Blog Coordinator

« The Consequence Argument | Main | Neo-Aristotelian Dispositional Compatibilism »

11/11/2014

Comments

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

Charles -

1. You write "it seems that the only modal accounts of the ability to do otherwise are compatibilist accounts." Do you mean "it seems that the only modal accounts of the ability to do otherwise THAT REFER TO ACTUAL, REAL ABILITIES THAT HUMANS POSSESS are compatibilist accounts"? It seems that we could adopt incompatibilist accounts of the ability to do otherwise, but, on those accounts, humans do not have the ability. Do you agree with this much?

2. The problem with counterfactual reasoning is that it refers to fiction. And fiction is not empowering. When people talk about free will, they are concerned with what they can actually obtain, in this world, and not what happens in fictitious other worlds that they will never access.

So if we're strictly defining abilities in terms of counterfactuals, then we're basing abilities on fantasy worlds that humans can never access. Why would we want that kind of ability? I am interested in knowing what I can do in the real world, not in fantasy land.

Imagine you said: "You have the ability to make a billion dollars." I would say, "great, give it to me!" And you said, "oh no, silly, you don't actually make a billion dollars in the real world. You make a billion dollars in a fantasy world, where the laws of nature are different, and the past is different. You will never make a billion dollars in this actual world. But you have the ability to make a billion dollars in this world!"

The incompatibilist is going to find your reliance on counterfactual worlds not very reassuring. If we reside in one real world, with one fixed past, and one fixed set of natural laws, then we're interested in what happens in this world - not fantasy worlds that we never access.

Kip,
Thanks for the response. No, I don't mean to add the qualifier "that refer to actual, real abilities that humans possess." That is a separate issue. I do think this is one problem with incompatibilist accounts of the ability to do otherwise...either they seem to add no additional control or they are the sorts of abilities that actual real human beings cannot possess. Yet, that is not my concern here.

Instead, my worry is that there are problems with an incompatibilist interpretation of 'abilities' regardless of whether we can possess them or not. My target here is not just libertarians, but skeptics and semi-compatibilists as well. The challenge is whether any concept of the ability to do otherwise can restrict our attention to only cases where the laws and the past are held fixed. "Ability" is a modal concept...it is about what an agent can do (in some sense of 'can').

As a side note, I don't endorse a counterfactual account of abilities but instead am a dispositional compatibilist. Yet, that's irrelevant to the present concern. The present concern is that modal reasoning is a species of counterfactual reasoning so any coherent concept of 'ability' must be consistent with our capacity to reason counterfactually.

As far as the claim that counterfactual reasoning refers to fiction and therefore is problematic, you're clearly making a mistake here. All modal reasoning requires considering other ways a world might have been. Consider a bad argument for the claim that we can never do otherwise than what we in fact do. I didn't have eggs for breakfast. Could I have? Well, you didn't so you couldn't. Instead, in order to consider whether I could have had eggs for breakfast I need to think about fictional worlds. Ones in which I did not have eggs for breakfast. Next, I need to ask whether I could have made such worlds actual. Reasoning about counterfactuals requires considering fictional worlds. Reasoning about causation, requires considering fictional worlds. Reasoning about laws of nature, requires considering fictional worlds. Reasoning about metaphysical necessity, requires reasoning about fictional worlds. In fact, all reasoning about anything other than what actually happened requires considering fictional worlds. So all philosophical reasoning and almost all scientific reasoning requires considering fictional worlds.

What the incompatibilist claims is that the only relevant fictions to be considering are those that have the same laws of nature and the same past as our own. Yet, when we explore what we are doing when we reason modally we recognize that this restriction makes no sense. Modal reasoning isn't independent of counterfactual reasoning but instead is a species of counterfactual reasoning. So, any alleged modal concept that removes from us our capacity to reason counterfactually is an incoherent concept. The problem with accounts of the ability to do otherwise that are not compatibilist accounts is that (at least so far) there are none.

Charles,

I am not so sure that the only conceptions of ability are compatibilist. Consider your typical moral choice. Person X has reasons to murder and not murder. X sweats over the decision, which is close, but finally chooses to murder. According to the compatibilist, X had the ability to not murder, based on an analysis of what would happen if the past/laws had been different. According to the incompatibilist, X had the ability to not murder, even if the past and laws were different. I don't see anything about the concept of "ability" per se that precludes abilities from being analyzed with the same past and laws.

Consider: with past P and laws L, the compatibilist says that X will take action A, and has an ability to do otherwise (~A) based on an analysis of nearby worlds and/or laws (for example). With P and L fixed, the incompatibilist says that X may take action A or !A, and the ability to do otherwise is based on an analysis of that set (A and !A). I don't see the dilemma here.

