Does Consciousness Matter?

These days, I am by trade a neuroethicist (no, I don’t know what that means either). This brings me into day to day contact with scientists. When I mention that I am a free will specialist, they tell me that Libet showed there could be no such thing. Of course we all know that Libet did no such thing. They are a whole bunch of criticisms out there in the literature which are each sufficient on their own to show that Libet did not succeed in that. Al Mele has shown that there is no reason to identify whatever mental event that subjects choose in these experiments with a conscious intention; Daniel Dennett has shown that fine-grained judgments of simultaneity of the kind Libet requires cannot be reliable; Owen Flanagan has questioned the ecological validity of the experiments… and so on. But what is the motivation for refuting Libet? Elsewhere, I have argued that despite the problems with Libet’s experiments, the results are precisely what we should expect, given any kind of broad physicalism. Intentions are neural events; the impulses from the event of intention formation ought to be registered later than the event itself (given that it takes time for impulses to travel across the brain). So: why the effort put into refuting an experiment that, for all its genuine problems, is aimed at demonstrating something we ought to expect?

Put another way, why think that consciousness matters for free will/moral responsibility? I actually do think that consciousness matters, though I doubt the fine-grained timing of conscious states matters. But this is not a view that we ought to simply assume. Indeed, it is a view that is now increasingly rejected. Nomy Arpaly, Tim Scanlon, Angela Smith and George Sher have each published books or articles rejecting the need for agents to be conscious of the reasons for which we act in order to be responsible for our actions. Christopher Suhler and Pat Churchland have argued for a similar conclusion, on very different grounds. If we are swayed by them, the Libet argument won’t even have a prima facie plausibility as a threat to free will.

 So my question is: why think that consciousness matters for free will/moral responsibility? Most of the arguments that occur to me off the top of my head seem question begging. One is an argument from the deliberative perspective. We might think that the deliberative perspective is bounded by what is in awareness, and argue as follows: my actions express the quality of my will. But my quality of will is a function of how I see things when I deliberate. If I fail to see that I harm you, then I do not express ill-will toward you, nor even indifference. So what I am (directly) responsible for is a function of how things stand in my deliberation, and this is a function of what I am aware of. The problem with this argument, obviously, is that it assumes, without argument, that it is only in consciously responding to reasons that I express the quality of my will. Similar arguments might be constructed turning on a Frankfurtian identification view, or a control view: they also seem to beg the question by assuming that conscious identification is what matters, or conscious control is what matters.

 I myself am attracted to an argument that Sher considers, and rejects, in Who Knew? The argument goes like this: we might think it is unfair to hold agents responsible for failing to take into consideration facts outside the searchlight of their awareness because in blaming an agent responsible for an action, we commit ourselves to the claim that the agent should have done something other than what they did , but he could intentionally have done something else only if the reasons we believe he ought to have considered featured in his deliberative perspective. Sher claims that this argument fails because it commits those who hold it to the claim “that no demand can influence any agent’s deliberations unless the agent has access to the demand”, where the sense of access is one to be understood with reference to the searchlight of consciousness. As Sher points out, this demand is false: our deliberation is constantly influenced by normative demands of which we are not conscious. But does the claim that it is unfair to demand that agent’s respond to demands outside the searchlight, really commit us to denying this? That our behavior is included (all the time) by facts outside the searchlight of awareness does not show that we can fairly be blamed for failing to respond to these demands.

 Suppose that d is a demand to which I ought to respond, but that d is outside of my searchlight. Nevertheless, unconscious mechanisms are responsive to d (perhaps this can be shown behaviorally, via priming). Given d, I ought to Ф, but I fail to Ф. Two claims: first, it may be true of me that I failed to Ф because d was outside the searchlight; second, I could not deliberately settle whether I responded to d.

I suppose Sher would concede the first, but he may not be moved by the second. I think attention to how facts which fall outside the searchlight influence behavior will help motivate it. Demands like this cannot enter into the content of our intentions (where an intention is the content of X in the phrase “… in order to X”). When we are influenced by such demands, either we act for a veridical intention (as when we act in order to comply with an experimenter’s instruction  < to post the card through the slot >) or we confabulate an intention (as when we act to choose the best item, without realizing that there is no best item). In either case, we remain unaware of the content of the demand, and of how it influences our behavior. It influences us either by influencing the content of our deliberative perspective, without becoming the content, or by stealth. It may be helpful to us in achieving our goals, or it may be detrimental; in neither case does the demand come up for review, and we have no opportunity to endorse or moderate the influence it has on us.

 At very least, Sher must concede that there is a difference between the way in which demands influence us when they are inside and outside the searchlight. These differences seem morally relevant, in that only when the demands are inside the searchlight can be endorse, question or reject their influence on us.

 That’s me. What about you (dear reader)? Do you think consciousness matters for moral responsibility? Why?

 

Comments

[George Ortega]

You state your thesis concisely at your ending as follows: (For clarity, I’ve replaced your term “searchlight” with it’s formal synonym, “consciousness,” corrected your typo. “be endorse” to read “we endorse,” and restored symmetry for the word “difference.”) --

“…there is a difference between the way in which demands influence us when they are inside and outside of consciousness. This difference seems morally relevant, in that only when the demands are conscious can we endorse, question or reject their influence on us.”

From the perspective you describe, consciousness does, indeed, matter for free will/moral responsibility, however it is not consciousness, per se, but rather the causal process that governs both conscious and unconscious thought that more generally explains why free will is an illusion and why moral responsibility is not correctly attributable to human beings.

In other words, while ONE determining factor for free will and attribution of personal moral responsibility is, in fact, whether or not we are conscious of the determinants of our behavior, as you rightly assert, THE MORE GENERAL factor (more general because it applies to both conscious and unconscious thought) is the causality that govern both.

You, of course, acknowledge this causality by describing human behavior as arising from “demands.” Once you establish demands as CAUSING human behavior, you, by extension, acknowledge that those demands have their own causes, and that those causes have their own preceding causes. Obviously the causal history that emerges spans back in time to before the agent had either a conscious or an unconscious mind, and, indeed, even before the person possessing the self with those two minds was born.

So while you are, of course, correct that consciousness does matter for free will/moral responsibility, it matters most essentially because it, like all else, is governed by causality. I redirect the matter to causality not because your consciousness imperative is in any way incorrect, but rather because omitting causality as the essential consideration unnecessarily complicates and confuses the matter.

[Eddy Nahmias]

Thanks for posting these interesting questions Neil. I'll try to say something later about why I think consciousness is essential for free will. But for now, I'll just post something relevant here I was planning to post as a thread. It's a recent Science article about how little conscious processes do (yes, yet another of these, and of course it starts by paying homage to Libet). The abstract is here (hopefully you can get full access if you want it): http://www.sciencemag.org/cgi/content/abstract/329/5987/47

One interesting thing to note is that this review article is titled "The Unconscious Will" but it never mentions free will or discusses how this research threatens free will. Yet the two media discussions of it I've seen so far both present the research as a threat to free will, including Time's article titled "Think You're Operating on Free Will? Think Again": http://www.time.com/time/printout/0,8816,2000994,00.html

I think these media moves represent a pretty natural (folksy) move to make from no causal role for consciousness to no free will, presumably because it's pretty natural (folksy) to think free will requires some causal role for consciousness (but what role?).

[Josh Shepherd]

Thanks for the links Eddy, and for the post Neil. These issues have absorbed a good bit of my summer reading.

Regarding the question ‘why think that consciousness matters for free will/moral responsibility?’ it seems to me that we should have some idea of how consciousness might contribute to action before we worry overmuch about whether it matters. And on this issue, Ezequiel Morsella has some good stuff (links to a lot of his publications on his website), and I’m also fond of Bongers & Dijksterhuis’s article in the Oxford Handbook of Human Action. The issues here get fairly involved, but some neuroscientists are making decent progress towards a story of how consciousness figures in at least bodily actions. How consciousness figures in mental actions is a thornier matter. Al Mele's recent book and some recent articles treat 'conscious deciding,' and these are certainly worth reading.

Of course, some think that consciousness is epiphenomenal, and some think that while consciousness is important for overt action, mental actions like deciding are always performed by sub-personal mechanisms. These issues are tough. Here I’ll just say that it strikes me as unlikely that consciousness, which seems to be a computationally expensive adaptation, would evolve for no reason, and would not contribute in important ways to both overt and mental actions.

Regarding Neil’s point that many arguments for consciousness’s relevance flirt with begging the question, it does seem as though the tie between consciousness and FW/MR is in some sense fundamental. That is to say, it might be that if an agent does not experience some action-relevant conscious states, moral responsibility is not attributable to that agent (is such a non-conscious system even an agent?). If this is right the question becomes, as Eddy noted, just what (causal) role(s) conscious states play for agency.

Regarding Sher’s argument, Neil, I agree that whether certain demands are within or beyond the searchlight of awareness is morally relevant, and that whether some demand was within or beyond an agent’s searchlight in a specific case might make a moral difference. But of course the difference might be complicated. We can, as you know, be responsible for unintentional actions and omissions. If the fact that I’m about to drown is beyond your searchlight as you stroll the beach with your pet marmot and observe me flailing about, this might say something about the kind of agent you are in general, regardless of how we judge your omission to save me from drowning.

[Manuel Vargas]

Hi Neil-

This is a really interesting question.

I wonder if the right answer is going to be similar to the one we might give if someone asked "Is consciousness required to be a good and loving partner of someone else?"

In reply to the last question, I can imagine one answer being something like this: "well, strictly speaking, it might not be required, but in the ordinary world it sure is useful. We might be able to imagine a good and loving partner that unreflectively had all and only whatever (non-conscious) dispositions and the like that we associate with being a good and loving partner. And, indeed, much of the time in the real world good and loving partners act unreflectively or unconsciously in ways that are partly constitutive of their being good and loving. However, short of the best of all possible worlds, consciousness is going to give you a way to compensate for our all-to-frequent shortcomings in the good-and-loving dispositions department."

So: although we might be able to imagine responsible agents without consciousness, in the world we find ourselves in consciousness is a big help in the "being responsible" business.

(Lots of details would need to get filled in, of course, but I imagine you can see how the story might go, and I'm betting you have thoughts about why this story won't work!)

[Dan Speak]

I think this is in the ballpark with Manuel's thoughts above. My suspicion is that consciousness is essential to responsibility because of its (perhaps contingent) roll in establishing what speaks for me among all the other stuff in my psychic stew. Some of those unconscious mechanisms the social scientists are uncovering have my agential authority and some of them don't. In Frankfurt's terminology, I identify with some and not others. Conscious intentionality is one really profound way of identifying myself and stopping the otherwise endless reverberations of desires, thoughts, impulses, subconscious execution mechanisms, etc.

[Neil Levy]

Thanks for the links, Eddy and Josh. Ezequiel Morsella is new to me. His work seems to give an empirical backing to Global Workspace Theory, which is something I'm attracted to. Of course Dijksterhuis has lots of relevant stuff, but it needs to be read in conjunction with two recent (perhaps still unpublished) review articles by Baumeister and colleagues on the functional role of consciousness; they find lots of evidence for its causal role as an information integrator. I don't see what's exciting about the Science article. John Bargh did similar work a decade ago. What am I missing?

Dan, I am sympathetic to the claim that only some states speak for me, and to the claim that consciousness plays a role in isolating the relevant states. But why think this? Here's my story. Suppose the global workspace theory is true. Then conscious states typically reflect more of my values and attitudes than unconscious. I also think that endorsement itself causes a propagation of relevant attitudes. So conscious attitudes are better reflective of my practical identity than unconscious.

[Tom Clark]

I think Josh has it right that the first order of business is to see how consciousness might contribute to action. This requires having some idea of what consciousness is and how it relates to its physical and functional correlates. Certainly the neural processes *associated* with consciousness (for instance those constituting the global workspace) seem necessary for certain sorts of higher level control and flexible behavior, but it isn't clear that consciousness itself, as distinct from those processes, is playing an additional necessary role in controlling action. If consciousness isn't distinct from those processes, then it isn't playing an additional causal role. But if it is, then the problem of mental causation arises: how do specifically mental states, e.g., phenomenal feels and/or reportable intentional content, add causal power to what their physical correlates do in controlling action?

Of course if it turns out that consciousness per se doesn't have its own special control function, that doesn't mean the end of responsibility, it only means that complex social biological mechanisms like us are the sorts of things that can and must be held responsible (http://www.naturalism.org/glannon.htm ). Consciousness then becomes simply a subjective marker for the operation of certain control capacities and behavioral abilities that make us legitimate targets of responsibility practices. Why legitimate targets? Perhaps it’s (at least partially) because it’s only those capacities and abilities that enable us to be responsive to the socially and linguistically communicated prospect of being held responsible.

Btw, Josh, I'd say that it wasn't consciousness that evolved, but rather its neural correlates. It's *those* that are computationally expensive, not the subjective feel that for some reason accompanies them. Pain don't weigh a thing: http://www.naturalism.org/privacy.htm

[Neil]

Tom, you seem to begging the question against an identity thesis. If consciousness *is* a certain set of neural properties (or whatever), then *consciousness* evolved. Even on broad physicalism, the claim seems okay: compare 'we mustn't say bipedalism evolved; rather, a certain gene cluster evolved'. On non-reductive physicalism, it will be true that consciousness plays a certain role and false that the neural correlates play that role; on reductive physicalism, it will be true that consciousness plays a certain role just when its neural correlates play a certain role. You don't get to solve the problem that easily!

[Josh Shepherd]

You're right to hone in on my loose language, Tom. But I agree with Neil. Since I'd rather not debate issues like whether to identify consciousness with neural events or processes (I'd also rather not engage in the related debate pressed recently by Alva Noe and others on whether the search for neural correlates of consciousness is misguided -- i.e., whether consciousness requires broader processes, albeit ones which necessarily involve the brain) I spoke loosely about the evolution of consciousness. Eventually precision will be called for here, but much of the precision will depend on empirical answers we don't yet have.

It seems that we can focus for now on the functional role(s) of mental states which involve consciousness. To do so, of course, involves one tangentially in debates about the definition of consciousness. Such is life.

[Tom Clark]

Neil,

I'm not sure I'm begging the question against an identity thesis, only showing its implications for whether consciousness per se matters. On the non-reductive view, consciousness does something over and above what its neural correlates do in controlling action. The question then is how does consciousness contribute it's extra causal/control effect on how limbs and lips move, and I haven't seen any good hypotheses about that. If consciousness is identical to some set of neural correlates, it doesn’t contribute additional control, it’s simply the subjective private aspect of being that sort of control system. Since the nature of consciousness is an open question, seems to me the best we can do at the moment is consider whether its neural correlates matter for responsibility, and the pretty obvious answer is that they do.

[Noah Te Stroete]

I am wondering if someone could direct me to some philosophical/psychology papers (hopefully available on philpapers.org because I don't have access to any journals) that explains why free will (whether its of the libertarian variety or compatibilist) is still a possible true fact about the world. It seems that certain philosophers here are pretty adamant that determinism is true and that it is incompatible with free will. Personally I have not been swayed in either direction in the free will/determinism debate, but I would like to read about it some more so I can understand why the possibility is still on the table. Anyone's help is much appreciated.

Regards,

Noah

[Chandra Sripada]

Hi Neil,

Thanks for the thought-provoking post.

I believe the following two claims:

1. There is a strong intuition that when actions have their source in psychological states that are not conscious, the agent is not responsible, or responsible to an importantly lesser degree.

2. There is no essential connection between consciousness and moral responsibility

I am not an intuition skeptic (I believe intuitions reveal something about essential relations between concepts). So for me (1) and (2) are in tension. I resolve this tension by positing a ‘hidden variable’ that our intuitions are tracking. I buy the Arpaly/Smith/Scanlon ‘aretaic’ view of responsibility, so here is how I see the hidden variable playing out:

3. When an action has its source in psychological states that are fully conscious, the action is more likely to reflect the agent’s basic evaluative standpoint, i.e., the action is more likely to be attributable to the agent (Note: there are important exceptions to this principle, but it holds as a general rule)

4. An agent is moral responsible for actions that are attributable to the agent.

In short, the tension between (1) and (2) arises because of contingent co-variation between actions that arise from conscious states and actions that are attributable to the agent (in the aretaic sense). But the take home point is that (2) can and should be maintained.

The foregoing explanation can be empirically tested. Of course, you can not directly ask subjects ‘What are your intuitions tracking, sir?’. But you can record patterns of covariation between elicited judgments using surveys presented to hundreds of subjects. The testable hypothesis that follows from the explanation is the following: In cases in which the sources of an agent’s actions are unconscious, people’s intuitions about diminished responsibility will be significantly correlated with judgments about the attributability of the action to the agent. Or, perhaps I am wrong and people’s intuitions are driven by a more ‘volitionist’ face of responsibility and thus track the fact that the agent’s deliberations are in some way suboptimal. I wonder if you think this is an empirical test worth doing?

[Josh Shepherd]

Chandra,

All very nicely put, I think. Personally, I think your suggestion is worth testing. It might also be good to run a study sensitive to my above suggestion, of whether a person 'just like us' except for the fact that they did not possess consciousness could ever be the subject of a responsibility attribution. Maybe the notion is incoherent, but it'd be interesting if folks found such a person less responsible than one whose action followed from unconscious machinations.

[John Fischer]

Chandra,

I'm not sure what you take the relationship to be between an action's "reflecting the basic evaluative standpoint" of an agent and the agent's being morally responsible for it. For instance, I believe that an agent can be fully morally responsible for a weak-willed act, such that the act does not (presumably) "reflect the basic evaluative standpoint" of the agent. That's a standard problem for so-called "normative" approaches to moral responsibility. You say that an agent is moral[ly] responsible for actions that are attributable to an agent; I note that you do not use the "iff" construction. Is this because it is obvious that agents can be morally responsible for certain actions that do not reflect the agent's basic evaluative orientation? And, if so, doesn't this leave open moral responsibility for acts that originate from unconscious sources?

By the way, you say that the approach to moral responsbility to which you are attracted is in part due to Arpaly; but, as I read her, Arpaly is strongly inclined to saying that we should in fact hold agents morally responsible for actions that flow from unconsious sources; indeed, it is one of her basic critiques of other approaches to moral responsibility that they tend to focus on fully conscious, deliberate actions.

[Neil Levy]

Chandra, contingent relations can be reliably instantiated - it might be that the relation is reliable enough that we are unjustified in holding agents morally responsible in the absence of consciousness.
Josh, the kind of consciousness in question here is an informational state, not phenomenal consciousness. Agents just can't lack thus state *globally*.

[Josh Shepherd]

Neil,
I think you're responding to my suggestion to Chandra. There I was intending to suggest something about what the folk view of the relationship between consciousness and moral responsibility might be. It might be that if someone never experienced conscious states, then the folk are unlikely to attribute moral responsibility to them, no matter what they do. Such a thought experiment is sullied by the likely possibility that a person without consciousness would be unable to perform the kinds of actions that consciousness enables -- those that require flexibility, complicated ongoing monitoring of action sequences, and so on -- especially if globally broadcast information can't but be conscious. Since many might fail to make the informational state/phenomenal state distinction you did, I was wondering whether there might be a more fundamental link in the folk view between consciousness and MR.

[Josh Shepherd]

Noah,

Since I'm not sure what papers you might have access to through philpapers, it might be helpful to you to explore the contents of the books linked on this page. Anything of the books written or edited by John Fischer would be a help to you, as would many of the papers in the Oxford Handbook of Free Will.

Also, the Stanford Encyclopedia of Philosophy has some good entries - the entry on compatibilism might explain why many think free will is still a live option, even if determinism holds.

[Neil]

John, I have never seen a prima facie plausible argument for the claim that we are responsible for weak willed actions. Instead it simply seems to be assume. Admittedly I will be hard to persuade since I am sceptical about the existence both of moral responsibility and weakness of the will. Anyway, there seems to be a difference between an act that the agent judges they ought not to perform, and performs anyway, and an action that does not reflect the agent's practical standpoint. With regard to some of the first, the mental states constitutive of the agent might have played a causal role in the production of the act, whereas that is not the case with regard to the second. If we were to find that this was false with regard to some or all akratic acts, that would be (further) reason not to hold agents responsible for them (in fact, on my account of actions performed despite sincere judgment, that is exactly what happens).

[John Fischer]

Hmm. Well, Neil, I'm not sure exactly what to say. Yes, I simply assume that there are weak-willed actions, and (by definition) they are free; further, I assume that the agents who perform such free actions are at least candidates for moral responsibility. I'm not sure what a "prima facie plausible argument" is supposed to look like here, or why we need one. Roughly, though, I assume that, given that the relevant epistemic condition is met, that agents can be morally responsible for free actions. Further, I suppose that weak-willed actions are free actions. Well--is that prima facie plausible?

I think that agents sometimes freely do what they don't think they ought (all things considered) to do. It would be a miracle, though, if I could convince a skeptic, especially on a blog.

My main point was to question the assumption that one can be morally responsible ONLY for those actions that "reflect one's evaluative perspective". Perhaps my problem is that I don't really know what the quoted phrase means.

[Neil]

My scepticism about weakness of the will is not scepticism that people don't act contrary to their best judgments: it is scepticism that the phrase 'weakness of will' and equivalents pick out something that is interesting at the functional level. So we can identify token acts about which to disagree (and that's a good start). It is, I think, always worth the exercise to try to defend a claim that everyone takes for granted: we don't, in any case, want to take the claim that akratic acts are free to be a primitive (do we?) though it is not hard to find people who define akrasia as something like 'free intentional action that conflicts with the agent's best judgment'. But your point about actions that reflect one's evaluative perspective is independent of all this, and you are of course right that a great deal needs to be said to cash out this notion. I tried to make a start on my comment above. More fully: I assume that some of the mental states of person's are more constitutive of the identity of the person that others (here 'identity' is being used in Marya Schechtman's 'characterization' sense). An act is reflective of their practical agency (as I prefer to say) iff enough of these mental states play the right kind of causal role in production of the action. Obviously that is really, really, really vague (how much is enough? What is the right kind of causal role?) I think the details can be filled in, but you are entitled to scepticism...I do think, though, that this kind of account runs orthogonal to the akratic/non-akratic distinction and can play a role in an account of moral responsibility.

[Chandra Sripada]

@Josh: Your suggestion to study Zombie responsibility is interesting! I suspect the ordinary people would have a hard time understanding such a case for just the reasons you say.

@John: I am not trying to argue that only consciously motivated actions can be free. Rather, I am trying to diagnose the source of people’s intuition that unconsciously motivated actions are not free. I am proposing these folk intuitions track aretaic factors, such as that the action does not reflect the person’s own basic evaluative standpoint and is thus not attributable to the agent. I think attributability is strongly connected to MR, but would not want to use terms like necessary or sufficient since I think you are right about the availability of counterexamples. I will have to try to get my own views clearer on how best to formulate the relationship, and re-read how other attributability theorists formulate their positions.

You are right that Arpaly would in many cases hold agents morally responsible for actions that flow from unconscious sources. But in many other cases she wouldn’t. On page 16 of Unprincipled Virtue, she says ‘It would be a mistake to endorse the sort of folk-Freudianism that automatically assigns the suppressed desire the status of ‘truer self’”, and I think she would endorse the same sentence with ‘unconscious motivation’ substituted for ‘suppressed desire’. Sometimes unconscious motivation clearly reflects the agent’s basic evaluative standpoint – in these cases the agent is responsible. In other cases, it doesn’t, and the agent isn’t responsible. I think the folk, given only limited information that a person does a hurtful act based on motivations that are not made available to consciousness, make a quick judgment that the action deviates from that person’s own basic evaluative standpoint. That is why there is an intuition that absence of consciousness diminishes MR. In short, attributability approaches could help illuminate *both* why unconscious motivations are compatible with MR in some cases, and the absence of consciousness diminishes MR in others.

Finally, briefly, I do not think weakness of will is a counterexample to attributability approaches. This is because I am a weakness of will skeptic, but for different reasons than Neil. I follow the old Gary Watson from his 1977 paper and think a weak willed action cannot be distinguished from an irresistible impulse. My skepticism (and I think the old Watson’s too) takes the form of an error theory. If weak willed action = ‘Free, intentional action that deviates from one’s own all things considered best judgment’, the term does not refer to anything.

@Neil: There is a reliable connection between ‘being a car’ and ‘being on the ground’, though nothing about the meaning of CAR entails the latter. The two are linked in that the latter is a contingent generalization about the former. If there is a relationship between ‘being morally responsible for an action’ and ‘the sources of the action being conscious’, I think the relationship is a contingent generalization in this way. From your remarks I take it you agree.

[John Fischer]

Wow--I didn't know there were so many "weakness of the will skeptics" out there. Well, that DOES make for fewer problems for one's theory. But at what cost? Perhaps it is just me, but I would think that it is an adequacy constraint on any theory of moral responsibility (and free will) that it makes room for genuine weakness of the will. But of course you skeptics are in a grand tradition going back to Socrates. Frankly, though, I think Aristotle (and the older, more mature Gary Watson) had it right about weakness of the will.

This is, of course, an ancient dispute, and not easily dismissed or resolved, especially in a context like this. I have my own views about how weakness of the will is to be analyzed, but I don't really know how original they are, and I can't develop them here.

[John Fischer]

One more brief thought (for now). Reasonable people can certainly disagree about whether there can be genuine weakness of the will (understood as intentional and free action against one's all-things-considered judgment about what's best to do). But isn't it a strike against a theory of responsibility that it requires that there be no weakness of the will (so understood)? That is, doesn't it seem like a problem--at least a prima facie problem--for a theory that it doesn't allow for a phenomenon that is widely and not unreasonably thought to exist? In general, I take it that it is at least a consideration against a theory when it requires one to take a stand on a highly contentious issue.

Of course, one might have a decisive argument there there are no genuine acts weakness of the will. But in the absence of such an argument, it seems to be at least a strike against any theory of responsibility that does not at least make room for the possibility of weakness of the will. (Granted, it takes three strikes to be out, but...)

[Neil Levy]

'Sir, we know our will is free, and there's an end on't' . That Johnson expressed a widely held view doesn't make the consequence argument any less compelling. Generally speaking, the more cognitive science I read the less impressed I am by common sense.

[Noah Te Stroete]

Are the skeptics about weakness of will or akrasia skeptical because the notion of akrasia seems self-contradictory to you? I'm just trying to understand why you are skeptical. Is the argument this:

An agent performs an act out of weakness of will if the act goes against her better judgment.

"Better judgment" means the best alternative consciously held at the moment prior to the act.

If an agent performs an act it is because it *is* the best alternative that is consciously held at the moment.

An agent cannot or simply would never act contrary to her best consciously held (at the time of the act) judgment.

Therefore, one cannot have weakness of will.

If the argument does not go like this, then how does it go?

Personally, I have experienced weakness of will. I have done things out of a strong desire despite my better judgment. Perhaps the skeptics are saying that one doesn't really have a choice in the face of these irresistible desires? If so, how is this different from a causally determined act that is incompatible with free will?

[George Ortega]

John,

Your argument implies that a weak willed act is to punished, while a strong willed act is to be rewarded. These judgments can ONLY arise by viewing the acts from a causal perspective; otherwise the judgments would be arbitrary. Bad character, or some manner of immorality, causes the weak willed act, while good character, or some manner of morality, causes the strong willed act. But once you rely on a causal description of ANY phenomenon, including a human act, you are accepting causality as the complete process responsible for the change that manifests. Any alternative becomes a self-defeating argument.

You accept that a human being caused the bad character, or immorality, that caused the weak willed act. I presume you would not object to extending this causal chain beyond the weak willed act to its effect, and then further extending the causal chain to what that effect subsequently causes, onward into the indefinite future. But you object to causality moving back beyond a human being’s will at the exact moment of the act.

Your own argument for free will, however, defeats your case. Let’s say the effect of a weak willed act is a second weak willed act. A person knows he should not steal, but he does anyway. His having stolen causes him to hide from the law even though he knows he should turn himself in. Your forward moving causal perspective, upon which the justice of your punishment relies, holds that this second act was caused by the first act. But you also assert that this second act was freely willed (you hold it a weak willed act, “by definition.”) Causality, and logic, do not allow for an act to be at once caused and freely willed.

[Neil]

Serious question Noah: how do you know when you are suffering from weakness of will? When someone says that they believe that lying is wrong, but lies habitually, we often are tempted to say 'you don't really believe that lying is wrong'. People can be wrong about what they think they think. When I act despite thinking that I believe that I ought not, I have two pieces of evidence for what I really believe. Why should I prefer the one over the other? This gives us grounds for a traditional kind of scepticism about WoW. Not my grounds, as it happens:my grounds are set out here:

http://oxford.academia.edu/NeilLevy/Papers/122186/Resisting-Weakness-of-the-Will

[John Fischer]

Ok, my all-things-considered judgment is that I shouldn't comment again on this post, due to diminishing marginal returns, and so forth. But here goes (and no one has a gun to my head)!!

George: As far as I can see, my argument does NOT entail that weak-willed acts should be punished while a strong-willed act should be rewarded. I didn't give much of an argument, but I don't see why the considerations I invoked would have this consequence.

Neil: That was a bit of a low blow, if I may say this (between friends). Of course I know that common-sense is problematic, and you surely know that I take the Consequence Argument extremely seriously--indeed, I am inclined to accept its conclusion, despite also emphasizing that common sense has it that we are often free to do otherwise (and we can't rule out causal determinism).

If one were being a bit fairer to my position (and I realize this is a blog,etc.), one would note that I have always emphasized a "wide-reflective equilibrium" methodology. On this sort of approach, sometimes common-sense loses out (or at least gets refined and adjusted).

But frankly, whereas I see a strong argument in the Consequence Argument, I do not see a similarly compelling argument that there cannot be weakness of the will. If I did, then presumably I would adjust my pretty strong view that genuine weakness of the will is indeed possible (and lamentably prevalent, to boot).

So, Neil, mere invocation of the lovely quote from Samuel Johnson does not touch my position. But surely you knew this! Surely you knew that I depart from common-sense in some pretty radical ways--ways which make me arguably one of the more "revisionist" theorists on the contemporary scene, to borrow a term from Manuel Vargas. But unlike some, I need a reason to depart from reliably replicable considered judgments.

Finally, here's a question for Chandra and Neil and any other akrasia-curmudgeons out there: do you also believe that there cannot be genuine cases of agents freely acting against what they believe they morally ought to do? I'm just curious; I realize that this question is different from the question of whether an agent can freely act against what she takes to be all-things-considered the thing to do. (At least, they are different questions, if one does not think morality trumps or provides overriding reasons.)

Actually, it would be really nice to believe that no one ever freely does what she takes to be morally wrong. And it would also be nice to believe that no one ever freely does what he takes to be against what he has reason to do, all-things-considered. But so far I don't see any decisive reason to deny the manifest reality that we are just not that cool; psychic reality, and its agential sequelae, are much more complicated and messy, even if the reality doesn't fit so neatly with one's favorite theories.

[Neil]

John, I didn't mean the argument to be a blow of *any* sort - I apologise for any offense. I really just meant to gesture toward a question: how seriously should we take common sense? These are big issues. Common sense has a variety of sources, and attention to these sources leads me to think that its propensity to track truth is not all that high. I too am a fan of wide reflective equilibrium. But I take it that it is a constraint on an acceptable reflective equilibrium that it comports with the science (that's what the "wide" refers to), and I think we need to defer to science on this question. I think we have a good deal of scientific evidence on it.

As Chandra noted, his version of scepticism is different from mine. I believe that there are (many) cases of agents voluntarily acting contrary to their resolutions: if we follow philosophers like Richard Holton, and we think voluntariness is sufficient for WoW, we will think that shows that there are cases of WoW. I don't follow Richard, but because of issues in action classification, not regarding whether actions like this exist. I believe that the category 'WoW' does not cut our psychological nature at its joints. I base this conclusion on scientific work. I also believe that these actions are not free, because I believe that voluntariness is not sufficient for freedom. This view is based on a prioristic argument. I am prepared to concede the word 'free', and fall back on moral responsibility instead.

I had a referee on a paper on the topic 'is WoW free' respond (in full): 'this paper is ridiculous. WoW is free by definition'. I don't think that responses like that one are a good way to advance debate (not, of course, that I would accuse you of that).

[John Fischer]

Neil,

Neil:

No offense taken.

And, btw, I wasn't the referee!

Also, I promise I'll do my homework (when I'm not enjoying Boulder, CO this summer) and read up on your (and Chandra's) sceptical worries about weakness of the will.

[Chandra Sripada]

My grounds for skepticism about weakness of will are the 'traditional grounds' that Neil describes. I do think one can build a more systematic case for this kind of skeptical position than has been done yet in the philosophical literature. I build a brief argument for this position in section 4 of the paper available here:

http://sitemaker.umich.edu/sripada/files/willpower.pdf

Here is the gist. I think we come to believe in weak-willed actions through the phenomenology of temptation. There is a feeling of 'uncompelledness' associated with giving in to temptation. But this phenomenology is misleading (and the cognitive science of action and self-control help us to see more precisely how it misleads). So we are in error when we think there are weak-willed actions. Cases that we label as involving weakness of will actually involve a certain kind of irresistible impulse. Much more needs to be said to fill in this picture and make it more plausible than it might initially seem, and I do this in a manuscript in preparation. I am also trying to address problems for my position raised by Neil’s PPR paper, which does a wonderful job of synthesizing philosophical and empirical considerations. But that’s the gist of my position and I do agree that since it contradicts common sense, I’ve got to do more work to get people to buy into it.

I do think one can freely deviate from one’s judgments about what one morally ought to do. Moral considerations are not overriding. They provide only pro tanto reasons. I get this conclusion on a priori grounds – no cog sci required!

[Manuel Vargas]

Hi Noah-

This is a followup on your question about what to read.

In addition to Josh's excellent advice, I'd recommend tracking down any of a number of books. Kane's /Contemporary Intro to Free Will/ is well-regarded. I'm a (admittedly partly self-interested) fan of /Four Views on Free Will/. There are also a lot of great anthologies out there that have classic compatibilist defenses in them (e.g., stuff by Frankfurt, Watson, Fischer, Wallace, Wolf, etc.). You might be able to find inexpensive used copies of anthologies by Watson, Pereboom, Fischer, Kane or others that have many of these essays.

Not completely visible without subscription, but maybe useful anyway:
http://www.oxfordbibliographiesonline.com/display/id/obo-9780195396577-0047

I should warn you that a number of terrific survey articles are behind paywalls or in anthologies (e.g., Watson's "Free Action, Free Will"; Fischer's (1999) survey; Levy and McKenna's Phil Compass piece), so if you just rely on PhilPapers or other free venues, you'll be missing a lot of stuff. Sadly, too much great stuff is not as widely accessible as it ought to be.

[Noah Te Stroete]

Thank you for the reading advice, Josh and Manuel.

[Eddy Nahmias]

Two quick possibilities addressing Neil's original question:

1. Surely, there are deep connections (contingent, conceptual, folk intuitional, or all of the above) between the self and consciousness, so if free will is understood (as it should be, at least in part) as a capacity for self-control, then free will requires that some of the conscious processes associated with the self are able to control some of one's actions. And if attributions of responsibility are aimed at those actions produced by one's self, then they likely trace back to some conscious process of the self. Ideas along these lines have surfaced in previous comments (and I'm not adding any precision to these vague connections).

2. But I haven't seen here (and rarely see anywhere) discussion of another connection between free will and consciousness, one that goes through language. In all the literature on non-conscious processing, you never see studies showing that people can process language non-consciously. Subliminal primes are limited to one or two words (and even a negation can't be processed subliminally--"not happy" does not prime sadder behavior). Good luck getting someone to understand a paragraph of philosophy or War and Peace without both attentional consciousness and phenomenal consciousness (this may be a contingent thing; I don't want to raise Zombies from the grave where they should be locked away).

So, assume language requires consciousness. Surely, language is also crucial for important aspects of the type of free will humans have. Thinking in words seems crucial for many types of complex deliberation, planning, and decision making. Some of that linguistic/symbolic thinking gets externalized in conversations we have with others as we try to make important decisions (or in lists or external memory resources). Language also seems crucial for enhancing memory and for constructing a narrative self, both of which seem to allow types of moral reasoning and deliberation that would be (practically?) impossible without language. The studies from Libet and Wegner and Bargh *may* show that conscious processes are not required for some (very) simple behaviors or to prime certain goals or thoughts. But it seems unlikely that research will show that conscious processes are causally irrelevant to linguistic processes or that linguistic processes are irrelevant for some of our more complex future-directed planning and behavior we do (and then, if we normally have such capacities, we may be held responsible for failing to deploy them when we could have and should have).

Any thoughts?

[Neil Levy]

Interesting thought, Eddy. Here's a complication. Unconscious linguistic processing is limited in the way you suggest on the input side. But a great deal of language production is handled by unconscious processes. How much we don't - given the limitations on the input side, this is hard to test experimentally.

[Eddy Nahmias]

That's an important point, Neil. I was thinking mostly about the importance of linguistic or symbolic thinking, the sort of inner monologues we engage in when we are deliberating about important decisions (or writing philosophy papers!). And there the importance of language seems to be in providing a "scaffolding" for holding ideas and reasons up for consideration and comparison. The production of the linguistic (or quasi-linguistic) representations is presumably largely done through non-conscious processes. But what seems important to me is the conscious consideration of those ideas and representations.

So, to oversimplify, it seems to me (from my own phenomenology, which yes, I know is defeasible evidence) that when I'm thinking about a complicated decision (like buying a used car, as I am doing this morning!), consciousness serves as a sort of screen over which lots and lots of information passes. That information is often linguistic in form. While that information is on the screen, I can consider its importance to my decision-making, compare and contrast it with other information, discard it, try to remember it, etc. (Insert Global Workspace Model here?) And as I said in previous post, some of this process gets externalized on my computer screen and in conversations with others, like my wife (although she likes to make these decisions with a lot less of this tedious deliberation than I do). Presumably, there must be non-conscious processes that are responsible for producing ("sending") all the linguistic information to the "conscious screen". But the neural processes that subvene (or are identical to) the conscious processes are then, presumably, playing an important causal role in what happens next with that information, and that is (one of the places) where the (compatibilist) free will activity is doing its job.

[Thomas Nadelhoffer]

Eddy,

When you talk about the causal efficacy of consciousness, you talk about what *you* do with information that is displayed on the so-called screen of consciousness. In short, you consider the information, weigh it, contrast it, etc. Once you're done deliberating, you form proximal and distal intentions, etc. But where is this *you* in this picture that's doing all this cool stuff with the information? In order for this model to work, wouldn't you need to be something distinct from but with access to this information? If you are somehow distinct--i.e., if you are not part of what informationally plays out on the screen of consciousness--then is this homunculur you a distinct form of consciousness above and beyond the screen? Keeping with your computer analogy, which part of the computer is you? The whole computer, the hardware, the software, the graphics, etc? If, on the other hand, you just are the collective constellation of life-long conscious experiences--i.e., *you* just are the information that has been, is, and will be present in consciousness--then where do you get a volitional handle on the process of acting in such a way that it makes sense to say you acted freely?

[Eddy Nahmias]

Well, I don't want to push the analogy too far, certainly not if it's going to inspire homunculur objections. I don't see any problems with talking about some parts or processes in a system controlling other parts or processes in that system, so all I mean to suggest is that some of the conscious processes have a causal influence on what happens in and with the body in such a way that it's proper to say the causal processes have some control over what happens. So, there is no dualistic "me" that views the information and controls it. Nonetheless, harkening back to my point 1 in earlier post, there are aspects of the system that I (and others) identify *me* with more than others, generally the aspects that are consciously accessible and can be shared through language, though my self--even my "deep self"--includes much more than just my conscious processes.

Look, dualism might be an easy and even natural way to interpret our experience of consciousness and agency. But there are other ways to explain the phenomenological data (and relevant intuitions) that do not require eliminativism about the relevant data and intuitions.

[Thomas Nadelhoffer]

Eddy,

Is Wegner (or anyone else for that matter) wed to the view that conscious experiences per se are causally inert? Instead, aren't they merely wed to the more narrow claim that the conscious will is an illusion--i.e., that appeareances to the contrary, the conscious self is epiphenomenal? If so, then showing that conscious experiences play some role in the etiology of action isn't enough to refute the view under consideration. What you need is to show that *you*--however you conceive of it--are at least sometimes both free and volitional. In order to do this, you need to show that the "user" of the information being displayed on the screen of consciousness is (a) distinct from the information on the screen, and (b) able to do something with it.

In short, in order for your strategy to work, you need to say something about where you think the *you* ends and other conscious processes and experiences begin because I think you are going to need this account if you want to use the "screen of consciousness" line to refute epiphenomenalism about conscious will rather than epiphenomenalism about consciousness more generally. After all, the *you* is ultimately supposed to put both the free and will into agency. Simply showing that conscious experiences play a role in the etiology of action doesn't mean that they are being put into action by *you* rather than non-conscious or sub-conscious processes that are insufficient for grounding free will.

[Josh Shepherd]

Thomas,

I'm obviously not speaking for Eddy here, but I find your suggestion interesting. It raises some difficult issues. Because it seems an open question which processes should count as constitutive of the 'conscious self,' and muddying the waters here is whether the notion of a conscious self as distinct from the agent considered as a whole is worth trying to ditinguish. Certainly our self-concept and our body-schema often influence action without passing before the screen of consciousness as they do so.
Would it be enough to identify a set of processes which generate and control action, and then to show that an important subset of these processes require consciousness to do their job?

[George Ortega]

Thomas,

The matter goes even beyond implications relating to the parameters and orientation of conscious processes. Ruud Custers and Henk Aarts' July 2, 2010 review in the journal Science that presents a now substantial body of empirical evidence that unconscious processes are routinely responsible for decisions our conscious mind nonetheless ascribes to itself presents another formidable challenge for free will. If this unconscious, which never sleeps, can AT ANY TIME be the decider of a human action, and can even determine what the desires determining that action are to be, how can a believer in free will claim WITH ANY DEGREE OF ASSURANCE that ANY decision anyone ever makes is a free will decision rather than an unconsciously willed decision? Reducing the assertion that any specific decision has been consciously and freely willed to no more than a guess, or possibility, also substantially reduces the weight of such a position.

The believer in free will is now in the unenviable position of needing to cite EVEN ONE decision that can be definitively shown to have been made by the conscious will, and could not just as likely have been made by unconscious processes.

[Eddy Nahmias]

Here's an argument you seem to be advancing, Thomas:
1. To have free will requires that "conscious will" plays a role in action.
2. What it means for "conscious will" to play a role in action is that there is some "conscious you" (Nichols and Knobe call it the "executive self") that is both distinct from and has some control over the psychological processes within one's head (whether they be conscious or non-conscious).
3. There is no clear way to make sense of the homuncular idea suggested in premise 2 and/or the empirical evidence suggests it is false.
4. So, free will is an illusion.

But premise 2 is false. And I don't think my undeveloped thoughts in earlier posts commit me to it. And I don't think ordinary people have beliefs, intuitions, or phenomenology that *commit* them to it, though their beliefs, intuitions, and phenomenology might allow for (perhaps even suggest) a philosophical theory that suggests such dualism. But as I said earlier, I think the folk beliefs, intuitions, and phenomenology also allow for other theories that save (much of) the phenomena rather than eliminating them.

George, I posted a link to the article you cite earlier in this thread. I see nothing in that article that suggests the burden is shifted to someone who wants to argue for a certain type of compatibilist free will, one that says some of our complex decisions involve a (real, not rationalizing) role for conscious processes, such as conscious deliberation and planning. None of the research on non-conscious processes comes close to showing that people can write a lecture, decide which college to go to, plan a wedding, etc. without any conscious processes playing a causal role.

You and Thomas (and Wegner, etc.) use this phrase "conscious will" as if its meaning is transparent. Whether conscious will is an illusion depends on how one defines it.

[Thomas Nadelhoffer]

Eddy,

Look, you were the one speaking in terms of the computer user--i.e., *you*--deliberating making decisions, forming plans, etc. based on the information being displayed on the screen of consciousness. I was simply asking you to say a little bit about what/where you think this *you* is. It's not enough to show that conscious experience plays some causal role in the transition from sensory input to behavioral output if you want to refute a Wegner-style epiphenomenalism about the conscious will. After all, we get sensory input/behavioral outputs without the conscious will. You need to show that *you* are the user of the computer rather than just another piece of the software, hardware, etc. Now, you claim the "executive self" view is (a) false, and (b) incompatibile with lay conceptions of the self. I think you're wrong on both fronts. Set aside our disagreement about (b) for now--although our latest data on dualism ought not give you much optimism on this front!--and let's focus on (a). On your reading, the executive self view has a disinctness component and a control component. Which of these two elements are you denying? I, for one, think you need both to make your computer user/screen of consciousness analogy work. If there's no user, the entire computer analogy breaks down.

[Thomas Nadelhoffer]

Eddy,

One more thing: I was using "conscious will" in the way Wegner uses is since I assumed he was your target. So, for present purposes, let's stick with his usage.

[Neil Levy]

I'm not sure what this debate is about anymore. I never understood what Wegner was supposed to be denying that could possibly matter. He often quotes Searle as his target, denying the latter's claim that the experience of will causes action. But no one apart from Searle ever thought that. Consciousness isn't the kind of thing that can start causal chains: it isn't an event. Here's my model for its nattering: deliberation actually alters (non-defiantly) what the subpersonal mechanisms which initiate action do: it is causally efficacious. In the absence of conscious deliberation, the factors which are causally efficacious are subpersonal only. Because they are narrower, they are less reflective of the agent. The agent is the whole shebang. What do you reject here, Eddy? What do you object to here, Thomas?

[Neil Levy]

Um, non-deviantly, not defiantly. Blame it on auto correct.

[George Ortega]

Eddy,

My point was not that Custer and Aarts’ review challenges the role conscious processes play in decisions, which may, not incidentally, be COMPLETELY limited to the acquisition of input data for the unconscious to decide by. My point was that the findings they review reduce ANY claim of ANY decision having been freely willed to a mere guess or possibility. Such decisions NOW have an equal likelihood of having been made by the unconscious.

Moreover, even a decision that might hypothetically arise via collaboration between conscious and unconscious processes cannot have been freely willed because the unconscious, over which we have no immediate control, would be acting as a necessary condition to the decision.

To address your last point, I’m using the term “conscious will” to denote the ability that conscious processes would, hypothetically, have to independently make a decision. Conscious will is clearly illusory if the decision was made independently by unconscious processes or through consciousness/unconscious collaboration.

[Matt King]

I was planning to comment earlier in this thread, but got sidetracked by other things. Now I'm less certain of just how what I was going to say fits in.

But the gist is this: I'm skeptical that the conscious self is more "me" than my unconscious self. (I often feel that discussions about consciousness and responsibility end up turning into a game of "pronouns" - as Thomas' questions for Eddy have illustrated.) So - in short - I'm not sure why *I* can't just be the computer; no separate user required.

But I agree with Neil that we need to more explicitly engage in the question of what our theories require regarding consciousness. Peter Carruther's and I have a forthcoming paper in JMP (titled, "Moral responsibility and consciousness") on this very issue. We argue that Real Self views (e.g., Frankfurt; Watson) are committed to the existence of a conscious self, and we show that there's good reason to think there may be no such thing. We also suggest some ways theories of responsibility might proceed in the absence of a conscious self.

[George Ortega]

The conscious self and the unconscious self are both equally “me,” however, because we cannot freely control either, they are aspects of a larger me that includes the decision making properties.

You can’t “just be the computer, no separate user required” because you can’t freely control what you decide. If you had that ability, you would be much happier and more ethical.

Consciousness is not an aspect of the determined will/free will question because extra-conscious factors, or the causal past, is what controls human behavior.

Demonstrating that a personal consciousness is a construct not separate from all else in the universe would clearly defeat all arguments for a free will.

No true personal theory of responsibility could arise because personal responsibility requires free will. Neither can we properly attribute responsibility to the universe that we cannot yet/ever know.