of course, Hume did not believe the antecedent. Thus, these are not real problems for him. That said, Hume would have no difficulty thinking that both God and human beings were responsible for their actions, were God the cause of everything. Again, interesting questions arise!

[Kip Werking]

Joe:

I want reiterate that I'm grateful for professional philosophers like you and John taking the time to address my comments.

And I agree the passage is difficult. I haven't even had much time to review it, so it remains quite possible that my interpretation is flawed.

I want to say, though, about your recent post: you focus on the statement:

"(2) If they are criminal, we must conclude that God isn’t perfect after all, and must accept that he is the ultimate author of guilt and moral wickedness in all his creatures."

This quotation is from what I call passage B.

As my previous comment tries (however inartfully) to show: the quotation in B corresponds to the following quotation in what I call passage A:

"So we have to conclude either that those actions are not criminal, or that
God and not man is accountable for them."

And the corresponding discussion of this problem, and why Hume thinks it's unresolvable, is found in passage C2.

Here are some points about the above:

1. Passages A and B are similar in that they both seem to limit themselves to considering "criminal" conduct.

2. But passages A and B differ in that, while you're right to note that B says nothing about humanity's innocence, passage A does.

3. So, to what extent are the arguments for (2) supposed to be the same in passages A, B and C? Well, that's the million dollar question.

I think it's clear that these passages or supposed to correspond to each other. First, in A, Hume paraphrases two arguments his critics might make. Then, in B, he briefly summarizes them again. And then, in C1 and C2, he addresses each argument in turn.

Thus, it remains open, I think, whether the differences between A (which mentions humanity's innocence) and B and C2 (which focus only on God's guilt) are inadvertent, or intentional. I think it's at least plausible that Hume regarded the claim about humanity's innocence in A as true, or unanswerable, but did not mention it because of:

1. inadvertence
2. concern about his contemporaries and readers reactions
3. concern about its apparent inconsistency with his larger view

But I also think it's possible that:

4. The claim about humanity's innocence in A is something that Hume would only attribute to his critics, even if he agrees with them that his views reaches the impious conclusion about God's guilt

So, when you say: "I do not see, in any of these passages, a worry about whether human beings are responsible for their actions if God is the cause of everything", the answer (as far as I can tell) is that the passage in found in A, and is supposed to correspond to (2) in B and C2. But nobody can know for certain, right?

[Mark Smeltzer]

Kip,

(A) says either (P) or (Q), right?

Q is something his critics would never desire to affirm (viz., that God is the author of moral evil and is accountable for all evil), and thus Hume is motivating his critics to affirm (P) (viz., that there really aren't any criminal actions). Since his critics will find Q logically impossible, Hume is doing his best to motivating them to affirm (P) (which is something they will still be very relucant to do).

Hume also invites them to attempt denying Q by explain why 1) God can guiltlessly be the ultimate cause of human crimes and 2) humans should be held accountable for their criminal actions. However, Hume states that this proposal seems impossible to defend in the quote that Joe provided:

... it isn’t possible to explain clearly how God can be the ultimate cause of all the actions of men without being the author of sin and moral wickedness.

In other words, your interpretation seems to stake a lot on the significance of P and Q in Hume's own thought, where in reality P and Q are only siginificant to Hume insofar as they represent tools he can use to motivate his critics to accept his Hard Compatibilism.

[Kip Werking]

Mark,

I reread the Liberty and Necessity section of the Enquiry again, and I think we read the section very differently.

For one, you first say that "Hume is motivating his critics to affirm (P) (viz., that there really aren't any criminal actions)". But Hume's entire discussion of this problem is directed to convincing them that human actions can be criminal, and elicit outrage from victims, regardless of philosophical meditations.

"Your interpretation seems to stake a lot on the significance of P and Q in Hume's own thought, where in reality P and Q are only siginificant to Hume insofar as they represent tools he can use to motivate his critics to accept his Hard Compatibilism."

You are right that I'm using Hume's discussion of his critics' views to pry into Hume's own views on this topic, and that the extent to which this can properly be done is limited (as many have pointed out in this thread).

But Hume's discussion of Q is not directed to persuading religious critics to adopt his views. Instead, he seems to be doing the opposite. The impression Hume's discussion of Q gives is:

"Oh, by the way, religious person, even though I've argued so fervently that my view of freedom is compatible, and necessary, for morality, I haven't removed all of the objections to this claim. I don't agree with you that P, that all human actions aren't criminal. That's just wrong. But this other argument of yours, Q, that God would be the author of all sin... there's something to that argument! I think you may be right about that."

Now, what allows Hume to say this? I can suggest at least:

1. Candor
2. He doesn't believe in a conventional (Christian) God, and so worries about conflicts between Christianity and his own views don't make him lose sleep at night.

[Mark Smeltzer]

Kip,

Would you at least agree that Hume is presenting a dual-pronged reducio to his critics?

[Kip Werking]

Mark,

I don't quite agree with that: my reading is that Hume is paraphrasing two reductios that his critics pose to him, rather than vice versa (see, for example, the last two sentences of the paragraph beginning "I pretend not to have obviated...").

Further, while Hume dismisses the first reductio as largely without merit, he regards the second as unanswerable.