A Peeve and A Proposal

My recent experience with refereeing for journals has had an inordinate commonality. I’m seeing a greater percentage of counterfactual analyses playing a role in papers than I can ever recall. I wouldn’t be surprised if my experience is idiosyncratic, and perhaps I notice these things more than others since counterfactual analyses are a pet peeve of mine–it is a sign of the lack of quality control in our discipline, since it has been known for more than two decades how precarious such analyses are.

My proposal is that no article containing a counterfactual account of anything should be allowed without an explicit explanation of how that account avoids the conditional fallacy described by Shope in his 1979 Journal of Philosophy piece (or whatever emendation of Shope’s results are favored). Upon encountering such an analysis, it is usually an easy task to construct counterexamples to the account, but something more is needed than simply rebutting all the counterexamples one can think of or that are brought to one’s attention. Responsible authors ought to be able to say why their particular counterfactual account can avoid the recipe that can be elicited from Shope’s piece (with an important precursor in Chisholm’s underappreciated “The Problem of Empiricism,” though Chisholm doesn’t attempt to generalize to all counterfactual theories).


A Peeve and A Proposal — 13 Comments

  1. Jonathan, Can you summarize the “counterfactual fallacy”
    pointed out by Shope? Is the same sort of problem
    that comes up in the Frankfurt-type cases and also
    the context of “passive” powers, such as mallegability
    solubility in water, etc.?

  2. I think it is the same sort of problem (e.g., as in Kripke’s ‘killer yellow’ and finks). It’s talked about alot in ethical theory. Any analysis or definition in terms of a subjuntive conditional has the problem (I’m not sure it is really a fallacy…it’s just a mistake). Anything with some P on the LHS and also something of the form (If Q were the case, then R would be case) in the RHS. The problem crops up when P’s truth depends on whether Q, the antecedent, is true, and not just on whether if Q were true, R would be true. One of Shope’s examples was the full-informational account of a person’s good: something is good for a person iff were she fully informed, then she would desire it. Shope noted that gathering information is sometimes good for you, but not something you would desire were you fully informed. So whether something is good for a person can’t depend on whether her fully informed self would desire it. After all, it sometimes depends also on whether she is indeed fully informed. That’s the gist of it, I think.

  3. Thanks, Robert. Yes, in my book The Metaphysics of Free Will, I briefly talk about these problems in the context of ethical theory, such as Kantian ethical theory. Various people like Barbara Herman define the role of the Categorical Imperative in terms of simple counterfactuals, but the problems then emerge.

    I dubbed these contexts, these “black holes in metaphysical space”, Schizophrenic Situations. They are characterized by weird “swerves” in hypothetical scenarios. The Frankfurt-type cases are special cases of such situations. Has anyone done a more systematic study of how similar problems pop up in so many diverse contexts in philosohpy?

    Cheers! This is a great blog!

  4. I like your name for the situations, very apt. Shope himself (in those 2 or 3 papers in the 70’s, but especially the ’78 paper in JPhil) is the only person I can think of who has done anything like a systematic look at the formal problem, and shown how it crops up in a variety of areas. His own interest seemed to be in how it makes a problem for epistemology–for instance, Chisholm’s view that a proposition is more reasonable than another when, were he rational and his concerns purely intellectual, he would choose the one rather than the other. (Shope’s counter: the proposition “my concerns are purely intellectual”).

  5. Just got back from the doctor and noticed Robert is doing an admirable job of answering John’s question about the formal character of the inference. Robert’s got a very nice piece in Ethics on the problem as it surfaces for virtue theories. The problem concerns obligations for self-improvement, because if you were fully virtuous, you wouldn’t need to improve.

  6. Both work, though Shope’s original was the propositions that my concerns are purely intellectual. Presumably that is not a reasonable proposition for you to believe, but you would prefer it rather than the reverse, were your concerns purely intellectual. With your version, it is reasonable for you to believe that your concerns are not purely intellectual, but that is not what you would believe were your concerns purely intellectual.

  7. I’ve posted something on The Garden of Forking Paths,
    with a link to this post, as I think the Gardeners,
    preocuppied as they are with Frankfurt-type cases,
    may be interested.

  8. Jon: Insofar as counter-examples to conditional analyses really can be generated by a simple “recipe,” yet their authors are oblivious to that problem, that’s something to be peeved about.
    But the term “conditional fallacy” has a funny way of expanding so that any conditional analysis which is subject to counter-examples of a certain very broad type, whether these c-e’s can be generated by a simple “recipe” or not, are guilty of committing the dreaded “conditional fallacy.” The term seems, at least to me, to have so expanded in Shope’s own use of it. And I’m not sure, but I see hints, that you use the term broadly, too, suggesting that authors should not only anticipate the “easy” to generate counter-examples, but, at least with respect to Shope-ish c-e’s, should show upfront that there are none, easy or not.
    When is a counter-example generated by a “simple recipe”? I have no analysis (conditional or otherwise!) of this to offer, but here’s what strikes me as a decent rule of thumb: If, upon encountering an analysis, you can immediately start speaking, saying something along the lines of, “So, if this analysis commits the ‘conditional fallacy’, there’s a case where the satisfaction of the conditional’s antecedent” … “, and a few seconds and a few sentences later you’ve semi-automatically located the counter-example, then it’s at least plausible that the analysis really does get shot down by a simple recipe. Or, where the analysis is itself fairly complex, if you can say, “Wait, give me just a minute” … ,” and you quickly and semi-automatically come up with the case, then again, there may well be a simple recipe at work. But when you have to sit down with paper and pencil, and work through several variations of some cases before you come up with what looks like a counter-example to the analysis in question, I think that’s a good sign you are no longer following any simple “recipe,” but are just locating a counter-example of a certain type (in this case, of a broad, “Shope-ish” variety).
    When “conditional fallacy” is used broadly, as roughly described above, I don’t like your proposal. If the analyzer has dealt with potential counter-examples that they’ve thought of, that are out there in literature, and that have brought to their attention, including all the potential Shope-ish c-e’s that really can be generated by a simple “recipe,” yet you are able to come up with a new Shope-ish counter-example they hadn’t anticipated, why not just say that the analysis turns out to be subject to counter-example, or more specifically a counter-example of a certain very broad Shope-ish type, after all, instead of saying it commits the “conditional fallacy”? That seems an overly broad and unfair use of “fallacy.” (Why not say of any analysis of any type that turns out to be subject to any type of counter-example, no matter how sophisticated, that it’s guilty of the “being incorrect fallacy”? After all, it’s not just counterfactual analyses, but analyses of rich, philosophically important concepts generally that our discipline’s experience has shown to be “precarious.”) And to demand of any conditional analysis that it explain how it avoids the “fallacy,” where “conditional fallacy” is used broadly, comes too close to the injunction: “You must not only deal with the counter-examples that you really are responsible for coming up with, but, with regard to c-e’s of a certain very broad type, but not others, you must tell us upfront why we should believe that there are no new, unanticipated counter-examples to your analysis.”

  9. I think I agree with Keith it is very difficult to see Shope’s, and Shope-like objections, as stemming from any single ‘mistake’. (Though I don’t think a mistake needs to be discoverable by recipe in order to be a fallacy.) They do all stem from the use of a conditional analysis, but it is clearly not true that every conditional analysis is subject to the problems he notes. For instance, the typical ‘fixer’ for the full information view is to say that what’s good for a person is not what he himself would want under full information, but what he himself would want for his less informed self. There are objections to this fixer, but it doesn’t make the mistake the first analysis did. Shope himself goes to great lengths to try to squeeze out a statement of the ‘fallacy’, but I don’t find it very pursuasive.

  10. Keith–you’re right to distinguish the two kinds of counterexamples, and it is also the case that the recipe in question is not easy to state. I’m not sure I can do it, but the proposals I have in mind all depend on the insight Robert speaks of, where the truth value of the antecedent of the counterfactual and the item being analyzed or explained are clearly not independent. So here the recipe is fairly simple, where Q counterfactually implies R is supposed to analyze P: figure out a specific example in which Q implies ~P, and then tinker to get R true; or the more common one, found in Chisholm and later in Plantinga, where you imagine a case in which Q is true but P is not, and yet the case as imagined entails ~R.

    These “recipes” don’t always yield a counterexample in just a few seconds, and if it takes some work to get the counterexample, then I think you’re right that this isn’t a reason to be peeved. In almost every case, though, here’s what I see: a counterfactual proposal, maybe a reply to some simple argument or some explanation about how the proposal is effective to explain the problem addressed, but no discussion or apparent awareness of how easy it is to find these kinds of counterexamples. I know it’s harder to find problems with one’s own work than it is to find problems with the work of others, but the standard ways of undermining counterfactual accounts are too much available in the literature for authors to ignore.

Leave a Reply

Your email address will not be published. Required fields are marked *