After MacIntyre: Is and Ought

This follows up on yesterday's discussion.  Thanks to Hodges for getting me started on this, to Milos for reminding me of MacIntyre, and to Peter for agreeing with me so far.

Are there any valid arguments that satisfy the following conditions:  (i) The premises are all factual  in the sense of purporting to state only what is the case; (ii) the conclusion is normative/evaluative?  Alasdair MacIntyre gives the following example (After Virtue, U. of Notre Dame Press, 1981, p. 55):

1. This watch is inaccurate.

Therefore

2. This is a bad watch.

MacIntyre claims that the premise is factual, the conclusion evaluative, and the argument valid.  The validity is supposed to hinge on the functional character of the concept watch.  A watch is an artifact created by an artificer for a specific purpose: to tell time accurately.  It therefore has a proper function, one assigned by the artificer.  (Serving as a paperweight being an example of an improper function.)  A good watch does its job, serves its purpose, fulfills its proper function. MacIntyre tells us that "the concept of a watch cannot be defined independently of the concept of a good watch . . ." and that "the criterion of something's being a watch and something's being a good watch . . . are not independent of each other." (Ibid.)  MacIntyre goes on to say that both sets of criteria are factual and that for this reason arguments like the one above validly move from a factual premise to an evaluative conclusion.

Speaking as someone who has been more influenced by the moderns than by the ancients, I don't see it.  It is not the case that both sets of criteria are factual in the sense defined above. The criteria of something's being a good watch already contains evaluative criteria.  For if a good watch is one that tells time accurately, then that criterion of chronometric goodness involves a standard of evaluation.  If I say of a watch that it is inaccurate, I am not merely describing it, but also evaluating it.  MacIntyre is playing the following game, to put it somewhat uncharitably.

He smuggles the evaluative attribute good into his definition of 'watch,' forgets that he has done so thereby generating the illusion that his definition is purely factual, and then pulls the evaluative rabbit out of the hat in his conclusion.  It is an illusion since the rabbit was already there in the premise.  In other words, both (1) and (2) are evaluative.  So, while the argument is valid, it is not a valid argument from a purely factual premise to an evaluative conclusion.

So if the precise question is whether one can validly move from a purely factual or descriptive premise to an evaluative conclusion, then MacIntyre's example fails to show that this is possible.

But I am being tendentious on purpose for didactic reasons.  I grant that it is not perfectly evident that values and facts are mutually exclusive.  I think what MacIntyre needs is the idea that some statements are both factual and evaluative.  If (1) is both, the MacIntyre gets what he wants.

Is Man a Functional Concept?

But suppose one would be wrong to reject the (1)-(2) counterexample and that, with respect to functional concepts, the move from fact to value is logically kosher.  Then this discussion is relevant to ethics, the normative study of human action,  only if man is a functional concept.  Aristotle maintains as much:  man qua man has a proper function, a proper role, a proper 'work' (ergon).  This is a proper function he has essentially, by his very nature, regardless of whatever contingent roles a particular human may instantiate, wife, father, sea captain.  Thus, " 'man' stands to 'good man' as 'watch' stands to 'good watch' . . . ." (56)  Now if man qua man has a proper and essential function, then to say of a particular man that he is good or bad is to imply that he has a proper and essential function.  But then to call a man good is also to make a factual statement.  (57)

The idea is that being human is a role that includes certain norms, a role that each of us necessarily instantiates whether like it or not.  There is a sort of coalescence of factual individual and norm in the case of each human being just as, in Aristotle's ontoloogy, there is a sort of coalescence of individual and nature in each primary substance. 

But does man qua man have a proper role or function?  The moderns fight shy of this notion.  They tend to  think of all roles, jobs, and functions as freely adopted and contingent.  Modern man likes to think of himself as a free and autonomous individual who exists prior to and apart from all roles.  This is what Sartre means when he says that existence precedes essence:  Man qua man has no pre-assigned nature or essence or proper function: man as existing individual makes himself what ever he becomes.  Man is not God's artifact, hence has no function other than one he freely adopts.

Although Aristotle did not believe in a creator God, it is an important question whether an Aristotle-style healing of the fact-value rift requires classical theism as underpinning. MacIntyre seems to think so. (Cf. p. 57)

Interim Conclusion

If the precise question is whether one can validly move from a purely factual or descriptive premise to an evaluative conclusion, I have yet to see a clear example of this.  But one ought to question the strict bifurcation of fact and value.  The failure of entailment is perhaps no surprise given the bifurcation.  The Aristotelian view, despite its murkiness, remains a contender.

Related articles

From 'Is' to 'Ought'? Help for Hodges

Comments

[Tully Borland]

"But one ought to question the strict bifurcation of fact and value."

I agree, since I tend to fancy the medieval doctrine of the equivalence of being and goodness. But here might be another example to motivate the pervasiveness of value in "factual" statements. Let's try this out:
1. It is snowing.
2. Thus, "it is snowing" is true (from p one can deduce that "p is true").
3. Thus, one ought to believe p and disbelieve ~p (since one ought to believe what is true and disbelieve contradictions).

Any assertion in ordinary contexts (of what is believed) carries with it an implicit norm that it is to be believed because it is true.
If that's right, then to think that there are many factual claims from which no norm can be deduced would be to fall into a pretty radical skepticism and perhaps even incoherence.

[Bill Vallicella]

Very interesting, Tully. And thanks for commenting. But there is a problem. The justification for the move from (2) to (3), which is given in parentheses, belongs in the premise set. Thus you are assuming the normative principle that one ought to believe what is true and disbelieve contradictions.

Thus you haven't given an example of a valid argument all of the premises of which are purely factual, and the conclusion of which is normative. You have given a valid argument to a normative conclusion from a mixed premise set.

[Bill Vallicella]

And then there is the great question that Nietzsche was the first to bring front and center, the question of the value of truth. Why is truth a value? Why ought I believe what is true? Ought I believe what is true even if the truth is not life-enhancing?

"One ought to believe what is true." How justify this? Why is it never morally permissible to disbelieve what is true?

"One ought to do the right thing" is a near-tautology. "One ought to believe that is true" is not close to being a tautology unless one assumes that the true is the good by way of belief. (Wm James)But then how justify the Jamesian dictum?

[Bill Vallicella]

There is also an issue concerning doxastic voluntarism.

If I ought to believe that p and disbelieve that ~p, and 'ought' implies 'can,' then I can both believe and not believe. But the example is from sense perception. Presumably one cannot do otherwise than believe it is snowing if one is standing in the middle of a snowstorm.

So if I know that it is true that it is snowing by sense perception, and cannot help having this true belief, then it seems to follow that it is not the case that I ought to believe that it is snowing.

[Tully Borland]

All good points, Bill. The reason for the parenthetical remarks in the argument I gave was (a) to be clear about where I'm cheating if I'm cheating but also (b) because I think the argument I gave sans parentheses is valid. Granted, it's not formally valid without the parenthetical propositions (it would be a bad example if I were teaching modern logic), but it does seem to me that if the first premise is true then 2 and 3 must be true. Is that cheating? I don't know. Would using modus ponens be cheating? Implicit in using modus ponens is that I think it is a correct form of inference; it's a good form of inference. Am I smuggling in the proposition "'p and (if p, then q), then necessarily q' is right" if a make a purported argument from factual premises to a normative conclusion? Maybe not, but there is some sense, it seems to me, in which I am assuming normativity in affirming the conclusion based on the argument. (Of course since Hume was willing to call into question induction maybe upon further thought he would call into question deduction and abduction as well. In which case, it may not be above board of Hume--or others like him--to ask for an argument from factual premises to a normative conclusion).

"One ought to believe what is true." How [to] justify this? Why is it never morally permissible to disbelieve what is true?

I wasn't trying to suggest that one morally ought to believe what is true (I'm unsure about that). I'd say moral "oughts" are perhaps a species of a more general normative "ought" (which perhaps is a species...). [Anselm understand truth in terms of rightness and I think there's something to be said for that.] So I was looking for some factual proposition which by its essence entails some normative notion or other. Maybe in this case it's a an "epistemic ought" of rationality or maybe proper function with respect to that segment of my noetic faculties aimed at the production of true belief. Whatever it is, it seems to me that in the simple act of making an argument, I'm committed to there being a sense in which at least one of the assertions ought to be believed. Since we can't escape normativity in our reasoning, the Humean objection loses a lot of its force. Even in the following argument, we're assuming some some normative fact, even if not explicitly stated in the premises:
1. 2+2=4
2. If 2+2=4 then squares aren't round.
3. Thus squares aren't round.

[Bill Vallicella]

>> I think the argument I gave sans parentheses is valid. Granted, it's not formally valid without the parenthetical propositions (it would be a bad example if I were teaching modern logic), but it does seem to me that if the first premise is true then 2 and 3 must be true. Is that cheating?<<

To be charitable, we can say that your argument is semantically valid although not syntactically valid, in which case it is like this one:

Tully is a brother
Therefore
Tully is a sibling.

This is syntactically invalid but semantically valid. But is it built into the meaning of 'true' that whatever is true ought to be believed, in the way it is built into the meaning of 'brother' that every brother is a sibling? I should think that *Whatever is true ought to be believed* is not an analytic proposition, but a synthetic necessary proposition. But then what grounds the connection?

[Tully Borland]

Thanks for hashing this out with me. One last attempt....
Let me try to simplify things with another shorter argument, ignoring for the moment "whatever is true ought to be believed":
1. 2+2=4.
2. Thus, one ought to believe that 2+2=4 or the Steelers won a Super Bowl in the 70's.

2 follows from 1 + the employment of addition. Formally, the "one ought to believe that" complicates matters; but not only that (p v q) but that (p v q ought to be believed given p) seems also to be implied by p and the use of addition. Now maybe if a terrorist puts a gun to my head and says that if I believe 2 on the basis of 1 then I'll be shot, then there's a sense in which I shouldn't affirm the argument, but to me this would be a case of conflicting norms (and I should pray that by some miracle my brain is quickly rewired).

But then have I derived a normative fact from a factual one without any implicit premises? Here's a thought: the move from the norm is justified by the justification for (1) and the process of reasoning itself and not by any other assumed propositions.

[Bill Vallicella]

Well, I don't see that (2) follows from (1) by Addition. I also don't see how the disjunctive conclusion counts as a normative statement.

(2) is equivalent to

2* If the Steelers didn't win a Super Bowl in the 70s, then one ought to believe that 2 + 2 = 4.

(2*), however, is not a categorical normative statement due to its conditional form.

Recall that what we are looking for is a valid argument from purely factual premises to a categorically normative conclusion.

[Tully Borland]

Bill, I spoke falsely when I said "for the last time". I'm going to try this again because I'm either confused or not being clear enough or both. And, yes, I understand what we're looking for.

My (2) was ambiguous. I meant that one ought to believe the disjunctive statement p v q, not simply p (as you interpreted it).

From p by addition we can derive p v q. But you don't think we can further derive "one ought to believe p v q"? But if I'm offering you
1. p
2. Thus, p v q
then am I not giving an argument for p v q and thereby presenting reasons for p v q such that it's to be believed?

But I have my own worries about that first example so maybe here's a better one. From p by the law of non-contradiction we can derive ~p. So I offer you the following argument:
1. p
2. Thus, ~p.

Shouldn't one also conclude from that argument that one should not believe~p? Is that not implied as well by p? You might say, "No, Tully, you're smuggling in a conditional premise "if p and the law of non-contradiction, then one shouldn't believe ~p." But I'm suggesting that this needn't be the case. I'm suggesting that, in the same way that we don't invoke the proposition "if (p then q) and p, thus necessarily q" when employing (the inferential process) modus ponens, so too we needn't invoke "if p and the law of non-contradiction, then one shouldn't believe ~p" in concluding that one shouldn't believe ~p.

[Bill Vallicella]

I'm afraid I am not following you at all.

It is true that snow is white. Does it follow that I ought to believe that snow is white? Well, if I know that snow is white, then I already believe it. How can I be under a moral obligation to believe a proposition that I know?