(Notice there is a subtlety here: some compatibilists, like Nahmias I believe, hinge free will on the past not being necessary. Others don't have whether the past is necessary or contingent, like Fischer I believe. I wonder if this distinction makes a difference for your argument?)

Perhaps you are arguing that any incompatibilist conception of ability must be irrational, because (A) something must be different about P or L to justify making a different choice or (B) there can only be one rational choice for any given set of P and L. But the literature is full of examples of how multiple decisions can be rational or justified, given the same P and L.

This returns me to my earlier point: even if it is natural or common to analyze ability in terms of counter-factuals, why is it a good idea to do that?

Consider your statement: "you're clearly making a mistake here. All modal reasoning requires considering other ways a world might have been."

On my preferred conception of ability, ability should only refer to things that you will actually ever do in the real world. The fact that, if the world had been slightly different, you might have done differently is *irrelevant*, because those counterfactual worlds are fictitious and inaccessable. Under a very simple understanding of determinism, we will only ever do what the past and laws predict - there is only one single outcome for my life. That life is my entire ability - a set of one single outcome.

You might say that I have the ability to type a different email than I type now. Or to eat a different dinner than I ate tonight. Or to see Big Hero 6 at the movies, instead of Interstellar (which I saw on Tuesday). And your statements would be intuitive and common sense, but they would also be - I submit - wrong. It was popular and common sense to think that the sun revolved around the earth, and that human beings were created according to the Bible instead of Darwin. Sometimes common sense is massively wrong, and that happened, I think, in the realm of free will and personal responsibility.

If you think I have the ability to do otherwise, even in a deterministic world, then isn't it curious that I never actually do otherwise? Not once? Not a single time? If I had an ability to do X, don't you think that I would ever do it? My entire life? Even if I really really really wanted to? And yet I never will, no matter how hard I try, forever I am doomed to doing just exactly what P and L predict - and I cannot change P or L. What I am able to do is live out the exact single future that determinism predicts for me, and not one iota more. And if anyone doubts that, I dare them to exercise their ability to do otherwise, just one single time. I double dog dare you.

Hi Charles,

This is a really fascinating post and I am disposed to agree with a lot of what you say. Let me ask for some clarification on your view.

Are you suggesting that all modal notions are rooted in counterfactuals, as if they are a kind of primitive from which the full range of modal notions can be built?

Also, what do you say about this case: A quantum physicist is talking about a photon in a slit experiment. He says, “The photon went left, but it could have gone right.” How do you interpret this could statement? It seems the ‘could’ used here holds fixed the past and (fundamentally probabilistic) laws. If so, then isn’t it the case that even you need to allow the cogency of modal notions that are not constitutionally iffy?

Hi Charles,

I have a question that is, in part, similar to Chandra's. Consider The Simple View: S is able to X iff S's Xing is compossible with the past and the laws (a view you allude to). It's crazy to deny that The Simple View is a modal view---it's a view stated in terms of metaphysical possibility and that's just the end of the story (because it's analytic that possibility-claims are modal-claims). So I say: The Simple View is undeniably a modal theory of ability and it entails incompatibilism. So it's just false that there are not any incompatibilist modal views.

You might be saying (but I'm not sure): all modal claims are, at bottom, counterfactual claims. If so, we can rewrite The Simple View as a counterfactual claim (we know, after all, that it is a modal claim)--but then the incompatibilist has a counterfactual theory of ability (which is what you say he can't have). If The Simple View cannot be rewritten as a counterfactual claim, then I would say that it's a counterexample to the thesis that all modal claims are counterfactual claims. (I'm not sure that you endorse this latter thesis because you just say that modal reasoning is counterfactual reasoning; but I'm not sure how to make sense of this if modal claims are not reducible or equivalent to counterfactual claims.)

What part(s) of the above do you disagree with?

Charles: I just wanted to expand on Chandra’s question a bit, too -- and I agree that this is a very interesting post!

You write: “the only modal accounts of the ability to do otherwise are compatibilist accounts.” But Al’s “Agents’ Abilities” (2003) and Randy’s “Dispositions, Abilities to Act, and Free Will: The New Dispositionalism” (2009) discuss libertarian accounts of abilities, and any libertarian account would require the extension principle (Fischer 1994: the requisite abilities are constrained by the past and laws) which is a modal principle.

You could, as Chandra suggests, think that they’re not counterfactual accounts – not modal in a specific sense. Or you could be making a broader claim: that libertarian accounts of abilities fail, maybe because they fail to incorporate counterfactuals into their analyses. Or you could be saying something else!

