Is mere design disturbing? Should it be?

Consider the following situation.

Suppose a cosmic being sets up the initial conditions and/or laws of a deterministic world so that your entire life is predetermined.  The being knows exactly what you will do, before you do it, before you are born, and sets everything up so that you do exactly that.  The being could be God, but need not be divine.  A very powerful alien or computer would suffice.

Consider further: the being could even tell you he designed your life story.  He could say “I defined every detail of your life!  There’s nothing you can do that I didn’t predict.”  The being could write down, on a card, a prediction of what you will do, every five minutes, and then, after five minutes go by, show you the card.  And the being would always be right.  Each and every card would perfectly predict your life story.  The being could even write all of them before you are born.  This could really aggravate you, and there would be nothing you could do to stop him.

To me, the above could not be more violent to free will.

  To me, free will is the opposite of the above.  To me, the fact that free will is in conflict with the above is not an artifact or after-effect of free will’s nature.  It is, rather, a basic, essential aspect of what “free will” means in the first place.  So if someone thinks that free will is compatible with the above, then they are just confused about what “free will” means.

People see free will as protecting them from the above, because the above possibility (rightly or wrongly) disturbs them.  In the above scenario, it seems like we are just puppets and God (or whoever) is the puppetmaster.  Sure, we are fancy puppets.  We have fancy compatibilist bells and whistles.  But none of those bells and whistles prevents someone from designing and predicting our entire life story, like Shakespeare writing a play.

So, imagine my surprise when I met compatibilists who believe that the above is perfectly consistent with free will.  The following quote from Harry Frankfurt is typical, although I think that, if pressed, virtually every compatibilist scholar today would defend a similar view:

A manipulator may succeed, through his interventions, in providing a person not merely with particular feelings and thoughts but with a new character. That person is then morally responsible for the choices and the conduct to which having this character leads. We are inevitably fashioned and sustained, after all, by circumstances over which we have no control. The causes to which we are subject may also change us radically, without thereby bringing it about that we are not morally responsible agents. It is irrelevant whether those causes are operating by virtue of the natural forces that shape our environment or whether they operate through the deliberate manipulative designs of other human agents (2002, p. 27).

What shocks me about these statements is that the compatibilist doesn’t seem to recognize the danger identified above.  All of the incompatibilist’s worries and fears about being a puppet, being controlled, being ultimately out of control, seem to leave the compatibilist cold.

I couldn’t understand how the compatibilist could be so blind, it seems, to that danger.  But then it occurred to me: maybe the compatibilist is distinguishing between control and mere design.

Return to the deterministic scenario.  If God sets up the initial conditions so that your life follows his predetermined story, is he really controlling you?  Or just designing you?

I think that’s a fundamental distinction that has to be explored: control versus mere design.  Everyone can agree that control is disturbing.  But is design disturbing?  Should it be?

My own thinking is that people tend to find them both disturbing and that they should find them both disturbing.  The reason is that, everything a Bad Guy wants to accomplish with control, he can also accomplish by design.

Suppose a Bad Guy wants Joe to rob a bank on January 13, 2019.  One thing he can do is wait until January 13, 2019 and then intervene to control ordinary Joe to rob the bank.

But the Bad Guy, if he has amazing powers, can also just design Joe from scratch to rob the bank on January 13, 2019.

The compatibilist will reply: but only the control involves bypassing.  Ordinary Joe was not a criminal, then you made him a criminal.  So, of course, you undermined his free will.  In contrast, designed Joe was bad from the start.  You didn’t bypass anything, you just had him express his criminal character.

That response strikes me, and most incompatibilists, as entirely unsatisfactory.  But I’m not entirely satisfied, yet, with my articulation of why I’m not convinced.

Gardeners, do you think that the compatibilist distinction between control and mere design works?  Why or why not?  Do you think design should be as disturbing as control?  Why or why not?

Comments

[John Fischer]

I admit: I do not find mere design, qua design, problematic or disturbing. Now if you add to design that the designer causes (or even deterministically causes), this is also not eo ipso disturbing, because not all causal sequences are created equal. It all depends on the kind of causal sequence in question, on my view; what does not matter, in my opinion, is whether there is an agent with specific intentions at the beginning of the causal sequence.

I developed this sort of view in my review in MIND of Al Mele's book, especially Mele's example of the Goddess Diana.

[Grant Rozeboom]

One question ought to be: In what way is the state of affairs you describe the "opposite" of free will? I guess I'm not sure what you think the opposite of being designed and/or controlled by an alien is (other than _not_ being designed and/or controlled by an alien). Surely it need not be true that, for us to exercise free will, we somehow design ourselves! So, the next question is: If there is pressure created by your case, why is this sort of design -- agent induced or agent-controlled design -- more anti-free than regular old cultural, parental, and biological (CPB) design?

A natural answer might be that, in the super alien case, _all_ of your actions are preset, in some sense, whereas in garden-variety CPB design, it doesn't seem to be true that all of our actions are predetermined by our design. But it seems to me that this is just as much an epistemological fact as it is a metaphysical one. That is, what might be intuitively unnerving about the alien case is that we know, with near certainty, that all of our actions are predetermined. But when we are just CPB designed, we have no way of knowing if/how 100% of our actions are predetermined.

And so we are led to a third set of questions: What if the alien was only right 70% of the time? And what if social scientists and biologists became really good at predicting the actions of agents on the basis of their CPB design -- say, they were 70% accurate? Would it then be true that we were equally free (or unfree) in both the alien-design and CPB design cases? If so, then it is not the metaphysical fact of design that is lessening our freedom. Or do you think it must be the case that the designer and the predictor are the same entity in order for freedom to be lessened? In that case, CPB design, even if scientists could use it to predict 100% of our actions, would not be sufficient for lessening our freedom. In either case, it is not the metaphysical of design, by itself, that is sufficient for counteracting free willing.

[Jonathan Farrell]

This comment is intended to say something about which notions of design we are (or perhaps should be) talking about here. I don't say anything about the distinction between design and control but despite this, what I say does seem to me to be on topic. (So although it might look like I'm changing the subject/hijacking the thread I don't think I am, and I hope that I'm not (and apologise if I am!)).

I think Grant is right to point out the importance of epistemic factors in these sorts of cases. But I wonder if the real work (in encouraging people to say that designed agents are not free) isn't being done by another element: whether the designed person is designed with some purpose in mind.

