requiem for certainty

Williams on Internal and External Reasons

with 12 comments

I’m not sure I entirely understand Bernard Williams’s views on internal and external reasons.  Fortunately, I’m not sure that most people who read his work (even publish on it) understand those views either.  (Maybe that’s just further blindness on my part but I think his work is rather more complex than is usually admitted.)  I do think that what I understand of his account usefully connects to some of his thoughts about history and how we can best make sense of ourselves and others.  (I proceed with the caveat that what follows is just notes and ramblings and may well be misguided [but I am committed to using this blog to just experiment more rather than to ‘be right’ or ‘show off’].)

Williams’s views are about reasons for actions and when we take reasons for actions to be explanatory of actions.  Here is a key claim in his 1979 piece on the matter: “nothing can explain an agent’s (intentional) actions except something that motivates him to act” (107).  This is reasons internalism, the idea that reasons for actions are internally related to being motivated to act.  The contrast view is reasons externalism, which suggests that sometimes there are reasons for actions which are normatively binding but which have no internal relations to agents’ motivations.

The force of reasons externalism is that we can say of some agent A that they have a reason to X even if they themselves do not feel the force of that reason in the sense of not being motivated to X (or not exhibiting evidence of being so motivated, etc.).  This view seems amenable to moral universalism, which enables us to rather easily make claims to the effect that A is wrong to ~X.  A on this case is irrational.  But the downside of this view is that we are putting ourselves in a position to know what is irrational or rational for agents in very different circumstances (and hence with very different motivational sets) from our own — we may not want to put ourselves in such a position (cf. Williams 1979, 110).

In later work (including Truth and Truthfulness (2002) and “Realism and Moralism in Political Argument” in In the Beginning was the Deed) Williams discusses historical understanding.  One idea there is that we are not in a good position to normatively evaluate cultures in drastically different positions from our own.  We may, to use his examples, endorse liberalism as the right kind of political regime for we moderns, but to do so we need not claim that liberalism in all places and all times was the most rational view.  We can accept that ancient tribe (AT) did not have reasons to be liberal (L).  We can accept this without saying that AT was irrational and that only we liberals who L are rational.

Let’s say we want to explain the acts of AT.  What are the reasons that explain AT’s choosing to ~L?  Well, the internal reasons theorist says that the reasons for AT’s actions are ones that have to be internally connected to AT’s motivation set (which, says Williams, is a very complex ensemble including desires, dispositions, emotions, tendencies, etc., cf. 1979, 105). The internal reasons theorist will not be inclined to pass judgment on AT and say that AT had reason to L and acted irrationally in choosing to ~L.  Now if AT had our motivation set and operated under the constraints that we operate under around here just now, then we very well might (and probably should) say that.  But the point is that AT had different motivations.  The external reasons theorist says, by contrast, that AT should have L’d because AT had reasons to L even if these reasons are not to be found among (or by consequence of) AT’s motivation set.

Now our reasons for acting (i.e., to L) are internal to our motivation to act as we do.  The same holds for AT, members of which act as they do for reasons internal to their motivations.  In cases where motivation sets are different, it just would not make sense to expect (or to insist) that the same reasons for acting hold.

This bears on a crucial question concerning what we do when we confront a group (such as AT) that acts differently than we do, and perhaps in ways that are (by our lights) wrong, perhaps because blatantly and problematically illiberal.  If AT is intolerant with respect to something for which we think they should be tolerant, then what do we do?

The external reasons theorist is all set up to say ‘oh they are irrational, so you point that out to them’.  But this theorist faces the problem of explaining why ‘pointing out reasons’ is a valuable activity given that reasons do not stand in intimate relationship to motivations.  Williams writes: “There are of course many things that a speaker may say to one who is not disposed to [L] when the speaker thinks that he should be, as that he is inconsiderate, or cruel, or selfish, or imprudent; or that things, and he, would be a lot nicer if he were so motivated. Any of these can be sensible things to say. But one who makes a great deal out of putting the criticism in the form of an external reason statement seems concerned to say that what is particularly wrong with the agent is that he is irrational. It is this theorist who particularly needs to make this charge precise: in particular, because he wants any rational agent, as, such, to acknowledge the requirement to do the thing in question” (1979, 110).  Or in another piece: “”There are many things I can say about or to this man: that he is ungrateful, inconsiderate, hard, sexist, nasty, selfish, brutal, and many other disadvantageous things. I shall presumably say, whatever else I say, that it would be better if he were nicer to her. There is one specific thing the external reasons theorist wants me to say: that the man has a reason to be nicer” (1995, 39). The external reasons theorist thinks we can explain the actions of AT in terms that AT itself did not (indeed cannot) come to recognize.  AT may never have the proper motivations but the reasons that should compel AT nevertheless hold.  AT is irrational but is too entrenched in their motivation set (slave to their passions) to act well.  Hence, we hold AT responsible for ~L even though psychologically (motivationally) there may be no way for them to be motivated to L.

The internal reasons theorist will ask us to take a different approach to AT and their failure to L.  If AT fails to L, this may be a failure by our lights, but what good (Williams wants to ask) is it to call it a failure that they too should acknowledge?  In what sense should they, or could, they acknowledge this?  We should aim for historical understanding, but this will in many instances be a matter of historical imagination rather than a matter of rational representation.  For we cannot fully understand their reasons in our terms since they do not line up to our motivations and hence what we could take as a reason.  Understanding their reasons in their terms would better explain their actions but without putting us in a position to morally castigate them.   “We need to invoke some values, no doubt, in interpreting other people’s preferences and in construing their rationality, but the values that we have to invoke are their values rather than ours” (“Values, Reasons, Persuasion”, p. 114 in Williams, PHD).

This suggests that historical understanding looks different when we apply it to radically different others than when we apply it to ourselves.  Is this an odd result for Williams to be endorsing?  Perhaps.  But it has the advantage of not bartering with a conception of an invariant trans-historical rationality that is (or should be?) available to all tribes at all times.  Rationality is something that is internal to a tribe.  Thus, we make sense of our own tribe in a more internal-ish kind of way that is capable of generating a normative tug.  But we make sense of other tribes in a more external-ish kind of way that might help us understand why they acted as they did given their motivation set, and even if it does our understanding will not have any normative tug for us.  We might understand why AT was despicably ~L but we do not do this in a way that would suggest for us that we understand why ~L would be rational or would be right for us.  Do we understand them?  Yes but only externally, not from the inside.  And since all reasons for acting are internal reasons, we are at best giving a kind of rough explanation of their action which is not in any ways up to the task of feeling the force of those actions or the rationality that guides them, or the normativity that may be implicit in them.

Here is an attempt at summary (updated 11/10): since Williams rejects reasons externalism, he rejects the idea that we can stand in a position to normatively evaluate a radically different tribe, just because that would require that we impute to them reasons which we accept, but which we recognize they cannot be brought to accept (given their motivational set). However, since he affirms reasons internalism, he would also hold that our lack of ability to normatively evaluate the radically other tribe in no way impugns our ability to normatively evaluate our tribe (or similar but not identical tribes). To judge another tribe radically different from your own you have to buy into the idea that there are reasons for acting which you have access to, which you accept that they did not have access to, but which they should have had access to.  This, for Williams, is a tough pill to swallow, because its implications are decidedly anti-historicist.

So what does the internal reasons theorist do when facing an awful AT?  Well, if AT is historically distant, we can measure the distance between and try to learn about how we get from AT to us today (UT).  This is the relativism of distance (see Ethics and the Limits of Philosophy).  And if AT is contemporary with UT, then we can try to explore AT’s motivation set to see what we might construct so as to appeal to them in the hopes of changing their action via the package of motivations and reasons that make their actions what they are.

Doesn’t the internal reasons theorist sell the moral philosopher short?  Because if reasons internalism holds then the moral philosopher cannot show that AT is wrong (at least not in every case of this type).  But, Williams wants to argue, what would be the point of that anyway?  We can show why UT would be wrong to ~L.  But why think it important to show that AT is wrong to ~L?  What’s the use of blaming AT for ~L?  To shore up our own view that we (UT) are right to L?  But if we want to shore up our own view, shouldn’t we explicate the rationality of that position, rather than assuming it’s rightness and negatively cutting down others who disagree?  Isn’t it better to know ourselves better than to know that they are wrong?

A going background concern for Williams: it is more important for us to understand ourselves than most moral philosophers have typically appreciated.

See also a recent article on related topics (and more) by Miranda Fricker, whose work I generally find quite smart: <>.


Written by Colin Koopman

November 10, 2010 at 8:33 am

12 Responses

Subscribe to comments with RSS.

  1. Are you familiar with the notion of reason that Brandom has been floating? Eg. in last year’s “Truth in Philosophy”. It’s a normative conception of reason (it follows that it’s basically and radically externalist, though Brandom doesn’t draw this entailment), but it also eschews a God’s-eye view of what is or is not reasonable. Instead, it treats reason as a matter of autonomy and recognition: we have to count as reasonable to the people whom we count as reasonable. Brandom credits the move to Kant and Hegel (though personally I think it’s Fichte who first hits on the social aspect).


    November 10, 2010 at 2:38 pm

    • Yes, Brandom’s work is great, but I don’t fully agree with his full-scale linguistic turn approach. The book is Reason in Philosophy. And the title Truth in Philosophy is Barry Allen’s 1993 book. (Both, by the way, worked with Rorty, fwiw.)

      There are a number of different internal-external debates in contemporary philosophy and Brandom’s (there) is different from Williams’s (here), which is to say, the debates themselves don’t line up. I don’t fully understand all of it but here’s a crack:

      Internal Reasons v. External Reasons – is a debate in the theory of action over whether or not one can have reasons for acting that are part of (or not) one’s motivational set. The Humean internalist says that reasons always engage motivations. The Kantian externalist says you (should) have access to (moral) reasons for acting even if you don’t feel their tug.

      Motivational Internalism v. Motivational Externalism – is a debate in moral psychology over whether or not moral convictions are intrinsically motivating. The motivationalist externalist says they are not. The internalist says they are. From what I understand, almost everyone today is an internalist, to the effect that if agent A accepts that P is moral then this acceptance itself is a motivation for A to act according to P.

      Internalism and Externalism about Justification – is a debate over whether the justificatory status of a belief hinges on factors that are not immediately available to a believer (externalism) or only factors that are fully available to a believer’s consciousness (internalism). In other words, is there something outside of the believer to which the justifiedness of the belief is answerable? Externalist says yes (perhaps external reality). Internalist says no (we are answerable only to what is available to us). Brandom is an externalist here.

      There’s also semantic internalism and externalism.

      So many ways to be inside or outside!

      Colin Koopman

      November 10, 2010 at 6:28 pm

      • Argh, yes, of course, “Reason in Philosophy”. Slip of the tongue — er, fingers?

        I appreciate the differences between these various controversies, but I think maybe Brandom’s position in RiP actually speaks to two levels of controversy. That is, he’s arguing for a semantic view, according to which the meaning of a sentence is to be understood in terms of the inferential commitments the speaker is obliged to take on when asserting it (which may or may not be apparent to the speaker when making the assertion).

        But on this view, the speaker’s mastery over the meaning of his words can only be recognized according to whether he accepts the entailments his audience recognizes him as bound to accept in light of the assertions he has made. And these entailments can, according to Brandom, include practical inferences as well as theoretical ones. Doesn’t that mean it’s a form of reason externalism, too? The speaker’s audience, in evaluating the meaning of the speaker’s words, can hold the speaker to reasons external to what the speaker might take himself to have committed to. The speaker, taking his audience to be rational, offers himself as a candidate for their determination of rationality; the audience must apply the same standards of rationality to the speaker as they take to justify their own standing as rational.

        The point I was trying to make in the first place is that, if this is a form of reason externalism, it isn’t susceptible to some of your reservations about externalism. For example, it doesn’t seem to me to suffer from the distastefulness of “putting ourselves in a position to know what is irrational or rational” for people of other cultures. We can acknowledge that our social practices of reason-querying, reason-giving, and reason-criticizing are institutions with a history, and they may only imperfectly instantiate the universal ideals of rationality. We can study the analogous practices of ancient people or foreign cultures with a view to understanding what *they* counted as reasons in *their* practices of reason-querying etc. But I don’t want to foreclose on the possibility of one culture criticizing another, nor would I have to on a Brandomish view, since it retains something of universalism, but without the God’s-eye view. It’s a sort of perspectival universalism, where each person makes their own determinations about reasonableness from their own situation, but recognizes themselves as having to do so as one member of a universal community of people who count one another as rational.


        November 10, 2010 at 8:34 pm

  2. Okay. But the ‘speaker’ in most of Williams’s examples are not ‘other cultures around here just now’ but those in the distant past or future. They aren’t really speakers that can be modeled on Brandom’s inferentialist semantic externalism, are they? Can we hold past speakers to commitments which we judge in the present to be implications of their past assertions? But that’s a side point.

    Because I’m still not sure that it’s the same debate going on here. Williams’s reasons internalism is a view about explanations for action couched in terms of reason. To say that agent A had a reason R to do P is to say that A could have been brought to accept R in virtue of their motivational set MS. If, given their MS, A could not have been brought to accept R as a reason for P, then Williams wants to deny that R can count as a reason for A to P. The reasons externalist (as I understand the argument) still wants to be able to say that R is a reason for A to P.

    Brandom’s view seems to me to not engage motivational sets, but is rather about all the other collateral commitments that are part of an agent’s ‘commitment set’ CS. If agent A’s CS (on our scorecard) implies (again on our scorecard) R, then we are entitled to hold A as committed to R, even if A denies R. Is that how you read Brandom, too? If so, then we agree.

    But things diverge insofar as CS is a purely rationalist notion whereas MS is much more empiricist. Brandom takes his lead from Sellars and Hegel. Williams takes his lead from something closer to Hume. Brandom is thinking about (with respect to these issues) epistemic justification (even if he does want to say that some of our commitments implicate practical outputs in the form of actions, but this always seems to me an after-thought for his project, though I may be wrong). Williams is thinking about (at least with respect to this distinction) moral motivation.

    Williams doesn’t want us to land ourselves in relativism. The claim is more of the order of, “what’s the use of judging the ancient tribe wrong?” It seems to me that he can very well say “not much use at all” but still be committed to our being able to judge ourselves right or wrong.

    Colin Koopman

    November 11, 2010 at 12:48 am

  3. Regarding the side point, I suppose what we’re talking about is asymmetrical recognition (where eg. L can recognize AT as having been rational or irrational, but AT never had the opportunity to assess the rationality of L). I don’t know of anywhere that Brandom considers that at all. But it doesn’t seem to me to be precluded by his notion of rational recognition.

    I agree with your reading of Brandom. I’m more sympathetic to his way of talking about reasons, so I guess I don’t really see the importance of the motivation set. I mean, I’m not sure I see how it would change the picture to admit it among the commitment set.

    I suppose a lot hangs on the “could have been brought to accept” part. I take it that we assume, by calling R a reason in the first place and attributing it to A, that A is rational, ie. liable to having her actions appraised in terms of reasons, and responsible to reasons. If A simply cannot or will not be persuaded to do P on grounds R, then either R is not a reason after all, or A is not rational after all.

    As for the broader issue: The question, “What’s the use of judging AT wrong?”, is related to another question, namely “What’s the use of judging AT right?” And the usefulness of judging AT right or wrong is a function of whether we can learn anything from them. If they acted rationally and rightly, then maybe we can learn something from them; if they didn’t, then certainly we can’t. (I’m thinking here of Rousseau’s paeans to Rome and Sparta.) Those seem to me to be intelligible questions, and they require us to treat AT as if they are responsible (if not actually responsive) to reasons we recognize as reasons.


    November 12, 2010 at 12:42 am

  4. sorry to intrude into this technical conversation but if I might add a side note the label of “rational” largely strikes me as an honorific akin to Dewey’s “intelligent” and perhaps the question is not so much who is rational but with whom can we reason, how and on what.


    November 12, 2010 at 2:43 pm

    • Yeah, one of the things I appreciate most about Brandom’s project is that he rethinks reason in terms of norms of recognition. That makes it — well, I wouldn’t say “honorific” per se, since it also has some determinate content, but definitely “evaluative”. The point is that you can’t disentangle the question “What ought we to count as reasons?” from the question “Whom ought we to count as responsive to reasons?”


      November 15, 2010 at 3:03 pm

      • except that this “it also has some determinate content” is questionable in terms of actual/folk practices, for me Brandom (not unlike Habermas) is engaging in a prescriptive enterprise and not a descriptive one and so ends up in all of the
        post-Wittgensteinian dilemmas around rule following.


        November 16, 2010 at 1:57 am

      • Is it that B3 (that’s B-cubed, i.e. Bob B. Brandom) has a prescriptive project? Or is it that he has a project that is trying to account for the very possibility of normativity itself?

        Given that we have an idea of normativity, i.e. of correctness in conceptual usage, how is it that this is possible?

        Few would deny that we are normative creatures. But few have been able to give a good non-foundational account of this.


        November 17, 2010 at 2:00 am

  5. It’s not prescriptive. But neither is it an account OF normativity. It’s an account OF meaning, in terms of normativity.

    The ideal is to be able to account for the meaning of a word or expression in terms of inference relationships: a word’s meaning is the contribution it makes to determining, for a sentence containing it, what entails it and what it entails in turn. These entailments can be practical (injunctions to act) as well as cognitive (other sentences).


    November 17, 2010 at 4:27 am

    • I always read it as an account of the normativity of concept use, first, and then a semantics (with an implicit account of pragmatics) second.

      Why normativity first? Because when you ask what B3 is primarily aiming to develop an account of, it is I think ‘sapience’ in contrast to ‘sentience’ and I take sapience to be a condition that is subjective to normative constraint. I think Brandom wants to explain, more than anything else, what makes us sapient, or what makes it possible that we are sapient. With this being his goal, he invokes an account of semantics to do most of the heavy lifting.

      As an aside: I would say that it is implicitly prescriptive.

      colin koopman

      November 17, 2010 at 6:48 am

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: