Normative Because False!?#

In what is meant to be  “a contribution of major importance to a unified theory of probability and utility” Jeffrey (The Logic of Decision) says about Bayesian decision theory that

Indeed, it is because logic and decision theory are woefully inadequate as descriptions that they are of interest as norms. (p.167)

Now,  here’s a worry I presented yesterday in the seminar and that I’d like to present again, so that other people may consider it and that the ones that heard it can see why it’s worrisome. There are, at least, two questions the claim above prompts:

1) If theory T is woefully inadequate as a description of phenomena F, and yet it is meant to be a normative theory of F, couldn’t it be that it makes absurd demands about F?

2) If it is in virtue of theory T’s woeful descriptive inadequacy towards F that T is an interesting normative theory of F, wouldn’t it be the case that false descriptive theories turn out to be interesting normative theories?

Here are some examples that illustrate both worries. Consider the following theory about human decision making behavior called ‘T1’

T1: For any subject and any decision D at any time t, subject S should fly up into the sky at time t2 if S reaches decision D at time t1.

T1 is woefully inadequate as a description of human decision behavior. If we accept Jeffrey’s diagnosis about  Bayesian Decision Theory, it is in virtue of this inadequacy that T1 is an interesting normative theory of decision making! But you may object, of course, that we have independent reasons to reject T1. The demands that it presents just don’t make sense at all! Point taken. But it will not get us out of trouble.

Consider the ‘ought’ implies ‘can’ principle which seems, to my mind, to offer a very sensible limitation on ANY normative theory: if subject S ought to do X in context C it must be the case that S is able in C to do X. Now suppose that, for some or other reason, whenever humans are in real decision making contexts (e.g., when they are decision where to go for dinner, or whether to watch a movie or do a reading) as opposed to consciously working out in a formal epistemology seminar, they cannot, because of their cognitive architecture, compute an algebra. Decision theory will still claim that humans should distribute their credences over some or other algebra. So it seems that Decision Theory will be claiming that humans ought to do what they cannot do. That seems absurd. Decision Theory should be a bit more informed about human psychology.

So far so good for the first worry. The second one is just as bad (if not worse). Suppose we accept Jeffrey’s claim that it is precisely because a theory T is woefully inadequate as a description that it is interesting as a normative theory. If so, then we seem to get the following argument:

P1 Theory T is woefully inadequate as a description of F.

P2 If T is woefully inadequate as a description of F then T is an interesting normative theory of F.

C Theory T is an interesting normative theory of F.

Now consider theories that give woefully inadequate descriptions of the phenomena they intend to theorize about. Here’s a big group of such theories: the false ones. Consider, in particular, Aristotle’s physics and astronomy. They are both woefully inadequate theories of the behavior of physical objects and planets. So we may run the argument:

P1 Aristotle’s astronomy is woefully inadequate as a theory of the behavior of planets.

P2 If Aristotle’s astronomy is woefully inadequate as a theory of the behavior of the planets, then it is an interesting normative theory of the behavior of the planets.

C Aristotle’s theory is an interesting normative theory of the behavior of the planets.

Now that does seem like a bad result, doesn’t it? That’s why we should worry about Jeffrey’s claim that

Indeed, it is because logic and decision theory are woefully inadequate as descriptions that they are of interest as norms.

The fact that a given theory T about phenomena F delivers a woefully inadequate description of F gives no good reason to take it to be an interesting normative theory of F. Jeffrey is simply wrong about this. And if he is right about decision theory being so woefully inadequate, then we should worry a bit about its theoretical value. In general, I think, all normative theories should worry about being so woefully inadequate!

Advertisements

13 Responses to Normative Because False!?#

  1. Aaron Boyden says:

    You don’t give a lot of context. It seems that while he might have phrased it better, I could easily imagine him meaning either of the following (non-absurd) things, and there could be others:

    1) He may mean that the specific features which make decision theory descriptively inadequate (perhaps features like its relative simplicity?) are strengths for a normative theory.

    2) He may mean that the descriptive inadequacy of decision theory means it could only possibly be of normative interest (his phrasing is definitely poor if this is what he means, but not so much that I find this interpretation unbelievable).

  2. dtlocke says:

    I’m with Aaron. There’s a much more charitable and much, much more plausible reading:

    “Indeed, it is because logic and decision theory are woefully inadequate as descriptions that they are of interest [mainly/only] as norms.”

    Emphasis on “mainly” or “only”.

  3. Edú says:

    Hi Aaron and Dustin,

    Here’s some more context. Jeffrey is discussing the relevance of including research on neuropsychology to understand how humans take on evidential information. He claims it would be relevant

    “If we were using the theory of preference to describe and interpret the behavior of de facto Bayesians: of agents who in fact behave as if they had preference rankings that satisfy the hypothesis of the uniqueness theorem (…) One day Bayesian robots may be built; but at present there are no such creatures, and in particular, human beings are not de facto Bayesians. Bayesian decision theory provides a set of norms for human decision making; but it is far from being a true description of our behavior. (…) Indeed, it is because logic and decision theory are woefully inadequate as descriptions that they are of interest as norms.”

    Two claims are suggested here:

    (i) The context does not help in any of the ways you may want to help Jeffrey. It seems that it, rather, confirms his point about Decision Theory being “Woefully” inadequate.

    (ii) Neither of the interpretations you so kindly offer would actually help him.

    Aaron’s second interpretation according to which the descriptive inadequacy of a given theory T “means” that T could “only possibly be of normative interest” won’t stop the claim from overgenerating: we’ll still have Aristotle’s Astronomy as a very interesting theory of how planets should behave.

    And the same goes for Dustin’s account. Even if we take Jeffrey to be saying that it is because T is inadequate that it is “mainly” of interest as a norm has the result that false theories, in virtue of their being false, are of interest mainly as norms. Saying that Aristotle’s theory of the planets is of interest mainly as a normative theory of planetary behavior is just as bad as saying that it is an interesting theory of how planets should behave (without mainly/only).

    I honestly don’t see how this helps!

  4. Steve C. says:

    Hey Edu, interesting thought. It definitely does seem as if Jeffrey’s claim is poorly worded. So, here’s a different suggestion from charity that occurs to me. I wonder if it might be implicit in Jeffrey’s discussion that (i) logic and decision theory are of interest with respect to human behavior, and (ii) the only two viable possibilities are: they are of interest as descriptive of human behavior, or they are of interest as prescriptive of human behavior. If all of that is implicit, then there is good sense to be made of Jeffrey’s comment, I would think. He would just be carrying out a disjunctive syllogism (with certain suppressed premises).

    (Also, it seems as if “of interest” admits of two interpretations. Logic and decision theory could be interesting with respect to human behavior, or they could be of interest to certain parties with respect to human behavior. If the latter, perhaps Jeffrey is himself engaged in charitable interpretation.)

    Well, I’m no scholar of Jeffrey, so I’ll leave it to you to tell me whether that interpretation flies.

  5. Steve Nayak-Young says:

    Steve C.’s suggestion is tempting, but I’m not sure whether it would work. If Jeffrey’s quote implies the descriptive claim that some or all people find logic and decision theory interesting, then it doesn’t follow that because it’s a false description of human behavior, it’s only of interest as normative.

    Even if logic/decision theory is a false description of human cognitive behavior (I agree that it is), many/most of the people who find it interesting might not be hip to the falseness, so they might still find it interesting as a description (which they believe to be true), but wouldn’t find it interesting as normative if they ceased to believe it was a true description.

    To illustrate, think of flat-earthers. They think/thought that it’s true that the Earth is flat, and they find that descriptive claim interesting, but if you convinced them that the Earth is round, they wouldn’t be interested in the absurd normative claim that the Earth *should* be flat.

    So how do we convince all these benighted folks that brains are “round”? ;-)

  6. Edú says:

    Hi all,

    Thanks to Steve C and Steve N’s comments. I think Steve C’s proposal is worth considering. Here’s some more context in case it helps. What Jeffrey says about Logic here is:

    “Similarly, deductive logic provides a set of norms for human deductive reasoning, but cannot usefully be reinterpreted as a description of human reasoning (unless we define reasoning as to exclude everything that violates the rules of deductive logic, in which case the description becomes true but fragmentary and uninteresting).”

    This seems to bring some support to Steve C’s comment that Jeffrey is trying to offer a disjunctive syllogism. The problem is that this very same argument runs into Steve N’s problems: we can use it to revive Aristotle’s woefully wrong theory of the planets. This theory is of interest either as descriptive or as normative. It’s totally false, so it must be of normative interest. The result is clearly ridiculous. Not only should the earth be flat, but species should have been created by God.

    How are we going to stop this from happening? Bayesian Decision theorists (and perhaps logicians) may want to claim something like “the dictates of reason” requires that humans should behave according to Decision Theory. But if we accept this petitio, I don’t see how we are going to block arguments of the form: “The dictates of Religion, God, Astrology, Numerology, and what not, make it so that the Earth should be flat, or what not.”

    I’m pretty sure the only way to stop this really bad result is to put some useful constraints against which you can compare normative proposals. And I don’t see any other useful constraint but the way the world actually is (i.e., empirically based limits). There’s no other way to defend it without begging the question (and opening the door for any false theory to be of normative interest).

  7. jdmitrig says:

    I don’t see how

    “If T is woefully inadequate as a description of F then T is an interesting normative theory of F.”

    follows from the claim that

    “it is because logic and decision theory are woefully inadequate as descriptions that they are of interest as norms.”

    This might follow if you insert a “solely” before “because.” But without this, or some other qualification to the effect that the *only* reason probabilism is of interest as a normative theory is its descriptive inadequacy, I don’t see how the conditional above follows.

    It seems to me that what Jeffrey meant to be affirming in the passage you quote was the converse of the conditional you’re trying to attribute to him; that is:

    (P2*) if T is of interest as a normative theory, then it is descriptively inadequate.

    That’s why he has the comment about us not needing a normative theory which tells us to breathe in and breathe out every few seconds. Such a norm is not of interest because it is perfectly descriptively adequate (this follows from the contrapositive of (P2*)). We don’t need a rule like that to guide our behavior because our behavior already accords with a rule like that.

    You may worry that probabilism demands behavior of agents that it’s not reasonable to expect, given our cognitive limitations. I think that’s an interesting point, and one worth debating. But I don’t see the relationship between that debate and the passage quoted above. All that passage commits Jeffrey to is (P2*). And while I agree with Jason that (P2*) is probably false – it certainly doesn’t entail that Ptolemaic astronomy is an interesting normative theory.

  8. Edú says:

    Hi Dimitri,

    Thanks for your interesting comments. I think you’re right in that the conditional I propose doesn’t straightforwardly follow from Jeffrey’s quote. But I also think the one doesn’t even indirectly follow from it and that it also has terrible results. The claim says:

    “It is because L and D are woefully inadequate as descriptions that they are of interest as norms.”

    Now here’s what I was reading:

    I1:
    It is in virtue of their being inadequate as descriptions that L and D are of interest as norms.

    I1 seems to me to be a pretty literal rephrasing of Jeffrey’s claim. Now, if we do accept I1, it seemed to me, one must implicitly accept I2:

    I2:
    Any theory that is descriptively false will have all that’s needed to be an interesting normative theory. Since it is “in virtue of” the former that the latter is normatively interesting.

    Now, from I2 the conditional I proposed does seem to follow:

    I3: If T is false then T is interesting as a norm.

    Your proposed reading does seem like a good alternative. You want us to read I4 from Jeffrey:

    I4: If T is normative then T is descriptively false.

    The reason why I think your conditional is wrong is because it puts the explanatory powers on the wrong side. For if I4 is true, I5 follows:

    I5: Any normative theory has everything that it needs to be descriptively false. Since it is in virtue its being normative that any given theory is False.

    This much follows from your claim. But it seems to me quite clear that Jeffrey is not asserting anything close to I5. It is pretty obvious from his use of “because” that “descriptive inadequacy” explains “normative interest” and not the other way around.

    “It is because L and D are woefully inadequate as descriptions that they are of interest as norms.”

    So your proposed reading simply doesn’t follow from Jeffrey’s text. Unless we take him to be highly uncooperative (or to be incompetent in the use of the English word “because”) there’s no way we can read him as saying what you want.

    So that’s not the right reading. Notice, however, that even if we accept such reading, we still get very ridiculous results. If your proposal is right, then if a given theory T makes interesting normative claims it has to be false. In your words:

    *If T is of interest as a normative theory, then it is descriptively inadequate.*

    Suppose we now have String Theory to be our theory T. And suppose String Theory happens to be making a normative claim of the following form:

    C: If T is right, then we “should” find the Higgins-B particle in our Geneva experiments.

    If your reading is right we’re in trouble. On your reading, for C to be an interesting claim, String Theory must be false. For example, there should be no Higgins-B particle. However, it seems to me that for C to be true at all, there better be adequate descriptive power. So we’re in a dilemma: since C is a normative claim it is an interesting claim in so far as it is false. Or the other way around, C is a truthful claim only in so far as it is a noninteresting claim.

    Thus, even in your alternative (not supported) reading of Jeffrey we run into serious trouble. That all interesting normative theories turn out to be false seems like a terrible claim to make. If that’s what Jeffrey is claiming of decision theory, and if it’s true, then decision theory doesn’t seem that useful anymore.

    Furthermore, notice that we can still build up Ptolemaic astronomy as an interesting normative view. Take PA* to be what results from changing Ptolemaic Astronomy from a descriptive theory into a normative one. So if it used to say things like “Earth is at the center, the planets move this way” it now says things like “Earth should be at the center, the planets should move this way”.

    PA* is a very simple, parsimonious, coherent, clear theory. So it’s really interesting. Furthermore, it’s normative. It has a terrible predictive power. But that’s not a vice, but rather a virtue. For, remember, its a normative theory and, given (P2*) all normative theories are descriptively inadequate. So the fact that PA* doesn’t adequately describe most of the planetary phenomena shouldn’t scare you. In fact, it should be a plus on behalf of PA*. For it verifies the Jeffrey-Dimitri claim that “If T is normatively interesting, then it is descriptively inadequate”.

    Once more, I don’t see how we’re helping Jeffrey here.

  9. jdmitrig says:

    In your jump from l1 to l2, you again stick an implicit “solely” in front of the “in virtue of” clause.

    Anything with Y has everything it needs to be X.

    does not follow from

    It is in virtue of Y that some (particular) thing is X.

    while it *may* follow from

    It is *solely* in virtue of Y that some (particular) thing is X.

    The first merely makes Y a necessary condition for X. The second makes Y a sufficient condition for X.

    ———–

    Also, I wasn’t claiming that (P2*) was the *only* thing that Jeffrey was asserting. I was just saying that it was the only conditional he was asserting (and not the converse conditional, as you were saying).

    You’re right that the descriptive inadequacy is *explaining* the normative interest. However (and this dovetails with my comment above the line) just because Y explains X doesn’t mean that Y is a sufficient condition for X. So I can simultaneously accept that Y explains X while denying that If something is Y then it is X – while affirming that if something is X then it is Y. There’s no tension between those three propositions.

    So just because the descriptive inadequacy is *explaining* the normative interest doesn’t mean that you have to accept any conditional relationship between the two – much less so that you have to accept *your* proposed conditional that ‘if something is descriptively inadequate, then it is of normative interest.’

    ———-

    Let’s run the whole conversation with a different subject matter – just to get some perspective. Imagine that instead of descriptive inadequacy and normative interest, we were talking about my grandmother’s new batch of taffy and it’s popularity. Suppose that one of the (many) things that one looks for in taffy is thickness. Also, of course, one likes taffy that is sweet, not covered in oil, large enough to be seen, etc. So thickness is clearly not the *only* thing that makes taffy good, but it’s *one* of the things that makes taffy good, and my grandmother’s taffy happens to be quite thick.

    Now suppose Jeffrey shows up and says:

    “It is precisely because the taffy is so thick that it is so popular.”

    Now, we can both agree that, in this case, Jeffrey is using the thickness to *explain* the popularity. That’s settled. But we could go further, taking a strong reading of this statement and assuming that Jeffrey is affirming (1) as well:

    (1) If taffy is popular, then it is thick.

    That’s what I read Jeffrey as saying – though I disagree with him (this is not the “Jeffrey-Dmitri” claim).

    Or, we could read him as saying this:

    (2) If taffy is thick, then it is popular.

    Of course, then we could go to town with all the awful consequences which follow from this. Jeffrey’s committed to the popularity of thick taffy which is covered in oil! How silly of him!

    I hope you agree with me that, in this case, what’s silly isn’t Jeffrey’s claim but the proposed interpretation. Perhaps you could tell me, then, how this case is different from the one we’ve been discussing.

  10. Steve C. says:

    (Disclaimer: I haven’t been reading the comments closely.)

    Edu,
    I had one other thought about a charitable reading of Jeffrey:

    “Indeed, it is because logic and decision theory are woefully inadequate as descriptions that they are of interest [if at all] as norms.” (p.167)

    Could he have meant this?

  11. Edú says:

    Hi Dimitri,

    I don’t see why I need the solely. If Jeffrey is making a causal claim that can be phrased as in I1 by using “in virtue of” then I2 follows. If it is in virtue of the rain and your being unprotected that you got wet, then rain and unprotectedness will always be enough for you to get wet. If it is in virtue of T’s falsity that T is normatively interesting, then falsity will always be enough for normative interest.

    If the “in virtue of” is not the same as “sufficient cause” then I think Jeffrey is very uncooperative.

    —–

    No matter what is true about the above, you have to admit that your interpretation of Jeffrey simply doesn’t make sense when compared to the text. You take him to claim that it is because decision theory is normative that it is descriptively inadequate. He’s saying literally the opposite!

    ——

    I don’t clearly see your analogy with taffy. But you can let go the conditional in (P2) and STILL get the same bad results. Here’s the reinterpretation:

    P1 Theory T is woefully inadequate as a description of F.

    P2 It is at least in virtue of T’s woeful inadequacy to describe F that T is an interesting normative theory of F.

    C There is at least one reason to think that T is an interesting normative theory of F.

    Notice the conclusion has changed (to adapt negative emotions against the alleged implicit “solely”). C, to my eyes, is still preposterous. It entails two terrible things:

    a) That there’s at least one good reason to think that Intelligent design describes the way the world should be; and

    b) That this “good reason” is simply the fact that Intelligent Design is woefully inadequate to describe that same world that it rules upon.

    ——-

    Steve C:

    This actually sounds like a more suitable interpretation. That is, it does reduce damage for Jeffrey. The price, however, seems high: it presents Decision Theory as a theory that does such a bad descriptive job that it’s only interest is the normative one.

    Another problem with this proposal is that it faces the same old problem: accepting this claim opens the door for theoretical hell (e.g., we can claim that Intelligent Design is of interest, if at all, as a normative theory). I’ll try to phrase this in a different way so that it can be clearer.

    No matter how we interpret Jeffrey here, one thing will always be the case: we can use his argument to “save” any theory (no matter how terribly false it may be) in exactly the same way in which he saves Decision Theory after admitting that it’s woefully inadequate. You may want to add as many hedges as possible, avoid the conditional (following Dimitri) put “at leasts” and get rid of “solelys” and that bad result will always be there.

    I think there’s an easier way to save Decision Theory here: reject Jeffrey’s claim! It is NOT because it is woefully inadequate as a description that Decision Theory is an interesting normative theory. It’s descriptive inadequacy is not even one among many potential causes here. It is, if at all, a vice and not a virtue of the theory.

    There’s another way to defend the theory: to show that it’s NOT woefully inadequate. I don’t know how and whether this can be done. But judging by the high degree of a prioricity that lurks around here, I doubt it’s going to be an easy task.

  12. dtlocke says:

    P1. If the literal interpretation of Jeffrey’s sentence is the correct one, then Jeffrey is a moron.
    P2. Jeffrey is not a moron.
    C. The literal interpretation of Jeffrey’s sentence is not the correct one.

    Can we get back to doing philosophy now?

  13. Edú says:

    Oops! I forgot to add:

    Dear Dustin: you’re being utterly oblivious of the series of arguments that have been presented here. Specially the very last ones which show that:

    P1. No matter how you interpret Jeffrey, we end making ridiculous defenses of stupid theories.
    P2. If a given philosophical view ends up making ridiculous defenses of stupid theories, then it’s a ridiculous view.
    C. Jeffrey is claiming something ridiculous. No matter how you interpret him.

    Can we stop defending philosophers and doctrines as if they were priests and churches?

%d bloggers like this: