Blog Coordinator

« Overturning the Hierarchy | Main | Thanks to John Martin Fischer! »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Hi John,

Interesting post! And interesting set of issues. Thanks for posting on this!

As you know, I'm a fan of these sorts of 'initial design' arguments.

You ask: "Is it really the case that because God had certain intentions in creating the world 5 billion years ago, a particular agent cannot act freely now and be morally responsible?"

My own answer is: No. I don't think it would be *because* God had those intentions that a particular agent could not now act freely and responsibly. That, I think, would be implausible, for the sorts of reasons you mention. Rather, it would be *because* of the *deterministic causes* these intentions produce that the agent isn't free or responsible.

This is what I think (and what I think the defender of the Zygote Argument, and similar such arguments, ought to think). Suppose you see Jones do something bad. Everything seems 'normal' in the relevant ways. You think: he's blameworthy. But wait: now you find out that everything Jones ever did was part of God's "script" for his life, as it were -- for everything Jones did, God had intentionally put in place causes sufficient to bring it about the he did those very things. Now, what I think is this: this is enough to *show* that Jones isn't free and responsible. Intuitively, once I find this out, I'm no longer inclined to think Jones can really be fairly blamed for what he's done.

Now, this is *not* to say that "because God had those intentions [which are sufficient to bring about the relevant outcomes] a particular agent could not now act freely". Rather, it's to say that God's having had those intentions (which were sufficient in the given way) is good *evidence*, so to speak, that Jones couldn't be free or responsible. For how could freedom and responsibility be consistent with one's entire life being the pre-determined outcome of someone else's plan? *That*, I think, seems implausible. But what would *make* Jones unfree, in this scenario, if anything does, are not God's intentions per se, but, again, the deterministic causes they require (or bring about).

I think the core issue is just this: once you find out that everything Jones did was the causally determined result of God's plan for his life, are you still inclined to think that he can fairly be blamed for doing those things? For me, I've just always had the powerful intuition that he can't be. But I don't imagine that you've got this intuition!


Some responses:

1. Of course, the proponents of the Initial Design Argument (IDA) agree with you that whether the being has intentions makes no difference. After all, these proponents generalize from the design scenario to the situation of mere determinism! And in the general case of mere determinism, those intentions are not posited. So of course the proponents of the IDA agree with you that the answer be the same in both cases: free will, or no free will.

2. Your post speaks in terms of "acting freely" or being "morally responsible." Let's use the specific term, "free will." What is free will? Where did the term come from? As far as I understand, one important aspect of free will's origin was a response to the specific threat that you outline! This is the "Free Will Defense"! From this perspective, it should be no surprise that God, in the scenario you outline, would undermine free will. As a semi-compatibilist, you can probably agree with me about this?

3. I think you're right to distinguish between IDA type arguments, and manipulation arguments more generally. The IDA more clearly invokes concerns/questions about personal identity. As I've explained earlier, the problem of free will is really a problem of constitutive luck. So I think that manipulation arguments (like Pereboom's 4 case argument) only work to the extent that they raise concerns about constitutive luck. In that case, they work effectively in the same way as the IDA.

It seems to me that creation arguments are supposed to work by providing us with help in seeing something that is there anyway, with or without the creation, given certain other assumed facts (e.g., determinism). The idea, I take it, is that determinism precludes responsibility, and we can see this fact more readily if we consider a deterministic creation story. The idea, as I understand it, ISN'T that the creation itself (with the creator's intention) makes a difference to whether the created agents are responsible. It is supposed just to make more readily discernible their lack of responsibility.

I think of its role as like that of staining samples in biology. There are, for example, cell structures that can't be seen unless the sample is stained. Adding the stain doesn't add the structures, and removing the stain doesn't itself remove them. The role is an epistemic one. So, it seems to me, is the role that creation (and the creator's intention) is supposed to play in these creation scenarios.

Whether it DOES allow us to see something about determinism is a question I leave open. My suggestion concerns only the role it's supposed to play.

There is a serious worry about initial design arguments that is parallel to worries about standard manipulation cases that I think people need to take more seriously. Randy suggests they are like a stain that leaves the underlying phenomenon unchanged, but let’s you see its fine structure more clearly. I think they ALTER the phenomenon in important ways.

In order for initial design arguments to work, cases involving design (‘Design Cases’) must be matched to cases of normal free and morally responsible agency (‘Normal cases’) with respect to whatever factors compatibilists think are important for freedom and MR. But I doubt the matching is successful. There is a *compatibilist factor* that is (subtly) altered between Normal cases and Design cases, thus defeating the argument. Here is what I mean:

Patrick provides the case of Normal Jones who does something bad (and there are no compatibilist-defeaters present in the case). In the Designed Jones case, God designed a world a million years ago and intended for Jones to do just this deed. In response to the Normal Jones case, we intuitively think Jones is free and MR. In the Designed Jones case, we intuitively think Jones is NOT free and MR, or is less free and MR. Why the difference in intuitions? Well, says Patrick and other incompatibilists, the Designed Jones case provides evidence for determinism, which in turn undermines freedom.

But here is another plausible explanation for why our intuitions differ: If a creature can be designed thousands of years ahead of time to execute someone else’s prior intentions with such incredible fidelity, that creature is surely simple, clunky, stimulus-bound, reflexive, cartoonishly mechanical (feel free to insert your favorite pejorative descriptions here). If we understand the cases *this* way, the factor that differentiates Normal and Design involves conditions for *compatibilist* freedom and MR. So the argument no longer provides any evidence for incompatibilism.

The ready incompatibilist response is to stipulate that Designed Jones is not a simple clod, but rather he is a sophisticated agent who enjoys all the rich and varied factors that support compatibilist freedom and MR: reasons-responsiveness, mesh between first- and higher-order attitudes, his actions express his deep self, not deceived or normatively ignorant/incompetent, and so on. But once you build all of that in, and if you *really* spend some time *carefully* envisioning the case, there is simply no longer an intuition that Designed Jones fails to be free and MR.

In short, either you get the ‘simpleton’ reading of the Design case, which generates the relevant intuition, but does NOT support incompatibilism. OR you get the correct reading of the Design case, in which case nothing intuitively follows.

Of course, the issue isn't whether there's a certain intuition or not; it's whether Designed Jones is morally responsible. And apparently we disagree among each other about that.

Might it help, at this point, to ask of those on each side: What do you take being responsible to amount to?

Is this something like how you see the above argument working, John?

Call it "Fischer's Three Case (Initial Design) Argument":

Case 1: Suppose that five billion years ago Fred--an independently existing being--sees God create the universe, and God tells him that he has specific intentions about each individual--specific intentions that each individual behave as he actually does throughout his life. Nevertheless, after creation, it is "hands off" for God--no more direct interventions. And God also endows at least some of his human creatures with all of the properties typically thought to ensure acting freely--mechanism ownership, reasons-responsivenss, or whatever one's favorite freedom-conferring properties may be.

Case 2: Suppose that Barney (Fred's friend) let's him know that in fact the creating individual he saw was not God, but God's brother, Schmod. It turns out that Schmod had NO intentions about future creatures, including humans. But otherwise everything else is exactly the same as in Case 1--down to the physics and psychology of the situation in which our agent acts from his own, moderately reasons-responsive mechanism.

Case 3: Suppose that the universe was created five billion years ago by some confluence of non-agential forces (i.e., the Big Bang). But otherwise everything else is exactly the same as in Case 1--down to the physics and psychology of the situation in which our agent acts from his own, moderately reasons-responsive mechanism.

Now it seems plausible that, in Case 3, someone acting from a moderately reasons-responsive mechanism is responsible for what she does. And it does not seem plausible that adding the creation story in Case 2 makes a difference to this judgment. Thus, the mere presence of determinism and design do not undermine responsibility. So it does not seem plausible that God's intentions in Case 1 undermine responsibility in virtue of pointing to the presence of determinism.

This is compatible with Patrick's intuition--once you find out that everything Jones did was the causally determined result of God's plan for his life you no longer think he can be fairly blamed for what he does. The presence of God's intentions in Case 1 may undermine responsibility in that case, but not because they point to some underlying fact (i.e., determinism). There may be something else about them that threatens responsibility.

Is this something like what you have in mind, John? Do others find it persuasive?

I think Chandra is right, and I made this sort of case here:
(and in a paper I presented last year at Scott Sehon's conference), but I'll try to make the basic point here:

1) Philosophers may think they are able to get pure intuitions about these cases, but why think we are not subject to the same psychological processes that likely influence ordinary people's intuitions about such cases (and that can be tested, as Chandra has done and I'm trying to do)?

Humans evolved to be agent detectors and manipulation avoiders. There is every reason to think we will respond to cases of intentional design/manipulation of one agent by another differently than we respond to cases of (deterministic) causation by natural forces. The different responses are likely to be driven by lots of factors, including our difficulty in holding fixed the sophisticated compatibilist capacities of the manipulated or designed agent (as Chandra suggests and as I develop in that post).

2) But if one wants a more principled (philosophical) explanation for the different responses to these cases, I think a plausible argument can be made that an intentional designer or manipulator diminishes the designed/manipulated agent's opportunities in ways that determinism does not. Put simply (but hopefully better in a paper someday), a determined agent has at least some type of access to many nearby possible worlds in which she makes different choices--worlds with very slightly different pasts, perhaps differences that are not even discernible at the macro level (or laws, but yuk to the latter option). But a designed/manipulated agent (e.g., Ernie) does not. Intuitively, such cases suggest we hold fixed the designer's goals, in which case the nearby worlds are ones where Ernie either still does what designer wants (even if his compatibilist capacities are bypassed) or where Ernie does not exist (because the designer created someone else to achieve his goals). I think we would have the same sort of intuitions about different possibilities (nearby worlds) regarding a stipulated deterministic process (e.g., die rolls) and the same process designed by a powerful agent who wants a particular outcome (e.g., roll) to occur.

If one starts dropping out the designer's power or goals, these intuitions should diminish (though this may not "jump to mind" because of our psychology--point 1). The upshot: we should be soft-line compatibilists about intentional design or manipulation arguments, but hard(er)-line as the cases are cunningly altered to drop the crucial feature that triggers our intuitions.

Thanks John!

There may be a freedom relevant difference between intending that Wilma do A and intending that she A and then arranging things so that A is done by her. "Hands off" after the initial creative act may mean that all the necessary arrangements for her A-ing have been made, that they were somehow implicit in God's fiat. In that case, the tremendous lapse of time between that act and her A-ing would not nullify the charge of manipulation. On the other hand, let's suppose that God merely intends that Wilma A, doing nothing to bring it about. Is this even possible: that is, for God to intend without being a manipulator? This is the crux of an earlier dispute I had with Mr. Werking: is it possible for GOD to intend that I A without being responsible for my A-ing? He said at the time something like "The connection is a little too tight in this case": it's not, as I was suggesting, like me intending that my son do his homework.

Chandra and Eddy:

You argue that, if we have the intuition that designed agents lack free will (or responsibility), it is only because we are thinking of the agents as stupid/simple/reflective robots. And you argue that, if only we imagined the agents as having the full range of compatibilist powers, then we would agree that the agent has free will.

In response: it is not the agent's being stupid/simple/reflexive that creates the intuition of lacking freedom. It is the focus on the preordained/preexisting/fated/destined aspect of the story.

You can see this by focusing on an agent that we know satisfied your preferred compatibilist criteria. Imagine, say, John F. Kennedy. Everyone agrees that John F. Kennedy satisfied the compatibilist criteria of your choice. Now imagine, additionally, the following:

A. A godlike being, Diana, exists.
B. Diana created this world and made it deterministic.
C. Because the world is deterministic, Diana could predict exactly what would happen 4 billion years later, based on the initial conditions of the world.
D. Diana, rather perversely, desired for JFK's life to play out exactly as it did. The election, Marilyn Monroe, the Bay of Pigs, the Cuban Missile Crisis, the Dallas drive leading to his death...
E. Diana set the initial conditions of the universe so that JFK's life would play out exactly as she desired. It was guaranteed that JFK would live that exact life, and no other life.
F. Diana has a special TV, with a slider along the time axis, that allows someone to watch JFK's entire life, before she creates this universe. And these observers can move the slider, and watch every moment of JFK's life, like a tape, before it even existed.

From this perspective, I think it is clear that: 1. we start to lose the intuition that JFK had free will (or responsibility) and 2. this is not because we think of JFK as stupid or reflective. It is because, instead, of the fact that JFK's life was fated, predetermined, and would never have been different... has been brought into vivid focus.

Eddy offers a second argument about "nearby possible" worlds. I completely disagree. This disagreement reflects a deep misunderstanding about "possible worlds" semantics on either my part, or Eddy's.

The designed agent does not have more "nearby possible worlds" than the merely determined agent in any way relevant to the free will debate. That is why philosophers like John emphasize that the designer is "hands off." (See above.)

Let me give you an example to make a point. Consider the Game of Life. You can play the game here (need JAVA installed):

The Game of Life world is analogous to the deterministic world. In the design scenario, someone like Diana sets the initial conditions so that a particular life story plays out (like JFK's life). In the merely deterministic scenario, something else does... randomly. But in both cases, the agent has equal access to "nearby possible worlds" - the agent's entire life is fixed according to those initial conditions, and it doesn't matter who/what sets them.

@Randy: “Of course, the issue isn't whether there's a certain intuition or not; it's whether Designed Jones is morally responsible.”

I disagree. Intuitions matter! I suspect there are serious differences in our respective theories of/approaches to philosophical theory construction that need to be addressed.

My view is that intuitions (defined roughly as spontaneous, direct, quasi-perceptual intellectual seemings) serve as a *distinctive* and *independent* kind of evidence in philosophical theorizing. It counts in favor of a philosophical theory when it accounts for our intuitions in hypothetical cases; it is a cost when the theory does not capture our intuitions. Of course, intuitions are not the only kinds of evidence and they certainly are not dispositive. Arguments certainly matter and other factors (simplicity, elegance…) as well. But intuitions make an *independent* contribution along side these other methods of supporting a philosophical theory. This picture may or may not be right, but it is certainly widespread in analytic philosophy. It is basically a restatement of the reflective equilibrium picture that many (including John) endorse. So if there is a disagreement here, it would be useful for you or others to say exactly where you might disagree. But if you grant this picture, then the arguments that Eddy and I have adduced matter. Here is why:

The Initial Design Argument relies at least in part on intuitions about cases such as Design Jones. Our intuitions that Jones is not free/MR is taken by incompatibilists to support their views. I have tried to argue that the intuition exists only to the extent that that it tracks subtle compatibilist factors that creep into Design cases. Remove these factors and the intuition--that is, the spontaneous intellectual seeming--that Jones is unfree/not MR disappears. I grant our reflective theory about MR may still count him as unfree/MR, but the point is that the *intuition* disappears. Eddy agrees and adds his own candidate compatibilist factors that the intuition in fact responds to, the removal of which makes the intuition disappear. If these claims are right, then the status of the intuition as evidence for incompatibilist views is defused. More specifically, the role of the intuition as a putative *independent* source of evidence for incompatibilism is seriously weakened. This certainly does not settle the matter. There are a lot of other kinds of evidence we all might marshal in favor of our respective views of moral responsibility. But something has been achieved. Defusing the intuition in Design cases (i.e., showing this intuition responds to compatibilist factors) is progress, albeit incremental, for the compatibilist.

"Is it really plausible to suppose that, whereas when Fred thought that God created the universe, he should think that the agent couldn't be morally responsible, but now that he knows that it was merely his pesky brother, Schmod, everything is cool?"

It seems as if we might be at a dialectical stalemate. As Kip points out above, why are you construing this as an argument against IDAs as opposed to compatibilism?

There are no relevant differences between the two cases but that's a point against compatibilism - not against IDAs.

Chandra and Eddy, you raise some really important and hard questions about how to do philosophy. I don't have thoroughly worked out views on this, but here are a few thoughts.

I certainly don't think that the nature of quasi-perceptual intellectual seemings is what is at issue in the moral responsibility debate. Sometimes our judgments do and ought to go against how things seem to us (think about some cases of visual illusion). The question isn't whether Designed Jones SEEMS responsible; the question is whether he is.

"The Initial Design Argument relies at least in part on intuitions about cases" -- An argument, in the sense employed in philosophy, is an abstract structure of propositions. Intuitions, I suppose, are psychological events are states. Things of the first sort don't depend in any way I can see on things of the second sort.

Perhaps the idea is that we must rely on intuition to judge one or another of the premises of the argument. But (taking intuition to be what Chandra suggested) we don't; as I pointed out, judgment sometimes does and ought to diverge from how things seem. We do have to use our cognitive faculties in order to make such a judgment; but this isn't news.

Here's a way of understanding reflective equilibrium. Consider a set of propositions P, Q, R,... about some philosophically important subject matter. Some of them are propositions we affirm when we consider particular cases. Some are general principles we affirm. (For wide reflective equilibrium, we add propositions we affirm in our broader view of the world, not just those concerning the subject matter in question.) Now we seek coherence among these propositions, which consists of not just logical consistency, but also mutual support, explanation, etc. To achieve this we might abandon some of the propositions and add some new ones.

In this process, we give a certain weight to propositions we affirm when we consider particular cases. I wouldn't say that we do so because we take cognitive seemings to be evidence. I'd say we exhibit a certain degree of trust in our ability to make judgments. We might find reason to have less trust in that ability in certain limited spheres, but thoroughgoing distrust would undermine intellectual inquiry altogether.

I gather you think you've shown that judgments about the responsibility of agents in deterministic settings (but only the incompatibilist judgments?) are not to be trusted. I'd like to hear more about how you think you've shown this.

(By the way, I don't myself concur in such judgments; I'm not an incompatibilist about responsibility. But I'm picky about arguments, even when they're offered in support of a view I accept.)

Here's a kind of argument that wouldn't cast doubt on propositions affirmed by judgment of particular cases:

"Some individuals (who have been selected for their lack of philosophical sophistication) judge that P on the basis of one or another stupid confusion. Therefore (probably) philosophers who judge that P do so on the basis of one or another of these same stupid confusions. Therefore (probably) not P."

I don't think it helps this bad argument to add:

"When I disabuse myself of these stupid confusions, I judge that not-P."

I'm certain this bad argument ISN'T your argument. But we might make progress if you would spell out more clearly how your argument concerning Designed Jones goes.

Chandra: for what it's worth (and I don't suppose it's worth too much), when I--a hopelessly recalcitrant incompatibilist, to be sure--keep firmly in mind that the designed agent has all the relevant compatibilist features, I have no inclination to think the agent is free or morally responsible. Those features, to borrow a line, aren't robust enough to ground moral responsibility. But I suppose I can see how someone with compatibilist sympathies might think they are sufficient and thus might think that the designed agent could be free and morally responsible. What I can't find plausible is that the deterministic design has no effect on the agent's freedom or responsibility, which is why I have suggested that (shameless self-promotion ahead) compatibilists should concede that being deterministically caused by factors beyond one's control can diminish an agent's freedom and responsibility even though it may not eliminate these things completely. This is mitigating soft compatibilism, and in my all-too-biased opinion it is the best form of compatibilism on the market.

'preordained/preexisting/fated/destined aspect of the story'

The acts in question simply did not exist at the time of the universe's creation, so that description is off the table. As for the others, the question is, does 'in the cards mean,' no freedom on the part of those playing the hands? Yes, they were going to occur- does that mean there was no sense in which they were freely performed, that they were not up to the agents in question- does it imply that they were merely puppets on God's strings? The answer is far from obvious. Here we have something akin to the problem of Freedom and Foreknowledge, which lots of philosophers have thought to have solved.

'D. Diana, rather perversely, desired for JFK's life to play out exactly as it did. The election, Marilyn Monroe, the Bay of Pigs, the Cuban Missile Crisis, the Dallas drive leading to his death...

E. Diana set the initial conditions of the universe so that JFK's life would play out exactly as she desired. It was guaranteed that JFK would live that exact life, and no other life.'

If E, then there is obviously no freedom at at all.

If D, then yes there is freedom unless D is taken to imply E. Can an omniscient/omnipotent being desire something without insuring that it comes to pass. I.e., is D tantamount to E?, which was Professor Fischer's original question.

This is a very interesting thought-experiment with great comments, and thanks to all and especially to John for this.

The example's assumption that at least some humans would be endowed with compatibilist "responsible-apt" features seems to me to be the sticking point. Intuitions break one way or another on whether this assumption is seen as decisive. Incompatibilists fixed on the deterministic metaphysics also assumed here think that that point is decisive, and (semi-)compatibilists think it is not--something else really matters. And this is really how it should be given the history of the two traditions. Does this constitute dialectical stalemate? Not necessarily I believe: it only locates where intuition diverges. I'll argue that the compatibilist (of some stripe) is better off (by some measure of "better").

One factor that favors compatibilism and is typically overshadowed by the metaphysics usually emphasized in these kinds of examples is a *pragmatic* moral concern. Let's set aside for the purposes of argument moral nihilism and the like. Then there is a posited need to ground ordinary moral practices like assigning responsibility. If we add in specific metaphysical concerns as incompatibilists do at this point, then questions about assigning responibility are potentially endlessly derailed. But then consider some possible pragmatic points about getting derailed here: we might never have good evidence about the truth of the determinism of human nature, we might never have good reasons to think God/Schmod exists, etc, etc. So could we then think that it is simply just more workable to focus on compatibilist concerns since those usually rely on epistemically accessible properties of people involved in (our presupposed) moral activity? I think this is a leg-up for at least certain forms of (semi-)compatibilism. They are doable (though I think Fischer, Watson, Wolf and some others are more accessible epistemically than, say, Frankfurt's--behaviorally demonstrated criteria trump merely psychological ones pragmatically, and cut down on the "mind-reading" of assessing responsibility).

From the pragmatic perspective--abetted by rejection of the inscrutable metaphysical problems thrust on compatibilism from the incompatibilist side--some form of compatibilism rules. So I'd say. Further, that puts an onus on incompatibilist naysayers: why is the pragmatic perspective wrong here?

Hi Randy,

Here's a way of taking on board some of the message of Chandra's argument without putting intuitions at the centre of philosophical methodology or even at the centre or any specially privileged place in the IDA. There's an enormous amount of evidence that people have different argumentative standards depending on whether they agree with the conclusion of the argument or not. That is, if I am antecedently convinced that p, other things equal I require rigour or less plausible premises before I think that an argument that p is convincing. This effect is stronger for people with less training in argument, but the effect is still there even for experts within their domain of expertise. There is every reason to think that this kind of bias will be triggered by finding 'p' intuitive. Now that does not entail that we can't come to think that p is false, though intuitive. But it does counsel an epistemic humility. And it does suggest that intuitions matter.

@Randy, your expanded remarks are very helpful, and it is clear to me now that my metaphilosophical views hardly differ from yours. I take intuitions to be a species of judgment, and didn’t mean to suggest otherwise. That said, there is one place where we DO disagree. You say “I gather you think you've shown that judgments about the responsibility of agents in deterministic settings (but only the incompatibilist judgments?) are not to be trusted.” No, this is not my view. I think I can present the structure of my argument more clearly by invoking the idea of ‘minimal pairs’. Here is a go at it (sorry this is so long)…

Philosophers often present paired cases, such as Normal Jones and Design Jones, as a way of supporting their favored theory. But intuitive judgments about these cases can only serve to support one theory (e.g., incompatibilism) over another when the cases are minimal pairs: The cases must be matched in all relevant respects but for the single difference of interest. So for example one couldn’t put forward Normal Jones versus Design Jones*, where Design Jones* was designed by Diana millions of years ago, AND he has an irresistible impulse to do what Diana intended him to do.

It is OBVIOUS that Normal Jones and Design Jones* fail to be minimal pairs. Sometimes it is not so obvious. Here is a great example from Ronald Butler in 1977 that illustrates this (Thomas has a great paper on this case):

"If Brown in an ordinary game of dice hopes to throw a six and does so, we do not say that he threw the six intentionally. On the other hand if Brown puts one cartridge into a six-chambered revolver, spins the chamber as he aims it at Smith and pulls the trigger hoping to kill Smith, we would say if he succeeded that he had killed Smith intentionally. How can this be so, since the probability of the desired result is the same?"

Call Butler’s first scenario ‘Dice’ and his second scenario ‘Revolver’. Are they minimal pairs with respect to moral factors? That is, is the ONLY relevant difference between the two cases that the action is morally repugnant in ‘Revolver’ but not in ‘Dice’? It is not at all obvious. My own view is that there is a second hidden/implicit difference (a non-moral difference) between Dice and Revolver that actually accounts for our differing intuitive judgments.

Notice that in making this point, I am NOT arguing our intuitions in these cases are untrustworthy. I am saying this: If Dice and Revolver were minimal pairs with respect to moral factors, then our intuitions to the cases should support some theory of intentional action with moral factors as a component, and indeed many philosophers have argued just this. But once we recognize the cases are not minimal pairs and there are other hidden/implicit differences between them, then the intuitions no longer support this ‘moral view’ of intentional action. The intuitions themselves remain trustworthy (i.e., the intuitions that the agent intentionally phi’s in Revolver but does not intentionally phi in Dice). When we find out the cases are not minimal pairs, what changes is not the trustworthiness of the intuitions, but rather the philosophical theory the intuitions should be taken to support.

I am taking a very similar approach to Normal Jones and Design Jones. I claim they are not minimal pairs because there are subtle compatibilist defeaters that are present in Design Jones. This does not represent a ‘stupid confusion’ as you put it ---rather tendentiously I might add :) :) --- any more than initially taking Dice and Revolver to be minimal pairs represents a stupid confusion. If Normal Jones and Design Jones are not minimal pairs as I contend, our intuitions in Normal Jones and Design Jones remain just as trustworthy as before; what changes is that the intuitions can no longer be taken to support incompatibilism.

Both Eddy and Chandra seem to be suggesting that there's a morally relevant difference between designed Jones and normal Jones. I don't see it. Can someone help me see what it is more clearly? Why aren't the two minimal pairs?

Chandra: Suppose it specified, in presenting the design argument, that Designed Jones has when he acts (on the occasion in question) features that various compatibilists take to suffice for being morally responsible for what one does. (When Mele presents the zygote argument, he specifies this sort of thing. I present a similar argument in my dissertation, and in doing so I specify this sort of thing. This is the standard way of presenting this kind of argument.) What I characterized as a stupid confusion (OK, let's call it silly) was reading this kind of presentation and then--right away--taking it that Designed Jones lacks the very features that it's explicitly specified he has.

I'd read some of your earlier comments as saying that those who judge Designed Jones to lack responsibility are guilty of this kind of confusion. I apologize for the misunderstanding.

I'm puzzled now why you say Designed Jones and Normal Jones aren't matched in relevant respects. The point of specifying the factors I mentioned is to make it out that they are.

@Justin, we are not saying there is a MORALLY-relevant difference. Just that there are relevant differences (i.e., non-moral) basically of the kind that are proposed in soft-line responses to standard manipulation cases.

Let me flip your question around. Why believe that Normal Jones and Design Jones ARE minimal pairs? I think some philosophers may operate with a kind ‘transparency assumption' that says we can simply LOOK at the text of two cases and instantly decide whether they are minimal pairs with respect to some philosophically-important feature. ‘Dice’ and ‘Revolver’ are supposed to remind us that the transparency assumption fails. Most everyone tends to think that the cases differ only in terms of the moral valence of the action. Thomas (in his The Butler Problem Revisited paper) identifies at least a second hidden feature that differs between the cases. A wonderful paper by Kim Davies from 1981 identifies a third feature – the supporting intention in Revolver is ‘global’ in a way that implies control, which is lacking in ‘Dice’. I have a Deep Self interpretation of the cases that either builds on Davies, or might be considered a 4th difference. My point: Given the text of two hypothetical cases, it is not at all transparent whether they are minimal pairs.

Given the strangeness of Initial Design cases and how difficult they are to fully cognize, I think proponents of the IDA should have to say much more to defend the claim that cases like Normal Jones and Design Jones are indeed minimal pairs. The argument I hear from Kip (i.e., his John Kennedy slider case) and Justin is: ‘When I vividly imagine Initial Design cases and try to mentally carefully remove any compatibilist soft-line defeaters from the scenario, the intuition that the agent is not free and not MR persists’. I am not convinced by arguments like this. Indeed this is precisely the argumentative form that Randy describes as a “bad argument” in the final section of his previous comment.

I wish I had more time now to participate in this very interesting discussion. Patrick Todd and I seem to interpret my zygote argument in the same way. I might add that I see the argument as something that can help make some commitments of compatibilism vivid to people who aren't immersed in the literature. As I said in *Free Will and Luck*, I certainly wasn't aiming to persuade experienced compatibilists with the argument, and I myself am not persuaded by it. Also, as I pointed out, committed incompatibilists didn't need any persuading. Still, the argument has already prompted some very good articles and discussion, and that may continue for a while. By the way, in connection with the vividness bit, I think Randy Clarke's "stain" image is very nice.

Justin: Of course I can't speak for Eddy and Chanda - they are quite capable of doing that themselves. But one might think that we should not try to identify differences and similarities across cases just from the inside, as it were. Eddy claims (plausibly) that there are features of the design case that might be expected to trigger different mental modules to the ones triggered by the non-design case. The difference is in us, rather than in the case. Compare cheater detection versions of the Wason selection task to formally equivalent versions which are not in the realm of social contract. They are formally equivalent, but there is good reason to think that we treat them as different because we process them in quite different ways.

Is it really the case that because God had certain intentions in creating the world 5 billion years ago, a particular agent cannot act freely now and be morally responsible? We assume he acts from his own, appropriately reasons-responsive mechanism. Why doesn't he act freely? Why isn't he morally responsible? How can someone's--even God's--intentions 5 billion years ago, with no further intervention, affect the suite of factors that intuitively are relevant to the agent's moral responsibilty now? Doesn't that set of factors include the physics and psychology of the situation, as it were, and not the intentions of some distal agent--5 billion years ago?

I'd break it into two questions. The first is, taking determinism to be the case, and say the god thingie can really compute future events well - well then it can just determine events 5 billion years into the future.

The other question is to 'simply' question why such a god, in it's intentions in how it sets up things, is somehow moral (by ones individual perspective) in regard to how they set it up.

Otherwise the question's difficulty comes from deriving that this god thingies intentions are ones which somehow already cover moral responsibility (and then taking it he can set out the future as it determines, even 5 billion years ahead - which funnily enough, isn't as questionable a claim)

@Randy, I agree that Initial Design cases typically come with an omnibus catch-all clause ‘The agents are alike in all other respects’. But is it realistic to rely on these catch-all clauses to ensure proper matching between the relevant cases? For example, what if I supplement ‘Dice’ and ‘Revolver’ with a catch-all clause that says ‘but for the differing moral valences of their respective actions, the cases are alike in all other philosophically-relevant respects.’ I suspect the catch-all would do no work at all. Intuitions in the two cases would continue to differ, and despite the catch-all, subtle *non-moral* differences in the cases of the kind identified by Thomas, Davies, and me would still account for the differing intuitions.

Why is it that catch-all clauses are not very effective at generating proper matching? I don’t have a good theory. The explanation offered by Neil (who in turn credits Eddy) seem on the right track. I don’t have a worked out theory to offer.

Neil speaks well for me (and himself)! Randy's intriguing stain analogy might be apt if the cognition involved in assessing responsibility were as seemingly simple as visual perception, but I suspect it is more complex. To the extent that it is relatively simple, I would guess this is the sort of cognition we evolved: when something bad happens, look for any agents who caused it to blame and punish the crap out of. A more complex rule might then develop: look for the agent most in charge (the chief)--that's the agent who has the power to ensure the bad thing happened; taking out that agent is most likely to prevent more bad stuff from happening in the future. An even more sophisticated rule would involve mitigating responsibility for agents who could not resist the causal force of the chief.

These basic cognitive mechanisms are being triggered in manipulation and design cases. The idea that they are not influencing us as we consider such cases is implausible. But before someone (e.g., Kip) suggests that this is evidence that our (compatibilist) responsibility judgments are cognitive illusions, or (e.g., Randy) that some people are making stupid or silly mistakes when they consider these cases, it'd be nice to hear what cognitive mechanisms are reliable in these debates and why *they* are the ones that suggest incompatibilist judgments.

Again, I think there are principled (reliable) reasons to trust judgments that distinguish deterministic James from designed James. For instance, I think we are pretty good at intuitively picking out the nearest possible worlds (i.e., the most viable alternative possibilities, both past and future). And I think we are properly picking out different nearby possible worlds for det-James and designed-James, as I suggested earlier. Powerful manipulators or designers (chiefs!) are able to ensure that James will do what they want by causing him to do it. Their knowledge and power is vast enough to cover their bets--i.e., to ensure James does what they want in nearby possible worlds too.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan