Monday, April 30, 2007

The Trembling Stoic

On a dispositionalist view of belief (which I elaborate in this essay and this encyclopedia entry), to believe some proposition P is just to act and react, both in one's outward behavior and one's inward feelings, as though P were the case. One difficulty for this sort of view is what to do when someone seems to sincerely, wholeheartedly endorse some proposition -- and hence, we might say, believe it -- and yet does not pervasively act and react as though that proposition were true.

Excluding cases of deception (self or other) or unusual cognitive background (such as strange accompanying beliefs and desires), such cases seem to come mainly in two varieties:

(1.) Trembling Stoic cases. The Stoic sincerely judges, both alone in his study and out in the world in discussions with others, unhesitantly and unreservedly, that death is not bad. Yet he trembles before the sword, and he fears what the doctor will say. Or: A liberal professor sincerely professes that all the races are intellectually equal (and has, let's suppose, the scientific evidence to prove it), yet reveals implicitly in her behavior and reactions a subtle but persistent racism when it comes to matters of intelligence. Or: Someone comes to believe that God and Heaven exist, but does not transform her behavior accordingly.

These cases needn't involve self-deception or insincerity. One may perfectly well realize the need to reform. Nor are such cases necessarily matters of "weakness of will" in the face of acute, impulsive desires: Our liberal professor, for example, need have no particular desire to treat other races as intellectually inferior.

(2.) Momentary forgetfulness cases. I get an email saying a bridge I normally take to work is closed. Yet the next day, I find myself headed toward the bridge, rather than my intended alternate route, until at some moment (maybe only after seeing the bridge) I recall the previous day's email. Or: The trashcan used always to be under the sink, now it's by the fridge. I still reach under the sink half the time, though, when I go to throw something away.

Aaron Zimmerman, Tori McGeer, and Ted Preston have emphasized the importance of such cases to me in evaluating dispositionalism about belief (Tori and Ted being sympathetic to dispositionalism, Aaron less so).

My response to such cases is to distinguish broadly dispositional belief from momentary occurrences of sincere judgment. Then I'll "bite the bullet" on belief: The Stoic, the absent-minded driver, do not fully and completely believe, respectively, that death is not bad, that the bridge is closed (neither do they fully and completely believe the opposite). Although there are moments when they make sincere judgments to that effect, it takes a certain about of work and self-regulation to allow such judgments fully to inform one's habitual everyday behavior. Until that work is done, they don't fully believe.

(I recognize that it may grate a bit to say that I don't believe that the bridge is closed, as I'm driving toward it, especially since it seems right for me to say, in retrospect, "I knew the bridge was closed!" Here's a brief discussion of that particular issue.)


Brad C said...

Hi Eric,

Interesting issue. In case you have not read it, you might be interested in Saul Traiger's paper about the trembling Stoic mentioned by Hume - the one who trembles when put in a cage over a precipice. Traiger traces reactions to the example in various modern philosophers - in particular he is interested in how it was used in arguments about the nature of (and relation between) reason and passion. If interested, you can get a copy here:

Anonymous said...

I wonder if someone has argued in the following way against the dispositional account of beliefs...

Not believing that X, might in some cases mean same behavior as believing that not X.
If that is the case, and if it is the case that not believing that X, is not same as believing that not X, it would mean that merely acting as believing that not X, might not in fact mean believing that not X (but instead not believing that X).

That not believing that X is different from believing that not X, I think is obvious, as not believing that X, doesn't even require for one to know about X, or to have wondered about X.

As an example, take A)"not believing that there is a Cartesian demon" vs. B)"believing that there is no Cartesian demon". I think that for a person that haven't heard/thought about Cartesian demon possibility, A would be the case but not B. However, it might be argued that in both cases person will be disposed to act in same way.

Eric Schwitzgebel said...

Thanks for that suggestion, Brad. I hadn't heard of that -- I'll definitely look it up!

And thanks, Tanasije, also. I agree that there's an important difference between believing not-P and not believing P. I do think that there are important dispositional differences though. For example, part of the steretype for believing not-P would be saying (under appropriate conditions) "P is not true". One can fail to believe that P without having that disposition. Likewise, believing not-P stereotypically involves being prone to feel surprise should one discover that P is true. One might fail to believe that P without possessing that disposition.

I would suggest that not believing P is just failing to match to any significant degree the stereotype for believing that P. One might also, as it happens, match the stereotype for believing not-P; but one might equally well not -- for example if you are neutral between P and not P or don't have the conceptual apparatus necessary to consider the matter.

That's my thought, at least. Believing that not-P is only *one* way of failing to believe that P.

(And, by the way -- as is implied by the remarks above -- on my dispositional approach to belief one cannot have simultaneous contradictory beliefs: The stereotypes for believing P and for believing not-P have incompatible compontents so that one cannot fully match one stereotype and at the same time also match the other.)

Justin Tiwald said...

Hi Eric,

Very interesting. As in your past posts on dispositionalism, my comparatively pre-theoretical intuitions tend to be with yours. But I wonder what might you say to someone who thinks the dispute about occurrent and dispositional accounts trades on equivocal senses of "belief." Take the bridge case: why couldn't we say that you believe (in the occurrent sense) that the bridge is open while at the same time believe (in the dispositional sense) that it's closed?

Or maybe this is precisely what you mean to say. But I took your careful phrasing of "momentary occurrence of sincere judgment" as an artful way to avoid saying "belief." Just curious!

Neil said...

You might also consider delusions. Famously, Capgras sufferers are not disposed to behave very consistently with their professed belief (I actually think that the dispositionalist gets the case right: there is no fact of the matter about what the sufferer believes).

Anonymous said...


I’m glad you put up another post on this interesting topic. I started interacting with your essays on this for my M.A. thesis, but had to cut this section out due to time and space limitations. Here’s one of my half-baked thoughts:

The meaning of ‘belief’ is rooted in our linguistic practices; we take the world to be so and so through conscious linguistic practice (or thought) or through conscious simulated modeling. ‘Belief’ is rooted in this linguistic practice and phenomenal simulation; belief therefore becomes a conscious standard that we must ‘live up to’ (just as self-narratives, in the form of storyworlds (Herman), provide standards for us to constantly live up to). But there is by definition large limitations to our conscious descriptions and simulations, and we should expect there to be a global discrepancy between these ad hoc abilities and the far more complex and sophisticated working of our unconscious mind. It is after all primarily our unconscious mind that does the ‘living up to’ throughout the day. And I think there might be a minor slip in your published essay that illustrates this point:

“Does Ellen believe that all Spanish nouns ending in ‘a’ are feminine? Some of her dispositions accord with that belief.”

Here it seems we have identified the belief as Ellen’s taking the world made true -- which includes phenomenal simulations modeling the world -- by “all Spanish nouns ending in ‘a’ are feminine” as the actual world; the dispositions either accord or don’t accord *with this belief*. The belief is the pre-existing standard that we judge Ellen’s behavior with; it therefore seems mildly inaccurate to say that stereotypical behavior just is what we reference by the word ‘belief’.

So, our behavioral dispositions are judged according to our beliefs, and our beliefs are references, directly and indirectly, to our occurent judgments and simulations. (This is also supported by my other commitment: that the ‘content’ of our beliefs are always indeterminate until fixed on an occurent judgment.)

Yes? No? Thanks!

Clark Goble said...

The thing that I always found interesting what the relationship between beliefs and counterfactuals. If you don't have an experience which manifests your belief ala the Stoic, would we say they believe or don't believe?

Eric Schwitzgebel said...

Thanks for the comments, Neil, Michael, and Clark!

Neil: I agree that delusions are an interesting class of cases in this connection; and I'm sympathetic to your idea that we shouldn't regard (some) delusions are full-fleged beliefs. Tim Bayne and Vaughan Bell, and their co-authors, have written interesting things on delusions and dispositional approaches to belief. Do you know their work?

Micheal: I really like the idea that we have obligations both to ourselves and others to "live up to" our beliefs -- this is something Tori McGeer has been writing on a bit, and which I develop a little in my "Acting Contrary to Our Professed Beliefs" (unpublished, but available on my homepage). We profess, to ourselves and/or to the world, that we believe that P. Now we are obliged to corral our behavioral dispositions to accord with that.

I'm not sure I entirely understand your analysis of the Ellen case, though; or exactly what the relationship is between belief, occurrent judgment, and simulation. I'm sure it's just that you're trying to compress a lot into a short space out of politeness!

Clark: I'm definitely into the counterfactuals here. You don't need to actually have the experience for its potentiality to be central to whether you believe. Many beliefs we act on hardly at all. It's just that I *would* do such-and-such *if* such-and-such.

Anonymous said...


Thanks; I’ll be more direct about Ellen. As I understand your view, a belief just is the disposition to act and react consistent with some set of stereotypical behavior, a set of behavior that satisfies the condition “as though P were the case”. As you say in this post: “ . . . to believe some proposition P is just to act and react, both in one's outward behavior and one's inward feelings, as though P were the case.”

But I’m claiming that the natural meaning of ‘belief’ or ‘believe that P’ does not reference the dispositions that we might think somehow consistent with P being the case – until we change the meaning, adding addition terms or technicality: e.g., “do you REALLY believe”?, or “faith without works is dead, even the demons believe . . . “. The natural meaning of ‘belief’ standing alone (right now in our culture anyway, so I say) is to hold the linguistic assertion ‘that P’ as true, regardless of what might or might not be behavior consistent with holding that P true. (This is why I think Paul’s letters are typically misinterpreted by modern theologians by the way.)

So, ‘belief’ is grounded in the linguistic expression ‘that P’ and not on the rationality of behavior. We only ‘believe that there is a dog on the mat’ if we are able to assert ‘there is a dog on the mat’. Otherwise, there would be no ‘that p’ to our belief ‘that p’, and hence no belief to begin with. And I was wondering if this is why you said the following in your essay:

“Does Ellen believe that all Spanish nouns ending in ‘a’ are feminine? Some of her dispositions accord with that belief.”

What does ‘that belief’ refer to in the last sentence? If the belief is reference to Ellen’s behavior to “act according to that p”, then her dispositions that accord with ‘that p’ JUST ARE Ellen’s ‘belief’. But here you set Ellen’s ‘belief’ up as the standard that determines the appropriate dispositions; translating: her belief accords with her belief, which doesn’t seem right. I’m wondering if you naturally used ‘belief’ hear to refer only to Ellen’s limited, conscious thinking or asserting ‘that p’ and not the sophisticated dispositions of Ellen’s unconscious mind.


Eric Schwitzgebel said...

Thanks, Michael, now I see better where you're coming from!

I see two issues: One is my treatment of the Ellen case in particular. The other is the broader issue about the role of linguistic vs. other dispositions in belief ascription.

On the Ellen case I think that if you grant my overall view of belief I am at worst guilty of putting things infelicitously.

On the more general issue, though, I think you're right to see a substantial disagreement between us. I don't want to stand language at the center of belief. For one thing, it leaves us in an odd position regarding non-linguistic creatures. I think much of our cognition really is fairly similar to that of apes.

Cases in which our linguistic dispositions point one way and the rest of our behavior points another way will be the test cases between our views. You'll say we believe; I'll say that we're "in between" (since if there's a consistent tendency to linguistically disavow P, that person can't fully and completely match the stereotype for believing that P). The question is: Which is the better way to think of such cases?

Juliet sincerely avows that all the races are intellectually equal; she's persistently racist in her behavior. Ellen says that all Spanish nouns ending in "a" are feminine; but she correctly says "el anarchista". You'll say, I think, that Juliet and Ellen do believe what they avow, and then you'll tell some other story about the rest of their behavior. I'll say that they are in an in-between state of belief with respect to these propositions.

It's a practical question which is the better approach, but for my money what we do and should care about, in matters of belief, is the overall arc of the person's dispositions and not just (or primarily) her linguistic ones. The minority kids in Juliet's class talk about what she believes because they care, not just what she'll say, but about whether she'll be biased against them.

Anonymous said...


Thank you for the thoughtful follow up. I think if you granted me my view on how we *do* use ‘belief’, I might be able to grant you your view of how we *should* use ‘belief’. Your use of belief seems more Hebraic, referencing also the habits and emotions, playing a creedal role, whereas I’m trying to account for the Hellenistic ‘that p’ role this concept plays in our culture – particularly philosophical culture! However, I’m arguing that without the ‘that p’ we don’t have what we call ‘belief’, and your point about apes might reveal that my proposed compromise won’t work. It seems to me that the statement ‘the ape believes that bananas are yummy’ would be metaphorical, as-if. This seems intuitive to me; but if not to you, consider the problem of attributing beliefs to an ape: you will always have to translate the cognitive processes (that, I agree, are similar to ours) into linguistic/propositional form. In my example above, we have to translate conscious and unconscious processes or states of an ape by the words ‘banana’ and ‘yummy’. But these are linguistic items that apes do not have access to. Only language users can be said to literally ‘believe’ in so far as we require the ‘that p’ to make such attributions. If this is not convincing enough, consider the trouble of less sophisticated monkeys, or dolphins, and then go on down the line. At what point is it metaphorical to ascribe beliefs to more primitive intentional systems? (think Dennett and Searle here) I would say that the metaphorical processes begin as soon as we reach beyond our own language using kin.


Eric Schwitzgebel said...

Interesting thoughts, Michael. I'm inclined to agree with you that there's a strand in our culture, especially our philosophical culture, that is more language-oriented in its use of 'believes' (and 'thinks').

Does the monkey believe/think that bananas are yummy? that his mate is in the canyon? that the alpha would slug him if he went too close? I grant that there are divergent intuitions here, but I don't accept your argument. That we have to use propositional language to describe the profile or content of the attitude doesn't mean there is anything intrinsically linguistic about the attitude. We have to use numbers to describe the heights of mountains, but mountain peaks aren't thereby inherently numerical.

Sentential/propositional ascriptions serve as an index, I'd say, rather than mirroring the internal structure. (I owe this point to Davidson, but I diverge from Davidson in that I think that for languageless mammals, and not just language-speaking humans, we need to deploy a propositional index to characterize their dispositional patterns.)

At some point, of course, behavior is too simple to require propositional, belief-stereotype structures to characterize; and then belief talk is merely metaphorical. It's a matter of degree, but I'd put adult mammals on the believing side of this divide.

Our concepts don't match a monkey's concepts, of course. But neither do my concepts completely match the concepts of an ancient Chinese philosopher, or anyone else for that matter. My concept of "morality" isn't quite the same as his concept of "de" or "yi". Yet I can say, accurately enough, that Mencius believed that allowing people to starve was immoral.

Chris said...

At the risk of trying to enter a discussion only at closing time (and sorry for the length)...

You don't need to actually have the experience for its potentiality to be central to whether you believe. Many beliefs we act on hardly at all. It's just that I *would* do such-and-such *if* such-and-such.

But surely that's contingent on a lot of things beyond the 'belief' itself - in particular, the agent's general emotional and physical capabilities, which to varying degrees aren't things easily (perhaps even possibly) in her control. So, you seem to accept that if a person's emotional dispositions aren't up to making good her beliefs, then she doesn't truly hold the beliefs in question. Well, OK, maybe (though see below). But what about the case of a dyslexic who sincerely believes 'bat' begins with a 'b', but frequently writes 'dat' instead? Such a person then only gets it 'right' with a great deal of mental effort. The difference between her failing and succeeding to live up to her professed belief in 'bat' beginning with a 'b', then, is not in the quality of the belief, but in the degree of conscious willpower or mental effort extended.

But furthemore, presumably the general context will partly determine how much willpower is needed. Compare working alone in a private office, say, to taking some sort of class for the first time and needing to write on the whiteboard in front of everyone. If in the second she ends up making far more spelling mistakes than she usually does, surely one would say it was her nerves that caused the matter, not her failing to truly believe that 'bat' is spelt with a 'b' and so forth. (In my view, the 'trembling stoic' case obscures this by having the belief itself to be in nervelessness.)

Eric Schwitzgebel said...

Those kinds of cases are, I agree, not as naturally handled on my account as on some others. Accounts that emphasize a core, inner state behind belief deal more easily with cases of this sort but also I think less easily with many in-between cases and cases in which we ascribe belief on the basis of patterns of behavior regardless of sincere avowals.

My technique for handling cases of this sort is to refer to what I call "excusing conditions" -- conditions under which the failure to act according to the stereotype interferes not at all with the appropriateness of ascribing the belief. Physical disability is a prototypical case of an excusing condition in this sense, and I would say that clear pathologies like dyslexia fit into the same category. Simple, ordinary weakness of will may or may not be an excusing condition in the relevant sense.

I confess that I'm not entirely satisfied with my account of excusing conditions. I do have a little section on it, though, in my 2002 essay on belief (available from my homepage) if you're interested in more detail.

Chris said...

Thanks for the reply. Having a look at the essay you suggested I do, does what you say there imply that the believer herself is supposed to know all (or even most) of the 'excusing' clauses, even if they are to be deemed 'relatively simple and folksy' as you put it? (By the by, I'm not really sure what that formulation is supposed to mean - for the ability to intuitively grasp the implications of a proposition, I guess, will considerably vary between a skilled analytic philosopher like yourself, say, and an 'ordinary Joe'.)

My (new-ish) worry off of the back of this is then the following: I presume one must allow that experience may modify a belief; yet your position, it seems to me, frequently won't allow saying a belief was there to be modified in the first place, as the belief-'changing' event is the test by which one judges the person had the belief in question. The experience of meeting an unforseen implication of a belief, then, may (I want to say) cause the belief to be modified *in that instant*, and so what you say about 'phenomenal dispositions' in pp.258-9 of the Nous paper doesn't apply (I think...).

Eric Schwitzgebel said...

Shoot, Chris, I hope you didn't feel obliged to go read that essay just now! I'd hate for people to start feeling obliged to read my essays before commenting on my blog. I'd lose too many interesting commentators that way, I suspect!

Regarding knowledge of the excusing clauses: On my account, it is the ascriber who knows them, and only implicit knowledge is necessary. "Ordinary Joe" can't formulate them explicitly, but some conditions, were he to see them realized, would lead him to rescind his belief ascription or at least back away from it a bit, while others wouldn't impact his ascription. By being disposed not to alter his judgment of the appropriateness of the belief in light the failure to manifest the relevant dispositions under certain conditions, he reveals an implicit sense of those as excusing conditions. That's my thought.

I'm not sure I understand the concern you raise in the second paragraph. I say this because it seems to me there's an easy way around it, as I interpret it: What is relevant are the *dispositions*, not the actual cognitive, phenomenal, or behavioral events. So even before the event occurs that would have caused you to change your mind -- maybe merely entertaining the thought explicitly would be enough to change your mind in some cases -- you are still disposed in a variety of ways. Probably it wouldn't be right to say that you are disposed to respond in the belief-appropriate way in exactly the circumstances that would cause a belief change (though there may be room for this on some metaphysics of dispositions), but only in odd cases would that be any but a small range of the total spectrum of relevant dispositions. In cases where it's more than just a narrow spectrum -- such as when you'd abandon it as soon as you thought explicitly about it -- I'll probably be comfortable enough to say it's an in-between case.

Would that work? Maybe it would help to have a more fully developed example in hand....

Anonymous said...


Anonymous said...