Here are four cases, each of which involve an agent being designed by a Super Intelligent Designer, Sid.

Case 1. Sid designs A by creating the zygote that develops into A from atoms of carbon, hydrogen, and all the rest that are lying around Sid's cosmic lab. Sid makes sure that the zygote and the womb-like environment it develops in are such that the zygote will develop into a normal human being. When A is born Sid gives her to some parents who bring her up. Sid has Laplacian Demon-like abilities to predict what will befall A (and everything else in the universe): he can know the state of the entire universe (including the state of the zygote) at the moment A's zygote is created; and he can make perfect and instantaneous calculations as to all future states of the entire universe. However, Sid isn't interested in all that future stuff, so never does the calculations. A goes on to have a normal life, involving lots of actions that she apparently freely performs, including one instance when she robs a bank.

Case 2. Sid designs B, much as in case 1, except that here Sid does do all the predictions (but only after the zygote has been designed, A's life begun, and when it is impossible for Sid to interfere on what unfolds). As with A, B lives a normal life involving many apparently free actions, including one bank robbery. This time, however, Sid knows that (and exactly how) each of these actions will occur.

Case 3. Sid designs C much as he designed A and B, except that this time Sid has a plan. By using his Laplacian skills, he designs C's zygote so that, at a particular time, C will rob a bank. C goes on to live a normal life that appears to involve lots of free action. In particular, it involves C appearing (at least to those who do not know about Sid) to freely rob a bank at the appointed hour.

Case 4: Sid designs D as he does A, B and C. As in case 3, Sid has a plan, and designs D so that (Sid hopes) D will perform the bank robbing at some specified time. In this case, however, Sid has a nap before he starts on designing D. He dreams that has he performed the Laplacian calcuations and so knows exactly which features he must give the zygote in order that D will grow up to rob the bank: he must place some particular base pair in a particular place on the zygote's DNA. On waking, Sid doesn't realise that his memory of doing the calculations is only a memory of a dream, so he designs a zygote with just those features mentioned in the dream. (The other features he assigns at random as far as this is compatible with the resulting zygote developing into a normal human). As it happens, the features that Sid dreamt of are completely irrelevant to what occurs in the future - whatever base pair Sid had put in that place on the zygote's DNA, the future would have been exactly the same. As it turns out, D robs the bank, and in a way that, to those ignorant of Sid's role, appears to be free.

All four cases involve designed agents (and none involve directly-controlled agents). Only cases 2 and 3 involve the perfect prediction of the relevant agent robbing a bank. And only 3 and 4 involve the agent being designed _so as_ to rob the bank. For what it's worth, here are my responses to the four cases, and some suggestions as to why I have these responses.

I'm not tempted to say that the bank robbing in case 1 is unfree, so it doesn't look as if it is mere design that is doing the intuition pumping. Nor am I tempted to say that the robbing in case 2 is unfree, so it doesn't seem to be mere design plus prediction that is doing the work for me. I'm happy to allow that the robbing in case 4 is free, so an agent's being designed to perform an action that they then perform doesn't seem relevant to whether or not the action is free. In case 3, however, when the agent is designed to perform the action and they perform it because of the way they were designed, I do feel the pull that might lead someone to deny that this robbing is freely performed by the agent. I suspect it is not a coincidence that it is only in case 3 that I might say that Sid performed (or played a part in performing) the action of robbing the bank (through C). So perhaps what's at work here is the intuition that (at least in cases like this) if one agent (Sid) freely performed the bank robbery, no other agent (such as C) can have freely performed it. Another relevant features is that, whether or not C is controlled by Sid in case 3, it certainly seems that C is used. Perhaps we are reluctant to attribute free agency to agents when their actions result from their being used by other agents.

[John Fischer]

Jonathan,

Thanks for your thoughtful post. You say:

In case 3, however, when the agent is designed to perform the action and they perform it because of the way they were designed, I do feel the pull that might lead someone to deny that this robbing is freely performed by the agent.

I grant that many people will have the intuition that the agent is not acting freely or morally responsible in such a case. But, again, I do not share this view. The way I look at it, we all might fit the description you give, since it is after all possible that a providential God created the universe. After all, he might have created this work along the lines of the Molinist picture, knowing in advance what every creature would freely do in every context, and creating this world because overall it is the best package, as it were.

Of course, I do not necessarily wish to defend Molinism, or any particular account of God or providence. I just want to point out that it might well be just as you describe it, and yet we don't hesitate to make responsibility attributions in the actual world. Additionally, I don't think we should stop making such attributions, upon reflection on the possibility that a providential God created us. That, in itself, doesn't make me think we are not morally responsible agents.

[Kip]

Thanks for your responses, Grant and Jonathan. Let me address some of your points in turn.

Grant, you ask why I think the above is the opposite of free will. Here are some answers:

1. I come to the free will debate with the pre-philosophical intuition or understanding that un-predictability is a basic or almost-essential aspect of what "free will" means. This is an aspect of free will that seems to be almost entirely ignored by compatibilists.
2. I believe that the historical roots of free will support this understanding: free will seems to be the historical reaction to the idea of determinism and the idea of a Laplacian god-figure predetermining our lives. In response to the threat of determinism, or to a Laplacian god-figure, people would say "yes, that would be a problem, except we have free will, so that situation doesn't apply to us." That is, people largely thought of themselves as exceptions to the otherwise deterministic ordering of the world, thereby overcoming the dangers/problems associated with that ordering, and they called that power "free will."

Surely it need not be true that, for us to exercise free will, we somehow design ourselves!

On the contrary, I think free will may, directly or indirectly, imply that we do design ourselves. That is, more or less, what Galen Strawson's argument implies.

And the reason is very clear: if we don't design ourselves, then someone else did (or random chance did), and that's really disturbing. Everything we do follows from our design. So if our design was out of our control, then our entire lives seem, ultimately, out of our control. And people hate the idea of being out of control.

So, the next question is: If there is pressure created by your case, why is this sort of design -- agent induced or agent-controlled design -- more anti-free than regular old cultural, parental, and biological (CPB) design?

I don't think it is. Like Mele's Zygote Argument, we would use a no-difference principle to say that the design scenario is no different than mere determinism, random chance or "regular old cultural/parental/biological design." The conclusion, therefore, is that we don't have free will, regardless of whether there was a real designer.

Jonathan,

I'm very glad that you agree that there is an intuition, in case 3, that free will has been undermined. I think this intuition is common. I think it is the same intuition behind the TNR and beta principles in the Consequence and Basic Argument. I think it is the same intuition that incompatibilists refer to, when they say that incompatibilism is the default or prephilosophical view.

[I understand that, when you remove the designer, compatibilists can get the folk to give compatibilist answers. I believe that this is because that designer makes clear freedom-undermining factors that are present in both situations, and not because the designer adds something extra in conflict with free will. As compatibilists agree, the designer doesn't undermine compatibilist free will, so the folk shouldn't give compatibilist answers in either the designed or no-design cases. But they do and they will in the design case.]

You guess as to why we feel that Case 3 undermines free will. You suggest that it is because the agent is being used.

My suggestion is this: people don't want to be controlled and they see free will as protecting them from control. But design is just another type of control. It doesn't matter whether the designer starts from your zygote, and sets you on a path to rob a bank, or whether the designer intervenes on the bank robbery day, and literally forces your hands to rob the bank. Either way, he is controlling you.

In the design case, you have an intermediate kind of control, a proximate kind of control that is consistent with the designer's meta-control. Your intermediate control fits with the designer's larger control and over-arching plan.

But that doesn't satisfy people. That just means that you are a sophisticated puppet, not that you aren't a puppet.

Free will was not meant to, and is not understood as, providing you with just intermediate control as you live out a life that someone else wrote for you. Free will is understand as the power to originally define your own life, and your own path, regardless of what anyone says. It involves a kind of ultimate freedom, which is immune to the design cases that I describe above. That's why the design cases give us the distinct impression that agent doesn't have free will.

[Paul Torek]

"The being could write down, on a card, a prediction of what you will do, every five minutes, and then, after five minutes go by, show you the card."

I wrote a story, at my website, where the characters are confronted with just this sort of prediction. I try to take the sting out of it.

[Justin Coates]

Kip, what do you think about a case like this?

Suppose indeterminism is true. Now suppose (pace the Ockhamist) that God cannot know (or infallibly predict) what you will freely do. Further suppose that God allows you to have a few self-forming actions (in Kane's sense) that set your character. Well, now that your character is reasonably well formed, God--being omnipotent--can manipulate the circumstances that you find yourself in for the rest of your life. Given the facts about your character produced by the SFAs, along with God's deductive abilities, God can now predict everything you will do for the rest of your life. Now imagine that God can do this on the basis of a few SFAs that happened before you turn 10. So every action you perform from the age of 10 until your death is manipulated in this way, and therefore, predicted by God (he could even write them down on cards or whatever).

Does this type of predictability worry you?

[Kip]

Justin,

I do not begin to see how any kind of indeterminism helps. I am hardcore on that point. Because of that, views like Kane's strike me as very strange. I think that, at bottom, all libertarians want (or should want) the magical, impossible kind of free will that G. Strawson describes, and that it is misguided for any libertarian to think that indeterminism can secure that. To the extent that they offer something different, I agree with Gary Watson and Richard Double that they are avoiding the real issue and offering a watered-down version of free will.

I need to also note the following: I don't necessarily think that design ought to be as disturbing as some people think (although I do find it disturbing). I agree with compatibilists that there is a sense in which it isn't disturbing. I just disagree with compatibilists that this enlightened sense is how free will is defined. In my view, people defined free will to be a magical or libertarian kind of power, to escape determinism-types of threats. Whether they were right to regard determinism as threatening or not is irrelevant, the definition they made of free will remains the same. Any definition of free will that is based on a more enlightened and reflective notion of how we can still be free and responsible, despite determinism, is so revisionist in my view that it no longer qualifies as "free will."

In view of the above, let me answer your question: no, I don't think the SFAs would help, because I don't think any form or number of SFAs, or indeterminism, will ever help. I'm adamant about that. But I'm also not sure that design is as disturbing as we initially think it is. It's disturbing in the sense that it conflicts with an important part of our self-image, which is that we are the exclusive originators of our life stories, so that nobody else could have written our life story. But it's not disturbing in the sense that design harms us or bypasses our intentions and desires.

In other words, I think the compatibilist's exclusive focus on bypassing ignores another danger that determinism poses: the danger to our self-image as exclusive originators of our life stories.

[Kip]

John,

Thanks for your responses. Let me ask you a few brief questions:

1. You understand free will as an umbrella term. Would you, along those lines, agree that the design scenario is incompatible with a common notion of free will, even if it is compatible with another notion of free will (tied to moral responsibility)?

2. Do you believe, as I do, that an important pre-theoretic or pre-scientific part of our self-image, is that we are the exclusive originators of our life stories, so that another person couldn't write our life story? You talk about the importance of "writing our own life stories" with respect to moral responsibility in your own work. I think that the origination intuition is common and powerful, and helps explain the incompatibilist's insistent that TNR and Beta-like principles are true, as well as the emphasis of skeptics like Honderich and Strawson on origination and ultimacy.

3. If you agree that it is part of our self-image, do you agree that it's part of what free will is defined as securing, or intended to secure?

4. If you and Eddy and others don't find design per se disturbing, why would so many folk find it disturbing? Are they just confused about the design, confusing it with manipulation or intervention? Or do you think that my theory above, about design conflicting with our self-image as exclusive originators of our life stories, helps to explain why it disturbs them?

[John Fischer]

Ok, Kip, here are the answers to my test:

1. If the designer also causally determines the agent's activity, then the agent may not have freedom to do otherwise (regulative control), although he may still act freely (exhibit guidance control).

2. When I act freely, I write sentences in the story of my life. I'm not sure what the "exclusive originator" means--I doubt if I buy into that, interpreted in certain ways, whereas I probably accept it, interpreted more reasonably (!)

3. Not sure. But I do think that when I act freely, I "initiate" my behavior or am the "source" of it; I just don't believe that anything compels us to interpret these notions "strictly" (so as to imply indeterminism).

4. I agree that the design issue worries many people, and that they are not confused. It is a good way of distinguishing those of us who are "natural compatibilists" from those of us who aren't. I think that reasonable people can certainly disagree about such matters, and I think the design worry brings out the incompatibilist intuition nicely.

Ok, what grade did I get?

[Jonathan Farrell]

John,

I agree with you that the kind of design present in case 3 is compatible with C freely robbing the bank. I just wanted to register that, despite this, I do feel the pull towards denying this (and this is not so with regard to cases 1, 2, and 4).

Kip,

You say that "people don't want to be controlled." I think that this is usually true. But there could be a least two reasons why people don't want to be controlled. It might be that people just don't want to be controlled, regardless of what else is the case. Or it might be that people don't want to be controlled because, when you are controlled, it follows that you do not control yourself; and it might be that it is not controling themselves that people don't really want.

I'd deny that design generally entails control. Seem my remarks about my cases 1, 2, and 4 above. (Also, presumably someone designed my shoes, but they don't control them (I do).) But lets assume for the sake of argument that to be a designed-for agent like C in my case 3 is thereby to be controlled by one's designer. Does it follow from this that the agent does not control themselves? If the answer is 'no,' and if lacking control is what people don't want, then case 3 looks like a situation people don't want to be in. If the answer is 'yes,' then whether it is lacking control or being controlled that people don't want, case 3 looks like a situation people don't want to be in. But, whichever way we answer the question, I think we need some arguments for why that's the right way to answer it. (For the record, I think it's lacking control that we don't want, and, as my reply to John above suggests, I think the answer to the question is 'no.')

To put the point in terms of your puppet analogy: people don't want to be like puppets. But it's not clear that designed-for agents like C are like puppets in the relevant ways (i.e. have the features that puppets have that are responsible for people not wanting to be like puppets).

[Kip]

John,

Ok, I am enjoying this perverse reversal of roles. (In asking you questions, I was only trying to stir up some conversation in this thread).

First, let me explain what I mean by the "exclusive originator" of our life stories. What I mean is this: we create our life stories, and nobody else. Nobody has the power to say how my life will go, because only I have that power. Another way to think of this is, "the buck stops here." I control my life, but a designer doesn't control (or design) my life, because the buck stops with me.

Another way to think of this idea is to say that another person cannot, in principle, predict what I will do until I decide. They can't predict what I will do, until I decide, not because of some technical inability on their part. Not because they lack enough knowledge. Rather, they can't predict what I will do because, before I decide, there simply is no fact of the matter as to what I will do. There is only a fact of the matter after I decide, because I am the original creator of that decision. It cannot be traced above, and before, me. The buck stops with me.

[The adjective "exclusive" is redundant, because origination implies exclusivity. I added it just to be extra clear that people think of themselves as originating their life stories in a way incompatible with cosmic design.]

That said, I give you an A. You would have made an A+ if you had said "Wow! The Zygote Argument is right; if someone wrote my entire life story, and I'm just living out someone else's script, then I can't have free will." Sadly, that will have to wait for at least a few more years.

More seriously, I'm very glad to see you acknowledge that design issues worry many people and *not* because they are confused. I think many or most compatibilists think the folk are just confused. But, if I am right that the ability to be the exclusive originator of our life stories is an important part of our self-image, then they're not confused, at least not entirely.

[John Fischer]

Kip,

Whew, I was worried you'd give me a "C" for "compatibilist".

[Wolf]

In an indeterministic universe there is no such thing as a seed or design that is guaranteed to lead to a desired state. So the design aspect of your story requires a deterministic universe. What is control supposed to be? Having the robber function in a fashion independent of whatever law governs the universe?

[Tamler Sommers]

John, you wrote:

"I agree that the design issue worries many people, and that they are not confused. It is a good way of distinguishing those of us who are "natural compatibilists" from those of us who aren't."

Is the implication here that there's some degree of relativism when it comes to the compatibility question? Or do you think that the unconfused people who aren't natural compatibilists would come around in a different way (artificial compatibilists?)

[John Fischer]

Tamler,

Well, as you know, I adopt something like a WRE (wide reflective equilibrium) methodology here. I just think people come to the questions with different initial inclinations, although there is also considerable overlap. In my experience talking to people, giving papers, and so forth, I find that there just seem to be a bunch of people for whom the "design possibility" is a major drawback of compatibilism. For others, it may be a prima facie worry, but upon reflection they do not believe that this scuttles compatibilism. I do not know whether any rational person, simply in virtue of her rationality, could be brought to compatibilism; more likely, this is like other areas of philosophy where the best we can do is to systematize our intuitive judgments as a whole, seeking a "wide" reflective equilibrium, where this includes a consideration of a broad range of cases (hypothetical and actual), and a broad range of principles and theories, including skeptical worries about the very methodology of relying on intuitions.

[Tamler Sommers]

John, thanks. Just making sure that the non-confused incompatibilist hadn't already gone through the WRE process...

[Eddy Nahmias]

I'm a compatibilist who thinks there are principled differences between determinism and design. And I want there to be, because I think design or manipulation (at least of most sorts) is intuitively threatening to free will, and I don't want (or think) it to be the case that determinism, properly understood, is intuitively threatening to free will. So, I'd like to have a response to the 'no difference' premise of design/manipulation arguments (like Mele's zygote argument or the cases Kip suggests). Here's an underexplained gesture towards an answer with the hope others will help it graduate to a stab at an answer.

Free will requires certain types of abilities to do otherwise--not all-in (or unconditional) abilities to do otherwise holding fixed everything, but abilities to do otherwise had conditions been different such that the agent had reasons to do otherwise. Determinism does not conflict with such abilities. My intuition is that design and manipulation, intuitively understood, do conflict with such abilities.

If Diana has a goal for my life (e.g., to rob a bank at a particular time and place), then I think the best way to understand such a case is that she will do what it takes to ensure that that goal is achieved. That means that, even if in the 'actual sequence' I rob the bank in accord with my desires and values and rational deliberation, if we jiggle the situation to test for relevant possibilities, we will find that I do not have the abilities to do otherwise that I have in the deterministic scenario. If the situation were altered so that my rational deliberations would lead me to choose not to rob the bank, Diana will cut off that possibility (either by manipulating me at the time of choice to ensure I still rob the bank or by choosing not to create me in the first place).

If, to deal with this plausible way of understanding the original cases, we then try to bring the design scenario closer to the determinism scenario by suppressing these differences--e.g., by making Diana crazy or disinterested, so she could care less what would happen in the alternative possibilities, or by making Diana a computer randomly assembled to design me with no particular goal in mind (or no mind!)--then I think the intuition that such design conflicts with free will increasingly drains away (as it should). The ordinary intuition that manipulation threatens free will is, of course, most clearly driven by real-world cases, which typically involve bypassing the manipulated agent's rational capacities and involve a manipulator who will do what it takes to ensure that his (usually nefarious) goals are attained. (This all connects up with Jonathan's point about the importance of Sid in his case C having the intention to make sure the bank gets robbed and the fact that Sid then becomes an apt target of blame, etc.)

[Kip]

Eddy,

I find your response a little maddening (in a good way). So let me begin by qualifying everything I to say here, by first saying that I have tremendous respect for you, your neurticompatibilism, and your experimental philosophy results (which are more impressive, I think, than just about any other experimental philosophy in this area).

This is what I find challenging about your response:

1. you seem to deliberately put in elements into the stimulus that are not there. That is, you seem to deliberately attack a straw man; and
2. you seem to not appreciate the degree to which the designer ("Diana") doesn't care at all what happens in near-by worlds.

Let me address these points in turn.

Although I could have made this more clear (and I believe the Zygote Argument does make it clear), the agent in these scenarios is otherwise a perfectly normal human being. He could be, as John Fischer noted above, you or me or any other person. We have no way of knowing whether our lives are designed in this way.

That is, the stimulus does not mention, or even imply, that the counter-factual abilities of the agent are in any way compromised. There is no fail-safe mechanism. There is no compromise of rationality. There is no compromise of compatibilist "free will."

This brings me to my second point. Why is there no fail safe mechanism? Because Diana knows that she cannot fail. That's what determinism means: the future, given the past, will necessarily happen. As such, Diana could not care less what happens in nearby worlds. Mere design, itself, is sufficient for her to guarantee that the agent will rob the bank. Fail-safe mechanism are, therefore, gratuitous, unnecessary, overkill, redundant...

Here is what I would imagine a conversation between Eddy and Diana to go like:

Eddy: Diana, I see that you've designed Ernie to rob the bank on January 10.
Diana: That's right.
Eddy: Well, Diana, if something is just slightly different in the past, then Ernie might not rob the bank, so you better put in a fail-safe mechanism to ensure that Ernie actually does rob the bank.
Diana: But the past isn't slightly different. The past is exactly the past. It's not going to change. Therefore, given determinism, Ernie will definitely rob the bank. It's guaranteed. I have no need of fail-safe mechanisms.

In view of the above, I hope you can see how bizarre your reaction to the design scenario is to me. The original scenario suggests a particular problem: that our lives are entirely predictable and predetermined, even though our compatibilist capacities are untouched, so that we are just living out a story that somebody else wrote. Instead of addressing that concern, you've injected into the stimulus a different problem, nowhere mentioned in the original description (a bypassing fail-safe mechanism), and then proceeding to show why that undermines free will. I can't help but feel that you're just attacking a straw man.

[Eddy Nahmias]

Well, Kip, we've been down this road before and I doubt I'll convince you (perhaps becausee I am wrong!). When trying to understand the actual world, it is often appropriate to ask about possible worlds (or counterfactual situations). Here's an example: in trying to understand what free will actually is (or whether it exists), it is entirely appropriate to ask about counterfactual situations ... such as what would be the case if we were designed by someone who wanted us to rob a bank!

I'm not deliberately adding features to the thought experiment that are not there in order to create a straw man. I'm trying to flesh out the thought experiment to see how best to understand it (e.g., what features it has) and how it might be pumping people's intuitions. Even if Diana in fact knows that in the actual world Ernie will rob the bank, it is not logically necessary that he robs the bank, so it is entirely legitimate to wonder what would happen in nearby possible worlds where Ernie does not rob the bank.

And my contention is that those possible worlds (where Ernie does otherwise) are "farther away" from the designed world than they are from a deterministic world, precisely because Diana would presumably use her power to cut off those possibilities *were it (contra the actual scenario) necessary*. Of course, one could offer more information in the scenario to clarify that Diana would not ensure a particular outcome, but then I think one loses the feature that is driving the intuition that design or manipulation undermines free will.

[Mark Young]

The question is not whether Diana needs a failsafe mechanism, I think. The original story says that Diana "sets everything up so that you do exactly that." That implies that Diana considered how things needed to be set up to guarantee that outcome -- that she considered what the outcomes of various possible "initial conditions and/or laws of [the] deterministic world" would be.

I think Eddy is wrong to say that Diana would manipulate him "at the time of choice" in order to prevent his rational deliberations from leading him to not rob the bank. Diana does "what it takes to ensure that that goal is achieved" at time zero -- when the laws are set up. In her deliberations at that time, she must work thru the consequences of various options and "cut off" any that lead to the state she does not desire.

And so it's not simple determinism that's the threat here, it's the fact that we have some particular person using her free will -- whether it's a compatibilist free will because she's subject to deterministic laws of her own, or a libertarian free will because she stands "outside" somehow -- using her free will to manipulate me into some action she wants me to take. That is the principled difference between determinism and design.

[Kip]

Eddy,

Yes, we've been down this road before, and perhaps only because I'm wrong. I like to always leave that possibility open.

And my contention is that those possible worlds (where Ernie does otherwise) are "farther away" from the designed world than they are from a deterministic world,

I agree that worlds where Diana sets up a failsafe mechanism are farther away from a merely deterministic world than the merely-designed world is from a merely-deterministic one.

Of course, one could offer more information in the scenario to clarify that Diana would not ensure a particular outcome, but then I think one loses the feature that is driving the intuition that design or manipulation undermines free will.

Well, yes, you would lose the intuition. But, again, you're beginning to deviate from the thought experiment.

Your comment above doesn't persuade me to abandon my position. It doesn't, for example, seem to address Diana's point in the dialogue I wrote above. So we may have to agree to disagree for now.

[Matt King]

Kip,

I think this cases illustrate a certain dialectical problem in the literature. And this may be part of what Eddy's suggesting as well.

I *think* Eddy's point is simply that, as compatibilists, we already think determinism is compatible with free will. Now you've given us a case illustrating what you're calling "mere design", and you think it ought to be worrisome to everyone.

The dialectical problem is that if cases such as this one are going to pose a problem for compatibilists all by themselves, it seems two conditions must obtain:
(1) The cases involves undermined freedom (or responsibility); and,
(2) There's no (relevant) difference between the cases and simple deterministic worlds.

This is familiar territory (as you've noted) - we've all seen arguments that employ a No-Diff principle for this reason.

But here's the snag. The more plausible you make the case that there is *NO* relevant difference between the given case and a simple deterministic world, the less grip it's going to have on compatibilists. After all, we're already comfortable with saying that free will is consistent with determinism. If you come along and say "Here's a case where the agent isn't free and it is *JUST* like a deterministic world", you can hardly be surprised that we won't be apt to budge. One's modus ponens is another's modus tollens.

If a compatibilist finds the case plausibly does involve undermined freedom (or responsibility) this is likely because he or she has found a relevant difference (thus - e.g. - Eddy's suggestion that Diana 'design' seems to imply that she has a goal or plan for Eddy, something "the universe" by itself does not).

Incompatibilists are apt to find design cases compelling, because they pull on those very intuitions that impel people towards incompatibilism in the first place; and vice versa for compatbilists.

This is perhaps a *VERY* long-winded way of agreeing with John's point that cases by themselves are extremely unlikely to carry the day.

[Kip]

Matt,

First, I think Eddy is focusing more on the possibility of failsafe mechanisms (undermining counterfactual powers) rather than the designer's mere intent or goal-seeking.

Everyone seems to agree that the whether the designer is goal oriented shouldn't increase or decrease an agent's freedom. That is, almost everyone agrees with a version of the No-Difference principle (assuming there are no failsafe mechanisms).

So, that does leave us with a possible rhetorical stalemate. I believe Neil Levy put it this way (but I could be wrong):

The two things are relevantly the same:

1. Diana designs the universe so that every moment of my life is predetermined according to her plan, although I still satisfy the compatibilist criteria of your choice (still reasons-responsive, etc.).
2. Nobody designs my life, and I come to live my life through the blind, organic evolution of the world and my childhood. The universe, and my compatibilist powers, are exactly the same as in 1, but not by design.

People like me want to say: 1 is obviously in conflict with free will, so you're mistaken to think that 2 isn't.

But, you're right to note, the compatibilist can reply: 2 is obviously consistent with free will, and so you're mistaken to think that 1 isn't.

How do I answer the compatibilist? I don't have any proof. The answer strikes me as bizarre, yes. But perhaps that only represents the idiosyncratic intuitions and prejudices that I bring to the debate.

It seems to me that the word "obviously" is much less appropriate in the compatibilist's response, than in mine. The pull that the designed agent's life isn't free isn't hard to feel. Many compatibilists will acknowledge feeling it. They just somehow wiggle, and think, their way out of it.

All of the above reinforces my view that the compatibilist position (which says we're wrong to think the designed, predetermined life isn't free) is using a revisionist notion of free will that has drifted so far from the common understanding of the term, that it no longer counts as the real thing.

In contrast, the anti-realist doesn't need to do any wiggling or thinking. The conflict, at least the prima facie conflict, is plain as day. But, as I said, this isn't a mathematical proof.

[Paul Torek]

Kip,

I think you've misinterpreted Eddy's failsafe mechanism. The way I read Eddy, Diana's design-intent is the failsafe mechanism. Ernie robs the bank in the actual scenario due to his keen desire for money. One of the formative experiences, let's say, leading to this desire, is that Ernie found a gold nugget as a child and got lots of attention for it.

Consider an alternate scenario where that gold nugget never existed. Well, Diana would just arrange some other experience that would lead to an equally keen desire for money, and Ernie would still rob the bank. She would do so, because of her design-intent. Under a wide range of conditions, Ernie still robs the bank, and Diana's design is the fail-safe that guarantees this.

Now, in my view, this is still a no-difference difference. Eddy's making a very understandable mistake to see such external control as incompatible with freedom. (In the real world, it always is, but logical possibility roams wider.) But that's another argument.

[Eddy Nahmias]

Perhaps changing the subject a bit, does anyone know if anyone has made the connection between a designer case (like Diana in the zygote argument) and a *global* Frankfurt case? The connection seems like it's ripe for the making, but I don't know if it's been made already. John?

[James Gibson]

Eddy,

Sometimes the global Frankfurt case is connected to an omniscient being capable of knowing future contingents. Dave Hunt's work in various places runs the Frankfurt cases through this sort of story.

But there is an important difference between the two cases, or so it seems to me that it is important: in the designer case, the designer has intervened in the actual sequence whereas the the Frankfurt case is supposed to leave the intervener as merely a viewer of Jones, one who acts only in the counterfactual cases.

Whether this difference suggests that designed-Jones cannot be responsible but merely-watched-Jones can be responsible is a point to be debated. Still, the noted difference does suggest an important distinction in the elements of the narratives.

[Kip]

Eddy,

I may be completely wrong but I think that maybe, given your interest in the combination of design scenarios and Global Frankfurt cases, you may be interested in my post on Brittle People. But I'm not very up-to-date on what Global Frankfurt cases are...

[Kip]

Paul,

Four things are starting to get blurred:

1. Diana designs Ernie's to rob the bank. But if things had been different, Diana might not have designed Ernie's life to rob the bank.
2. Diana designs Ernie's life and, if the world had been different, would also have designed Ernie's life to rob the bank, across a wide range of possible worlds.
3. Diana designs Ernie's life to have failsafe mechanisms, which are not actuated in the real sequence, but would be activated in non-actual sequences, to ensure that he robs the bank.
4. Diana intervenes in Ernie's life and forces him to rob the bank.

I was only talking about 1. Eddy started talking about design and failsafe mechanisms, and I thought he was talking about 3.

In response, I think you are saying that, instead, Eddy was talking about 2.

Suppose Eddy was talking about 2. Here is my response:

1 is just as disturbing as 2. The thought that every aspect of our life was predetermined and setup and predicted by us, that ultimate control resides in someone or something outside of us, is disturbing to most people. Even most compatibilists feel an initial pull that such a thought is disturbing.

It doesn't matter whether (1) the designer only would have designed our lives in the real sequences and would not have designed our lives in non-actual sequences or (ii) would have designed our lives in all possible sequences. Non-actual sequences are irrelevant. Just looking at the real sequence, there is already design, which is already disturbing, regardless of what would have happened if things had been different. Things weren't different. This is another example, I think, of where compatibilists are far too enamored with counter-factual possibilities.

[Sofia Jeppsson]

Matt, I think you pointed to something crucial here. I can relate to some discussions I've had on B F Skinner's book Walden 2. Robert Kane thinks compatibilists have a problem, since they'd have a hard time explaining what's wrong freedom-wise with that community, and adds that some compatibilists "bite the bullet" and claim that the W2 citizens ARE free after all. I'd say they are free, but I couldn't see that there is even a bullet to bite in that claim.

I talked that over with a colleague. He hadn't read the book himself, so I had to describe the scenario for him. First he thought the scenario was really worrisome. But he assumed that some people were controlling the others, sort of like in "brave new world", so I explained that there's no dictatorship, and a very equal society to that. I remember there were more worries, but they all rested on some initial misunderstanding of the scenario. When I'd eventually made things really clear on what the scenario looks like - it's equal, no censorship, no dictatorship, everyone can travel freely between W2 and the otside world, most people also HAVE spent quite a lot of time in the outside world and can compare it to W2, nobody is forced to remain a citizen of W2 - to sum up: The only difference one would experience between being a W2 citizen and being an ordinary person, would be that a W2 citizen is happier, healthier, more virtuos and more rational, and so is everyone around you.
When that much was finally settled he had the same reaction as myself - if such a society existed in the real world, we'd happily sign up for some social engineering and end up with new and improved selves.

And I think that's the natural reaction for somebody who has compatibilist intuitions before hearing the case. If you're an incompatibilist from before, then those cases is just another way of bringing out what you think is wrong with an ordinary deterministic setting.

[Mark Young]

Suppose that Diana, our cosmic designer, has not designed Joe to rob a bank, but rather for him to design a second deterministic world -- a world wherein Betty (a being with all the powers and properties of compatibilist "free will") robs a bank.

Joe, the sub-cosmic designer, can do the card trick on Betty. Betty, the sub-cosmically designed, can do nothing to falsify Joe's predictions.

It is suggested that there is no morally significant difference between control and design -- which would imply that Joe is controlling Betty's life. Joe designed the world Betty is in, and made her so that she would "choose" to rob the bank.

But that doesn't seem right to me. Diana is (supposedly) in control of Joe's life. How can Joe control Betty's life if he can't even control his own? It looks like it's Diana who's "really" controlling Betty's life. Joe is just an instrument of Diana's plan.

But that raises the question -- where did Diana come from? Did she have a creator? Presumably not -- and if she did, we simply regress one level, and ask where her creator came from.

So if she wasn't created, she must have come about naturally. If she came about naturally, then the nature that produced her was either deterministic or indeterministic.

But if Diana came about from a nature that is deterministic, then she is no more the controller of Joe than Joe is of Betty. The point of the story is that there is no difference between us being designed by Diana, and us being the products of a natural deterministic world. What's true for us must be true for her, and so she has no control over our lives.

On the other hand, if the nature that produced Diana is indeterministic, then the story relies on there being some form of indeterminism at work. And that's a problem because what has then been "proved" is that free will and moral responsibility are incompatible with universal determinism assuming there is a being with free will around.

And note that the story requires more than just a chance setting of laws at the beginning of time. Diana must be capable of making and executing a plan not determined by the laws of the universe -- else the story undermines itself. The story would lose its strength if Diana just threw the world together at random.

Comments?

[Kip]

Mark,

But if Diana came about from a nature that is deterministic, then she is no more the controller of Joe than Joe is of Betty. The point of the story is that there is no difference between us being designed by Diana, and us being the products of a natural deterministic world. What's true for us must be true for her, and so she has no control over our lives.

I agree with essentially everything you say. Diana and Ernie have the same amount of ultimate control over their lives: zero.

[Ben Burgis]

i wonder how much work the causal determinism is doing in the "mere design" scenario. What if, instead, the cosmic designer was (a) omniscient, as in traditional religious views, and (b) designed human beings to be capable of routinely fulfilling all of the *libertarian* conditions for free will, a la whatever one's favorite libertarian theory might be? Given (a), the designer would know, before any particular sperm-egg pair into contact, all of the decisions that the resulting person would make over the course of their life (even if those decisions were radically undetermined by any kind of physical chain of cause and effect and fulfilled all sorts of extra libertarian conditions). The designer would only allow sperm-egg pairs to come into contact when the resulting person would be one who would make exactly the decisions the designer wanted them to make.

Under these circumstances, you'd have a designer who'd be capable of doing the writing-stuff-down-on-cards-and-showing-the-cards-to-you-five-minutes-later-bit, and who would be more than just a passive omniscient observer, since the designer would in some sense be the reason that the cards said what they did. Given that the aggravating-ness, and the result-of-design-iness, is identical in this scenario (despite the absence of causal determinism), would this do just as much violence to free will?

[Mark Young]

Kip wrote "I agree with essentially everything you say. Diana and Ernie have the same amount of ultimate control over their lives: zero."

What I said was that the designer/manipulator (Diana) either violates the laws of the (deterministic) universe, or has no control over Joe/Ernie's and Betty's lives.

The implication is that a deterministic universe has no control over our lives. Since the intuition is driven by the impression that someone else has control of our lives, that explains why the story doesn't impress compatibilists.

Do you agree with that?

[Kip]

Mark,

No, I don't agree with that. The basic point is that design is a type of control.

To help you see that, consider two situations:

1. In world A, I want a robot to rob a bank on January 2. So I get the Roomba from Amazon.com, which is not designed to rob banks, and instead cleans the floor. Then, on January 2, I intervene to make the Roomba rob the bank.

2. In world B, I decide to just build a robot from scratch. I design the robot, from scratch, to rob the bank on January 2. So the robot has "January 2" and "bank-robbing-subroutine" written into its software. I release the robot into the public, well before January 2, knowing that the robot will follow my predetermined plan, seek out the bank, and rob it on January 2---because I designed it to do so.

To me, there is no relevant difference between the amount of control that the controller/designer has in these situations. Yes, there is a distinction between the two: in world A, the controller literally intervenes and forces the robot's hand. In world B, the designer sets up the plan ahead of time, through design, in an indirect way.

But that difference is not relevant. If you want to rob the bank, you can use either method with equal (total) success.

So I think it's a mistake to think that design doesn't not involve control, at least in some important sense.

[Mark Young]

Kip:

Design is a type of control. OK. So back to my variation on your story:

(1) Diana designed Joe.
(2) Diana controls Joe.
(3) Joe designed Betty.
(4) Joe controls Betty.

Is (4) true or false?

If false, why? Is it because (3) is false, due to the truth of (2)?

[Kip]

It's helpful to distinguish between ultimate and proximate/intermediate (PM) control. Joe PM controls Betty. Joe does not ultimately control Betty, because nobody ultimately controls anything.

[Paul Torek]

Mark Young writes,
"The story would lose its strength if Diana just threw the world together at random."

This seems well-taken. But I suppose Kip is arguing that Diana's planning suffices to make it non-random. And that this planning involves only proximate control, not ultimate control.

But let's step back to the evolution of Diana. That was a process without design, so I wonder if Mark would argue that in the grand scheme of things, the worlds of Joe and Betty are random, because Diana herself is an outcome of a random process.

Adapting Kip's terminology, let's say that the worlds of Joe and Betty are proximately non-random but ultimately random. Which one of these facts matters, for the intuitive pull of the case? Is merely proximate non-randomness enough to be intuitively threatening?

I'm inclined to say yes, thus siding with Kip. But I wonder if, paradoxically, that's my compatibilist bias talking. To wit: as a compatibilist, I see the "merely proximate" agents Diana and Joe as genuine agents. So of course, my evolutionarily honed instincts raise huge alarm bells when I think of these agents having control over some poor innocent victims. In the real world, whence those instincts came, being controlled by others is really bad news.

[Mark Young]

It's not just that nobody ultimately controls anybody; nothing ultimately controls anybody. If Diana is herself the result of a deterministic process, then she has no ultimate control over us. If Diana has ultimate control over us, then the postulated universe is not deterministic.

And it's not just ultimate control. The (deterministic) universe has no proximate control over us, either. To control requires some kind of intention (either immediate or derived). We have proximate control over ourselves -- we have intentions for ourselves that (normally) regulate our behaviour. A (super-powerful) manipulator has proximate control over us -- has intentions for us that regulate our behaviour.

It is possible for two people to have complete (not ultimate) control of one object -- so long as their intentions for that object are sufficiently compatible. When the super-manipulator works by designing the bank robber, the two people's intentions are in perfect accord. When the less-super-manipulator works by bypassing the robber's self-direction, the two people's intentions are not in accord.

Responsibility falls on the one(s) who intend the actions taken. In the latter case that is clearly the manipulator. In the former case it is both manipulator and robber. The proper conclusion (IMO) is that both are fully responsible for the robbery. I believe that people who want to mitigate the robber's responsibility are falling victim to a "lump of responsibility" fallacy -- that the amount of responsibility is fixed and must be divided between the two.

But to be clear, I am not making claims about free will. I am trying to explain why I (and perhaps other (semi?) compatibilists) think that there is a significant difference between design and manipulation. I think the distinction works (in the sense of making the story non-disturbing).

[Kip]

Mark Y:

You're adopting the "responsibility multiplies" answer to the problem. Paul Torek favors that response, too, if I recall.

Paul Russell also suggests that, in general, we can hold the bank robber responsible (as agents on the same level) but that the designer can't hold the agent responsible. Because doing so would be pretty absurd.

Those answers strike me as backwards rationalizations, or bandaids, to try to fix the problem that the dilemma finds. You think that skeptics are tricked into thinking that responsibility doesn't multiply. Perhaps. I just think that human attribution of freedom and responsibility is so riddled with biases and errors that, it's more likely, that the design scenario simply shows that we're not as free and responsible are we ordinarily think. It exposes that free will is just another positive illusion.

But I don't have a textbook proof that my interpretation is better than yours.

[Eddy Nahmias]

Kip, you ask if Ernie has more options in B [deterministic world] than A [designed world] even though we can stipulate that they are molecule-for-molecule identical. I think the answer is "yes" (again assuming that the designer actually has designs for Ernie). Rather than re-hashing my view once more, let me just offer an analogy that I think illustrates how two internally identical persons can have different options.

Freddy and Eddy are molecule-for-molecule identical. But Freddy, unlike Eddy, has a guardian angel, Diana, who watches over him. If something is about to severely harm or kill Freddy, Diana will intervene to protect him. Alas, Eddy would just get run over by the truck (at which point his molecules would no longer be identically situated to Freddy's!). As it turns out, Freddy and Eddy live identical lives, avoiding severe harms, and dying peacefully in bed.

I think Freddy and Eddy have different options--that different things are possible for them, that different possible worlds are closer or farther from them (e.g., the possible world where Eddy gets crushed by a truck is much closer than the one where Freddy does).

I won't try to make the connections to manipulation arguments here, but I think the connections are there to be made.

[Paul Torek]

Mark, I agree wholeheartedly with your latest response. As Kip correctly quotes me, responsibility multiplies. A look at the law and at everyday ethics instructs us that and how it multiplies, and we should consistently apply those observations here. But, doesn't this obviate the point you raised about who designed (or didn't design) the designer?

[Mark Young]

You think that skeptics are tricked into thinking that responsibility doesn't multiply.

That would imply that I think someone tricked them. I have no reason to think that, and no theory as to why anyone would. Plus I have no reason to believe that the problem is confined to skeptics.

I disagree that "the design scenario simply shows that we're not as free and responsible are we ordinarily think." I think our ordinary conceptions of freedom are anti-fatalistic, not anti-deterministic. Many people (not you) might get determinism and fatalism confused, and so think that determinism is a threat to their beliefs. It's easy to confuse the design scenario with a fatalistic scenario ("no matter what Ernie chooses, he will end up robbing the bank") -- a confusion encouraged by the card-playing sub-story. Even you write "there would be nothing you could do to stop [Diana]" -- which is false according to the design scenario (on a compatibilist conception of "can"). Such comments can bias the observers of the scenario in ways that they would reject if it were put to them directly.

[Mark Young]

"[D]oesn't this obviate the point you raised about who designed (or didn't design) the designer?"

When I raised the issue of who designed the designer, I was trying to show that the scenario was self-undermining unless the designer violated the deterministic laws of the universe. I was trying to approach the issue using the original premises.

I don't think that a designer would eliminate or even mitigate our responsibility. Responsibility (IMO) flows from epistemic facts more than metaphysical ones -- what we knew and when we knew it is way more important than the facts beyond our ken. Even in the design scenario, we don't know what plans the designer has for us, and we have no way of finding out. Thus those plans are meaningless with regard to our choices, and our responsibility for those choices.