Kip,
Thanks for the post. My argument is that we can learn from psychologists what it is to reason modally. Modal reasoning is a form of counterfactual reasoning. Nothing I am saying here defends a counterfactual account of agents abilities. Instead, the claim is that the very concept of agents abilities is a modal concept. So, if you ask someone to consider a question like whether we are able to do otherwise in a deterministic world, and then demand that they cannot employ the resources of counterfactual reasoning, you are telling them that they can't reason modally about abilities. Since the concept of an ability is a modal concept, demanding that agents do not employ the resources of modal reasoning when considering abilities makes no sense. Now, I guess there are two responses that seem very plausible to make against my argument. One is that 'ability' is not a modal concept. The second is to deny that modal reasoning is a species of counterfactual reasoning. The second response would require taking issue with what we are discovering from psychologists studying modal reasoning and deny that modal reasoning is a species of counterfactual reasoning. Personally, if I were to reject my argument I'd employ the second strategy. But, you seem to employ the first. "On my preferred conception of ability, ability should only refer to things that you will actually ever do in the real world." To be able to do something is to do that very thing. On your preferred account of abilities, it is not a modal concept. So, even if it were true that all modal reasoning is a species of counterfactual reasoning this in and of itself is no threat to your account of abilities. Yet, I doubt anyone would seriously adopt your account. Imagine an indeterministic world where I could freely choose to have eggs for breakfast or not. As it so happens, I don't. Could I have done otherwise? Well, if being able to do something means that I actually do it, then the answer is no. Showing that nobody has the ability to do otherwise in a deterministic or in an indeterministic world is very easy on your account. In fact, it follows by definition from your account. I doubt anyone in the free will literature, however, would take seriously the claim that being able to do something means the same thing as doing that very thing. Also, note, I am not arguing here that any particular account of the ability to do otherwise is the correct account. Instead, my argument is that no account that requires the laws and past be held fixed in a deterministic world is adequate. If it could be shown that any account that allows us to consider worlds with different laws or a different past than our own are untenable, then the ability to do otherwise is impossible regardless of whether determinism is true. I don't think accounts that allow us to consider worlds with different laws or a different past are problematic. Yet, nothing I argue for here defends these accounts. My point is that there is no other alternative that could allow agents the ability to do otherwise regardless of whether determinism is true.

Chandra,
Thanks for the comment. I'm not arguing that modality is grounded in counterfactuals. I think that would be a category mistake. Instead, what I am arguing is that modal reasoning is grounded in counterfactual reasoning. In fact, modal reasoning is a species of counterfactual reasoning.

As far as the double slit case, we are still engaging in counterfactual reasoning. If an electron were fired toward the double slit it would turn to the left is false. There are equally nearby worlds where the electron turns to the right. For the same reasons, it is false that it would turn to the right. So, it seems to me that in this case it is easy to explain what the electron could do by evaluating the relevant counterfactuals. (If the question is about the actual electron that went to the left, we are merely reasoning about counterfactuals with true antecedents. On Stalnaker's account, we'd end up with the same verdict as what I stated above. On Lewis's account, counterfactuals with true antecedents share the material conditionals truth table. Yet, Lewis agrees with Stalnaker that his account gives a counter intuitive result for this case and tries showing how he can still mimic Stalnakers results for this case. It is precisely cases like this that were central to the Lewis Stalnaker debate about counterfactuals.)

Do we hold the laws of nature and the past fixed when we evaluate the counterfactuals about the double slit experiment? No. Although frequently the closest worlds to our own will be ones that share our laws and past. Frequently, while doing science, we want to ignore the effects of interference from other forces in the area. So, frequently, science proceeds by considering simplistic models. Consider a model where all that exists is the electron and the experimental apparatus. Doing so allows us to ignore cases of interference. When we consider the case in this manner, we do not hold the laws of nature fixed. After all, on all the major accounts of laws of nature a world with less basic natural kinds than our own is a world with a different set of laws than our own. Now, instead of considering the case of reasoning with a simplistic model consider reasoning about a case where the electron is fired toward the apparatus in the actual world. In order for scientists to be able to employ simplistic models in their reasoning, it seems like we must say that the same reasoning process is being employed regardless of whether we are thinking of the case in a laboratory or in a simplistic model. Otherwise, scientists would have no right to employ simplistic models in their reasoning about the actual world. So, if our reasoning does not (and cannot) hold the laws of nature fixed in the simplistic model then the reasoning does not require us to hold the laws of nature fixed when considering the case in a laboratory setting. In both cases we are reasoning counterfactually about the situation and there is no requirement that the laws be held fixed.

Nevertheless, unlike counterfactual reasoning in a deterministic world ,sometimes the closest worlds to our own where the antecedent is true in an indeterministic world are worlds with the same laws of nature and the past as our own. Even in indeterministic worlds, we do not hold the laws of nature and the past fixed when we reason about cases like the double slit experiment. Otherwise, we would reach the conclusion that since worlds depicted by simplistic models do not share the same laws of nature as our own anything trivially follows in the simplistic model. The way we reason about modality in indeterministic worlds is counterfactually. Yet, sometimes the closest world to our own are worlds that share the same laws and past as our own when reasoning counterfactually about indeterministic worlds. The fundamental difference between reasoning modally in a deterministic world or an indeterministic world is that sometimes in an indeterministic world the closest worlds to our own where the antecedent is true are worlds with the same laws and past as our own. Yet, this is not always the case.

Nate,
Thanks for the post. First, you seem to be making a pretty serious mistake in the comment "it's a view stated in terms of metaphysical possibility and that's just the end of the story (because it's analytic that possibility-claims are modal-claims)." Your comment reminds me significantly of David Lewis's attack on David Armstrong's account of laws of nature. Lewis claimed that calling a relation a necessitation relation no more makes it a form of necessity than calling a man Armstrong makes his arms strong. We can state many non-modal concepts in terms of a possible worlds semantics. Being true is not a modal notion. Nor, is being true in the actual world. The mere fact that the second way of talking about being true employs modal language does not make the claim that P is true a modal claim. If what we are learning from the psychology of modal reasoning and definitions of other modalities in terms of counterfactual reasoning are right then all modal reasoning is a species of counterfactual reasoning. So, if a concept removes from us the ability to reason counterfactually it thereby removes from us the capacity to reason modally. Hence, it is not a modal concept. reasoning is not a modal concept....regardless of whether you use the word "possible" when talking about it.

Further, claiming that I am denying an analytic claim or that I am denying a metaphysically necessary claim is to say little more than that I am wrong and pound ones fist. What we are trying to figure out when we provide an analysis of agent's abilities is what it means to claim that an agent could have done otherwise. The incompatibilist claims one thing, the compatibilist claims another. Yet, since we are analyzing what it means to be able to do otherwise whichever of us is wrong is necessarily wrong and denying an analytic claim. I certainly admit that if what it means to be able to do otherwise is what the incompatibilist claims then I am denying an analytic claim. What I would like is an argument to show why I am wrong that consists of something more than merely noticing that we use the word 'necessary' when stating the simple view.

Joe,
Thanks for the post...I don't know what to say directly about each of these accounts but certainly should go back and consider them more closely.

I assume each of these accounts restricts our attention to worlds with the same laws and the same past as our own? If so, they are problematic accounts for determining whether an agent is able to do otherwise in a deterministic world. If all modal reasoning is a species of counterfactual reasoning, then (despite appearances) any concept that prohibits us from engaging in counterfactual reasoning in a deterministic world prohibits us from reasoning modally about a deterministic world. So, despite appearances, the concept is not a modal concept. We can certainly add all sorts of apparently modal language into such accounts, but this in and of itself does not make it a modal concept.

A concept that prohibits us from reasoning about it modally is not a modal concept regardless of how many times it employs the word 'necessary' and regardless of whether we can talk about it in terms of possible worlds. Further, reasoning counterfactually in deterministic worlds requires considering worlds with different laws and a different past than our own. Maybe I'm missing something, but I'm very confident about those parts of my argument.

As I see it, the contentious part of the argument is that all modal reasoning is a species of counterfactual reasoning. I think this gets incredibly strong support from work in the psychology of modal reasoning as well as from the manner in which other forms of modal reasoning can be defined in terms of counterfactuals. Yet, if that part is right, then it seems clear that any concept that does not allow us to engage with our capacity to reason counterfactually does not allow us to reason modally and therefore is not a modal concept.

Consider the claim, "It is metaphysically necessary that P is true in the actual world." Is this a modal claim? At first it might appear to be so, after all, it refers to metaphysical necessity. Nevertheless, the claim means nothing more than P is true. Clearly, the sentence P is true is not a modal claim. So, neither is the initial claim. There are actually an infinite number of ways to merely claim that P is true while employing the terms 'necessary' and 'possible'. Just because someone employs modal language in developing a concept does not mean that the concept produced is itself a modal concept.

Charles, as everyone else has said: great post.

One big issue about ability claims concerns a value side of the use of that term. Generally the use of that term connotes something positive (look up synonyms), though it is certainly not a contradiction to say that someone was able to deceive or harm someone else. But in a metaphysical sense of the positive attribution of a property to bring something about (that might have further axiological results as being labelled good or bad), such as being able to choose, do, falsify, etc. is an affirmative value of power. What seems to me to be one big part of the question bandied about here is where to locate the relevant axiological claim about the power itself: why are counterfactual claims valued as assertions of power as opposed to categorical ones (Lewis versus Austin, e.g.)? Why should one kind of metaphysical claim be valued over the other as the right claim?

Note for instance that Chandra's example has force against iffy accounts iff (irony here) quantum phenomena are actually metaphysically probabilistic. If some Bohmian account is correct then that drags counterfactuals back into the picture involving deterministic laws. But say that the indeterministic account of the photon's behavior is right. Would that settle some deep question about the photon's ability to go through either slit? Might I not say that the photon has indifferent metaphysical opportunities to go either way, even though by this metaphysically grounded probability there was no intrinsic ability to go one way as opposed to another (as manifesting a form of control)?

The question of what values might lurk in diverging ability claims needs airing and ironing out. I think this unresolved issue lends itself to a lot of talking past one another.

Alan,
I'm not sure what to think about the value aspect of ability terms. Frequently, it seems to me that the ability to do otherwise becomes a liability of doing otherwise. The reason I am interested in the ability to do otherwise is that I am concerned about what sort of control we can have over our world.

I don't see how Chandra's example has force against accounts that allow us to consider worlds with different laws or a different past than our own. It may be that in certain cases the closest worlds to our own have the same laws and past as our own. That seems to be what occurs when we think of the double slit experiment occurring in a laboratory. Yet, it is a mistake to think that the reasoning process we are going through when considering the double slit experiment is one where we hold the laws of nature and the past fixed. After all, we reason the same way about the double slit experiment when we consider it in a simplistic model when all we focus upon is a world that contains just the photon and the experimental apparatus. Yet, when we consider the simplistic model, the laws of the world being considered differ from our own (at least according to all the major accounts of laws of nature). So, it is a mistake to think that when we reason about the double slit experiment we hold the laws and the past fixed. Instead, we reason the same way we always do when we reason about modality. We reason counterfactually. In fact, thinking about the double slit experiment provides excellent reason to deny that any account of the ability to do otherwise that does not allow us to consider worlds with different laws or a different past than our own is a serious contender.

Hi Charles,

Is this the argument?

1. When we reason counterfactually, we don't hold fixed the past and the laws.
2. Modal reasoning is a species of counterfactual reasoning.
3. So, when reasoning modally, we don't hold fixed the past and the laws.
4. Incompatibilist accounts of the ability to do otherwise require us to hold fix the past and the laws when determining what an agent is able to do.
5. So, incompatibilist accounts of the ability to do otherwise require us to refrain from reasoning modally when determining what an agent is able to do.
6. But ability claims are modal claims.
7. So, incompatibilists accounts of the ability to do otherwise are false.

Let's agree that (4) and (6) are obviously true. And (2) seems plausible enough to me. But what's the reason to believe (1)? The motivation you give is the simplistic world case. But I don't think I see how that case supports the broader claim about reasoning counterfactually in general. Is the counterfactual question: "If the world only contained one electron with no forces acting on it, and the electron is in a certain place at t1, where will the electron be at t2?" If that's the question, it seems like you've built the differing laws into the antecedent. How exactly does that show that questions about *what would happen if an agent were to phi* don't require holding fixed the past and the laws? Isn't the very question at issue whether we should build the laws and the past into the antecedent of that counterfactual?

Neal,
Thanks for the comment...my concerns on this topic are relatively new and your post is very helpful on getting me clearer on what I want to be saying.

My reasons for discussing the simplistic world is not to defend 1 but to defend 2. It seems natural for someone to think that when we are reasoning about nomic necessity and nomic possibility we don't reason counterfactually. Instead, a natural, but wrong, way to think about nomic necessity is that something is nomically possible if and only if it occurs in some world with the same laws of nature as our own. This account is inconsistent with all the major views of laws of nature. So, even reasoning about nomic necessity is a species of counterfactual reasoning. In almost all normal cases the closest worlds when considering a counterfactual in indeterministic settings will be worlds with the same laws and the past as our own. Also, notice that the antecedent of the conditional I explore in the simplistic world does not say anything about the laws being different, nor is the antecedent nomically impossible. The point of the case is that if reasoning about nomic possibilities required holding the laws of nature fixed, then anything follows from reasoning with simplistic worlds. So, even reasoning about nomic necessity does not require holding the laws fixed. Also, note, that on all major accounts for evaluating counterfactuals 1 is true even in indeterministic worlds. Maintaining the same laws and the same past as the actual world counts highly in favor for the similarity of a world, but nomically possible antecedents can cause us to consider worlds with different laws of nature than our own.

While I think that 1 is true, I don't think I need it for the argument. Instead, I would claim 1* Counterfactual reasoning in deterministic worlds requires that we don't hold the past and the laws fixed. and 3* modal reasoning in deterministic worlds requires that we don't hold fixed the past and the laws. Now, we are trying to address the question of whether determinism is compatible with the ability to do otherwise. So, we need to think seriously about modal properties in a deterministic world. The incompatibilist needs to argue that we cannot do otherwise in a deterministic world, the compatibilist needs to argue that we can do otherwise in a deterministic world. Yet, if the incompatibiilst insists that we hold the laws and the past fixed, the incompatibilist has produced an account that does not allow us to reason modally about deterministic worlds. So, they cannot derive the result that the ability to do otherwise is inconsistent with determinism. I certainly think that 1 is true as well, but I don't think it is needed for the argument.

Charles,

That helps, thanks. My probably predictable follow-up is to ask why we should think that 1* is true, in all generality. Perhaps there are some cases of counterfactual reasoning that require us not to hold fixed the past and the laws. But why think that all cases are like that? In particular, why think that when we are wondering *what would happen if an agent were to phi*, we must not hold fixed the past and the laws? Incompatibilists are presumably going to think that reasoning counterfactually about human action is a special case of reasoning counterfactually.

Charles,

The issue is very subtle, and I might have misspoke.

A critical criterion for ability, I think, is not "you actually do X." Of course, you can do X and have the ability to do not X. I think we agree with at least this much.

The criterion, I think, instead, is that "the laws of physics, given the past, don't bar you from doing X." If the laws of physics bar you from doing X, then I don't think you have the ability to do X. Because I don't think we have the ability to change the past or to change the laws of physics or to violate the laws of physics.

Now, you might be right that most people wouldn't agree with me about this concept of ability. But most people deny that determinism is true. And most people (including me?) don't have a terribly sophisticated understanding of either determinism or ability. And, more importantly, sometimes people are just massively incorrect about certain important questions - like the question of the heliocentric theory or the question of evolution.

It wouldn't be that surprisingly if people massively overstated their abilities. It's well known, in the psych literature, that people maintain positive illusions about themselves, their abilities, and their level of control. If the world is actually deterministic, then I think our ability is limited to doing what we actually do. Not because "you actually X" is a criterion for the ability to do X, but because it becomes a criterion when combined with (A) determinism and (B) a fixed, unchanging past. That simply is what the physics tells us. People who define ability in terms of what deterministic agents do in counterfactual worlds are not, I think, sufficiently appreciating the consequences of determinism.

Hi Charles,

I agree with Chandra, this is a really fascinating post and I agree with you. I wish to put attention on your claim: "if you were asked to consider whether the ability to do otherwise is compatible with determinism without employing your capacity to reason morally you would be unable to do so". I think so, too. This claim is have a narrow relation to the use of ordinary language that is related to the common sense, too. How can your learn other language without explores ability of doing so? How can we type a different email than we type now without the ability to type any email? How can we think about common sense without think about non common sense, too? This not means that non common sense is in the backgrounds of thinking common sense, but that thinking something requires ability to think otherwise than we can do it, and this suggest an appeal to fiction, i.e., a type of accessible fiction.

I agree with you about counterfactuals and fiction. Why all counterfactuals should be inaccessible fictions? Actually, is possible to say, "All counterfactuals are inaccessible fictions" but, Is this an empirical claim? If it is, someone should present sufficient supporting evidence. If we haven't sufficient supporting evidence, then the claim that "all counterfactuals are inaccessible fictions " is an hypothesis. But, in doing so, we need the ability to think otherwise, i.e., the assumption that, as hypothesis, we may arrive to the proposition "some counterfactuals are inaccessible fictions". In effect, “all counterfactuals are inaccessible fictions” is a proposition that did not correspond to our empirical experience, is to think in other way than we can do it.

Well, if " All counterfactuals are inaccessible fictions" isn’t verifiable, then this claim is an empirical inaccessible fiction itself, while "some counterfactuals are inaccessible fictions", actually, is an empirical accessible fiction, or hypothesis, or assumption. Then, to say "All counterfactuals are inaccessible fictions" is a claim from the non common sense that is impossible to be supported with sufficient evidence. We can think propositions with impossible empirical evidence, but isn’t mean that these propositions be accessible.

So, we can think inaccessible fictions like "all counterfactuals are inaccessible fictions", and accessible fictions like "some counterfactuals are inaccessible fictions". We cannot prove the former proposition, while the second one yes. And likewise, we never cannot prove that common sense always fails. From my point of view, this isn't in disagree with your claim that incompatibilism can't offer a strong account for the ability to do otherwise, because all this reasoning presuppose the ability to think types of counterfactuals, both empirically accessible and inaccessible. Then, empirical thinking is a counterfactual process which is focused on the assumption of necessity based on experience and evidence.

I think that empirical counterfactual reasoning is a species of modal reasoning. Inaccessible fictions are impossible to present positive supporting evidence. These are empirically impossible claims, but we have an evidence of their impossibility that is found in their impracticable empirically mode to be verifiable. These claims, empirically, cannot be otherwise. We have this positive evidence: there are unverifiable claims. Then, to think about an empirical counterfactual impossibility brings us to a theoretical model of reasoning: some claims are empirically impossible (can't be otherwise), while other claims are empirically possible (can be otherwise). And this itself can teach us the way of empirical reasoning, and doesn’t be against common sense that, of course, could fail, like the scientific reasoning, but this doesn't made it constitutionally iffy.

For that, when a quantum physicist is talking about a photon in a slit experiment, and he says, “the photon went left, but it could have gone right”, beyond his supposition of a fixed past and a probabilistic laws, he assumes a empirical reasoning, i.e., that laws forged by empirical verifications could be different in not yet assumed conditions. Then, laws, although probabilistic, implies a type of counterfactual reasoning, an ability to think counterfactually. Laws could be otherwise, because we can obtain more data about their validity in not yet assumed conditions. They could be otherwise. Then, the scientific claims implies a counterfactual reasoning, which supports probabilistic laws.

Neal:

Assume that determinism is true and that I don't raise my arm. Consider the counterfactual "were I to raise my arm, I would have moved my limbs". Do you think that it's reasonable for the incompatibilist to think that the closest A-worlds are metaphysically impossible? This strikes me as a rather strange view, for exactly the reason that it would be strange to think that the closest A-worlds are metaphysically impossible for "were the boulder to fall, it would have been ..." (where determinism is true and the boulder doesn't fall). I would think that the default position is that the ability counterfactual is just a perfectly ordinary counterfactual. And we would need a very good argument for thinking that the seemingly ordinary ability-counterfactual is to be treated so differently than "had the boulder fell, it would have crushed the car."

Charles:

You're charging me with making a "serious mistake", but I just disagree with you about the word 'modal'. To clarify something: what I think is analytic is not The Simple View itself, but the claim that The Simple View is a modal view. I think philosophical usage of the term of art 'modal' settles this, but it's just a terminological quibble and is irrelevant to the substance of your post.

Kip,
I agree that there is something attractive about the idea that being able to do otherwise requires that we do otherwise in some world with the same laws and the same past as our own. I'm not so convinced by my compatibilist position that I have lost my incompatibilist intuitions along the way. If my argument works, then we ought to abandon accounts of ability that require the laws and the past be held fixed...even if we do find them intuitive. Yet, I certainly at times have the same intuitions you mention here. As I see it, the important thing to do is try to figure out what the cost of maintaining those intuitions is by exploring arguments.

Charles,

Thanks for your mind-stretching and fascinating post. I don't understand why you regard Neal's premise 2 as important to your argument. Who cares if the "natural, but wrong, way to think about nomic necessity : that something is nomically possible if and only if it occurs in some world with the same laws of nature as our own" is true? Same laws does not imply same laws + same past. And I think the Natural But Wrong view may well be true.

Why do I defend the Natural But Wrong view? Consider Lawrence Krauss's explanation of the quantum origin of the universe: http://www.nytimes.com/2012/03/25/books/review/a-universe-from-nothing-by-lawrence-m-krauss.html?pagewanted=all&_r=0 . Ignore the needless philosophical controversy caused by Krauss's sloppy use of language-that-philosophers-think-they-own, and focus on the fact that spacetime itself, on this theory, emerges according to quantum natural law. Here we have an empty universe, in a very deep sense without even spacetime, with natural laws! I have no idea if the best physics would say so - that's above my pay grade. But if some smart and knowledgeable physicists take it seriously, as they evidently do, then philosophers who fail to make room for it in their accounts of natural law are in peril.

Martin,
I think there are things that we can know without empirical evidence, so I wonder if some of what you claim in the comments is significantly stronger than anything I would accept. But, I think the important point you are making is that too frequently philosophers ignore empirical evidence when it exists. Philosophers tend to think that modal reasoning consists in conceiving of different ways our world might have been. Since conceivability is a guide to possibility, we can construct all the metaphysically possible worlds purely from conceivability arguments. From there, we construct other modal notions by placing restrictions upon which possible worlds we consider. I'm not sure how far back philosophers have had this conception, but it seems to be employed constantly throughout modern and contemporary philosophy. This conception of modal reasoning makes it seem that reasoning about metaphysical and nomic necessity is prior to counterfactual reasoning and that counterfactual reasoning is grounded in our capacity to reason about these modalities. One of the nice things about developments in science is that our conception of how modal reasoning works no longer has to be something we grasped at a priori. Instead, there is serious empirical work on modal reasoning. Taking it seriously reverses the order. Counterfactual reasoning is not grounded in reasoning about metaphysical and nomic necessities. Instead, reasoning about nomic and metaphysical necessities is grounded in counterfactual reasoning. I don't want to claim that there is nothing that we can learn a priori. But, when there is relevant empirical evidence philosopher's ought to pay attention to it.

Neal,
I agree with Nate's response. Suppose the ceiling is two inches above my head. Now, consider the following counterfactuals from a deterministic base world where I did not raise my hand. 1. If I raised my hand I would touch the ceiling. 2. If I raised my hand I would not touch the ceiling. My hunch is that 1 is true and 2 is false. Yet, if we need to evaluate the counterfactuals purely in terms of worlds with the same laws and past as our own, then both conditionals are trivially true.

Nate and Neal,
I retract my claim that the simple view is not a modal view. I wanted to thank both of you for getting me to rethink how I want to present this argument. So far, the only thing I wrote on this argument is this post so the idea is still very much in it's infancy. Yet, I think I can still get the main point I am trying to develop in the argument while saying something weaker.

In fact when Neal said that he was about to give me the predictable follow up I expected something different.
I don't think the mere fact that we employ modal language is enough to show that it is a modal view. Certainly, the surface structure of the standard expression of the view appears modal, my claim is that the deep structure is not modal. Since whether the concept is a modal concept does not depend upon how it is expressed it is the deep structure we ought to concern ourselves with for determining whether a concept is modal.

What concerns me is Neal's pressure on 1 caused me to realize it may be very important that I really only mean to be defending 1*. Yet, unless I defend 1, I cannot claim that incompatibilist views are not modal views. After all, all the normal reasoning we engage in when considering the these views will result in worlds with the same laws and the same past as our own being the closest worlds. I do think more simplistic worlds can be employed to defend 1, but that's a significantly more contentious claim than I hope I'll have to defend. So, since I admit that incompatibilist views allow counterfactual reasoning (or something similar enough) in indeterministic worlds my argument is not strong enough to deny that they are modal views.

Yet, I think the problem is that I overstated my case. What I ought to have claimed is that these views are irrelevant for determining whether we are able to do otherwise in a deterministic world. Claiming that we are able to do otherwise, or denying that we are able to do otherwise, are both modal claims about deterministic worlds. In order to figure out which of these claims is accurate we need to be able to employ modal reasoning. Yet, modal reasoning is a species of counterfactual reasoning. Since counterfactual reasoning in a deterministic world requires that we consider worlds with different laws or different pasts, the simple view prohibits us from modal reasoning about deterministic worlds. Since it prohibits from reasoning modally about deterministic worlds it is useless for telling us whether we can do otherwise in a deterministic world. So, I guess one way of putting the worry is that incompatibilist accounts of the ability to do otherwise are silent on whether an agent can do otherwise in a deterministic world. Incompatibilist accounts are not incompatibiilst accounts.

Paul,
I am not arguing that modality is grounded in counterfactuals. Counterfactuals are aspects of language, and I am tempted to think there are modal aspects of the world. What I am claiming is that modal reasoning is grounded in counterfactual reasoning. Why think this, because there is mounting empirical evidence for it coming from psychologists. Why should we take 2 seriously, because psychologists working on modal reasoning not only take it seriously but keep mounting evidence in its favor. While I am tempted to take very seriously what physicists say when I consider what grounds modality, it seems like we ought to take just as seriously what psychologists are telling us about modal reasoning. Further, it is very important to examine exactly what scientists are claiming. Scientists are amazingly skilled at producing fascinating empirical results. Interpreting these results, takes argumentation and conceptual clarity. It is at that point that the skills of a philosopher become very valuable.

Here's one way to make sense of the claim that there are laws in the world before there are any objects. This certainly is true on the Humean account. On the Humean account laws are merely the axioms of theorems of a complete description of all events throughout history that provides the best balance of simplicity and strength. Prior to any entities existing in our world, the laws were exactly as they are today. Yet, claiming that the laws were as they are before any entities existed is no to say the same thing as that the laws would have been the same had the world only consisted of a single stationary electron throughout all of history.

The comments to this entry are closed.

Books about Agency


3QD Prize 2014: Marcus Arvan